diff options
Diffstat (limited to 'drivers/gpu/drm')
164 files changed, 5925 insertions, 3469 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c index 9ada56c16a58..4c851fde1e82 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c @@ -840,6 +840,9 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, else if (type == CGS_UCODE_ID_SMU_SK) strcpy(fw_name, "amdgpu/polaris10_smc_sk.bin"); break; + case CHIP_POLARIS12: + strcpy(fw_name, "amdgpu/polaris12_smc.bin"); + break; default: DRM_ERROR("SMC firmware not supported\n"); return -EINVAL; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 31ca817f16b6..2201303b9262 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -73,6 +73,7 @@ static const char *amdgpu_asic_name[] = { "STONEY", "POLARIS10", "POLARIS11", + "POLARIS12", "LAST", }; @@ -1277,6 +1278,7 @@ static int amdgpu_early_init(struct amdgpu_device *adev) case CHIP_FIJI: case CHIP_POLARIS11: case CHIP_POLARIS10: + case CHIP_POLARIS12: case CHIP_CARRIZO: case CHIP_STONEY: if (adev->asic_type == CHIP_CARRIZO || adev->asic_type == CHIP_STONEY) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index ec4783881597..ba2816b3cd07 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -418,6 +418,13 @@ static const struct pci_device_id pciidlist[] = { {0x1002, 0x67CA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS10}, {0x1002, 0x67CC, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS10}, {0x1002, 0x67CF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS10}, + /* Polaris12 */ + {0x1002, 0x6980, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS12}, + {0x1002, 0x6981, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS12}, + {0x1002, 0x6985, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS12}, + {0x1002, 0x6986, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS12}, + {0x1002, 0x6987, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS12}, + {0x1002, 0x699F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_POLARIS12}, {0, 0, 0} }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c index fc592c2b0e16..95a568df8551 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c @@ -98,6 +98,7 @@ static int amdgpu_pp_early_init(void *handle) switch (adev->asic_type) { case CHIP_POLARIS11: case CHIP_POLARIS10: + case CHIP_POLARIS12: case CHIP_TONGA: case CHIP_FIJI: case CHIP_TOPAZ: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index a81dfaeeb8c0..1d564beb0fde 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -65,6 +65,7 @@ #define FIRMWARE_STONEY "amdgpu/stoney_uvd.bin" #define FIRMWARE_POLARIS10 "amdgpu/polaris10_uvd.bin" #define FIRMWARE_POLARIS11 "amdgpu/polaris11_uvd.bin" +#define FIRMWARE_POLARIS12 "amdgpu/polaris12_uvd.bin" /** * amdgpu_uvd_cs_ctx - Command submission parser context @@ -98,6 +99,7 @@ MODULE_FIRMWARE(FIRMWARE_FIJI); MODULE_FIRMWARE(FIRMWARE_STONEY); MODULE_FIRMWARE(FIRMWARE_POLARIS10); MODULE_FIRMWARE(FIRMWARE_POLARIS11); +MODULE_FIRMWARE(FIRMWARE_POLARIS12); static void amdgpu_uvd_idle_work_handler(struct work_struct *work); @@ -149,6 +151,9 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) case CHIP_POLARIS11: fw_name = FIRMWARE_POLARIS11; break; + case CHIP_POLARIS12: + fw_name = FIRMWARE_POLARIS12; + break; default: return -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 69b66b9e7f57..8fec802d3908 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -52,6 +52,7 @@ #define FIRMWARE_STONEY "amdgpu/stoney_vce.bin" #define FIRMWARE_POLARIS10 "amdgpu/polaris10_vce.bin" #define FIRMWARE_POLARIS11 "amdgpu/polaris11_vce.bin" +#define FIRMWARE_POLARIS12 "amdgpu/polaris12_vce.bin" #ifdef CONFIG_DRM_AMDGPU_CIK MODULE_FIRMWARE(FIRMWARE_BONAIRE); @@ -66,6 +67,7 @@ MODULE_FIRMWARE(FIRMWARE_FIJI); MODULE_FIRMWARE(FIRMWARE_STONEY); MODULE_FIRMWARE(FIRMWARE_POLARIS10); MODULE_FIRMWARE(FIRMWARE_POLARIS11); +MODULE_FIRMWARE(FIRMWARE_POLARIS12); static void amdgpu_vce_idle_work_handler(struct work_struct *work); @@ -121,6 +123,9 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) case CHIP_POLARIS11: fw_name = FIRMWARE_POLARIS11; break; + case CHIP_POLARIS12: + fw_name = FIRMWARE_POLARIS12; + break; default: return -EINVAL; diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index 84afaae97e65..d4452d8f76ca 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -2512,6 +2512,8 @@ static int dce_v10_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((amdgpu_crtc->cursor_width - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -2537,7 +2539,6 @@ static int dce_v10_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -2578,7 +2579,9 @@ static int dce_v10_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v10_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -2587,16 +2590,10 @@ static int dce_v10_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v10_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v10_0_show_cursor(crtc); @@ -2620,7 +2617,6 @@ unpin: static void dce_v10_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v10_0_lock_cursor(crtc, true); @@ -2628,10 +2624,6 @@ static void dce_v10_0_cursor_reset(struct drm_crtc *crtc) dce_v10_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v10_0_show_cursor(crtc); dce_v10_0_lock_cursor(crtc, false); diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 7a7fa96d2e49..1cf1d9d1aec1 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -167,6 +167,7 @@ static void dce_v11_0_init_golden_registers(struct amdgpu_device *adev) (const u32)ARRAY_SIZE(stoney_golden_settings_a11)); break; case CHIP_POLARIS11: + case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, polaris11_golden_settings_a11, (const u32)ARRAY_SIZE(polaris11_golden_settings_a11)); @@ -608,6 +609,7 @@ static int dce_v11_0_get_num_crtc (struct amdgpu_device *adev) num_crtc = 6; break; case CHIP_POLARIS11: + case CHIP_POLARIS12: num_crtc = 5; break; default: @@ -1589,6 +1591,7 @@ static int dce_v11_0_audio_init(struct amdgpu_device *adev) adev->mode_info.audio.num_pins = 8; break; case CHIP_POLARIS11: + case CHIP_POLARIS12: adev->mode_info.audio.num_pins = 6; break; default: @@ -2388,7 +2391,8 @@ static u32 dce_v11_0_pick_pll(struct drm_crtc *crtc) int pll; if ((adev->asic_type == CHIP_POLARIS10) || - (adev->asic_type == CHIP_POLARIS11)) { + (adev->asic_type == CHIP_POLARIS11) || + (adev->asic_type == CHIP_POLARIS12)) { struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(amdgpu_crtc->encoder); struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv; @@ -2528,6 +2532,8 @@ static int dce_v11_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((amdgpu_crtc->cursor_width - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -2553,7 +2559,6 @@ static int dce_v11_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -2594,7 +2599,9 @@ static int dce_v11_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v11_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -2603,16 +2610,10 @@ static int dce_v11_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v11_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v11_0_show_cursor(crtc); @@ -2636,7 +2637,6 @@ unpin: static void dce_v11_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v11_0_lock_cursor(crtc, true); @@ -2644,10 +2644,6 @@ static void dce_v11_0_cursor_reset(struct drm_crtc *crtc) dce_v11_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v11_0_show_cursor(crtc); dce_v11_0_lock_cursor(crtc, false); @@ -2822,7 +2818,8 @@ static int dce_v11_0_crtc_mode_set(struct drm_crtc *crtc, return -EINVAL; if ((adev->asic_type == CHIP_POLARIS10) || - (adev->asic_type == CHIP_POLARIS11)) { + (adev->asic_type == CHIP_POLARIS11) || + (adev->asic_type == CHIP_POLARIS12)) { struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(amdgpu_crtc->encoder); int encoder_mode = @@ -2992,6 +2989,7 @@ static int dce_v11_0_early_init(void *handle) adev->mode_info.num_dig = 6; break; case CHIP_POLARIS11: + case CHIP_POLARIS12: adev->mode_info.num_hpd = 5; adev->mode_info.num_dig = 5; break; @@ -3101,7 +3099,8 @@ static int dce_v11_0_hw_init(void *handle) amdgpu_atombios_crtc_powergate_init(adev); amdgpu_atombios_encoder_init_dig(adev); if ((adev->asic_type == CHIP_POLARIS10) || - (adev->asic_type == CHIP_POLARIS11)) { + (adev->asic_type == CHIP_POLARIS11) || + (adev->asic_type == CHIP_POLARIS12)) { amdgpu_atombios_crtc_set_dce_clock(adev, adev->clock.default_dispclk, DCE_CLOCK_TYPE_DISPCLK, ATOM_GCK_DFS); amdgpu_atombios_crtc_set_dce_clock(adev, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c index 44f024c9b9aa..809aa94a0cc1 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c @@ -1859,6 +1859,8 @@ static int dce_v6_0_cursor_move_locked(struct drm_crtc *crtc, struct amdgpu_device *adev = crtc->dev->dev_private; int xorigin = 0, yorigin = 0; + int w = amdgpu_crtc->cursor_width; + amdgpu_crtc->cursor_x = x; amdgpu_crtc->cursor_y = y; @@ -1878,6 +1880,8 @@ static int dce_v6_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((w - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -1903,7 +1907,6 @@ static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -1944,7 +1947,9 @@ static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v6_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -1953,16 +1958,10 @@ static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v6_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v6_0_show_cursor(crtc); @@ -1986,7 +1985,6 @@ unpin: static void dce_v6_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v6_0_lock_cursor(crtc, true); @@ -1994,10 +1992,6 @@ static void dce_v6_0_cursor_reset(struct drm_crtc *crtc) dce_v6_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v6_0_show_cursor(crtc); dce_v6_0_lock_cursor(crtc, false); } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index 30945fe55ac7..d2590d75aa11 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -2363,6 +2363,8 @@ static int dce_v8_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((amdgpu_crtc->cursor_width - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -2388,7 +2390,6 @@ static int dce_v8_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -2429,7 +2430,9 @@ static int dce_v8_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v8_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -2438,16 +2441,10 @@ static int dce_v8_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v8_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v8_0_show_cursor(crtc); @@ -2471,7 +2468,6 @@ unpin: static void dce_v8_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v8_0_lock_cursor(crtc, true); @@ -2479,10 +2475,6 @@ static void dce_v8_0_cursor_reset(struct drm_crtc *crtc) dce_v8_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v8_0_show_cursor(crtc); dce_v8_0_lock_cursor(crtc, false); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index d0ec00986f38..373374164bd5 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -139,6 +139,13 @@ MODULE_FIRMWARE("amdgpu/polaris10_mec.bin"); MODULE_FIRMWARE("amdgpu/polaris10_mec2.bin"); MODULE_FIRMWARE("amdgpu/polaris10_rlc.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_ce.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_pfp.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_me.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_mec.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_mec2.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_rlc.bin"); + static const struct amdgpu_gds_reg_offset amdgpu_gds_reg_offset[] = { {mmGDS_VMID0_BASE, mmGDS_VMID0_SIZE, mmGDS_GWS_VMID0, mmGDS_OA_VMID0}, @@ -689,6 +696,7 @@ static void gfx_v8_0_init_golden_registers(struct amdgpu_device *adev) (const u32)ARRAY_SIZE(tonga_golden_common_all)); break; case CHIP_POLARIS11: + case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, golden_settings_polaris11_a11, (const u32)ARRAY_SIZE(golden_settings_polaris11_a11)); @@ -903,6 +911,9 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev) case CHIP_POLARIS10: chip_name = "polaris10"; break; + case CHIP_POLARIS12: + chip_name = "polaris12"; + break; case CHIP_STONEY: chip_name = "stoney"; break; @@ -1768,6 +1779,7 @@ static int gfx_v8_0_gpu_early_init(struct amdgpu_device *adev) gb_addr_config = TONGA_GB_ADDR_CONFIG_GOLDEN; break; case CHIP_POLARIS11: + case CHIP_POLARIS12: ret = amdgpu_atombios_get_gfx_info(adev); if (ret) return ret; @@ -2682,6 +2694,7 @@ static void gfx_v8_0_tiling_mode_table_init(struct amdgpu_device *adev) break; case CHIP_POLARIS11: + case CHIP_POLARIS12: modearray[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | PIPE_CONFIG(ADDR_SURF_P4_16x16) | TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | @@ -3503,6 +3516,7 @@ gfx_v8_0_raster_config(struct amdgpu_device *adev, u32 *rconf, u32 *rconf1) *rconf1 |= 0x0; break; case CHIP_POLARIS11: + case CHIP_POLARIS12: *rconf |= RB_MAP_PKR0(2) | RB_XSEL2(1) | SE_MAP(2) | SE_XSEL(1) | SE_YSEL(1); *rconf1 |= 0x0; @@ -4021,7 +4035,8 @@ static void gfx_v8_0_init_pg(struct amdgpu_device *adev) cz_enable_cp_power_gating(adev, true); else cz_enable_cp_power_gating(adev, false); - } else if (adev->asic_type == CHIP_POLARIS11) { + } else if ((adev->asic_type == CHIP_POLARIS11) || + (adev->asic_type == CHIP_POLARIS12)) { gfx_v8_0_init_csb(adev); gfx_v8_0_init_save_restore_list(adev); gfx_v8_0_enable_save_restore_machine(adev); @@ -4095,7 +4110,8 @@ static int gfx_v8_0_rlc_resume(struct amdgpu_device *adev) RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK); WREG32(mmRLC_CGCG_CGLS_CTRL, tmp); if (adev->asic_type == CHIP_POLARIS11 || - adev->asic_type == CHIP_POLARIS10) { + adev->asic_type == CHIP_POLARIS10 || + adev->asic_type == CHIP_POLARIS12) { tmp = RREG32(mmRLC_CGCG_CGLS_CTRL_3D); tmp &= ~0x3; WREG32(mmRLC_CGCG_CGLS_CTRL_3D, tmp); @@ -4283,6 +4299,7 @@ static int gfx_v8_0_cp_gfx_start(struct amdgpu_device *adev) amdgpu_ring_write(ring, 0x0000002A); break; case CHIP_POLARIS11: + case CHIP_POLARIS12: amdgpu_ring_write(ring, 0x16000012); amdgpu_ring_write(ring, 0x00000000); break; @@ -4664,7 +4681,8 @@ static int gfx_v8_0_cp_compute_resume(struct amdgpu_device *adev) (adev->asic_type == CHIP_FIJI) || (adev->asic_type == CHIP_STONEY) || (adev->asic_type == CHIP_POLARIS11) || - (adev->asic_type == CHIP_POLARIS10)) { + (adev->asic_type == CHIP_POLARIS10) || + (adev->asic_type == CHIP_POLARIS12)) { WREG32(mmCP_MEC_DOORBELL_RANGE_LOWER, AMDGPU_DOORBELL_KIQ << 2); WREG32(mmCP_MEC_DOORBELL_RANGE_UPPER, @@ -4700,7 +4718,8 @@ static int gfx_v8_0_cp_compute_resume(struct amdgpu_device *adev) mqd->cp_hqd_persistent_state = tmp; if (adev->asic_type == CHIP_STONEY || adev->asic_type == CHIP_POLARIS11 || - adev->asic_type == CHIP_POLARIS10) { + adev->asic_type == CHIP_POLARIS10 || + adev->asic_type == CHIP_POLARIS12) { tmp = RREG32(mmCP_ME1_PIPE3_INT_CNTL); tmp = REG_SET_FIELD(tmp, CP_ME1_PIPE3_INT_CNTL, GENERIC2_INT_ENABLE, 1); WREG32(mmCP_ME1_PIPE3_INT_CNTL, tmp); @@ -5279,7 +5298,8 @@ static int gfx_v8_0_late_init(void *handle) static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *adev, bool enable) { - if (adev->asic_type == CHIP_POLARIS11) + if ((adev->asic_type == CHIP_POLARIS11) || + (adev->asic_type == CHIP_POLARIS12)) /* Send msg to SMU via Powerplay */ amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_SMC, @@ -5353,6 +5373,7 @@ static int gfx_v8_0_set_powergating_state(void *handle, gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, false); break; case CHIP_POLARIS11: + case CHIP_POLARIS12: if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG) && enable) gfx_v8_0_enable_gfx_static_mg_power_gating(adev, true); else diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index 45a573e63d4a..e2b0b1646f99 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -44,6 +44,7 @@ MODULE_FIRMWARE("radeon/tahiti_mc.bin"); MODULE_FIRMWARE("radeon/pitcairn_mc.bin"); MODULE_FIRMWARE("radeon/verde_mc.bin"); MODULE_FIRMWARE("radeon/oland_mc.bin"); +MODULE_FIRMWARE("radeon/si58_mc.bin"); #define MC_SEQ_MISC0__MT__MASK 0xf0000000 #define MC_SEQ_MISC0__MT__GDDR1 0x10000000 @@ -113,6 +114,7 @@ static int gmc_v6_0_init_microcode(struct amdgpu_device *adev) const char *chip_name; char fw_name[30]; int err; + bool is_58_fw = false; DRM_DEBUG("\n"); @@ -135,7 +137,14 @@ static int gmc_v6_0_init_microcode(struct amdgpu_device *adev) default: BUG(); } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + /* this memory configuration requires special firmware */ + if (((RREG32(mmMC_SEQ_MISC0) & 0xff000000) >> 24) == 0x58) + is_58_fw = true; + + if (is_58_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/si58_mc.bin"); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); err = request_firmware(&adev->mc.fw, fw_name, adev->dev); if (err) goto out; @@ -463,19 +472,11 @@ static int gmc_v6_0_gart_enable(struct amdgpu_device *adev) WREG32(mmVM_CONTEXT1_CNTL, VM_CONTEXT1_CNTL__ENABLE_CONTEXT_MASK | (1UL << VM_CONTEXT1_CNTL__PAGE_TABLE_DEPTH__SHIFT) | - ((amdgpu_vm_block_size - 9) << VM_CONTEXT1_CNTL__PAGE_TABLE_BLOCK_SIZE__SHIFT) | - VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK); + ((amdgpu_vm_block_size - 9) << VM_CONTEXT1_CNTL__PAGE_TABLE_BLOCK_SIZE__SHIFT)); + if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) + gmc_v6_0_set_fault_enable_default(adev, false); + else + gmc_v6_0_set_fault_enable_default(adev, true); gmc_v6_0_gart_flush_gpu_tlb(adev, 0); dev_info(adev->dev, "PCIE GART of %uM enabled (table at 0x%016llX).\n", @@ -754,7 +755,10 @@ static int gmc_v6_0_late_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0); + if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS) + return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0); + else + return 0; } static int gmc_v6_0_sw_init(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 0daac3a5be79..476bc9f1954b 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -46,6 +46,7 @@ static int gmc_v8_0_wait_for_idle(void *handle); MODULE_FIRMWARE("amdgpu/tonga_mc.bin"); MODULE_FIRMWARE("amdgpu/polaris11_mc.bin"); MODULE_FIRMWARE("amdgpu/polaris10_mc.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_mc.bin"); static const u32 golden_settings_tonga_a11[] = { @@ -130,6 +131,7 @@ static void gmc_v8_0_init_golden_registers(struct amdgpu_device *adev) (const u32)ARRAY_SIZE(golden_settings_tonga_a11)); break; case CHIP_POLARIS11: + case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, golden_settings_polaris11_a11, (const u32)ARRAY_SIZE(golden_settings_polaris11_a11)); @@ -225,6 +227,9 @@ static int gmc_v8_0_init_microcode(struct amdgpu_device *adev) case CHIP_POLARIS10: chip_name = "polaris10"; break; + case CHIP_POLARIS12: + chip_name = "polaris12"; + break; case CHIP_FIJI: case CHIP_CARRIZO: case CHIP_STONEY: diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index 1170a64a3184..034ace79ed49 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -60,6 +60,8 @@ MODULE_FIRMWARE("amdgpu/polaris10_sdma.bin"); MODULE_FIRMWARE("amdgpu/polaris10_sdma1.bin"); MODULE_FIRMWARE("amdgpu/polaris11_sdma.bin"); MODULE_FIRMWARE("amdgpu/polaris11_sdma1.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_sdma.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_sdma1.bin"); static const u32 sdma_offsets[SDMA_MAX_INSTANCE] = @@ -206,6 +208,7 @@ static void sdma_v3_0_init_golden_registers(struct amdgpu_device *adev) (const u32)ARRAY_SIZE(golden_settings_tonga_a11)); break; case CHIP_POLARIS11: + case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, golden_settings_polaris11_a11, (const u32)ARRAY_SIZE(golden_settings_polaris11_a11)); @@ -278,6 +281,9 @@ static int sdma_v3_0_init_microcode(struct amdgpu_device *adev) case CHIP_POLARIS10: chip_name = "polaris10"; break; + case CHIP_POLARIS12: + chip_name = "polaris12"; + break; case CHIP_CARRIZO: chip_name = "carrizo"; break; diff --git a/drivers/gpu/drm/amd/amdgpu/si_dpm.c b/drivers/gpu/drm/amd/amdgpu/si_dpm.c index 6c65a1a2de79..6e150db8f380 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dpm.c @@ -56,7 +56,6 @@ #define BIOS_SCRATCH_4 0x5cd MODULE_FIRMWARE("radeon/tahiti_smc.bin"); -MODULE_FIRMWARE("radeon/tahiti_k_smc.bin"); MODULE_FIRMWARE("radeon/pitcairn_smc.bin"); MODULE_FIRMWARE("radeon/pitcairn_k_smc.bin"); MODULE_FIRMWARE("radeon/verde_smc.bin"); @@ -65,6 +64,7 @@ MODULE_FIRMWARE("radeon/oland_smc.bin"); MODULE_FIRMWARE("radeon/oland_k_smc.bin"); MODULE_FIRMWARE("radeon/hainan_smc.bin"); MODULE_FIRMWARE("radeon/hainan_k_smc.bin"); +MODULE_FIRMWARE("radeon/banks_k_2_smc.bin"); union power_info { struct _ATOM_POWERPLAY_INFO info; @@ -3488,30 +3488,6 @@ static void si_apply_state_adjust_rules(struct amdgpu_device *adev, (adev->pdev->device == 0x6817) || (adev->pdev->device == 0x6806)) max_mclk = 120000; - } else if (adev->asic_type == CHIP_VERDE) { - if ((adev->pdev->revision == 0x81) || - (adev->pdev->revision == 0x83) || - (adev->pdev->revision == 0x87) || - (adev->pdev->device == 0x6820) || - (adev->pdev->device == 0x6821) || - (adev->pdev->device == 0x6822) || - (adev->pdev->device == 0x6823) || - (adev->pdev->device == 0x682A) || - (adev->pdev->device == 0x682B)) { - max_sclk = 75000; - max_mclk = 80000; - } - } else if (adev->asic_type == CHIP_OLAND) { - if ((adev->pdev->revision == 0xC7) || - (adev->pdev->revision == 0x80) || - (adev->pdev->revision == 0x81) || - (adev->pdev->revision == 0x83) || - (adev->pdev->revision == 0x87) || - (adev->pdev->device == 0x6604) || - (adev->pdev->device == 0x6605)) { - max_sclk = 75000; - max_mclk = 80000; - } } else if (adev->asic_type == CHIP_HAINAN) { if ((adev->pdev->revision == 0x81) || (adev->pdev->revision == 0x83) || @@ -3520,7 +3496,6 @@ static void si_apply_state_adjust_rules(struct amdgpu_device *adev, (adev->pdev->device == 0x6665) || (adev->pdev->device == 0x6667)) { max_sclk = 75000; - max_mclk = 80000; } } /* Apply dpm quirks */ @@ -7687,50 +7662,51 @@ static int si_dpm_init_microcode(struct amdgpu_device *adev) chip_name = "tahiti"; break; case CHIP_PITCAIRN: - if ((adev->pdev->revision == 0x81) || - (adev->pdev->device == 0x6810) || - (adev->pdev->device == 0x6811) || - (adev->pdev->device == 0x6816) || - (adev->pdev->device == 0x6817) || - (adev->pdev->device == 0x6806)) + if ((adev->pdev->revision == 0x81) && + ((adev->pdev->device == 0x6810) || + (adev->pdev->device == 0x6811))) chip_name = "pitcairn_k"; else chip_name = "pitcairn"; break; case CHIP_VERDE: - if ((adev->pdev->revision == 0x81) || - (adev->pdev->revision == 0x83) || - (adev->pdev->revision == 0x87) || - (adev->pdev->device == 0x6820) || - (adev->pdev->device == 0x6821) || - (adev->pdev->device == 0x6822) || - (adev->pdev->device == 0x6823) || - (adev->pdev->device == 0x682A) || - (adev->pdev->device == 0x682B)) + if (((adev->pdev->device == 0x6820) && + ((adev->pdev->revision == 0x81) || + (adev->pdev->revision == 0x83))) || + ((adev->pdev->device == 0x6821) && + ((adev->pdev->revision == 0x83) || + (adev->pdev->revision == 0x87))) || + ((adev->pdev->revision == 0x87) && + ((adev->pdev->device == 0x6823) || + (adev->pdev->device == 0x682b)))) chip_name = "verde_k"; else chip_name = "verde"; break; case CHIP_OLAND: - if ((adev->pdev->revision == 0xC7) || - (adev->pdev->revision == 0x80) || - (adev->pdev->revision == 0x81) || - (adev->pdev->revision == 0x83) || - (adev->pdev->revision == 0x87) || - (adev->pdev->device == 0x6604) || - (adev->pdev->device == 0x6605)) + if (((adev->pdev->revision == 0x81) && + ((adev->pdev->device == 0x6600) || + (adev->pdev->device == 0x6604) || + (adev->pdev->device == 0x6605) || + (adev->pdev->device == 0x6610))) || + ((adev->pdev->revision == 0x83) && + (adev->pdev->device == 0x6610))) chip_name = "oland_k"; else chip_name = "oland"; break; case CHIP_HAINAN: - if ((adev->pdev->revision == 0x81) || - (adev->pdev->revision == 0x83) || - (adev->pdev->revision == 0xC3) || - (adev->pdev->device == 0x6664) || - (adev->pdev->device == 0x6665) || - (adev->pdev->device == 0x6667)) + if (((adev->pdev->revision == 0x81) && + (adev->pdev->device == 0x6660)) || + ((adev->pdev->revision == 0x83) && + ((adev->pdev->device == 0x6660) || + (adev->pdev->device == 0x6663) || + (adev->pdev->device == 0x6665) || + (adev->pdev->device == 0x6667)))) chip_name = "hainan_k"; + else if ((adev->pdev->revision == 0xc3) && + (adev->pdev->device == 0x6665)) + chip_name = "banks_k_2"; else chip_name = "hainan"; break; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index 96444e4d862a..7fb9137dd89b 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -40,13 +40,14 @@ #include "smu/smu_7_0_1_sh_mask.h" static void uvd_v4_2_mc_resume(struct amdgpu_device *adev); -static void uvd_v4_2_init_cg(struct amdgpu_device *adev); static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev); static void uvd_v4_2_set_irq_funcs(struct amdgpu_device *adev); static int uvd_v4_2_start(struct amdgpu_device *adev); static void uvd_v4_2_stop(struct amdgpu_device *adev); static int uvd_v4_2_set_clockgating_state(void *handle, enum amd_clockgating_state state); +static void uvd_v4_2_set_dcm(struct amdgpu_device *adev, + bool sw_mode); /** * uvd_v4_2_ring_get_rptr - get read pointer * @@ -140,7 +141,8 @@ static int uvd_v4_2_sw_fini(void *handle) return r; } - +static void uvd_v4_2_enable_mgcg(struct amdgpu_device *adev, + bool enable); /** * uvd_v4_2_hw_init - start and test UVD block * @@ -155,8 +157,7 @@ static int uvd_v4_2_hw_init(void *handle) uint32_t tmp; int r; - uvd_v4_2_init_cg(adev); - uvd_v4_2_set_clockgating_state(adev, AMD_CG_STATE_GATE); + uvd_v4_2_enable_mgcg(adev, true); amdgpu_asic_set_uvd_clocks(adev, 10000, 10000); r = uvd_v4_2_start(adev); if (r) @@ -266,11 +267,13 @@ static int uvd_v4_2_start(struct amdgpu_device *adev) struct amdgpu_ring *ring = &adev->uvd.ring; uint32_t rb_bufsz; int i, j, r; - /* disable byte swapping */ u32 lmi_swap_cntl = 0; u32 mp_swap_cntl = 0; + WREG32(mmUVD_CGC_GATE, 0); + uvd_v4_2_set_dcm(adev, true); + uvd_v4_2_mc_resume(adev); /* disable interupt */ @@ -406,6 +409,8 @@ static void uvd_v4_2_stop(struct amdgpu_device *adev) /* Unstall UMC and register bus */ WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8)); + + uvd_v4_2_set_dcm(adev, false); } /** @@ -619,19 +624,6 @@ static void uvd_v4_2_set_dcm(struct amdgpu_device *adev, WREG32_UVD_CTX(ixUVD_CGC_CTRL2, tmp2); } -static void uvd_v4_2_init_cg(struct amdgpu_device *adev) -{ - bool hw_mode = true; - - if (hw_mode) { - uvd_v4_2_set_dcm(adev, false); - } else { - u32 tmp = RREG32(mmUVD_CGC_CTRL); - tmp &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK; - WREG32(mmUVD_CGC_CTRL, tmp); - } -} - static bool uvd_v4_2_is_idle(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -685,17 +677,6 @@ static int uvd_v4_2_process_interrupt(struct amdgpu_device *adev, static int uvd_v4_2_set_clockgating_state(void *handle, enum amd_clockgating_state state) { - bool gate = false; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG)) - return 0; - - if (state == AMD_CG_STATE_GATE) - gate = true; - - uvd_v4_2_enable_mgcg(adev, gate); - return 0; } @@ -711,9 +692,6 @@ static int uvd_v4_2_set_powergating_state(void *handle, */ struct amdgpu_device *adev = (struct amdgpu_device *)handle; - if (!(adev->pg_flags & AMD_PG_SUPPORT_UVD)) - return 0; - if (state == AMD_PG_STATE_GATE) { uvd_v4_2_stop(adev); return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index a79e283590fb..6de6becce745 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -791,15 +791,10 @@ static int uvd_v5_0_set_clockgating_state(void *handle, { struct amdgpu_device *adev = (struct amdgpu_device *)handle; bool enable = (state == AMD_CG_STATE_GATE) ? true : false; - static int curstate = -1; if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG)) return 0; - if (curstate == state) - return 0; - - curstate = state; if (enable) { /* wait for STATUS to clear */ if (uvd_v5_0_wait_for_idle(handle)) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index 6b3293a1c7b8..37ca685e5a9a 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -43,9 +43,13 @@ #define GRBM_GFX_INDEX__VCE_INSTANCE__SHIFT 0x04 #define GRBM_GFX_INDEX__VCE_INSTANCE_MASK 0x10 +#define GRBM_GFX_INDEX__VCE_ALL_PIPE 0x07 + #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR0 0x8616 #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR1 0x8617 #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR2 0x8618 +#define mmGRBM_GFX_INDEX_DEFAULT 0xE0000000 + #define VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK 0x02 #define VCE_V3_0_FW_SIZE (384 * 1024) @@ -54,6 +58,9 @@ #define FW_52_8_3 ((52 << 24) | (8 << 16) | (3 << 8)) +#define GET_VCE_INSTANCE(i) ((i) << GRBM_GFX_INDEX__VCE_INSTANCE__SHIFT \ + | GRBM_GFX_INDEX__VCE_ALL_PIPE) + static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx); static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev); static void vce_v3_0_set_irq_funcs(struct amdgpu_device *adev); @@ -175,7 +182,7 @@ static void vce_v3_0_set_vce_sw_clock_gating(struct amdgpu_device *adev, WREG32(mmVCE_UENC_CLOCK_GATING_2, data); data = RREG32(mmVCE_UENC_REG_CLOCK_GATING); - data &= ~0xffc00000; + data &= ~0x3ff; WREG32(mmVCE_UENC_REG_CLOCK_GATING, data); data = RREG32(mmVCE_UENC_DMA_DCLK_CTRL); @@ -249,7 +256,7 @@ static int vce_v3_0_start(struct amdgpu_device *adev) if (adev->vce.harvest_config & (1 << idx)) continue; - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, idx); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(idx)); vce_v3_0_mc_resume(adev, idx); WREG32_FIELD(VCE_STATUS, JOB_BUSY, 1); @@ -273,7 +280,7 @@ static int vce_v3_0_start(struct amdgpu_device *adev) } } - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, 0); + WREG32(mmGRBM_GFX_INDEX, mmGRBM_GFX_INDEX_DEFAULT); mutex_unlock(&adev->grbm_idx_mutex); return 0; @@ -288,7 +295,7 @@ static int vce_v3_0_stop(struct amdgpu_device *adev) if (adev->vce.harvest_config & (1 << idx)) continue; - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, idx); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(idx)); if (adev->asic_type >= CHIP_STONEY) WREG32_P(mmVCE_VCPU_CNTL, 0, ~0x200001); @@ -306,7 +313,7 @@ static int vce_v3_0_stop(struct amdgpu_device *adev) vce_v3_0_set_vce_sw_clock_gating(adev, false); } - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, 0); + WREG32(mmGRBM_GFX_INDEX, mmGRBM_GFX_INDEX_DEFAULT); mutex_unlock(&adev->grbm_idx_mutex); return 0; @@ -320,11 +327,12 @@ static unsigned vce_v3_0_get_harvest_config(struct amdgpu_device *adev) { u32 tmp; - /* Fiji, Stoney, Polaris10, Polaris11 are single pipe */ + /* Fiji, Stoney, Polaris10, Polaris11, Polaris12 are single pipe */ if ((adev->asic_type == CHIP_FIJI) || (adev->asic_type == CHIP_STONEY) || (adev->asic_type == CHIP_POLARIS10) || - (adev->asic_type == CHIP_POLARIS11)) + (adev->asic_type == CHIP_POLARIS11) || + (adev->asic_type == CHIP_POLARIS12)) return AMDGPU_VCE_HARVEST_VCE1; /* Tonga and CZ are dual or single pipe */ @@ -585,17 +593,17 @@ static bool vce_v3_0_check_soft_reset(void *handle) * VCE team suggest use bit 3--bit 6 for busy status check */ mutex_lock(&adev->grbm_idx_mutex); - WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(0)); if (RREG32(mmVCE_STATUS) & AMDGPU_VCE_STATUS_BUSY_MASK) { srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE0, 1); srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE1, 1); } - WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0x10); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(1)); if (RREG32(mmVCE_STATUS) & AMDGPU_VCE_STATUS_BUSY_MASK) { srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE0, 1); srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE1, 1); } - WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(0)); mutex_unlock(&adev->grbm_idx_mutex); if (srbm_soft_reset) { @@ -733,7 +741,7 @@ static int vce_v3_0_set_clockgating_state(void *handle, if (adev->vce.harvest_config & (1 << i)) continue; - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, i); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(i)); if (enable) { /* initialize VCE_CLOCK_GATING_A: Clock ON/OFF delay */ @@ -752,7 +760,7 @@ static int vce_v3_0_set_clockgating_state(void *handle, vce_v3_0_set_vce_sw_clock_gating(adev, enable); } - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, 0); + WREG32(mmGRBM_GFX_INDEX, mmGRBM_GFX_INDEX_DEFAULT); mutex_unlock(&adev->grbm_idx_mutex); return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index bf088d6d9bf1..c2ac54f11341 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -88,6 +88,7 @@ MODULE_FIRMWARE("amdgpu/polaris10_smc.bin"); MODULE_FIRMWARE("amdgpu/polaris10_smc_sk.bin"); MODULE_FIRMWARE("amdgpu/polaris11_smc.bin"); MODULE_FIRMWARE("amdgpu/polaris11_smc_sk.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_smc.bin"); /* * Indirect registers accessor @@ -312,6 +313,7 @@ static void vi_init_golden_registers(struct amdgpu_device *adev) break; case CHIP_POLARIS11: case CHIP_POLARIS10: + case CHIP_POLARIS12: default: break; } @@ -671,6 +673,7 @@ static int vi_read_register(struct amdgpu_device *adev, u32 se_num, case CHIP_TONGA: case CHIP_POLARIS11: case CHIP_POLARIS10: + case CHIP_POLARIS12: case CHIP_CARRIZO: case CHIP_STONEY: asic_register_table = cz_allowed_read_registers; @@ -994,6 +997,11 @@ static int vi_common_early_init(void *handle) adev->pg_flags = 0; adev->external_rev_id = adev->rev_id + 0x50; break; + case CHIP_POLARIS12: + adev->cg_flags = AMD_CG_SUPPORT_UVD_MGCG; + adev->pg_flags = 0; + adev->external_rev_id = adev->rev_id + 0x64; + break; case CHIP_CARRIZO: adev->cg_flags = AMD_CG_SUPPORT_UVD_MGCG | AMD_CG_SUPPORT_GFX_MGCG | @@ -1346,6 +1354,7 @@ static int vi_common_set_clockgating_state(void *handle, case CHIP_TONGA: case CHIP_POLARIS10: case CHIP_POLARIS11: + case CHIP_POLARIS12: vi_common_set_clockgating_state_by_smu(adev, state); default: break; @@ -1429,6 +1438,7 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) break; case CHIP_POLARIS11: case CHIP_POLARIS10: + case CHIP_POLARIS12: amdgpu_ip_block_add(adev, &vi_common_ip_block); amdgpu_ip_block_add(adev, &gmc_v8_1_ip_block); amdgpu_ip_block_add(adev, &tonga_ih_ip_block); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c index ee3e04e10dae..6316aad43a73 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c @@ -486,7 +486,7 @@ static int kfd_ioctl_dbg_register(struct file *filep, return status; } -static int kfd_ioctl_dbg_unrgesiter(struct file *filep, +static int kfd_ioctl_dbg_unregister(struct file *filep, struct kfd_process *p, void *data) { struct kfd_ioctl_dbg_unregister_args *args = data; @@ -498,7 +498,7 @@ static int kfd_ioctl_dbg_unrgesiter(struct file *filep, return -EINVAL; if (dev->device_info->asic_family == CHIP_CARRIZO) { - pr_debug("kfd_ioctl_dbg_unrgesiter not supported on CZ\n"); + pr_debug("kfd_ioctl_dbg_unregister not supported on CZ\n"); return -EINVAL; } @@ -892,7 +892,7 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = { kfd_ioctl_dbg_register, 0), AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_UNREGISTER, - kfd_ioctl_dbg_unrgesiter, 0), + kfd_ioctl_dbg_unregister, 0), AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_ADDRESS_WATCH, kfd_ioctl_dbg_address_watch, 0), diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_events.c index a6a4b2b1c0d9..6a3470f84998 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_events.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.c @@ -739,8 +739,10 @@ int kfd_wait_on_events(struct kfd_process *p, struct kfd_event_data event_data; if (copy_from_user(&event_data, &events[i], - sizeof(struct kfd_event_data))) + sizeof(struct kfd_event_data))) { + ret = -EFAULT; goto fail; + } ret = init_event_waiter(p, &event_waiters[i], event_data.event_id, i); diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index c02469ada9f1..85f358764bbc 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -23,7 +23,7 @@ #ifndef __AMD_SHARED_H__ #define __AMD_SHARED_H__ -#define AMD_MAX_USEC_TIMEOUT 100000 /* 100 ms */ +#define AMD_MAX_USEC_TIMEOUT 200000 /* 200 ms */ /* * Supported ASIC types @@ -46,6 +46,7 @@ enum amd_asic_type { CHIP_STONEY, CHIP_POLARIS10, CHIP_POLARIS11, + CHIP_POLARIS12, CHIP_LAST, }; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c index b0c63c5f54c9..6bb79c94cb9f 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c @@ -200,7 +200,7 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate) cgs_set_clockgating_state( hwmgr->device, AMD_IP_BLOCK_TYPE_VCE, - AMD_CG_STATE_UNGATE); + AMD_CG_STATE_GATE); cgs_set_powergating_state( hwmgr->device, AMD_IP_BLOCK_TYPE_VCE, @@ -218,7 +218,7 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate) cgs_set_clockgating_state( hwmgr->device, AMD_IP_BLOCK_TYPE_VCE, - AMD_PG_STATE_GATE); + AMD_PG_STATE_UNGATE); cz_dpm_update_vce_dpm(hwmgr); cz_enable_disable_vce_dpm(hwmgr, true); return 0; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c index 4b14f259a147..0fb4e8c8f5e1 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c @@ -1402,14 +1402,22 @@ int cz_dpm_update_vce_dpm(struct pp_hwmgr *hwmgr) cz_hwmgr->vce_dpm.hard_min_clk, PPSMC_MSG_SetEclkHardMin)); } else { - /*EPR# 419220 -HW limitation to to */ - cz_hwmgr->vce_dpm.hard_min_clk = hwmgr->vce_arbiter.ecclk; - smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, - PPSMC_MSG_SetEclkHardMin, - cz_get_eclk_level(hwmgr, - cz_hwmgr->vce_dpm.hard_min_clk, - PPSMC_MSG_SetEclkHardMin)); - + /*Program HardMin based on the vce_arbiter.ecclk */ + if (hwmgr->vce_arbiter.ecclk == 0) { + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetEclkHardMin, 0); + /* disable ECLK DPM 0. Otherwise VCE could hang if + * switching SCLK from DPM 0 to 6/7 */ + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetEclkSoftMin, 1); + } else { + cz_hwmgr->vce_dpm.hard_min_clk = hwmgr->vce_arbiter.ecclk; + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetEclkHardMin, + cz_get_eclk_level(hwmgr, + cz_hwmgr->vce_dpm.hard_min_clk, + PPSMC_MSG_SetEclkHardMin)); + } } return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c index dc6700aee18f..b03606405a53 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c @@ -95,6 +95,7 @@ int hwmgr_init(struct amd_pp_init *pp_init, struct pp_instance *handle) break; case CHIP_POLARIS11: case CHIP_POLARIS10: + case CHIP_POLARIS12: polaris_set_asic_special_caps(hwmgr); hwmgr->feature_mask &= ~(PP_UVD_HANDSHAKE_MASK); break; @@ -745,7 +746,7 @@ int polaris_set_asic_special_caps(struct pp_hwmgr *hwmgr) phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TablelessHardwareInterface); - if (hwmgr->chip_id == CHIP_POLARIS11) + if ((hwmgr->chip_id == CHIP_POLARIS11) || (hwmgr->chip_id == CHIP_POLARIS12)) phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_SPLLShutdownSupport); return 0; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_powertune.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_powertune.c index 26477f0f09dc..6cd1287a7a8f 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_powertune.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_powertune.c @@ -521,7 +521,7 @@ int smu7_enable_didt_config(struct pp_hwmgr *hwmgr) PP_ASSERT_WITH_CODE((result == 0), "DIDT Config failed.", return result); result = smu7_program_pt_config_registers(hwmgr, DIDTConfig_Polaris10); PP_ASSERT_WITH_CODE((result == 0), "DIDT Config failed.", return result); - } else if (hwmgr->chip_id == CHIP_POLARIS11) { + } else if ((hwmgr->chip_id == CHIP_POLARIS11) || (hwmgr->chip_id == CHIP_POLARIS12)) { result = smu7_program_pt_config_registers(hwmgr, GCCACConfig_Polaris11); PP_ASSERT_WITH_CODE((result == 0), "DIDT Config failed.", return result); result = smu7_program_pt_config_registers(hwmgr, DIDTConfig_Polaris11); diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c index e5812aa456f3..6e618aa20719 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c @@ -65,6 +65,7 @@ int smum_init(struct amd_pp_init *pp_init, struct pp_instance *handle) break; case CHIP_POLARIS11: case CHIP_POLARIS10: + case CHIP_POLARIS12: polaris10_smum_init(smumgr); break; default: diff --git a/drivers/gpu/drm/arm/malidp_drv.c b/drivers/gpu/drm/arm/malidp_drv.c index 32f746e31379..99fb0ab39191 100644 --- a/drivers/gpu/drm/arm/malidp_drv.c +++ b/drivers/gpu/drm/arm/malidp_drv.c @@ -22,7 +22,6 @@ #include <drm/drm_atomic_helper.h> #include <drm/drm_crtc.h> #include <drm/drm_crtc_helper.h> -#include <drm/drm_fb_helper.h> #include <drm/drm_fb_cma_helper.h> #include <drm/drm_gem_cma_helper.h> #include <drm/drm_of.h> @@ -256,6 +255,60 @@ static const struct of_device_id malidp_drm_of_match[] = { }; MODULE_DEVICE_TABLE(of, malidp_drm_of_match); +static bool malidp_is_compatible_hw_id(struct malidp_hw_device *hwdev, + const struct of_device_id *dev_id) +{ + u32 core_id; + const char *compatstr_dp500 = "arm,mali-dp500"; + bool is_dp500; + bool dt_is_dp500; + + /* + * The DP500 CORE_ID register is in a different location, so check it + * first. If the product id field matches, then this is DP500, otherwise + * check the DP550/650 CORE_ID register. + */ + core_id = malidp_hw_read(hwdev, MALIDP500_DC_BASE + MALIDP_DE_CORE_ID); + /* Offset 0x18 will never read 0x500 on products other than DP500. */ + is_dp500 = (MALIDP_PRODUCT_ID(core_id) == 0x500); + dt_is_dp500 = strnstr(dev_id->compatible, compatstr_dp500, + sizeof(dev_id->compatible)) != NULL; + if (is_dp500 != dt_is_dp500) { + DRM_ERROR("Device-tree expects %s, but hardware %s DP500.\n", + dev_id->compatible, is_dp500 ? "is" : "is not"); + return false; + } else if (!dt_is_dp500) { + u16 product_id; + char buf[32]; + + core_id = malidp_hw_read(hwdev, + MALIDP550_DC_BASE + MALIDP_DE_CORE_ID); + product_id = MALIDP_PRODUCT_ID(core_id); + snprintf(buf, sizeof(buf), "arm,mali-dp%X", product_id); + if (!strnstr(dev_id->compatible, buf, + sizeof(dev_id->compatible))) { + DRM_ERROR("Device-tree expects %s, but hardware is DP%03X.\n", + dev_id->compatible, product_id); + return false; + } + } + return true; +} + +static bool malidp_has_sufficient_address_space(const struct resource *res, + const struct of_device_id *dev_id) +{ + resource_size_t res_size = resource_size(res); + const char *compatstr_dp500 = "arm,mali-dp500"; + + if (!strnstr(dev_id->compatible, compatstr_dp500, + sizeof(dev_id->compatible))) + return res_size >= MALIDP550_ADDR_SPACE_SIZE; + else if (res_size < MALIDP500_ADDR_SPACE_SIZE) + return false; + return true; +} + #define MAX_OUTPUT_CHANNELS 3 static int malidp_bind(struct device *dev) @@ -266,6 +319,7 @@ static int malidp_bind(struct device *dev) struct malidp_drm *malidp; struct malidp_hw_device *hwdev; struct platform_device *pdev = to_platform_device(dev); + struct of_device_id const *dev_id; /* number of lines for the R, G and B output */ u8 output_width[MAX_OUTPUT_CHANNELS]; int ret = 0, i; @@ -286,7 +340,6 @@ static int malidp_bind(struct device *dev) memcpy(hwdev, of_device_get_match_data(dev), sizeof(*hwdev)); malidp->dev = hwdev; - INIT_LIST_HEAD(&malidp->event_list); res = platform_get_resource(pdev, IORESOURCE_MEM, 0); hwdev->regs = devm_ioremap_resource(dev, res); @@ -329,6 +382,23 @@ static int malidp_bind(struct device *dev) clk_prepare_enable(hwdev->aclk); clk_prepare_enable(hwdev->mclk); + dev_id = of_match_device(malidp_drm_of_match, dev); + if (!dev_id) { + ret = -EINVAL; + goto query_hw_fail; + } + + if (!malidp_has_sufficient_address_space(res, dev_id)) { + DRM_ERROR("Insufficient address space in device-tree.\n"); + ret = -EINVAL; + goto query_hw_fail; + } + + if (!malidp_is_compatible_hw_id(hwdev, dev_id)) { + ret = -EINVAL; + goto query_hw_fail; + } + ret = hwdev->query_hw(hwdev); if (ret) { DRM_ERROR("Invalid HW configuration\n"); diff --git a/drivers/gpu/drm/arm/malidp_drv.h b/drivers/gpu/drm/arm/malidp_drv.h index 9fc8a2e405e4..dbc617c6e4ef 100644 --- a/drivers/gpu/drm/arm/malidp_drv.h +++ b/drivers/gpu/drm/arm/malidp_drv.h @@ -15,12 +15,12 @@ #include <linux/mutex.h> #include <linux/wait.h> +#include <drm/drmP.h> #include "malidp_hw.h" struct malidp_drm { struct malidp_hw_device *dev; struct drm_fbdev_cma *fbdev; - struct list_head event_list; struct drm_crtc crtc; wait_queue_head_t wq; atomic_t config_valid; diff --git a/drivers/gpu/drm/arm/malidp_hw.c b/drivers/gpu/drm/arm/malidp_hw.c index 4bdf531f7844..488aedf5b58d 100644 --- a/drivers/gpu/drm/arm/malidp_hw.c +++ b/drivers/gpu/drm/arm/malidp_hw.c @@ -21,7 +21,7 @@ #include "malidp_drv.h" #include "malidp_hw.h" -static const struct malidp_input_format malidp500_de_formats[] = { +static const struct malidp_format_id malidp500_de_formats[] = { /* fourcc, layers supporting the format, internal id */ { DRM_FORMAT_ARGB2101010, DE_VIDEO1 | DE_GRAPHICS1 | DE_GRAPHICS2, 0 }, { DRM_FORMAT_ABGR2101010, DE_VIDEO1 | DE_GRAPHICS1 | DE_GRAPHICS2, 1 }, @@ -69,21 +69,21 @@ static const struct malidp_input_format malidp500_de_formats[] = { { DRM_FORMAT_NV12, DE_VIDEO1 | DE_VIDEO2, MALIDP_ID(5, 6) }, \ { DRM_FORMAT_YUV420, DE_VIDEO1 | DE_VIDEO2, MALIDP_ID(5, 7) } -static const struct malidp_input_format malidp550_de_formats[] = { +static const struct malidp_format_id malidp550_de_formats[] = { MALIDP_COMMON_FORMATS, }; static const struct malidp_layer malidp500_layers[] = { - { DE_VIDEO1, MALIDP500_DE_LV_BASE, MALIDP500_DE_LV_PTR_BASE }, - { DE_GRAPHICS1, MALIDP500_DE_LG1_BASE, MALIDP500_DE_LG1_PTR_BASE }, - { DE_GRAPHICS2, MALIDP500_DE_LG2_BASE, MALIDP500_DE_LG2_PTR_BASE }, + { DE_VIDEO1, MALIDP500_DE_LV_BASE, MALIDP500_DE_LV_PTR_BASE, MALIDP_DE_LV_STRIDE0 }, + { DE_GRAPHICS1, MALIDP500_DE_LG1_BASE, MALIDP500_DE_LG1_PTR_BASE, MALIDP_DE_LG_STRIDE }, + { DE_GRAPHICS2, MALIDP500_DE_LG2_BASE, MALIDP500_DE_LG2_PTR_BASE, MALIDP_DE_LG_STRIDE }, }; static const struct malidp_layer malidp550_layers[] = { - { DE_VIDEO1, MALIDP550_DE_LV1_BASE, MALIDP550_DE_LV1_PTR_BASE }, - { DE_GRAPHICS1, MALIDP550_DE_LG_BASE, MALIDP550_DE_LG_PTR_BASE }, - { DE_VIDEO2, MALIDP550_DE_LV2_BASE, MALIDP550_DE_LV2_PTR_BASE }, - { DE_SMART, MALIDP550_DE_LS_BASE, MALIDP550_DE_LS_PTR_BASE }, + { DE_VIDEO1, MALIDP550_DE_LV1_BASE, MALIDP550_DE_LV1_PTR_BASE, MALIDP_DE_LV_STRIDE0 }, + { DE_GRAPHICS1, MALIDP550_DE_LG_BASE, MALIDP550_DE_LG_PTR_BASE, MALIDP_DE_LG_STRIDE }, + { DE_VIDEO2, MALIDP550_DE_LV2_BASE, MALIDP550_DE_LV2_PTR_BASE, MALIDP_DE_LV_STRIDE0 }, + { DE_SMART, MALIDP550_DE_LS_BASE, MALIDP550_DE_LS_PTR_BASE, 0 }, }; #define MALIDP_DE_DEFAULT_PREFETCH_START 5 @@ -436,8 +436,8 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = { .irq_mask = MALIDP500_DE_IRQ_CONF_VALID, .vsync_irq = MALIDP500_DE_IRQ_CONF_VALID, }, - .input_formats = malidp500_de_formats, - .n_input_formats = ARRAY_SIZE(malidp500_de_formats), + .pixel_formats = malidp500_de_formats, + .n_pixel_formats = ARRAY_SIZE(malidp500_de_formats), .bus_align_bytes = 8, }, .query_hw = malidp500_query_hw, @@ -447,6 +447,7 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = { .set_config_valid = malidp500_set_config_valid, .modeset = malidp500_modeset, .rotmem_required = malidp500_rotmem_required, + .features = MALIDP_DEVICE_LV_HAS_3_STRIDES, }, [MALIDP_550] = { .map = { @@ -469,8 +470,8 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = { .irq_mask = MALIDP550_DC_IRQ_CONF_VALID, .vsync_irq = MALIDP550_DC_IRQ_CONF_VALID, }, - .input_formats = malidp550_de_formats, - .n_input_formats = ARRAY_SIZE(malidp550_de_formats), + .pixel_formats = malidp550_de_formats, + .n_pixel_formats = ARRAY_SIZE(malidp550_de_formats), .bus_align_bytes = 8, }, .query_hw = malidp550_query_hw, @@ -480,6 +481,7 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = { .set_config_valid = malidp550_set_config_valid, .modeset = malidp550_modeset, .rotmem_required = malidp550_rotmem_required, + .features = 0, }, [MALIDP_650] = { .map = { @@ -503,8 +505,8 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = { .irq_mask = MALIDP550_DC_IRQ_CONF_VALID, .vsync_irq = MALIDP550_DC_IRQ_CONF_VALID, }, - .input_formats = malidp550_de_formats, - .n_input_formats = ARRAY_SIZE(malidp550_de_formats), + .pixel_formats = malidp550_de_formats, + .n_pixel_formats = ARRAY_SIZE(malidp550_de_formats), .bus_align_bytes = 16, }, .query_hw = malidp650_query_hw, @@ -514,6 +516,7 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = { .set_config_valid = malidp550_set_config_valid, .modeset = malidp550_modeset, .rotmem_required = malidp550_rotmem_required, + .features = 0, }, }; @@ -522,10 +525,10 @@ u8 malidp_hw_get_format_id(const struct malidp_hw_regmap *map, { unsigned int i; - for (i = 0; i < map->n_input_formats; i++) { - if (((map->input_formats[i].layer & layer_id) == layer_id) && - (map->input_formats[i].format == format)) - return map->input_formats[i].id; + for (i = 0; i < map->n_pixel_formats; i++) { + if (((map->pixel_formats[i].layer & layer_id) == layer_id) && + (map->pixel_formats[i].format == format)) + return map->pixel_formats[i].id; } return MALIDP_INVALID_FORMAT_ID; diff --git a/drivers/gpu/drm/arm/malidp_hw.h b/drivers/gpu/drm/arm/malidp_hw.h index 087e1202db3d..00974b59407d 100644 --- a/drivers/gpu/drm/arm/malidp_hw.h +++ b/drivers/gpu/drm/arm/malidp_hw.h @@ -35,7 +35,7 @@ enum { DE_SMART = BIT(4), }; -struct malidp_input_format { +struct malidp_format_id { u32 format; /* DRM fourcc */ u8 layer; /* bitmask of layers supporting it */ u8 id; /* used internally */ @@ -58,6 +58,7 @@ struct malidp_layer { u16 id; /* layer ID */ u16 base; /* address offset for the register bank */ u16 ptr; /* address offset for the pointer register */ + u16 stride_offset; /* Offset to the first stride register. */ }; /* regmap features */ @@ -85,14 +86,18 @@ struct malidp_hw_regmap { const struct malidp_irq_map se_irq_map; const struct malidp_irq_map dc_irq_map; - /* list of supported input formats for each layer */ - const struct malidp_input_format *input_formats; - const u8 n_input_formats; + /* list of supported pixel formats for each layer */ + const struct malidp_format_id *pixel_formats; + const u8 n_pixel_formats; /* pitch alignment requirement in bytes */ const u8 bus_align_bytes; }; +/* device features */ +/* Unlike DP550/650, DP500 has 3 stride registers in its video layer. */ +#define MALIDP_DEVICE_LV_HAS_3_STRIDES BIT(0) + struct malidp_hw_device { const struct malidp_hw_regmap map; void __iomem *regs; diff --git a/drivers/gpu/drm/arm/malidp_planes.c b/drivers/gpu/drm/arm/malidp_planes.c index eff2fe47e26a..414aada10fe5 100644 --- a/drivers/gpu/drm/arm/malidp_planes.c +++ b/drivers/gpu/drm/arm/malidp_planes.c @@ -11,6 +11,7 @@ */ #include <drm/drmP.h> +#include <drm/drm_atomic.h> #include <drm/drm_atomic_helper.h> #include <drm/drm_fb_cma_helper.h> #include <drm/drm_gem_cma_helper.h> @@ -36,7 +37,6 @@ #define LAYER_V_VAL(x) (((x) & 0x1fff) << 16) #define MALIDP_LAYER_COMP_SIZE 0x010 #define MALIDP_LAYER_OFFSET 0x014 -#define MALIDP_LAYER_STRIDE 0x018 /* * This 4-entry look-up-table is used to determine the full 8-bit alpha value @@ -67,13 +67,14 @@ drm_plane_state *malidp_duplicate_plane_state(struct drm_plane *plane) return NULL; state = kmalloc(sizeof(*state), GFP_KERNEL); - if (state) { - m_state = to_malidp_plane_state(plane->state); - __drm_atomic_helper_plane_duplicate_state(plane, &state->base); - state->rotmem_size = m_state->rotmem_size; - state->format = m_state->format; - state->n_planes = m_state->n_planes; - } + if (!state) + return NULL; + + m_state = to_malidp_plane_state(plane->state); + __drm_atomic_helper_plane_duplicate_state(plane, &state->base); + state->rotmem_size = m_state->rotmem_size; + state->format = m_state->format; + state->n_planes = m_state->n_planes; return &state->base; } @@ -102,8 +103,10 @@ static int malidp_de_plane_check(struct drm_plane *plane, { struct malidp_plane *mp = to_malidp_plane(plane); struct malidp_plane_state *ms = to_malidp_plane_state(state); + struct drm_crtc_state *crtc_state; struct drm_framebuffer *fb; - int i; + struct drm_rect clip = { 0 }; + int i, ret; u32 src_w, src_h; if (!state->crtc || !state->fb) @@ -131,8 +134,17 @@ static int malidp_de_plane_check(struct drm_plane *plane, if ((state->crtc_w > mp->hwdev->max_line_size) || (state->crtc_h > mp->hwdev->max_line_size) || (state->crtc_w < mp->hwdev->min_line_size) || - (state->crtc_h < mp->hwdev->min_line_size) || - (state->crtc_w != src_w) || (state->crtc_h != src_h)) + (state->crtc_h < mp->hwdev->min_line_size)) + return -EINVAL; + + /* + * DP550/650 video layers can accept 3 plane formats only if + * fb->pitches[1] == fb->pitches[2] since they don't have a + * third plane stride register. + */ + if (ms->n_planes == 3 && + !(mp->hwdev->features & MALIDP_DEVICE_LV_HAS_3_STRIDES) && + (state->fb->pitches[1] != state->fb->pitches[2])) return -EINVAL; /* packed RGB888 / BGR888 can't be rotated or flipped */ @@ -141,6 +153,16 @@ static int malidp_de_plane_check(struct drm_plane *plane, fb->format->format == DRM_FORMAT_BGR888)) return -EINVAL; + crtc_state = drm_atomic_get_existing_crtc_state(state->state, state->crtc); + clip.x2 = crtc_state->adjusted_mode.hdisplay; + clip.y2 = crtc_state->adjusted_mode.vdisplay; + ret = drm_plane_helper_check_state(state, &clip, + DRM_PLANE_HELPER_NO_SCALING, + DRM_PLANE_HELPER_NO_SCALING, + true, true); + if (ret) + return ret; + ms->rotmem_size = 0; if (state->rotation & MALIDP_ROTATED_MASK) { int val; @@ -157,6 +179,25 @@ static int malidp_de_plane_check(struct drm_plane *plane, return 0; } +static void malidp_de_set_plane_pitches(struct malidp_plane *mp, + int num_planes, unsigned int pitches[3]) +{ + int i; + int num_strides = num_planes; + + if (!mp->layer->stride_offset) + return; + + if (num_planes == 3) + num_strides = (mp->hwdev->features & + MALIDP_DEVICE_LV_HAS_3_STRIDES) ? 3 : 2; + + for (i = 0; i < num_strides; ++i) + malidp_hw_write(mp->hwdev, pitches[i], + mp->layer->base + + mp->layer->stride_offset + i * 4); +} + static void malidp_de_plane_update(struct drm_plane *plane, struct drm_plane_state *old_state) { @@ -174,13 +215,8 @@ static void malidp_de_plane_update(struct drm_plane *plane, /* convert src values from Q16 fixed point to integer */ src_w = plane->state->src_w >> 16; src_h = plane->state->src_h >> 16; - if (plane->state->rotation & MALIDP_ROTATED_MASK) { - dest_w = plane->state->crtc_h; - dest_h = plane->state->crtc_w; - } else { - dest_w = plane->state->crtc_w; - dest_h = plane->state->crtc_h; - } + dest_w = plane->state->crtc_w; + dest_h = plane->state->crtc_h; malidp_hw_write(mp->hwdev, ms->format, mp->layer->base); @@ -189,11 +225,12 @@ static void malidp_de_plane_update(struct drm_plane *plane, ptr = mp->layer->ptr + (i << 4); obj = drm_fb_cma_get_gem_obj(plane->state->fb, i); + obj->paddr += plane->state->fb->offsets[i]; malidp_hw_write(mp->hwdev, lower_32_bits(obj->paddr), ptr); malidp_hw_write(mp->hwdev, upper_32_bits(obj->paddr), ptr + 4); - malidp_hw_write(mp->hwdev, plane->state->fb->pitches[i], - mp->layer->base + MALIDP_LAYER_STRIDE); } + malidp_de_set_plane_pitches(mp, ms->n_planes, + plane->state->fb->pitches); malidp_hw_write(mp->hwdev, LAYER_H_VAL(src_w) | LAYER_V_VAL(src_h), mp->layer->base + MALIDP_LAYER_SIZE); @@ -211,11 +248,12 @@ static void malidp_de_plane_update(struct drm_plane *plane, /* setup the rotation and axis flip bits */ if (plane->state->rotation & DRM_ROTATE_MASK) - val = ilog2(plane->state->rotation & DRM_ROTATE_MASK) << LAYER_ROT_OFFSET; + val |= ilog2(plane->state->rotation & DRM_ROTATE_MASK) << + LAYER_ROT_OFFSET; if (plane->state->rotation & DRM_REFLECT_X) - val |= LAYER_V_FLIP; - if (plane->state->rotation & DRM_REFLECT_Y) val |= LAYER_H_FLIP; + if (plane->state->rotation & DRM_REFLECT_Y) + val |= LAYER_V_FLIP; /* * always enable pixel alpha blending until we have a way to change @@ -258,7 +296,7 @@ int malidp_de_planes_init(struct drm_device *drm) u32 *formats; int ret, i, j, n; - formats = kcalloc(map->n_input_formats, sizeof(*formats), GFP_KERNEL); + formats = kcalloc(map->n_pixel_formats, sizeof(*formats), GFP_KERNEL); if (!formats) { ret = -ENOMEM; goto cleanup; @@ -274,9 +312,9 @@ int malidp_de_planes_init(struct drm_device *drm) } /* build the list of DRM supported formats based on the map */ - for (n = 0, j = 0; j < map->n_input_formats; j++) { - if ((map->input_formats[j].layer & id) == id) - formats[n++] = map->input_formats[j].format; + for (n = 0, j = 0; j < map->n_pixel_formats; j++) { + if ((map->pixel_formats[j].layer & id) == id) + formats[n++] = map->pixel_formats[j].format; } plane_type = (i == 0) ? DRM_PLANE_TYPE_PRIMARY : diff --git a/drivers/gpu/drm/arm/malidp_regs.h b/drivers/gpu/drm/arm/malidp_regs.h index 73fecb38f955..aff6d4a84e99 100644 --- a/drivers/gpu/drm/arm/malidp_regs.h +++ b/drivers/gpu/drm/arm/malidp_regs.h @@ -81,6 +81,10 @@ #define MALIDP_DE_SYNC_WIDTH 0x8 #define MALIDP_DE_HV_ACTIVE 0xc +/* Stride register offsets relative to Lx_BASE */ +#define MALIDP_DE_LG_STRIDE 0x18 +#define MALIDP_DE_LV_STRIDE0 0x18 + /* macros to set values into registers */ #define MALIDP_DE_H_FRONTPORCH(x) (((x) & 0xfff) << 0) #define MALIDP_DE_H_BACKPORCH(x) (((x) & 0x3ff) << 16) @@ -92,7 +96,10 @@ #define MALIDP_DE_H_ACTIVE(x) (((x) & 0x1fff) << 0) #define MALIDP_DE_V_ACTIVE(x) (((x) & 0x1fff) << 16) +#define MALIDP_PRODUCT_ID(__core_id) ((u32)(__core_id) >> 16) + /* register offsets and bits specific to DP500 */ +#define MALIDP500_ADDR_SPACE_SIZE 0x01000 #define MALIDP500_DC_BASE 0x00000 #define MALIDP500_DC_CONTROL 0x0000c #define MALIDP500_DC_CONFIG_REQ (1 << 17) @@ -125,6 +132,7 @@ #define MALIDP500_CONFIG_ID 0x00fd4 /* register offsets and bits specific to DP550/DP650 */ +#define MALIDP550_ADDR_SPACE_SIZE 0x10000 #define MALIDP550_DE_CONTROL 0x00010 #define MALIDP550_DE_LINE_COUNTER 0x00014 #define MALIDP550_DE_AXI_CONTROL 0x00018 diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c index cbd0070265c9..0bf32d6ac39b 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c @@ -431,15 +431,8 @@ static void atmel_hlcdc_fb_output_poll_changed(struct drm_device *dev) { struct atmel_hlcdc_dc *dc = dev->dev_private; - if (dc->fbdev) { + if (dc->fbdev) drm_fbdev_cma_hotplug_event(dc->fbdev); - } else { - dc->fbdev = drm_fbdev_cma_init(dev, 24, - dev->mode_config.num_crtc, - dev->mode_config.num_connector); - if (IS_ERR(dc->fbdev)) - dc->fbdev = NULL; - } } struct atmel_hlcdc_dc_commit { @@ -653,10 +646,13 @@ static int atmel_hlcdc_dc_load(struct drm_device *dev) platform_set_drvdata(pdev, dev); - drm_kms_helper_poll_init(dev); + dc->fbdev = drm_fbdev_cma_init(dev, 24, + dev->mode_config.num_crtc, + dev->mode_config.num_connector); + if (IS_ERR(dc->fbdev)) + dc->fbdev = NULL; - /* force connectors detection */ - drm_helper_hpd_irq_event(dev); + drm_kms_helper_poll_init(dev); return 0; diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c index 02b97bf64ee4..e7cd1056ff2d 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c @@ -1385,6 +1385,7 @@ int analogix_dp_bind(struct device *dev, struct drm_device *drm_dev, pm_runtime_enable(dev); + pm_runtime_get_sync(dev); phy_power_on(dp->phy); analogix_dp_init_dp(dp); @@ -1417,9 +1418,15 @@ int analogix_dp_bind(struct device *dev, struct drm_device *drm_dev, goto err_disable_pm_runtime; } + phy_power_off(dp->phy); + pm_runtime_put(dev); + return 0; err_disable_pm_runtime: + + phy_power_off(dp->phy); + pm_runtime_put(dev); pm_runtime_disable(dev); return ret; diff --git a/drivers/gpu/drm/cirrus/Kconfig b/drivers/gpu/drm/cirrus/Kconfig index ca3809851377..fc78c90ee931 100644 --- a/drivers/gpu/drm/cirrus/Kconfig +++ b/drivers/gpu/drm/cirrus/Kconfig @@ -7,3 +7,12 @@ config DRM_CIRRUS_QEMU This is a KMS driver for emulated cirrus device in qemu. It is *NOT* intended for real cirrus devices. This requires the modesetting userspace X.org driver. + + Cirrus is obsolete, the hardware was designed in the 90ies + and can't keep up with todays needs. More background: + https://www.kraxel.org/blog/2014/10/qemu-using-cirrus-considered-harmful/ + + Better alternatives are: + - stdvga (DRM_BOCHS, qemu -vga std, default in qemu 2.2+) + - qxl (DRM_QXL, qemu -vga qxl, works best with spice) + - virtio (DRM_VIRTIO_GPU), qemu -vga virtio) diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 43049882b811..9a08445a7a7a 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -1232,8 +1232,10 @@ int drm_atomic_helper_commit(struct drm_device *dev, if (!nonblock) { ret = drm_atomic_helper_wait_for_fences(dev, state, true); - if (ret) + if (ret) { + drm_atomic_helper_cleanup_planes(dev, state); return ret; + } } /* diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index 18c1b2cbfcdb..6cbd67f4fbc5 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -598,6 +598,8 @@ static void drm_dev_release(struct kref *ref) { struct drm_device *dev = container_of(ref, struct drm_device, ref); + drm_vblank_cleanup(dev); + if (drm_core_check_feature(dev, DRIVER_GEM)) drm_gem_destroy(dev); @@ -805,8 +807,6 @@ void drm_dev_unregister(struct drm_device *dev) if (dev->agp) drm_pci_agp_destroy(dev); - drm_vblank_cleanup(dev); - list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head) drm_legacy_rmmap(dev, r_list->map); diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c index 30c716ca236b..a8616b1a8d22 100644 --- a/drivers/gpu/drm/drm_modes.c +++ b/drivers/gpu/drm/drm_modes.c @@ -1480,6 +1480,13 @@ drm_mode_create_from_cmdline_mode(struct drm_device *dev, return NULL; mode->type |= DRM_MODE_TYPE_USERDEF; + /* fix up 1368x768: GFT/CVT can't express 1366 width due to alignment */ + if (cmd->xres == 1366 && mode->hdisplay == 1368) { + mode->hdisplay = 1366; + mode->hsync_start--; + mode->hsync_end--; + drm_mode_set_name(mode); + } drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V); return mode; } diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c index 060211ac74a1..93381454bdf7 100644 --- a/drivers/gpu/drm/drm_probe_helper.c +++ b/drivers/gpu/drm/drm_probe_helper.c @@ -115,25 +115,28 @@ static int drm_helper_probe_add_cmdline_mode(struct drm_connector *connector) #define DRM_OUTPUT_POLL_PERIOD (10*HZ) /** - * drm_kms_helper_poll_enable_locked - re-enable output polling. + * drm_kms_helper_poll_enable - re-enable output polling. * @dev: drm_device * - * This function re-enables the output polling work without - * locking the mode_config mutex. + * This function re-enables the output polling work, after it has been + * temporarily disabled using drm_kms_helper_poll_disable(), for example over + * suspend/resume. * - * This is like drm_kms_helper_poll_enable() however it is to be - * called from a context where the mode_config mutex is locked - * already. + * Drivers can call this helper from their device resume implementation. It is + * an error to call this when the output polling support has not yet been set + * up. + * + * Note that calls to enable and disable polling must be strictly ordered, which + * is automatically the case when they're only call from suspend/resume + * callbacks. */ -void drm_kms_helper_poll_enable_locked(struct drm_device *dev) +void drm_kms_helper_poll_enable(struct drm_device *dev) { bool poll = false; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; unsigned long delay = DRM_OUTPUT_POLL_PERIOD; - WARN_ON(!mutex_is_locked(&dev->mode_config.mutex)); - if (!dev->mode_config.poll_enabled || !drm_kms_helper_poll) return; @@ -146,14 +149,24 @@ void drm_kms_helper_poll_enable_locked(struct drm_device *dev) drm_connector_list_iter_put(&conn_iter); if (dev->mode_config.delayed_event) { + /* + * FIXME: + * + * Use short (1s) delay to handle the initial delayed event. + * This delay should not be needed, but Optimus/nouveau will + * fail in a mysterious way if the delayed event is handled as + * soon as possible like it is done in + * drm_helper_probe_single_connector_modes() in case the poll + * was enabled before. + */ poll = true; - delay = 0; + delay = HZ; } if (poll) schedule_delayed_work(&dev->mode_config.output_poll_work, delay); } -EXPORT_SYMBOL(drm_kms_helper_poll_enable_locked); +EXPORT_SYMBOL(drm_kms_helper_poll_enable); static enum drm_connector_status drm_connector_detect(struct drm_connector *connector, bool force) @@ -280,7 +293,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, /* Re-enable polling in case the global poll config changed. */ if (drm_kms_helper_poll != dev->mode_config.poll_running) - drm_kms_helper_poll_enable_locked(dev); + drm_kms_helper_poll_enable(dev); dev->mode_config.poll_running = drm_kms_helper_poll; @@ -474,8 +487,12 @@ out: * This function disables the output polling work. * * Drivers can call this helper from their device suspend implementation. It is - * not an error to call this even when output polling isn't enabled or arlready - * disabled. + * not an error to call this even when output polling isn't enabled or already + * disabled. Polling is re-enabled by calling drm_kms_helper_poll_enable(). + * + * Note that calls to enable and disable polling must be strictly ordered, which + * is automatically the case when they're only call from suspend/resume + * callbacks. */ void drm_kms_helper_poll_disable(struct drm_device *dev) { @@ -486,24 +503,6 @@ void drm_kms_helper_poll_disable(struct drm_device *dev) EXPORT_SYMBOL(drm_kms_helper_poll_disable); /** - * drm_kms_helper_poll_enable - re-enable output polling. - * @dev: drm_device - * - * This function re-enables the output polling work. - * - * Drivers can call this helper from their device resume implementation. It is - * an error to call this when the output polling support has not yet been set - * up. - */ -void drm_kms_helper_poll_enable(struct drm_device *dev) -{ - mutex_lock(&dev->mode_config.mutex); - drm_kms_helper_poll_enable_locked(dev); - mutex_unlock(&dev->mode_config.mutex); -} -EXPORT_SYMBOL(drm_kms_helper_poll_enable); - -/** * drm_kms_helper_poll_init - initialize and enable output polling * @dev: drm_device * diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c index ae2733a609ba..f503af462dad 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c @@ -117,9 +117,14 @@ static int etnaviv_iommu_find_iova(struct etnaviv_iommu *mmu, struct list_head list; bool found; + /* + * XXX: The DRM_MM_SEARCH_BELOW is really a hack to trick + * drm_mm into giving out a low IOVA after address space + * rollover. This needs a proper fix. + */ ret = drm_mm_insert_node_in_range(&mmu->mm, node, size, 0, mmu->last_iova, ~0UL, - DRM_MM_SEARCH_DEFAULT); + mmu->last_iova ? DRM_MM_SEARCH_DEFAULT : DRM_MM_SEARCH_BELOW); if (ret != -ENOSPC) break; diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index 31eafc546d6c..d69af00bdd6a 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -46,7 +46,8 @@ enum decon_flag_bits { BIT_CLKS_ENABLED, BIT_IRQS_ENABLED, BIT_WIN_UPDATED, - BIT_SUSPENDED + BIT_SUSPENDED, + BIT_REQUEST_UPDATE }; struct decon_context { @@ -141,12 +142,6 @@ static void decon_commit(struct exynos_drm_crtc *crtc) m->crtc_vsync_end = m->crtc_vsync_start + 1; } - decon_set_bits(ctx, DECON_VIDCON0, VIDCON0_ENVID, 0); - - /* enable clock gate */ - val = CMU_CLKGAGE_MODE_SFR_F | CMU_CLKGAGE_MODE_MEM_F; - writel(val, ctx->addr + DECON_CMU); - if (ctx->out_type & (IFTYPE_I80 | I80_HW_TRG)) decon_setup_trigger(ctx); @@ -315,6 +310,7 @@ static void decon_update_plane(struct exynos_drm_crtc *crtc, /* window enable */ decon_set_bits(ctx, DECON_WINCONx(win), WINCONx_ENWIN_F, ~0); + set_bit(BIT_REQUEST_UPDATE, &ctx->flags); } static void decon_disable_plane(struct exynos_drm_crtc *crtc, @@ -327,6 +323,7 @@ static void decon_disable_plane(struct exynos_drm_crtc *crtc, return; decon_set_bits(ctx, DECON_WINCONx(win), WINCONx_ENWIN_F, 0); + set_bit(BIT_REQUEST_UPDATE, &ctx->flags); } static void decon_atomic_flush(struct exynos_drm_crtc *crtc) @@ -340,8 +337,8 @@ static void decon_atomic_flush(struct exynos_drm_crtc *crtc) for (i = ctx->first_win; i < WINDOWS_NR; i++) decon_shadow_protect_win(ctx, i, false); - /* standalone update */ - decon_set_bits(ctx, DECON_UPDATE, STANDALONE_UPDATE_F, ~0); + if (test_and_clear_bit(BIT_REQUEST_UPDATE, &ctx->flags)) + decon_set_bits(ctx, DECON_UPDATE, STANDALONE_UPDATE_F, ~0); if (ctx->out_type & IFTYPE_I80) set_bit(BIT_WIN_UPDATED, &ctx->flags); diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 5196509e71cf..74ca2e8b2494 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -56,7 +56,9 @@ i915-y += i915_cmd_parser.o \ # general-purpose microcontroller (GuC) support i915-y += intel_uc.o \ + intel_guc_log.o \ intel_guc_loader.o \ + intel_huc.o \ i915_guc_submission.o # autogenerated null render state diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c index 7d33b607bc89..7311aeab16f7 100644 --- a/drivers/gpu/drm/i915/gvt/aperture_gm.c +++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c @@ -37,58 +37,38 @@ #include "i915_drv.h" #include "gvt.h" -#define MB_TO_BYTES(mb) ((mb) << 20ULL) -#define BYTES_TO_MB(b) ((b) >> 20ULL) - -#define HOST_LOW_GM_SIZE MB_TO_BYTES(128) -#define HOST_HIGH_GM_SIZE MB_TO_BYTES(384) -#define HOST_FENCE 4 - static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm) { struct intel_gvt *gvt = vgpu->gvt; struct drm_i915_private *dev_priv = gvt->dev_priv; - u32 alloc_flag, search_flag; + unsigned int flags; u64 start, end, size; struct drm_mm_node *node; - int retried = 0; int ret; if (high_gm) { - search_flag = DRM_MM_SEARCH_BELOW; - alloc_flag = DRM_MM_CREATE_TOP; node = &vgpu->gm.high_gm_node; size = vgpu_hidden_sz(vgpu); start = gvt_hidden_gmadr_base(gvt); end = gvt_hidden_gmadr_end(gvt); + flags = PIN_HIGH; } else { - search_flag = DRM_MM_SEARCH_DEFAULT; - alloc_flag = DRM_MM_CREATE_DEFAULT; node = &vgpu->gm.low_gm_node; size = vgpu_aperture_sz(vgpu); start = gvt_aperture_gmadr_base(gvt); end = gvt_aperture_gmadr_end(gvt); + flags = PIN_MAPPABLE; } mutex_lock(&dev_priv->drm.struct_mutex); -search_again: - ret = drm_mm_insert_node_in_range_generic(&dev_priv->ggtt.base.mm, - node, size, 4096, - I915_COLOR_UNEVICTABLE, - start, end, search_flag, - alloc_flag); - if (ret) { - ret = i915_gem_evict_something(&dev_priv->ggtt.base, - size, 4096, - I915_COLOR_UNEVICTABLE, - start, end, 0); - if (ret == 0 && ++retried < 3) - goto search_again; - - gvt_err("fail to alloc %s gm space from host, retried %d\n", - high_gm ? "high" : "low", retried); - } + ret = i915_gem_gtt_insert(&dev_priv->ggtt.base, node, + size, 4096, I915_COLOR_UNEVICTABLE, + start, end, flags); mutex_unlock(&dev_priv->drm.struct_mutex); + if (ret) + gvt_err("fail to alloc %s gm space from host\n", + high_gm ? "high" : "low"); + return ret; } @@ -168,6 +148,14 @@ void intel_vgpu_write_fence(struct intel_vgpu *vgpu, POSTING_READ(fence_reg_lo); } +static void _clear_vgpu_fence(struct intel_vgpu *vgpu) +{ + int i; + + for (i = 0; i < vgpu_fence_sz(vgpu); i++) + intel_vgpu_write_fence(vgpu, i, 0); +} + static void free_vgpu_fence(struct intel_vgpu *vgpu) { struct intel_gvt *gvt = vgpu->gvt; @@ -181,9 +169,9 @@ static void free_vgpu_fence(struct intel_vgpu *vgpu) intel_runtime_pm_get(dev_priv); mutex_lock(&dev_priv->drm.struct_mutex); + _clear_vgpu_fence(vgpu); for (i = 0; i < vgpu_fence_sz(vgpu); i++) { reg = vgpu->fence.regs[i]; - intel_vgpu_write_fence(vgpu, i, 0); list_add_tail(®->link, &dev_priv->mm.fence_list); } @@ -211,13 +199,14 @@ static int alloc_vgpu_fence(struct intel_vgpu *vgpu) continue; list_del(pos); vgpu->fence.regs[i] = reg; - intel_vgpu_write_fence(vgpu, i, 0); if (++i == vgpu_fence_sz(vgpu)) break; } if (i != vgpu_fence_sz(vgpu)) goto out_free_fence; + _clear_vgpu_fence(vgpu); + mutex_unlock(&dev_priv->drm.struct_mutex); intel_runtime_pm_put(dev_priv); return 0; @@ -317,6 +306,22 @@ void intel_vgpu_free_resource(struct intel_vgpu *vgpu) } /** + * intel_vgpu_reset_resource - reset resource state owned by a vGPU + * @vgpu: a vGPU + * + * This function is used to reset resource state owned by a vGPU. + * + */ +void intel_vgpu_reset_resource(struct intel_vgpu *vgpu) +{ + struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; + + intel_runtime_pm_get(dev_priv); + _clear_vgpu_fence(vgpu); + intel_runtime_pm_put(dev_priv); +} + +/** * intel_alloc_vgpu_resource - allocate HW resource for a vGPU * @vgpu: vGPU * @param: vGPU creation params diff --git a/drivers/gpu/drm/i915/gvt/cfg_space.c b/drivers/gpu/drm/i915/gvt/cfg_space.c index 711c31c8d8b4..4a6a2ed65732 100644 --- a/drivers/gpu/drm/i915/gvt/cfg_space.c +++ b/drivers/gpu/drm/i915/gvt/cfg_space.c @@ -282,3 +282,77 @@ int intel_vgpu_emulate_cfg_write(struct intel_vgpu *vgpu, unsigned int offset, } return 0; } + +/** + * intel_vgpu_init_cfg_space - init vGPU configuration space when create vGPU + * + * @vgpu: a vGPU + * @primary: is the vGPU presented as primary + * + */ +void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu, + bool primary) +{ + struct intel_gvt *gvt = vgpu->gvt; + const struct intel_gvt_device_info *info = &gvt->device_info; + u16 *gmch_ctl; + int i; + + memcpy(vgpu_cfg_space(vgpu), gvt->firmware.cfg_space, + info->cfg_space_size); + + if (!primary) { + vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] = + INTEL_GVT_PCI_CLASS_VGA_OTHER; + vgpu_cfg_space(vgpu)[PCI_CLASS_PROG] = + INTEL_GVT_PCI_CLASS_VGA_OTHER; + } + + /* Show guest that there isn't any stolen memory.*/ + gmch_ctl = (u16 *)(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_GMCH_CONTROL); + *gmch_ctl &= ~(BDW_GMCH_GMS_MASK << BDW_GMCH_GMS_SHIFT); + + intel_vgpu_write_pci_bar(vgpu, PCI_BASE_ADDRESS_2, + gvt_aperture_pa_base(gvt), true); + + vgpu_cfg_space(vgpu)[PCI_COMMAND] &= ~(PCI_COMMAND_IO + | PCI_COMMAND_MEMORY + | PCI_COMMAND_MASTER); + /* + * Clear the bar upper 32bit and let guest to assign the new value + */ + memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_1, 0, 4); + memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_3, 0, 4); + memset(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_OPREGION, 0, 4); + + for (i = 0; i < INTEL_GVT_MAX_BAR_NUM; i++) { + vgpu->cfg_space.bar[i].size = pci_resource_len( + gvt->dev_priv->drm.pdev, i * 2); + vgpu->cfg_space.bar[i].tracked = false; + } +} + +/** + * intel_vgpu_reset_cfg_space - reset vGPU configuration space + * + * @vgpu: a vGPU + * + */ +void intel_vgpu_reset_cfg_space(struct intel_vgpu *vgpu) +{ + u8 cmd = vgpu_cfg_space(vgpu)[PCI_COMMAND]; + bool primary = vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] != + INTEL_GVT_PCI_CLASS_VGA_OTHER; + + if (cmd & PCI_COMMAND_MEMORY) { + trap_gttmmio(vgpu, false); + map_aperture(vgpu, false); + } + + /** + * Currently we only do such reset when vGPU is not + * owned by any VM, so we simply restore entire cfg + * space to default value. + */ + intel_vgpu_init_cfg_space(vgpu, primary); +} diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c index 6c5fdf5b2ce2..47dec4acf7ff 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.c +++ b/drivers/gpu/drm/i915/gvt/gtt.c @@ -240,15 +240,8 @@ static inline int get_pse_type(int type) static u64 read_pte64(struct drm_i915_private *dev_priv, unsigned long index) { void __iomem *addr = (gen8_pte_t __iomem *)dev_priv->ggtt.gsm + index; - u64 pte; -#ifdef readq - pte = readq(addr); -#else - pte = ioread32(addr); - pte |= (u64)ioread32(addr + 4) << 32; -#endif - return pte; + return readq(addr); } static void write_pte64(struct drm_i915_private *dev_priv, @@ -256,12 +249,8 @@ static void write_pte64(struct drm_i915_private *dev_priv, { void __iomem *addr = (gen8_pte_t __iomem *)dev_priv->ggtt.gsm + index; -#ifdef writeq writeq(pte, addr); -#else - iowrite32((u32)pte, addr); - iowrite32(pte >> 32, addr + 4); -#endif + I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); POSTING_READ(GFX_FLSH_CNTL_GEN6); } @@ -1380,8 +1369,7 @@ static int gen8_mm_alloc_page_table(struct intel_vgpu_mm *mm) info->gtt_entry_size; mem = kzalloc(mm->has_shadow_page_table ? mm->page_table_entry_size * 2 - : mm->page_table_entry_size, - GFP_ATOMIC); + : mm->page_table_entry_size, GFP_KERNEL); if (!mem) return -ENOMEM; mm->virtual_page_table = mem; @@ -1532,7 +1520,7 @@ struct intel_vgpu_mm *intel_vgpu_create_mm(struct intel_vgpu *vgpu, struct intel_vgpu_mm *mm; int ret; - mm = kzalloc(sizeof(*mm), GFP_ATOMIC); + mm = kzalloc(sizeof(*mm), GFP_KERNEL); if (!mm) { ret = -ENOMEM; goto fail; @@ -1886,30 +1874,27 @@ static int alloc_scratch_pages(struct intel_vgpu *vgpu, struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops; int page_entry_num = GTT_PAGE_SIZE >> vgpu->gvt->device_info.gtt_entry_size_shift; - struct page *scratch_pt; + void *scratch_pt; unsigned long mfn; int i; - void *p; if (WARN_ON(type < GTT_TYPE_PPGTT_PTE_PT || type >= GTT_TYPE_MAX)) return -EINVAL; - scratch_pt = alloc_page(GFP_KERNEL | GFP_ATOMIC | __GFP_ZERO); + scratch_pt = (void *)get_zeroed_page(GFP_KERNEL); if (!scratch_pt) { gvt_err("fail to allocate scratch page\n"); return -ENOMEM; } - p = kmap_atomic(scratch_pt); - mfn = intel_gvt_hypervisor_virt_to_mfn(p); + mfn = intel_gvt_hypervisor_virt_to_mfn(scratch_pt); if (mfn == INTEL_GVT_INVALID_ADDR) { - gvt_err("fail to translate vaddr:0x%llx\n", (u64)p); - kunmap_atomic(p); - __free_page(scratch_pt); + gvt_err("fail to translate vaddr:0x%lx\n", (unsigned long)scratch_pt); + free_page((unsigned long)scratch_pt); return -EFAULT; } gtt->scratch_pt[type].page_mfn = mfn; - gtt->scratch_pt[type].page = scratch_pt; + gtt->scratch_pt[type].page = virt_to_page(scratch_pt); gvt_dbg_mm("vgpu%d create scratch_pt: type %d mfn=0x%lx\n", vgpu->id, type, mfn); @@ -1918,7 +1903,7 @@ static int alloc_scratch_pages(struct intel_vgpu *vgpu, * scratch_pt[type] indicate the scratch pt/scratch page used by the * 'type' pt. * e.g. scratch_pt[GTT_TYPE_PPGTT_PDE_PT] is used by - * GTT_TYPE_PPGTT_PDE_PT level pt, that means this scatch_pt it self + * GTT_TYPE_PPGTT_PDE_PT level pt, that means this scratch_pt it self * is GTT_TYPE_PPGTT_PTE_PT, and full filled by scratch page mfn. */ if (type > GTT_TYPE_PPGTT_PTE_PT && type < GTT_TYPE_MAX) { @@ -1936,11 +1921,9 @@ static int alloc_scratch_pages(struct intel_vgpu *vgpu, se.val64 |= PPAT_CACHED_INDEX; for (i = 0; i < page_entry_num; i++) - ops->set_entry(p, &se, i, false, 0, vgpu); + ops->set_entry(scratch_pt, &se, i, false, 0, vgpu); } - kunmap_atomic(p); - return 0; } @@ -2208,7 +2191,7 @@ int intel_vgpu_g2v_destroy_ppgtt_mm(struct intel_vgpu *vgpu, int intel_gvt_init_gtt(struct intel_gvt *gvt) { int ret; - void *page_addr; + void *page; gvt_dbg_core("init gtt\n"); @@ -2221,17 +2204,14 @@ int intel_gvt_init_gtt(struct intel_gvt *gvt) return -ENODEV; } - gvt->gtt.scratch_ggtt_page = - alloc_page(GFP_KERNEL | GFP_ATOMIC | __GFP_ZERO); - if (!gvt->gtt.scratch_ggtt_page) { + page = (void *)get_zeroed_page(GFP_KERNEL); + if (!page) { gvt_err("fail to allocate scratch ggtt page\n"); return -ENOMEM; } + gvt->gtt.scratch_ggtt_page = virt_to_page(page); - page_addr = page_address(gvt->gtt.scratch_ggtt_page); - - gvt->gtt.scratch_ggtt_mfn = - intel_gvt_hypervisor_virt_to_mfn(page_addr); + gvt->gtt.scratch_ggtt_mfn = intel_gvt_hypervisor_virt_to_mfn(page); if (gvt->gtt.scratch_ggtt_mfn == INTEL_GVT_INVALID_ADDR) { gvt_err("fail to translate scratch ggtt page\n"); __free_page(gvt->gtt.scratch_ggtt_page); @@ -2297,3 +2277,30 @@ void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu) for (offset = 0; offset < num_entries; offset++) ops->set_entry(NULL, &e, index + offset, false, 0, vgpu); } + +/** + * intel_vgpu_reset_gtt - reset the all GTT related status + * @vgpu: a vGPU + * @dmlr: true for vGPU Device Model Level Reset, false for GT Reset + * + * This function is called from vfio core to reset reset all + * GTT related status, including GGTT, PPGTT, scratch page. + * + */ +void intel_vgpu_reset_gtt(struct intel_vgpu *vgpu, bool dmlr) +{ + int i; + + ppgtt_free_all_shadow_page(vgpu); + if (!dmlr) + return; + + intel_vgpu_reset_ggtt(vgpu); + + /* clear scratch page for security */ + for (i = GTT_TYPE_PPGTT_PTE_PT; i < GTT_TYPE_MAX; i++) { + if (vgpu->gtt.scratch_pt[i].page != NULL) + memset(page_address(vgpu->gtt.scratch_pt[i].page), + 0, PAGE_SIZE); + } +} diff --git a/drivers/gpu/drm/i915/gvt/gtt.h b/drivers/gpu/drm/i915/gvt/gtt.h index b315ab3593ec..f88eb5e89bea 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.h +++ b/drivers/gpu/drm/i915/gvt/gtt.h @@ -208,6 +208,7 @@ extern void intel_vgpu_clean_gtt(struct intel_vgpu *vgpu); void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu); extern int intel_gvt_init_gtt(struct intel_gvt *gvt); +extern void intel_vgpu_reset_gtt(struct intel_vgpu *vgpu, bool dmlr); extern void intel_gvt_clean_gtt(struct intel_gvt *gvt); extern struct intel_vgpu_mm *intel_gvt_find_ppgtt_mm(struct intel_vgpu *vgpu, diff --git a/drivers/gpu/drm/i915/gvt/gvt.c b/drivers/gpu/drm/i915/gvt/gvt.c index 398877c3d2fd..e6bf5c533fbe 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.c +++ b/drivers/gpu/drm/i915/gvt/gvt.c @@ -201,6 +201,8 @@ void intel_gvt_clean_device(struct drm_i915_private *dev_priv) intel_gvt_hypervisor_host_exit(&dev_priv->drm.pdev->dev, gvt); intel_gvt_clean_vgpu_types(gvt); + idr_destroy(&gvt->vgpu_idr); + kfree(dev_priv->gvt); dev_priv->gvt = NULL; } @@ -237,6 +239,8 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) gvt_dbg_core("init gvt device\n"); + idr_init(&gvt->vgpu_idr); + mutex_init(&gvt->lock); gvt->dev_priv = dev_priv; @@ -244,7 +248,7 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) ret = intel_gvt_setup_mmio_info(gvt); if (ret) - return ret; + goto out_clean_idr; ret = intel_gvt_load_firmware(gvt); if (ret) @@ -313,6 +317,8 @@ out_free_firmware: intel_gvt_free_firmware(gvt); out_clean_mmio_info: intel_gvt_clean_mmio_info(gvt); +out_clean_idr: + idr_destroy(&gvt->vgpu_idr); kfree(gvt); return ret; } diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index 0af17016f33f..e227caf5859e 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -323,6 +323,7 @@ struct intel_vgpu_creation_params { int intel_vgpu_alloc_resource(struct intel_vgpu *vgpu, struct intel_vgpu_creation_params *param); +void intel_vgpu_reset_resource(struct intel_vgpu *vgpu); void intel_vgpu_free_resource(struct intel_vgpu *vgpu); void intel_vgpu_write_fence(struct intel_vgpu *vgpu, u32 fence, u64 value); @@ -375,6 +376,8 @@ void intel_gvt_clean_vgpu_types(struct intel_gvt *gvt); struct intel_vgpu *intel_gvt_create_vgpu(struct intel_gvt *gvt, struct intel_vgpu_type *type); void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu); +void intel_gvt_reset_vgpu_locked(struct intel_vgpu *vgpu, bool dmlr, + unsigned int engine_mask); void intel_gvt_reset_vgpu(struct intel_vgpu *vgpu); @@ -411,6 +414,10 @@ int intel_gvt_ggtt_index_g2h(struct intel_vgpu *vgpu, unsigned long g_index, int intel_gvt_ggtt_h2g_index(struct intel_vgpu *vgpu, unsigned long h_index, unsigned long *g_index); +void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu, + bool primary); +void intel_vgpu_reset_cfg_space(struct intel_vgpu *vgpu); + int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes); @@ -424,7 +431,6 @@ void intel_vgpu_clean_opregion(struct intel_vgpu *vgpu); int intel_vgpu_init_opregion(struct intel_vgpu *vgpu, u32 gpa); int intel_vgpu_emulate_opregion_request(struct intel_vgpu *vgpu, u32 swsci); -int setup_vgpu_mmio(struct intel_vgpu *vgpu); void populate_pvinfo_page(struct intel_vgpu *vgpu); struct intel_gvt_ops { diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 57fb8e3cbd1f..1d450627ff65 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -93,7 +93,8 @@ static void write_vreg(struct intel_vgpu *vgpu, unsigned int offset, static int new_mmio_info(struct intel_gvt *gvt, u32 offset, u32 flags, u32 size, u32 addr_mask, u32 ro_mask, u32 device, - void *read, void *write) + int (*read)(struct intel_vgpu *, unsigned int, void *, unsigned int), + int (*write)(struct intel_vgpu *, unsigned int, void *, unsigned int)) { struct intel_gvt_mmio_info *info, *p; u32 start, end, i; @@ -219,7 +220,7 @@ static int mul_force_wake_write(struct intel_vgpu *vgpu, default: /*should not hit here*/ gvt_err("invalid forcewake offset 0x%x\n", offset); - return 1; + return -EINVAL; } } else { ack_reg_offset = FORCEWAKE_ACK_HSW_REG; @@ -230,77 +231,45 @@ static int mul_force_wake_write(struct intel_vgpu *vgpu, return 0; } -static int handle_device_reset(struct intel_vgpu *vgpu, unsigned int offset, - void *p_data, unsigned int bytes, unsigned long bitmap) -{ - struct intel_gvt_workload_scheduler *scheduler = - &vgpu->gvt->scheduler; - - vgpu->resetting = true; - - intel_vgpu_stop_schedule(vgpu); - /* - * The current_vgpu will set to NULL after stopping the - * scheduler when the reset is triggered by current vgpu. - */ - if (scheduler->current_vgpu == NULL) { - mutex_unlock(&vgpu->gvt->lock); - intel_gvt_wait_vgpu_idle(vgpu); - mutex_lock(&vgpu->gvt->lock); - } - - intel_vgpu_reset_execlist(vgpu, bitmap); - - /* full GPU reset */ - if (bitmap == 0xff) { - mutex_unlock(&vgpu->gvt->lock); - intel_vgpu_clean_gtt(vgpu); - mutex_lock(&vgpu->gvt->lock); - setup_vgpu_mmio(vgpu); - populate_pvinfo_page(vgpu); - intel_vgpu_init_gtt(vgpu); - } - - vgpu->resetting = false; - - return 0; -} - static int gdrst_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, - void *p_data, unsigned int bytes) + void *p_data, unsigned int bytes) { + unsigned int engine_mask = 0; u32 data; - u64 bitmap = 0; write_vreg(vgpu, offset, p_data, bytes); data = vgpu_vreg(vgpu, offset); if (data & GEN6_GRDOM_FULL) { gvt_dbg_mmio("vgpu%d: request full GPU reset\n", vgpu->id); - bitmap = 0xff; - } - if (data & GEN6_GRDOM_RENDER) { - gvt_dbg_mmio("vgpu%d: request RCS reset\n", vgpu->id); - bitmap |= (1 << RCS); - } - if (data & GEN6_GRDOM_MEDIA) { - gvt_dbg_mmio("vgpu%d: request VCS reset\n", vgpu->id); - bitmap |= (1 << VCS); - } - if (data & GEN6_GRDOM_BLT) { - gvt_dbg_mmio("vgpu%d: request BCS Reset\n", vgpu->id); - bitmap |= (1 << BCS); - } - if (data & GEN6_GRDOM_VECS) { - gvt_dbg_mmio("vgpu%d: request VECS Reset\n", vgpu->id); - bitmap |= (1 << VECS); - } - if (data & GEN8_GRDOM_MEDIA2) { - gvt_dbg_mmio("vgpu%d: request VCS2 Reset\n", vgpu->id); - if (HAS_BSD2(vgpu->gvt->dev_priv)) - bitmap |= (1 << VCS2); + engine_mask = ALL_ENGINES; + } else { + if (data & GEN6_GRDOM_RENDER) { + gvt_dbg_mmio("vgpu%d: request RCS reset\n", vgpu->id); + engine_mask |= (1 << RCS); + } + if (data & GEN6_GRDOM_MEDIA) { + gvt_dbg_mmio("vgpu%d: request VCS reset\n", vgpu->id); + engine_mask |= (1 << VCS); + } + if (data & GEN6_GRDOM_BLT) { + gvt_dbg_mmio("vgpu%d: request BCS Reset\n", vgpu->id); + engine_mask |= (1 << BCS); + } + if (data & GEN6_GRDOM_VECS) { + gvt_dbg_mmio("vgpu%d: request VECS Reset\n", vgpu->id); + engine_mask |= (1 << VECS); + } + if (data & GEN8_GRDOM_MEDIA2) { + gvt_dbg_mmio("vgpu%d: request VCS2 Reset\n", vgpu->id); + if (HAS_BSD2(vgpu->gvt->dev_priv)) + engine_mask |= (1 << VCS2); + } } - return handle_device_reset(vgpu, offset, p_data, bytes, bitmap); + + intel_gvt_reset_vgpu_locked(vgpu, false, engine_mask); + + return 0; } static int gmbus_mmio_read(struct intel_vgpu *vgpu, unsigned int offset, @@ -974,7 +943,7 @@ static int sbi_data_mmio_read(struct intel_vgpu *vgpu, unsigned int offset, return 0; } -static bool sbi_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, +static int sbi_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes) { u32 data; @@ -1366,7 +1335,6 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, static int gvt_reg_tlb_control_handler(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes) { - int rc = 0; unsigned int id = 0; write_vreg(vgpu, offset, p_data, bytes); @@ -1389,12 +1357,11 @@ static int gvt_reg_tlb_control_handler(struct intel_vgpu *vgpu, id = VECS; break; default: - rc = -EINVAL; - break; + return -EINVAL; } set_bit(id, (void *)vgpu->tlb_handle_pending); - return rc; + return 0; } static int ring_reset_ctl_write(struct intel_vgpu *vgpu, diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c index faaae07ae487..0c9234a87a20 100644 --- a/drivers/gpu/drm/i915/gvt/kvmgt.c +++ b/drivers/gpu/drm/i915/gvt/kvmgt.c @@ -398,6 +398,7 @@ static int intel_vgpu_create(struct kobject *kobj, struct mdev_device *mdev) struct intel_vgpu_type *type; struct device *pdev; void *gvt; + int ret; pdev = mdev_parent_dev(mdev); gvt = kdev_to_i915(pdev)->gvt; @@ -406,13 +407,15 @@ static int intel_vgpu_create(struct kobject *kobj, struct mdev_device *mdev) if (!type) { gvt_err("failed to find type %s to create\n", kobject_name(kobj)); - return -EINVAL; + ret = -EINVAL; + goto out; } vgpu = intel_gvt_ops->vgpu_create(gvt, type); if (IS_ERR_OR_NULL(vgpu)) { - gvt_err("create intel vgpu failed\n"); - return -EINVAL; + ret = vgpu == NULL ? -EFAULT : PTR_ERR(vgpu); + gvt_err("failed to create intel vgpu: %d\n", ret); + goto out; } INIT_WORK(&vgpu->vdev.release_work, intel_vgpu_release_work); @@ -422,7 +425,10 @@ static int intel_vgpu_create(struct kobject *kobj, struct mdev_device *mdev) gvt_dbg_core("intel_vgpu_create succeeded for mdev: %s\n", dev_name(mdev_dev(mdev))); - return 0; + ret = 0; + +out: + return ret; } static int intel_vgpu_remove(struct mdev_device *mdev) diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c index 09c9450a1946..4df078bc5d04 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.c +++ b/drivers/gpu/drm/i915/gvt/mmio.c @@ -125,25 +125,12 @@ int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, uint64_t pa, if (WARN_ON(!reg_is_mmio(gvt, offset + bytes - 1))) goto err; - mmio = intel_gvt_find_mmio_info(gvt, rounddown(offset, 4)); - if (!mmio && !vgpu->mmio.disable_warn_untrack) { - gvt_err("vgpu%d: read untracked MMIO %x len %d val %x\n", - vgpu->id, offset, bytes, *(u32 *)p_data); - - if (offset == 0x206c) { - gvt_err("------------------------------------------\n"); - gvt_err("vgpu%d: likely triggers a gfx reset\n", - vgpu->id); - gvt_err("------------------------------------------\n"); - vgpu->mmio.disable_warn_untrack = true; - } - } - if (!intel_gvt_mmio_is_unalign(gvt, offset)) { if (WARN_ON(!IS_ALIGNED(offset, bytes))) goto err; } + mmio = intel_gvt_find_mmio_info(gvt, rounddown(offset, 4)); if (mmio) { if (!intel_gvt_mmio_is_unalign(gvt, mmio->offset)) { if (WARN_ON(offset + bytes > mmio->offset + mmio->size)) @@ -152,9 +139,23 @@ int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, uint64_t pa, goto err; } ret = mmio->read(vgpu, offset, p_data, bytes); - } else + } else { ret = intel_vgpu_default_mmio_read(vgpu, offset, p_data, bytes); + if (!vgpu->mmio.disable_warn_untrack) { + gvt_err("vgpu%d: read untracked MMIO %x(%dB) val %x\n", + vgpu->id, offset, bytes, *(u32 *)p_data); + + if (offset == 0x206c) { + gvt_err("------------------------------------------\n"); + gvt_err("vgpu%d: likely triggers a gfx reset\n", + vgpu->id); + gvt_err("------------------------------------------\n"); + vgpu->mmio.disable_warn_untrack = true; + } + } + } + if (ret) goto err; @@ -302,3 +303,56 @@ err: mutex_unlock(&gvt->lock); return ret; } + + +/** + * intel_vgpu_reset_mmio - reset virtual MMIO space + * @vgpu: a vGPU + * + */ +void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu) +{ + struct intel_gvt *gvt = vgpu->gvt; + const struct intel_gvt_device_info *info = &gvt->device_info; + + memcpy(vgpu->mmio.vreg, gvt->firmware.mmio, info->mmio_size); + memcpy(vgpu->mmio.sreg, gvt->firmware.mmio, info->mmio_size); + + vgpu_vreg(vgpu, GEN6_GT_THREAD_STATUS_REG) = 0; + + /* set the bit 0:2(Core C-State ) to C0 */ + vgpu_vreg(vgpu, GEN6_GT_CORE_STATUS) = 0; +} + +/** + * intel_vgpu_init_mmio - init MMIO space + * @vgpu: a vGPU + * + * Returns: + * Zero on success, negative error code if failed + */ +int intel_vgpu_init_mmio(struct intel_vgpu *vgpu) +{ + const struct intel_gvt_device_info *info = &vgpu->gvt->device_info; + + vgpu->mmio.vreg = vzalloc(info->mmio_size * 2); + if (!vgpu->mmio.vreg) + return -ENOMEM; + + vgpu->mmio.sreg = vgpu->mmio.vreg + info->mmio_size; + + intel_vgpu_reset_mmio(vgpu); + + return 0; +} + +/** + * intel_vgpu_clean_mmio - clean MMIO space + * @vgpu: a vGPU + * + */ +void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu) +{ + vfree(vgpu->mmio.vreg); + vgpu->mmio.vreg = vgpu->mmio.sreg = NULL; +} diff --git a/drivers/gpu/drm/i915/gvt/mmio.h b/drivers/gpu/drm/i915/gvt/mmio.h index 87d5b5e366a3..3bc620f56f35 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.h +++ b/drivers/gpu/drm/i915/gvt/mmio.h @@ -86,6 +86,10 @@ struct intel_gvt_mmio_info *intel_gvt_find_mmio_info(struct intel_gvt *gvt, *offset; \ }) +int intel_vgpu_init_mmio(struct intel_vgpu *vgpu); +void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu); +void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu); + int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa); int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, u64 pa, diff --git a/drivers/gpu/drm/i915/gvt/opregion.c b/drivers/gpu/drm/i915/gvt/opregion.c index 81cd921770c6..d9fb41ab7119 100644 --- a/drivers/gpu/drm/i915/gvt/opregion.c +++ b/drivers/gpu/drm/i915/gvt/opregion.c @@ -36,9 +36,9 @@ static int init_vgpu_opregion(struct intel_vgpu *vgpu, u32 gpa) vgpu->id)) return -EINVAL; - vgpu_opregion(vgpu)->va = (void *)__get_free_pages(GFP_ATOMIC | - GFP_DMA32 | __GFP_ZERO, - INTEL_GVT_OPREGION_PORDER); + vgpu_opregion(vgpu)->va = (void *)__get_free_pages(GFP_KERNEL | + __GFP_ZERO, + get_order(INTEL_GVT_OPREGION_SIZE)); if (!vgpu_opregion(vgpu)->va) return -ENOMEM; @@ -97,7 +97,7 @@ void intel_vgpu_clean_opregion(struct intel_vgpu *vgpu) if (intel_gvt_host.hypervisor_type == INTEL_GVT_HYPERVISOR_XEN) { map_vgpu_opregion(vgpu, false); free_pages((unsigned long)vgpu_opregion(vgpu)->va, - INTEL_GVT_OPREGION_PORDER); + get_order(INTEL_GVT_OPREGION_SIZE)); vgpu_opregion(vgpu)->va = NULL; } diff --git a/drivers/gpu/drm/i915/gvt/reg.h b/drivers/gpu/drm/i915/gvt/reg.h index 0dfe789d8f02..fbd023a16f18 100644 --- a/drivers/gpu/drm/i915/gvt/reg.h +++ b/drivers/gpu/drm/i915/gvt/reg.h @@ -50,8 +50,7 @@ #define INTEL_GVT_OPREGION_PARM 0x204 #define INTEL_GVT_OPREGION_PAGES 2 -#define INTEL_GVT_OPREGION_PORDER 1 -#define INTEL_GVT_OPREGION_SIZE (2 * 4096) +#define INTEL_GVT_OPREGION_SIZE (INTEL_GVT_OPREGION_PAGES * PAGE_SIZE) #define VGT_SPRSTRIDE(pipe) _PIPE(pipe, _SPRA_STRIDE, _PLANE_STRIDE_2_B) diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c index fd2b026f7ecd..7ea68a75dc46 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.c +++ b/drivers/gpu/drm/i915/gvt/scheduler.c @@ -350,13 +350,15 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) { struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler; struct intel_vgpu_workload *workload; + struct intel_vgpu *vgpu; int event; mutex_lock(&gvt->lock); workload = scheduler->current_workload[ring_id]; + vgpu = workload->vgpu; - if (!workload->status && !workload->vgpu->resetting) { + if (!workload->status && !vgpu->resetting) { wait_event(workload->shadow_ctx_status_wq, !atomic_read(&workload->shadow_ctx_active)); @@ -364,8 +366,7 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) for_each_set_bit(event, workload->pending_events, INTEL_GVT_EVENT_MAX) - intel_vgpu_trigger_virtual_event(workload->vgpu, - event); + intel_vgpu_trigger_virtual_event(vgpu, event); } gvt_dbg_sched("ring id %d complete workload %p status %d\n", @@ -373,11 +374,10 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) scheduler->current_workload[ring_id] = NULL; - atomic_dec(&workload->vgpu->running_workload_num); - list_del_init(&workload->list); workload->complete(workload); + atomic_dec(&vgpu->running_workload_num); wake_up(&scheduler->workload_complete_wq); mutex_unlock(&gvt->lock); } @@ -459,11 +459,11 @@ complete: gvt_dbg_sched("will complete workload %p\n, status: %d\n", workload, workload->status); - complete_current_workload(gvt, ring_id); - if (workload->req) i915_gem_request_put(fetch_and_zero(&workload->req)); + complete_current_workload(gvt, ring_id); + if (need_force_wake) intel_uncore_forcewake_put(gvt->dev_priv, FORCEWAKE_ALL); diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index 536d2b9d5777..7295bc8e12fb 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -35,79 +35,6 @@ #include "gvt.h" #include "i915_pvinfo.h" -static void clean_vgpu_mmio(struct intel_vgpu *vgpu) -{ - vfree(vgpu->mmio.vreg); - vgpu->mmio.vreg = vgpu->mmio.sreg = NULL; -} - -int setup_vgpu_mmio(struct intel_vgpu *vgpu) -{ - struct intel_gvt *gvt = vgpu->gvt; - const struct intel_gvt_device_info *info = &gvt->device_info; - - if (vgpu->mmio.vreg) - memset(vgpu->mmio.vreg, 0, info->mmio_size * 2); - else { - vgpu->mmio.vreg = vzalloc(info->mmio_size * 2); - if (!vgpu->mmio.vreg) - return -ENOMEM; - } - - vgpu->mmio.sreg = vgpu->mmio.vreg + info->mmio_size; - - memcpy(vgpu->mmio.vreg, gvt->firmware.mmio, info->mmio_size); - memcpy(vgpu->mmio.sreg, gvt->firmware.mmio, info->mmio_size); - - vgpu_vreg(vgpu, GEN6_GT_THREAD_STATUS_REG) = 0; - - /* set the bit 0:2(Core C-State ) to C0 */ - vgpu_vreg(vgpu, GEN6_GT_CORE_STATUS) = 0; - return 0; -} - -static void setup_vgpu_cfg_space(struct intel_vgpu *vgpu, - struct intel_vgpu_creation_params *param) -{ - struct intel_gvt *gvt = vgpu->gvt; - const struct intel_gvt_device_info *info = &gvt->device_info; - u16 *gmch_ctl; - int i; - - memcpy(vgpu_cfg_space(vgpu), gvt->firmware.cfg_space, - info->cfg_space_size); - - if (!param->primary) { - vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] = - INTEL_GVT_PCI_CLASS_VGA_OTHER; - vgpu_cfg_space(vgpu)[PCI_CLASS_PROG] = - INTEL_GVT_PCI_CLASS_VGA_OTHER; - } - - /* Show guest that there isn't any stolen memory.*/ - gmch_ctl = (u16 *)(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_GMCH_CONTROL); - *gmch_ctl &= ~(BDW_GMCH_GMS_MASK << BDW_GMCH_GMS_SHIFT); - - intel_vgpu_write_pci_bar(vgpu, PCI_BASE_ADDRESS_2, - gvt_aperture_pa_base(gvt), true); - - vgpu_cfg_space(vgpu)[PCI_COMMAND] &= ~(PCI_COMMAND_IO - | PCI_COMMAND_MEMORY - | PCI_COMMAND_MASTER); - /* - * Clear the bar upper 32bit and let guest to assign the new value - */ - memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_1, 0, 4); - memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_3, 0, 4); - memset(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_OPREGION, 0, 4); - - for (i = 0; i < INTEL_GVT_MAX_BAR_NUM; i++) { - vgpu->cfg_space.bar[i].size = pci_resource_len( - gvt->dev_priv->drm.pdev, i * 2); - vgpu->cfg_space.bar[i].tracked = false; - } -} - void populate_pvinfo_page(struct intel_vgpu *vgpu) { /* setup the ballooning information */ @@ -177,7 +104,7 @@ int intel_gvt_init_vgpu_types(struct intel_gvt *gvt) if (low_avail / min_low == 0) break; gvt->types[i].low_gm_size = min_low; - gvt->types[i].high_gm_size = 3 * gvt->types[i].low_gm_size; + gvt->types[i].high_gm_size = max((min_low<<3), MB_TO_BYTES(384U)); gvt->types[i].fence = 4; gvt->types[i].max_instance = low_avail / min_low; gvt->types[i].avail_instance = gvt->types[i].max_instance; @@ -217,7 +144,7 @@ static void intel_gvt_update_vgpu_types(struct intel_gvt *gvt) */ low_gm_avail = MB_TO_BYTES(256) - HOST_LOW_GM_SIZE - gvt->gm.vgpu_allocated_low_gm_size; - high_gm_avail = MB_TO_BYTES(256) * 3 - HOST_HIGH_GM_SIZE - + high_gm_avail = MB_TO_BYTES(256) * 8UL - HOST_HIGH_GM_SIZE - gvt->gm.vgpu_allocated_high_gm_size; fence_avail = gvt_fence_sz(gvt) - HOST_FENCE - gvt->fence.vgpu_allocated_fence_num; @@ -268,7 +195,7 @@ void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu) intel_vgpu_clean_gtt(vgpu); intel_gvt_hypervisor_detach_vgpu(vgpu); intel_vgpu_free_resource(vgpu); - clean_vgpu_mmio(vgpu); + intel_vgpu_clean_mmio(vgpu); vfree(vgpu); intel_gvt_update_vgpu_types(gvt); @@ -300,11 +227,11 @@ static struct intel_vgpu *__intel_gvt_create_vgpu(struct intel_gvt *gvt, vgpu->gvt = gvt; bitmap_zero(vgpu->tlb_handle_pending, I915_NUM_ENGINES); - setup_vgpu_cfg_space(vgpu, param); + intel_vgpu_init_cfg_space(vgpu, param->primary); - ret = setup_vgpu_mmio(vgpu); + ret = intel_vgpu_init_mmio(vgpu); if (ret) - goto out_free_vgpu; + goto out_clean_idr; ret = intel_vgpu_alloc_resource(vgpu, param); if (ret) @@ -354,7 +281,9 @@ out_detach_hypervisor_vgpu: out_clean_vgpu_resource: intel_vgpu_free_resource(vgpu); out_clean_vgpu_mmio: - clean_vgpu_mmio(vgpu); + intel_vgpu_clean_mmio(vgpu); +out_clean_idr: + idr_remove(&gvt->vgpu_idr, vgpu->id); out_free_vgpu: vfree(vgpu); mutex_unlock(&gvt->lock); @@ -398,7 +327,75 @@ struct intel_vgpu *intel_gvt_create_vgpu(struct intel_gvt *gvt, } /** - * intel_gvt_reset_vgpu - reset a virtual GPU + * intel_gvt_reset_vgpu_locked - reset a virtual GPU by DMLR or GT reset + * @vgpu: virtual GPU + * @dmlr: vGPU Device Model Level Reset or GT Reset + * @engine_mask: engines to reset for GT reset + * + * This function is called when user wants to reset a virtual GPU through + * device model reset or GT reset. The caller should hold the gvt lock. + * + * vGPU Device Model Level Reset (DMLR) simulates the PCI level reset to reset + * the whole vGPU to default state as when it is created. This vGPU function + * is required both for functionary and security concerns.The ultimate goal + * of vGPU FLR is that reuse a vGPU instance by virtual machines. When we + * assign a vGPU to a virtual machine we must isse such reset first. + * + * Full GT Reset and Per-Engine GT Reset are soft reset flow for GPU engines + * (Render, Blitter, Video, Video Enhancement). It is defined by GPU Spec. + * Unlike the FLR, GT reset only reset particular resource of a vGPU per + * the reset request. Guest driver can issue a GT reset by programming the + * virtual GDRST register to reset specific virtual GPU engine or all + * engines. + * + * The parameter dev_level is to identify if we will do DMLR or GT reset. + * The parameter engine_mask is to specific the engines that need to be + * resetted. If value ALL_ENGINES is given for engine_mask, it means + * the caller requests a full GT reset that we will reset all virtual + * GPU engines. For FLR, engine_mask is ignored. + */ +void intel_gvt_reset_vgpu_locked(struct intel_vgpu *vgpu, bool dmlr, + unsigned int engine_mask) +{ + struct intel_gvt *gvt = vgpu->gvt; + struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler; + + gvt_dbg_core("------------------------------------------\n"); + gvt_dbg_core("resseting vgpu%d, dmlr %d, engine_mask %08x\n", + vgpu->id, dmlr, engine_mask); + vgpu->resetting = true; + + intel_vgpu_stop_schedule(vgpu); + /* + * The current_vgpu will set to NULL after stopping the + * scheduler when the reset is triggered by current vgpu. + */ + if (scheduler->current_vgpu == NULL) { + mutex_unlock(&gvt->lock); + intel_gvt_wait_vgpu_idle(vgpu); + mutex_lock(&gvt->lock); + } + + intel_vgpu_reset_execlist(vgpu, dmlr ? ALL_ENGINES : engine_mask); + + /* full GPU reset or device model level reset */ + if (engine_mask == ALL_ENGINES || dmlr) { + intel_vgpu_reset_gtt(vgpu, dmlr); + intel_vgpu_reset_resource(vgpu); + intel_vgpu_reset_mmio(vgpu); + populate_pvinfo_page(vgpu); + + if (dmlr) + intel_vgpu_reset_cfg_space(vgpu); + } + + vgpu->resetting = false; + gvt_dbg_core("reset vgpu%d done\n", vgpu->id); + gvt_dbg_core("------------------------------------------\n"); +} + +/** + * intel_gvt_reset_vgpu - reset a virtual GPU (Function Level) * @vgpu: virtual GPU * * This function is called when user wants to reset a virtual GPU. @@ -406,4 +403,7 @@ struct intel_vgpu *intel_gvt_create_vgpu(struct intel_gvt *gvt, */ void intel_gvt_reset_vgpu(struct intel_vgpu *vgpu) { + mutex_lock(&vgpu->gvt->lock); + intel_gvt_reset_vgpu_locked(vgpu, true, 0); + mutex_unlock(&vgpu->gvt->lock); } diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 9d7b5a8c8dea..fa69d72fdcb9 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -159,8 +159,35 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) seq_printf(m, " (%sgtt offset: %08llx, size: %08llx", i915_vma_is_ggtt(vma) ? "g" : "pp", vma->node.start, vma->node.size); - if (i915_vma_is_ggtt(vma)) - seq_printf(m, ", type: %u", vma->ggtt_view.type); + if (i915_vma_is_ggtt(vma)) { + switch (vma->ggtt_view.type) { + case I915_GGTT_VIEW_NORMAL: + seq_puts(m, ", normal"); + break; + + case I915_GGTT_VIEW_PARTIAL: + seq_printf(m, ", partial [%08llx+%x]", + vma->ggtt_view.partial.offset << PAGE_SHIFT, + vma->ggtt_view.partial.size << PAGE_SHIFT); + break; + + case I915_GGTT_VIEW_ROTATED: + seq_printf(m, ", rotated [(%ux%u, stride=%u, offset=%u), (%ux%u, stride=%u, offset=%u)]", + vma->ggtt_view.rotated.plane[0].width, + vma->ggtt_view.rotated.plane[0].height, + vma->ggtt_view.rotated.plane[0].stride, + vma->ggtt_view.rotated.plane[0].offset, + vma->ggtt_view.rotated.plane[1].width, + vma->ggtt_view.rotated.plane[1].height, + vma->ggtt_view.rotated.plane[1].stride, + vma->ggtt_view.rotated.plane[1].offset); + break; + + default: + MISSING_CASE(vma->ggtt_view.type); + break; + } + } if (vma->fence) seq_printf(m, " , fence: %d%s", vma->fence->id, @@ -2325,10 +2352,40 @@ static int i915_llc(struct seq_file *m, void *data) return 0; } +static int i915_huc_load_status_info(struct seq_file *m, void *data) +{ + struct drm_i915_private *dev_priv = node_to_i915(m->private); + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + + if (!HAS_HUC_UCODE(dev_priv)) + return 0; + + seq_puts(m, "HuC firmware status:\n"); + seq_printf(m, "\tpath: %s\n", huc_fw->path); + seq_printf(m, "\tfetch: %s\n", + intel_uc_fw_status_repr(huc_fw->fetch_status)); + seq_printf(m, "\tload: %s\n", + intel_uc_fw_status_repr(huc_fw->load_status)); + seq_printf(m, "\tversion wanted: %d.%d\n", + huc_fw->major_ver_wanted, huc_fw->minor_ver_wanted); + seq_printf(m, "\tversion found: %d.%d\n", + huc_fw->major_ver_found, huc_fw->minor_ver_found); + seq_printf(m, "\theader: offset is %d; size = %d\n", + huc_fw->header_offset, huc_fw->header_size); + seq_printf(m, "\tuCode: offset is %d; size = %d\n", + huc_fw->ucode_offset, huc_fw->ucode_size); + seq_printf(m, "\tRSA: offset is %d; size = %d\n", + huc_fw->rsa_offset, huc_fw->rsa_size); + + seq_printf(m, "\nHuC status 0x%08x:\n", I915_READ(HUC_STATUS2)); + + return 0; +} + static int i915_guc_load_status_info(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; u32 tmp, i; if (!HAS_GUC_UCODE(dev_priv)) @@ -2336,15 +2393,15 @@ static int i915_guc_load_status_info(struct seq_file *m, void *data) seq_printf(m, "GuC firmware status:\n"); seq_printf(m, "\tpath: %s\n", - guc_fw->guc_fw_path); + guc_fw->path); seq_printf(m, "\tfetch: %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status)); seq_printf(m, "\tload: %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->load_status)); seq_printf(m, "\tversion wanted: %d.%d\n", - guc_fw->guc_fw_major_wanted, guc_fw->guc_fw_minor_wanted); + guc_fw->major_ver_wanted, guc_fw->minor_ver_wanted); seq_printf(m, "\tversion found: %d.%d\n", - guc_fw->guc_fw_major_found, guc_fw->guc_fw_minor_found); + guc_fw->major_ver_found, guc_fw->minor_ver_found); seq_printf(m, "\theader: offset is %d; size = %d\n", guc_fw->header_offset, guc_fw->header_size); seq_printf(m, "\tuCode: offset is %d; size = %d\n", @@ -2532,6 +2589,29 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_control_fops, i915_guc_log_control_get, i915_guc_log_control_set, "%lld\n"); +static const char *psr2_live_status(u32 val) +{ + static const char * const live_status[] = { + "IDLE", + "CAPTURE", + "CAPTURE_FS", + "SLEEP", + "BUFON_FW", + "ML_UP", + "SU_STANDBY", + "FAST_SLEEP", + "DEEP_SLEEP", + "BUF_ON", + "TG_ON" + }; + + val = (val & EDP_PSR2_STATUS_STATE_MASK) >> EDP_PSR2_STATUS_STATE_SHIFT; + if (val < ARRAY_SIZE(live_status)) + return live_status[val]; + + return "unknown"; +} + static int i915_edp_psr_status(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); @@ -2606,6 +2686,12 @@ static int i915_edp_psr_status(struct seq_file *m, void *data) seq_printf(m, "Performance_Counter: %u\n", psrperf); } + if (dev_priv->psr.psr2_support) { + u32 psr2 = I915_READ(EDP_PSR2_STATUS_CTL); + + seq_printf(m, "EDP_PSR2_STATUS_CTL: %x [%s]\n", + psr2, psr2_live_status(psr2)); + } mutex_unlock(&dev_priv->psr.lock); intel_runtime_pm_put(dev_priv); @@ -4553,6 +4639,7 @@ static const struct drm_info_list i915_debugfs_list[] = { {"i915_guc_info", i915_guc_info, 0}, {"i915_guc_load_status", i915_guc_load_status_info, 0}, {"i915_guc_log_dump", i915_guc_log_dump, 0}, + {"i915_huc_load_status", i915_huc_load_status_info, 0}, {"i915_frequency_info", i915_frequency_info, 0}, {"i915_hangcheck_info", i915_hangcheck_info, 0}, {"i915_drpc_info", i915_drpc_info, 0}, diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 4d22b4b479b8..4ae69ebe166e 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -49,6 +49,7 @@ #include "i915_trace.h" #include "i915_vgpu.h" #include "intel_drv.h" +#include "intel_uc.h" static struct drm_driver driver; @@ -315,6 +316,12 @@ static int i915_getparam(struct drm_device *dev, void *data, case I915_PARAM_MIN_EU_IN_POOL: value = INTEL_INFO(dev_priv)->sseu.min_eu_in_pool; break; + case I915_PARAM_HUC_STATUS: + /* The register is already force-woken. We dont need + * any rpm here + */ + value = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; + break; case I915_PARAM_MMAP_GTT_VERSION: /* Though we've started our numbering from 1, and so class all * earlier versions as 0, in effect their value is undefined as @@ -599,6 +606,7 @@ static int i915_load_modeset_init(struct drm_device *dev) if (ret) goto cleanup_irq; + intel_huc_init(dev_priv); intel_guc_init(dev_priv); ret = i915_gem_init(dev_priv); @@ -627,6 +635,7 @@ cleanup_gem: i915_gem_fini(dev_priv); cleanup_irq: intel_guc_fini(dev_priv); + intel_huc_fini(dev_priv); drm_irq_uninstall(dev); intel_teardown_gmbus(dev_priv); cleanup_csr: @@ -1114,7 +1123,7 @@ static void i915_driver_register(struct drm_i915_private *dev_priv) /* Reveal our presence to userspace */ if (drm_dev_register(dev, 0) == 0) { i915_debugfs_register(dev_priv); - i915_guc_register(dev_priv); + i915_guc_log_register(dev_priv); i915_setup_sysfs(dev_priv); /* Depends on sysfs having been initialized */ @@ -1158,7 +1167,7 @@ static void i915_driver_unregister(struct drm_i915_private *dev_priv) i915_perf_unregister(dev_priv); i915_teardown_sysfs(dev_priv); - i915_guc_unregister(dev_priv); + i915_guc_log_unregister(dev_priv); i915_debugfs_unregister(dev_priv); drm_dev_unregister(&dev_priv->drm); @@ -1314,6 +1323,7 @@ void i915_driver_unload(struct drm_device *dev) drain_workqueue(dev_priv->wq); intel_guc_fini(dev_priv); + intel_huc_fini(dev_priv); i915_gem_fini(dev_priv); intel_fbc_cleanup_cfb(dev_priv); @@ -1471,7 +1481,7 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) intel_display_set_init_power(dev_priv, false); - fw_csr = !IS_BROXTON(dev_priv) && + fw_csr = !IS_GEN9_LP(dev_priv) && suspend_to_idle(dev_priv) && dev_priv->csr.dmc_payload; /* * In case of firmware assisted context save/restore don't manually @@ -1484,7 +1494,7 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) intel_power_domains_suspend(dev_priv); ret = 0; - if (IS_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) bxt_enable_dc9(dev_priv); else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) hsw_enable_pc8(dev_priv); @@ -1692,7 +1702,7 @@ static int i915_drm_resume_early(struct drm_device *dev) intel_uncore_early_sanitize(dev_priv, true); - if (IS_BROXTON(dev_priv)) { + if (IS_GEN9_LP(dev_priv)) { if (!dev_priv->suspended_to_idle) gen9_sanitize_dc_state(dev_priv); bxt_disable_dc9(dev_priv); @@ -1702,7 +1712,7 @@ static int i915_drm_resume_early(struct drm_device *dev) intel_uncore_sanitize(dev_priv); - if (IS_BROXTON(dev_priv) || + if (IS_GEN9_LP(dev_priv) || !(dev_priv->suspended_to_idle && dev_priv->csr.dmc_payload)) intel_power_domains_init_hw(dev_priv, true); @@ -1728,25 +1738,9 @@ static int i915_resume_switcheroo(struct drm_device *dev) return i915_drm_resume(dev); } -static void disable_engines_irq(struct drm_i915_private *dev_priv) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - - /* Ensure irq handler finishes, and not run again. */ - disable_irq(dev_priv->drm.irq); - for_each_engine(engine, dev_priv, id) - tasklet_kill(&engine->irq_tasklet); -} - -static void enable_engines_irq(struct drm_i915_private *dev_priv) -{ - enable_irq(dev_priv->drm.irq); -} - /** * i915_reset - reset chip after a hang - * @dev: drm device to reset + * @dev_priv: device private to reset * * Reset the chip. Useful if a hang is detected. Marks the device as wedged * on failure. @@ -1776,12 +1770,15 @@ void i915_reset(struct drm_i915_private *dev_priv) error->reset_count++; pr_notice("drm/i915: Resetting chip after gpu hang\n"); - i915_gem_reset_prepare(dev_priv); + disable_irq(dev_priv->drm.irq); + ret = i915_gem_reset_prepare(dev_priv); + if (ret) { + DRM_ERROR("GPU recovery failed\n"); + intel_gpu_reset(dev_priv, ALL_ENGINES); + goto error; + } - disable_engines_irq(dev_priv); ret = intel_gpu_reset(dev_priv, ALL_ENGINES); - enable_engines_irq(dev_priv); - if (ret) { if (ret != -ENODEV) DRM_ERROR("Failed to reset chip: %i\n", ret); @@ -1816,6 +1813,7 @@ void i915_reset(struct drm_i915_private *dev_priv) i915_queue_hangcheck(dev_priv); wakeup: + enable_irq(dev_priv->drm.irq); wake_up_bit(&error->flags, I915_RESET_IN_PROGRESS); return; @@ -2326,7 +2324,7 @@ static int intel_runtime_suspend(struct device *kdev) intel_runtime_pm_disable_interrupts(dev_priv); ret = 0; - if (IS_BROXTON(dev_priv)) { + if (IS_GEN9_LP(dev_priv)) { bxt_display_core_uninit(dev_priv); bxt_enable_dc9(dev_priv); } else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { @@ -2411,7 +2409,7 @@ static int intel_runtime_resume(struct device *kdev) if (IS_GEN6(dev_priv)) intel_init_pch_refclk(dev_priv); - if (IS_BROXTON(dev_priv)) { + if (IS_GEN9_LP(dev_priv)) { bxt_disable_dc9(dev_priv); bxt_display_core_init(dev_priv, true); if (dev_priv->csr.dmc_payload && @@ -2549,8 +2547,8 @@ static const struct drm_ioctl_desc i915_ioctls[] = { DRM_IOCTL_DEF_DRV(I915_GEM_MMAP_GTT, i915_gem_mmap_gtt_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_SET_DOMAIN, i915_gem_set_domain_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_SW_FINISH, i915_gem_sw_finish_ioctl, DRM_RENDER_ALLOW), - DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling, DRM_RENDER_ALLOW), - DRM_IOCTL_DEF_DRV(I915_GEM_GET_TILING, i915_gem_get_tiling, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(I915_GEM_GET_TILING, i915_gem_get_tiling_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_GET_APERTURE, i915_gem_get_aperture_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GET_PIPE_FROM_CRTC_ID, intel_get_pipe_from_crtc_id, 0), DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_RENDER_ALLOW), diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 24eb16231938..244628065f94 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -79,8 +79,8 @@ #define DRIVER_NAME "i915" #define DRIVER_DESC "Intel Graphics" -#define DRIVER_DATE "20170109" -#define DRIVER_TIMESTAMP 1483953121 +#define DRIVER_DATE "20170123" +#define DRIVER_TIMESTAMP 1485156432 #undef WARN_ON /* Many gcc seem to no see through this and fall over :( */ @@ -1070,6 +1070,8 @@ struct intel_fbc { struct work_struct underrun_work; struct intel_fbc_state_cache { + struct i915_vma *vma; + struct { unsigned int mode_flags; uint32_t hsw_bdw_pixel_rate; @@ -1083,15 +1085,14 @@ struct intel_fbc { } plane; struct { - u64 ilk_ggtt_offset; const struct drm_format_info *format; unsigned int stride; - int fence_reg; - unsigned int tiling_mode; } fb; } state_cache; struct intel_fbc_reg_params { + struct i915_vma *vma; + struct { enum pipe pipe; enum plane plane; @@ -1099,10 +1100,8 @@ struct intel_fbc { } crtc; struct { - u64 ggtt_offset; const struct drm_format_info *format; unsigned int stride; - int fence_reg; } fb; int cfb_size; @@ -1155,6 +1154,9 @@ struct i915_psr { bool psr2_support; bool aux_frame_sync; bool link_standby; + bool y_cord_support; + bool colorimetry_support; + bool alpm; }; enum intel_pch { @@ -1810,6 +1812,7 @@ struct intel_pipe_crc { enum intel_pipe_crc_source source; int head, tail; wait_queue_head_t wq; + int skipped; }; struct i915_frontbuffer_tracking { @@ -2070,6 +2073,7 @@ struct drm_i915_private { struct intel_gvt *gvt; + struct intel_huc huc; struct intel_guc guc; struct intel_csr csr; @@ -2844,6 +2848,7 @@ intel_info(const struct drm_i915_private *dev_priv) #define HAS_GUC(dev_priv) ((dev_priv)->info.has_guc) #define HAS_GUC_UCODE(dev_priv) (HAS_GUC(dev_priv)) #define HAS_GUC_SCHED(dev_priv) (HAS_GUC(dev_priv)) +#define HAS_HUC_UCODE(dev_priv) (HAS_GUC(dev_priv)) #define HAS_RESOURCE_STREAMER(dev_priv) ((dev_priv)->info.has_resource_streamer) @@ -3102,10 +3107,10 @@ int i915_gem_throttle_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int i915_gem_madvise_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); -int i915_gem_set_tiling(struct drm_device *dev, void *data, - struct drm_file *file_priv); -int i915_gem_get_tiling(struct drm_device *dev, void *data, - struct drm_file *file_priv); +int i915_gem_set_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); +int i915_gem_get_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); void i915_gem_init_userptr(struct drm_i915_private *dev_priv); int i915_gem_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *file); @@ -3324,7 +3329,7 @@ static inline u32 i915_reset_count(struct i915_gpu_error *error) return READ_ONCE(error->reset_count); } -void i915_gem_reset_prepare(struct drm_i915_private *dev_priv); +int i915_gem_reset_prepare(struct drm_i915_private *dev_priv); void i915_gem_reset_finish(struct drm_i915_private *dev_priv); void i915_gem_set_wedged(struct drm_i915_private *dev_priv); void i915_gem_clflush_object(struct drm_i915_gem_object *obj, bool force); @@ -3361,11 +3366,6 @@ int i915_gem_object_attach_phys(struct drm_i915_gem_object *obj, int i915_gem_open(struct drm_device *dev, struct drm_file *file); void i915_gem_release(struct drm_device *dev, struct drm_file *file); -u64 i915_gem_get_ggtt_size(struct drm_i915_private *dev_priv, u64 size, - int tiling_mode); -u64 i915_gem_get_ggtt_alignment(struct drm_i915_private *dev_priv, u64 size, - int tiling_mode, bool fenced); - int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, enum i915_cache_level cache_level); @@ -3375,36 +3375,12 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev, struct dma_buf *i915_gem_prime_export(struct drm_device *dev, struct drm_gem_object *gem_obj, int flags); -struct i915_vma * -i915_gem_obj_to_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view); - -struct i915_vma * -i915_gem_obj_lookup_or_create_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view); - static inline struct i915_hw_ppgtt * i915_vm_to_ppgtt(struct i915_address_space *vm) { return container_of(vm, struct i915_hw_ppgtt, base); } -static inline struct i915_vma * -i915_gem_object_to_ggtt(struct drm_i915_gem_object *obj, - const struct i915_ggtt_view *view) -{ - return i915_gem_obj_to_vma(obj, &to_i915(obj->base.dev)->ggtt.base, view); -} - -static inline unsigned long -i915_gem_object_ggtt_offset(struct drm_i915_gem_object *o, - const struct i915_ggtt_view *view) -{ - return i915_ggtt_offset(i915_gem_object_to_ggtt(o, view)); -} - /* i915_gem_fence_reg.c */ int __must_check i915_vma_get_fence(struct i915_vma *vma); int __must_check i915_vma_put_fence(struct i915_vma *vma); @@ -3472,8 +3448,9 @@ int __must_check i915_gem_evict_something(struct i915_address_space *vm, unsigned cache_level, u64 start, u64 end, unsigned flags); -int __must_check i915_gem_evict_for_vma(struct i915_vma *vma, - unsigned int flags); +int __must_check i915_gem_evict_for_node(struct i915_address_space *vm, + struct drm_mm_node *node, + unsigned int flags); int i915_gem_evict_vm(struct i915_address_space *vm, bool do_idle); /* belongs in i915_gem_gtt.h */ @@ -3507,7 +3484,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv /* i915_gem_internal.c */ struct drm_i915_gem_object * i915_gem_object_create_internal(struct drm_i915_private *dev_priv, - unsigned int size); + phys_addr_t size); /* i915_gem_shrinker.c */ unsigned long i915_gem_shrink(struct drm_i915_private *dev_priv, @@ -3532,6 +3509,11 @@ static inline bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_objec i915_gem_object_is_tiled(obj); } +u32 i915_gem_fence_size(struct drm_i915_private *dev_priv, u32 size, + unsigned int tiling, unsigned int stride); +u32 i915_gem_fence_alignment(struct drm_i915_private *dev_priv, u32 size, + unsigned int tiling, unsigned int stride); + /* i915_debugfs.c */ #ifdef CONFIG_DEBUG_FS int i915_debugfs_register(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index dc00d9ae6d92..a07b62732923 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1696,12 +1696,7 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data, static unsigned int tile_row_pages(struct drm_i915_gem_object *obj) { - u64 size; - - size = i915_gem_object_get_stride(obj); - size *= i915_gem_object_get_tiling(obj) == I915_TILING_Y ? 32 : 8; - - return size >> PAGE_SHIFT; + return i915_gem_object_get_tile_row_size(obj) >> PAGE_SHIFT; } /** @@ -1754,6 +1749,29 @@ int i915_gem_mmap_gtt_version(void) return 1; } +static inline struct i915_ggtt_view +compute_partial_view(struct drm_i915_gem_object *obj, + pgoff_t page_offset, + unsigned int chunk) +{ + struct i915_ggtt_view view; + + if (i915_gem_object_is_tiled(obj)) + chunk = roundup(chunk, tile_row_pages(obj)); + + view.type = I915_GGTT_VIEW_PARTIAL; + view.partial.offset = rounddown(page_offset, chunk); + view.partial.size = + min_t(unsigned int, chunk, + (obj->base.size >> PAGE_SHIFT) - view.partial.offset); + + /* If the partial covers the entire object, just create a normal VMA. */ + if (chunk >= obj->base.size >> PAGE_SHIFT) + view.type = I915_GGTT_VIEW_NORMAL; + + return view; +} + /** * i915_gem_fault - fault a page into the GTT * @area: CPU VMA in question @@ -1830,26 +1848,9 @@ int i915_gem_fault(struct vm_area_struct *area, struct vm_fault *vmf) /* Now pin it into the GTT as needed */ vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, flags); if (IS_ERR(vma)) { - struct i915_ggtt_view view; - unsigned int chunk_size; - /* Use a partial view if it is bigger than available space */ - chunk_size = MIN_CHUNK_PAGES; - if (i915_gem_object_is_tiled(obj)) - chunk_size = roundup(chunk_size, tile_row_pages(obj)); - - memset(&view, 0, sizeof(view)); - view.type = I915_GGTT_VIEW_PARTIAL; - view.params.partial.offset = rounddown(page_offset, chunk_size); - view.params.partial.size = - min_t(unsigned int, chunk_size, - vma_pages(area) - view.params.partial.offset); - - /* If the partial covers the entire object, just create a - * normal VMA. - */ - if (chunk_size >= obj->base.size >> PAGE_SHIFT) - view.type = I915_GGTT_VIEW_NORMAL; + struct i915_ggtt_view view = + compute_partial_view(obj, page_offset, MIN_CHUNK_PAGES); /* Userspace is now writing through an untracked VMA, abandon * all hope that the hardware is able to track future writes. @@ -1878,7 +1879,7 @@ int i915_gem_fault(struct vm_area_struct *area, struct vm_fault *vmf) /* Finally, remap it using the new GTT offset */ ret = remap_io_mapping(area, - area->vm_start + (vma->ggtt_view.params.partial.offset << PAGE_SHIFT), + area->vm_start + (vma->ggtt_view.partial.offset << PAGE_SHIFT), (ggtt->mappable_base + vma->node.start) >> PAGE_SHIFT, min_t(u64, vma->size, area->vm_end - area->vm_start), &ggtt->mappable); @@ -2021,69 +2022,6 @@ void i915_gem_runtime_suspend(struct drm_i915_private *dev_priv) } } -/** - * i915_gem_get_ggtt_size - return required global GTT size for an object - * @dev_priv: i915 device - * @size: object size - * @tiling_mode: tiling mode - * - * Return the required global GTT size for an object, taking into account - * potential fence register mapping. - */ -u64 i915_gem_get_ggtt_size(struct drm_i915_private *dev_priv, - u64 size, int tiling_mode) -{ - u64 ggtt_size; - - GEM_BUG_ON(size == 0); - - if (INTEL_GEN(dev_priv) >= 4 || - tiling_mode == I915_TILING_NONE) - return size; - - /* Previous chips need a power-of-two fence region when tiling */ - if (IS_GEN3(dev_priv)) - ggtt_size = 1024*1024; - else - ggtt_size = 512*1024; - - while (ggtt_size < size) - ggtt_size <<= 1; - - return ggtt_size; -} - -/** - * i915_gem_get_ggtt_alignment - return required global GTT alignment - * @dev_priv: i915 device - * @size: object size - * @tiling_mode: tiling mode - * @fenced: is fenced alignment required or not - * - * Return the required global GTT alignment for an object, taking into account - * potential fence register mapping. - */ -u64 i915_gem_get_ggtt_alignment(struct drm_i915_private *dev_priv, u64 size, - int tiling_mode, bool fenced) -{ - GEM_BUG_ON(size == 0); - - /* - * Minimum alignment is 4k (GTT page size), but might be greater - * if a fence register is needed for the object. - */ - if (INTEL_GEN(dev_priv) >= 4 || - (!fenced && (IS_G33(dev_priv) || IS_PINEVIEW(dev_priv))) || - tiling_mode == I915_TILING_NONE) - return 4096; - - /* - * Previous chips need to be aligned to the size of the smallest - * fence register that can contain the object. - */ - return i915_gem_get_ggtt_size(dev_priv, size, tiling_mode); -} - static int i915_gem_object_create_mmap_offset(struct drm_i915_gem_object *obj) { struct drm_i915_private *dev_priv = to_i915(obj->base.dev); @@ -2666,13 +2604,52 @@ i915_gem_find_active_request(struct intel_engine_cs *engine) if (__i915_gem_request_completed(request)) continue; + GEM_BUG_ON(request->engine != engine); return request; } return NULL; } -static void reset_request(struct drm_i915_gem_request *request) +static bool engine_stalled(struct intel_engine_cs *engine) +{ + if (!engine->hangcheck.stalled) + return false; + + /* Check for possible seqno movement after hang declaration */ + if (engine->hangcheck.seqno != intel_engine_get_seqno(engine)) { + DRM_DEBUG_DRIVER("%s pardoned\n", engine->name); + return false; + } + + return true; +} + +int i915_gem_reset_prepare(struct drm_i915_private *dev_priv) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + int err = 0; + + /* Ensure irq handler finishes, and not run again. */ + for_each_engine(engine, dev_priv, id) { + struct drm_i915_gem_request *request; + + tasklet_kill(&engine->irq_tasklet); + + if (engine_stalled(engine)) { + request = i915_gem_find_active_request(engine); + if (request && request->fence.error == -EIO) + err = -EIO; /* Previous reset failed! */ + } + } + + i915_gem_revoke_fences(dev_priv); + + return err; +} + +static void skip_request(struct drm_i915_gem_request *request) { void *vaddr = request->ring->vaddr; u32 head; @@ -2687,20 +2664,74 @@ static void reset_request(struct drm_i915_gem_request *request) head = 0; } memset(vaddr + head, 0, request->postfix - head); + + dma_fence_set_error(&request->fence, -EIO); } -void i915_gem_reset_prepare(struct drm_i915_private *dev_priv) +static void engine_skip_context(struct drm_i915_gem_request *request) { - i915_gem_revoke_fences(dev_priv); + struct intel_engine_cs *engine = request->engine; + struct i915_gem_context *hung_ctx = request->ctx; + struct intel_timeline *timeline; + unsigned long flags; + + timeline = i915_gem_context_lookup_timeline(hung_ctx, engine); + + spin_lock_irqsave(&engine->timeline->lock, flags); + spin_lock(&timeline->lock); + + list_for_each_entry_continue(request, &engine->timeline->requests, link) + if (request->ctx == hung_ctx) + skip_request(request); + + list_for_each_entry(request, &timeline->requests, link) + skip_request(request); + + spin_unlock(&timeline->lock); + spin_unlock_irqrestore(&engine->timeline->lock, flags); +} + +/* Returns true if the request was guilty of hang */ +static bool i915_gem_reset_request(struct drm_i915_gem_request *request) +{ + /* Read once and return the resolution */ + const bool guilty = engine_stalled(request->engine); + + /* The guilty request will get skipped on a hung engine. + * + * Users of client default contexts do not rely on logical + * state preserved between batches so it is safe to execute + * queued requests following the hang. Non default contexts + * rely on preserved state, so skipping a batch loses the + * evolution of the state and it needs to be considered corrupted. + * Executing more queued batches on top of corrupted state is + * risky. But we take the risk by trying to advance through + * the queued requests in order to make the client behaviour + * more predictable around resets, by not throwing away random + * amount of batches it has prepared for execution. Sophisticated + * clients can use gem_reset_stats_ioctl and dma fence status + * (exported via sync_file info ioctl on explicit fences) to observe + * when it loses the context state and should rebuild accordingly. + * + * The context ban, and ultimately the client ban, mechanism are safety + * valves if client submission ends up resulting in nothing more than + * subsequent hangs. + */ + + if (guilty) { + i915_gem_context_mark_guilty(request->ctx); + skip_request(request); + } else { + i915_gem_context_mark_innocent(request->ctx); + dma_fence_set_error(&request->fence, -EAGAIN); + } + + return guilty; } static void i915_gem_reset_engine(struct intel_engine_cs *engine) { struct drm_i915_gem_request *request; - struct i915_gem_context *hung_ctx; - struct intel_timeline *timeline; - unsigned long flags; - bool ring_hung; if (engine->irq_seqno_barrier) engine->irq_seqno_barrier(engine); @@ -2709,22 +2740,7 @@ static void i915_gem_reset_engine(struct intel_engine_cs *engine) if (!request) return; - hung_ctx = request->ctx; - - ring_hung = engine->hangcheck.stalled; - if (engine->hangcheck.seqno != intel_engine_get_seqno(engine)) { - DRM_DEBUG_DRIVER("%s pardoned, was guilty? %s\n", - engine->name, - yesno(ring_hung)); - ring_hung = false; - } - - if (ring_hung) - i915_gem_context_mark_guilty(hung_ctx); - else - i915_gem_context_mark_innocent(hung_ctx); - - if (!ring_hung) + if (!i915_gem_reset_request(request)) return; DRM_DEBUG_DRIVER("resetting %s to restart from tail of request 0x%x\n", @@ -2734,34 +2750,8 @@ static void i915_gem_reset_engine(struct intel_engine_cs *engine) engine->reset_hw(engine, request); /* If this context is now banned, skip all of its pending requests. */ - if (!i915_gem_context_is_banned(hung_ctx)) - return; - - /* Users of the default context do not rely on logical state - * preserved between batches. They have to emit full state on - * every batch and so it is safe to execute queued requests following - * the hang. - * - * Other contexts preserve state, now corrupt. We want to skip all - * queued requests that reference the corrupt context. - */ - if (i915_gem_context_is_default(hung_ctx)) - return; - - timeline = i915_gem_context_lookup_timeline(hung_ctx, engine); - - spin_lock_irqsave(&engine->timeline->lock, flags); - spin_lock(&timeline->lock); - - list_for_each_entry_continue(request, &engine->timeline->requests, link) - if (request->ctx == hung_ctx) - reset_request(request); - - list_for_each_entry(request, &timeline->requests, link) - reset_request(request); - - spin_unlock(&timeline->lock); - spin_unlock_irqrestore(&engine->timeline->lock, flags); + if (i915_gem_context_is_banned(request->ctx)) + engine_skip_context(request); } void i915_gem_reset_finish(struct drm_i915_private *dev_priv) @@ -2788,12 +2778,16 @@ void i915_gem_reset_finish(struct drm_i915_private *dev_priv) static void nop_submit_request(struct drm_i915_gem_request *request) { + dma_fence_set_error(&request->fence, -EIO); i915_gem_request_submit(request); intel_engine_init_global_seqno(request->engine, request->global_seqno); } -static void i915_gem_cleanup_engine(struct intel_engine_cs *engine) +static void engine_set_wedged(struct intel_engine_cs *engine) { + struct drm_i915_gem_request *request; + unsigned long flags; + /* We need to be sure that no thread is running the old callback as * we install the nop handler (otherwise we would submit a request * to hardware that will never complete). In order to prevent this @@ -2802,6 +2796,12 @@ static void i915_gem_cleanup_engine(struct intel_engine_cs *engine) */ engine->submit_request = nop_submit_request; + /* Mark all executing requests as skipped */ + spin_lock_irqsave(&engine->timeline->lock, flags); + list_for_each_entry(request, &engine->timeline->requests, link) + dma_fence_set_error(&request->fence, -EIO); + spin_unlock_irqrestore(&engine->timeline->lock, flags); + /* Mark all pending requests as complete so that any concurrent * (lockless) lookup doesn't try and wait upon the request as we * reset it. @@ -2837,7 +2837,7 @@ static int __i915_gem_set_wedged_BKL(void *data) enum intel_engine_id id; for_each_engine(engine, i915, id) - i915_gem_cleanup_engine(engine); + engine_set_wedged(engine); return 0; } @@ -3397,7 +3397,7 @@ int i915_gem_set_caching_ioctl(struct drm_device *dev, void *data, struct drm_i915_gem_caching *args = data; struct drm_i915_gem_object *obj; enum i915_cache_level level; - int ret; + int ret = 0; switch (args->caching) { case I915_CACHING_NONE: @@ -3422,20 +3422,29 @@ int i915_gem_set_caching_ioctl(struct drm_device *dev, void *data, return -EINVAL; } - ret = i915_mutex_lock_interruptible(dev); + obj = i915_gem_object_lookup(file, args->handle); + if (!obj) + return -ENOENT; + + if (obj->cache_level == level) + goto out; + + ret = i915_gem_object_wait(obj, + I915_WAIT_INTERRUPTIBLE, + MAX_SCHEDULE_TIMEOUT, + to_rps_client(file)); if (ret) - return ret; + goto out; - obj = i915_gem_object_lookup(file, args->handle); - if (!obj) { - ret = -ENOENT; - goto unlock; - } + ret = i915_mutex_lock_interruptible(dev); + if (ret) + goto out; ret = i915_gem_object_set_cache_level(obj, level); - i915_gem_object_put(obj); -unlock: mutex_unlock(&dev->struct_mutex); + +out: + i915_gem_object_put(obj); return ret; } @@ -3485,7 +3494,7 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj, * try to preserve the existing ABI). */ vma = ERR_PTR(-ENOSPC); - if (view->type == I915_GGTT_VIEW_NORMAL) + if (!view || view->type == I915_GGTT_VIEW_NORMAL) vma = i915_gem_object_ggtt_pin(obj, view, 0, alignment, PIN_MAPPABLE | PIN_NONBLOCK); if (IS_ERR(vma)) { @@ -3544,11 +3553,10 @@ i915_gem_object_unpin_from_display_plane(struct i915_vma *vma) return; if (--vma->obj->pin_display == 0) - vma->display_alignment = 0; + vma->display_alignment = I915_GTT_MIN_ALIGNMENT; /* Bump the LRU to try and avoid premature eviction whilst flipping */ - if (!i915_vma_is_active(vma)) - list_move_tail(&vma->vm_link, &vma->vm->inactive_list); + i915_gem_object_bump_inactive_ggtt(vma->obj); i915_vma_unpin(vma); } @@ -3679,8 +3687,8 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, lockdep_assert_held(&obj->base.dev->struct_mutex); - vma = i915_gem_obj_lookup_or_create_vma(obj, vm, view); - if (IS_ERR(vma)) + vma = i915_vma_instance(obj, vm, view); + if (unlikely(IS_ERR(vma))) return vma; if (i915_vma_misplaced(vma, size, alignment, flags)) { @@ -3689,10 +3697,6 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, return ERR_PTR(-ENOSPC); if (flags & PIN_MAPPABLE) { - u32 fence_size; - - fence_size = i915_gem_get_ggtt_size(dev_priv, vma->size, - i915_gem_object_get_tiling(obj)); /* If the required space is larger than the available * aperture, we will not able to find a slot for the * object and unbinding the object now will be in @@ -3700,7 +3704,7 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, * the object in and out of the Global GTT and * waste a lot of cycles under the mutex. */ - if (fence_size > dev_priv->ggtt.mappable_end) + if (vma->fence_size > dev_priv->ggtt.mappable_end) return ERR_PTR(-E2BIG); /* If NONBLOCK is set the caller is optimistically @@ -3719,7 +3723,7 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, * we could try to minimise harm to others. */ if (flags & PIN_NONBLOCK && - fence_size > dev_priv->ggtt.mappable_end / 2) + vma->fence_size > dev_priv->ggtt.mappable_end / 2) return ERR_PTR(-ENOSPC); } @@ -4193,7 +4197,8 @@ static void assert_kernel_context_is_current(struct drm_i915_private *dev_priv) enum intel_engine_id id; for_each_engine(engine, dev_priv, id) - GEM_BUG_ON(!i915_gem_context_is_kernel(engine->last_retired_context)); + GEM_BUG_ON(engine->last_retired_context && + !i915_gem_context_is_kernel(engine->last_retired_context)); } int i915_gem_suspend(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 40a6939e3956..17f90c618208 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -97,7 +97,7 @@ * part. It should be safe to decrease this, but it's more future proof as is. */ #define GEN6_CONTEXT_ALIGN (64<<10) -#define GEN7_CONTEXT_ALIGN 4096 +#define GEN7_CONTEXT_ALIGN I915_GTT_MIN_ALIGNMENT static size_t get_context_alignment(struct drm_i915_private *dev_priv) { @@ -205,27 +205,6 @@ alloc_context_obj(struct drm_i915_private *dev_priv, u64 size) return obj; } -static void i915_ppgtt_close(struct i915_address_space *vm) -{ - struct list_head *phases[] = { - &vm->active_list, - &vm->inactive_list, - &vm->unbound_list, - NULL, - }, **phase; - - GEM_BUG_ON(vm->closed); - vm->closed = true; - - for (phase = phases; *phase; phase++) { - struct i915_vma *vma, *vn; - - list_for_each_entry_safe(vma, vn, *phase, vm_link) - if (!i915_vma_is_closed(vma)) - i915_vma_close(vma); - } -} - static void context_close(struct i915_gem_context *ctx) { i915_gem_context_set_closed(ctx); @@ -290,7 +269,7 @@ __create_hw_context(struct drm_i915_private *dev_priv, goto err_out; } - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) { i915_gem_object_put(obj); ret = PTR_ERR(vma); @@ -341,7 +320,7 @@ __create_hw_context(struct drm_i915_private *dev_priv, if (HAS_GUC(dev_priv) && i915.enable_guc_loading) ctx->ggtt_offset_bias = GUC_WOPCM_TOP; else - ctx->ggtt_offset_bias = 4096; + ctx->ggtt_offset_bias = I915_GTT_PAGE_SIZE; return ctx; @@ -456,7 +435,8 @@ int i915_gem_context_init(struct drm_i915_private *dev_priv) dev_priv->hw_context_size = 0; } else if (HAS_HW_CONTEXTS(dev_priv)) { dev_priv->hw_context_size = - round_up(get_context_size(dev_priv), 4096); + round_up(get_context_size(dev_priv), + I915_GTT_PAGE_SIZE); if (dev_priv->hw_context_size > (1<<20)) { DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size %d\n", dev_priv->hw_context_size); @@ -897,6 +877,26 @@ int i915_switch_context(struct drm_i915_gem_request *req) return do_rcs_switch(req); } +static bool engine_has_kernel_context(struct intel_engine_cs *engine) +{ + struct i915_gem_timeline *timeline; + + list_for_each_entry(timeline, &engine->i915->gt.timelines, link) { + struct intel_timeline *tl; + + if (timeline == &engine->i915->gt.global_timeline) + continue; + + tl = &timeline->engine[engine->id]; + if (i915_gem_active_peek(&tl->last_request, + &engine->i915->drm.struct_mutex)) + return false; + } + + return (!engine->last_retired_context || + i915_gem_context_is_kernel(engine->last_retired_context)); +} + int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; @@ -905,10 +905,15 @@ int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv) lockdep_assert_held(&dev_priv->drm.struct_mutex); + i915_gem_retire_requests(dev_priv); + for_each_engine(engine, dev_priv, id) { struct drm_i915_gem_request *req; int ret; + if (engine_has_kernel_context(engine)) + continue; + req = i915_gem_request_alloc(engine, dev_priv->kernel_context); if (IS_ERR(req)) return PTR_ERR(req); diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index 026ebc5a452a..a43e44e18042 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -231,7 +231,8 @@ found: /** * i915_gem_evict_for_vma - Evict vmas to make room for binding a new one - * @target: address space and range to evict for + * @vm: address space to evict from + * @target: range (and color) to evict for * @flags: additional flags to control the eviction algorithm * * This function will try to evict vmas that overlap the target node. @@ -239,18 +240,20 @@ found: * To clarify: This is for freeing up virtual address space, not for freeing * memory in e.g. the shrinker. */ -int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) +int i915_gem_evict_for_node(struct i915_address_space *vm, + struct drm_mm_node *target, + unsigned int flags) { LIST_HEAD(eviction_list); struct drm_mm_node *node; - u64 start = target->node.start; - u64 end = start + target->node.size; + u64 start = target->start; + u64 end = start + target->size; struct i915_vma *vma, *next; bool check_color; int ret = 0; - lockdep_assert_held(&target->vm->i915->drm.struct_mutex); - trace_i915_gem_evict_vma(target, flags); + lockdep_assert_held(&vm->i915->drm.struct_mutex); + trace_i915_gem_evict_node(vm, target, flags); /* Retire before we search the active list. Although we have * reasonable accuracy in our retirement lists, we may have @@ -258,18 +261,18 @@ int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) * retiring. */ if (!(flags & PIN_NONBLOCK)) - i915_gem_retire_requests(target->vm->i915); + i915_gem_retire_requests(vm->i915); - check_color = target->vm->mm.color_adjust; + check_color = vm->mm.color_adjust; if (check_color) { /* Expand search to cover neighbouring guard pages (or lack!) */ - if (start > target->vm->start) - start -= 4096; - if (end < target->vm->start + target->vm->total) - end += 4096; + if (start > vm->start) + start -= I915_GTT_PAGE_SIZE; + if (end < vm->start + vm->total) + end += I915_GTT_PAGE_SIZE; } - drm_mm_for_each_node_in_range(node, &target->vm->mm, start, end) { + drm_mm_for_each_node_in_range(node, &vm->mm, start, end) { /* If we find any non-objects (!vma), we cannot evict them */ if (node->color == I915_COLOR_UNEVICTABLE) { ret = -ENOSPC; @@ -285,12 +288,12 @@ int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) * those as well to make room for our guard pages. */ if (check_color) { - if (vma->node.start + vma->node.size == target->node.start) { - if (vma->node.color == target->node.color) + if (vma->node.start + vma->node.size == node->start) { + if (vma->node.color == node->color) continue; } - if (vma->node.start == target->node.start + target->node.size) { - if (vma->node.color == target->node.color) + if (vma->node.start == node->start + node->size) { + if (vma->node.color == node->color) continue; } } @@ -302,7 +305,7 @@ int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) } /* Overlap of objects in the same batch? */ - if (i915_vma_is_pinned(vma)) { + if (i915_vma_is_pinned(vma) || !list_empty(&vma->exec_list)) { ret = -ENOSPC; if (vma->exec_entry && vma->exec_entry->flags & EXEC_OBJECT_PINNED) diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index a5fe299da1d3..c66e90571031 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -184,7 +184,7 @@ eb_lookup_vmas(struct eb_vmas *eb, * from the (obj, vm) we don't run the risk of creating * duplicated vmas for the same vm. */ - vma = i915_gem_obj_lookup_or_create_vma(obj, vm, NULL); + vma = i915_vma_instance(obj, vm, NULL); if (unlikely(IS_ERR(vma))) { DRM_DEBUG("Failed to lookup VMA\n"); ret = PTR_ERR(vma); @@ -438,7 +438,7 @@ static void *reloc_iomap(struct drm_i915_gem_object *obj, memset(&cache->node, 0, sizeof(cache->node)); ret = drm_mm_insert_node_in_range_generic (&ggtt->base.mm, &cache->node, - 4096, 0, I915_COLOR_UNEVICTABLE, + PAGE_SIZE, 0, I915_COLOR_UNEVICTABLE, 0, ggtt->mappable_end, DRM_MM_SEARCH_DEFAULT, DRM_MM_CREATE_DEFAULT); @@ -851,8 +851,7 @@ eb_vma_misplaced(struct i915_vma *vma) WARN_ON(entry->flags & __EXEC_OBJECT_NEEDS_MAP && !i915_vma_is_ggtt(vma)); - if (entry->alignment && - vma->node.start & (entry->alignment - 1)) + if (entry->alignment && !IS_ALIGNED(vma->node.start, entry->alignment)) return true; if (vma->node.size < entry->pad_to_size) diff --git a/drivers/gpu/drm/i915/i915_gem_fence_reg.c b/drivers/gpu/drm/i915/i915_gem_fence_reg.c index 775059e19ab9..fadbe8f4c745 100644 --- a/drivers/gpu/drm/i915/i915_gem_fence_reg.c +++ b/drivers/gpu/drm/i915/i915_gem_fence_reg.c @@ -77,16 +77,17 @@ static void i965_write_fence_reg(struct drm_i915_fence_reg *fence, val = 0; if (vma) { - unsigned int tiling = i915_gem_object_get_tiling(vma->obj); - bool is_y_tiled = tiling == I915_TILING_Y; unsigned int stride = i915_gem_object_get_stride(vma->obj); - u32 row_size = stride * (is_y_tiled ? 32 : 8); - u32 size = rounddown((u32)vma->node.size, row_size); - val = ((vma->node.start + size - 4096) & 0xfffff000) << 32; - val |= vma->node.start & 0xfffff000; + GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma)); + GEM_BUG_ON(!IS_ALIGNED(vma->node.start, I965_FENCE_PAGE)); + GEM_BUG_ON(!IS_ALIGNED(vma->fence_size, I965_FENCE_PAGE)); + GEM_BUG_ON(!IS_ALIGNED(stride, 128)); + + val = (vma->node.start + vma->fence_size - I965_FENCE_PAGE) << 32; + val |= vma->node.start; val |= (u64)((stride / 128) - 1) << fence_pitch_shift; - if (is_y_tiled) + if (i915_gem_object_get_tiling(vma->obj) == I915_TILING_Y) val |= BIT(I965_FENCE_TILING_Y_SHIFT); val |= I965_FENCE_REG_VALID; } @@ -122,31 +123,24 @@ static void i915_write_fence_reg(struct drm_i915_fence_reg *fence, unsigned int tiling = i915_gem_object_get_tiling(vma->obj); bool is_y_tiled = tiling == I915_TILING_Y; unsigned int stride = i915_gem_object_get_stride(vma->obj); - int pitch_val; - int tile_width; - WARN((vma->node.start & ~I915_FENCE_START_MASK) || - !is_power_of_2(vma->node.size) || - (vma->node.start & (vma->node.size - 1)), - "object 0x%08llx [fenceable? %d] not 1M or pot-size (0x%08llx) aligned\n", - vma->node.start, - i915_vma_is_map_and_fenceable(vma), - vma->node.size); + GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma)); + GEM_BUG_ON(vma->node.start & ~I915_FENCE_START_MASK); + GEM_BUG_ON(!is_power_of_2(vma->fence_size)); + GEM_BUG_ON(!IS_ALIGNED(vma->node.start, vma->fence_size)); if (is_y_tiled && HAS_128_BYTE_Y_TILING(fence->i915)) - tile_width = 128; + stride /= 128; else - tile_width = 512; - - /* Note: pitch better be a power of two tile widths */ - pitch_val = stride / tile_width; - pitch_val = ffs(pitch_val) - 1; + stride /= 512; + GEM_BUG_ON(!is_power_of_2(stride)); val = vma->node.start; if (is_y_tiled) val |= BIT(I830_FENCE_TILING_Y_SHIFT); - val |= I915_FENCE_SIZE_BITS(vma->node.size); - val |= pitch_val << I830_FENCE_PITCH_SHIFT; + val |= I915_FENCE_SIZE_BITS(vma->fence_size); + val |= ilog2(stride) << I830_FENCE_PITCH_SHIFT; + val |= I830_FENCE_REG_VALID; } @@ -166,25 +160,19 @@ static void i830_write_fence_reg(struct drm_i915_fence_reg *fence, val = 0; if (vma) { - unsigned int tiling = i915_gem_object_get_tiling(vma->obj); - bool is_y_tiled = tiling == I915_TILING_Y; unsigned int stride = i915_gem_object_get_stride(vma->obj); - u32 pitch_val; - WARN((vma->node.start & ~I830_FENCE_START_MASK) || - !is_power_of_2(vma->node.size) || - (vma->node.start & (vma->node.size - 1)), - "object 0x%08llx not 512K or pot-size 0x%08llx aligned\n", - vma->node.start, vma->node.size); - - pitch_val = stride / 128; - pitch_val = ffs(pitch_val) - 1; + GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma)); + GEM_BUG_ON(vma->node.start & ~I830_FENCE_START_MASK); + GEM_BUG_ON(!is_power_of_2(vma->fence_size)); + GEM_BUG_ON(!is_power_of_2(stride / 128)); + GEM_BUG_ON(!IS_ALIGNED(vma->node.start, vma->fence_size)); val = vma->node.start; - if (is_y_tiled) + if (i915_gem_object_get_tiling(vma->obj) == I915_TILING_Y) val |= BIT(I830_FENCE_TILING_Y_SHIFT); - val |= I830_FENCE_SIZE_BITS(vma->node.size); - val |= pitch_val << I830_FENCE_PITCH_SHIFT; + val |= I830_FENCE_SIZE_BITS(vma->fence_size); + val |= ilog2(stride / 128) << I830_FENCE_PITCH_SHIFT; val |= I830_FENCE_REG_VALID; } diff --git a/drivers/gpu/drm/i915/i915_gem_fence_reg.h b/drivers/gpu/drm/i915/i915_gem_fence_reg.h index 22c4a2d01adf..99a31ded4dfd 100644 --- a/drivers/gpu/drm/i915/i915_gem_fence_reg.h +++ b/drivers/gpu/drm/i915/i915_gem_fence_reg.h @@ -30,6 +30,8 @@ struct drm_i915_private; struct i915_vma; +#define I965_FENCE_PAGE 4096UL + struct drm_i915_fence_reg { struct list_head link; struct drm_i915_private *i915; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index f698006fe883..e808aad203d8 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -23,10 +23,14 @@ * */ +#include <linux/log2.h> +#include <linux/random.h> #include <linux/seq_file.h> #include <linux/stop_machine.h> + #include <drm/drmP.h> #include <drm/i915_drm.h> + #include "i915_drv.h" #include "i915_vgpu.h" #include "i915_trace.h" @@ -99,12 +103,29 @@ static int i915_get_ggtt_vma_pages(struct i915_vma *vma); -const struct i915_ggtt_view i915_ggtt_view_normal = { - .type = I915_GGTT_VIEW_NORMAL, -}; -const struct i915_ggtt_view i915_ggtt_view_rotated = { - .type = I915_GGTT_VIEW_ROTATED, -}; +static void gen6_ggtt_invalidate(struct drm_i915_private *dev_priv) +{ + /* Note that as an uncached mmio write, this should flush the + * WCB of the writes into the GGTT before it triggers the invalidate. + */ + I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); +} + +static void guc_ggtt_invalidate(struct drm_i915_private *dev_priv) +{ + gen6_ggtt_invalidate(dev_priv); + I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); +} + +static void gmch_ggtt_invalidate(struct drm_i915_private *dev_priv) +{ + intel_gtt_chipset_flush(); +} + +static inline void i915_ggtt_invalidate(struct drm_i915_private *i915) +{ + i915->ggtt.invalidate(i915); +} int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, int enable_ppgtt) @@ -329,7 +350,7 @@ static int __setup_page_dma(struct drm_i915_private *dev_priv, return -ENOMEM; p->daddr = dma_map_page(kdev, - p->page, 0, 4096, PCI_DMA_BIDIRECTIONAL); + p->page, 0, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); if (dma_mapping_error(kdev, p->daddr)) { __free_page(p->page); @@ -353,7 +374,7 @@ static void cleanup_page_dma(struct drm_i915_private *dev_priv, if (WARN_ON(!p->page)) return; - dma_unmap_page(&pdev->dev, p->daddr, 4096, PCI_DMA_BIDIRECTIONAL); + dma_unmap_page(&pdev->dev, p->daddr, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); __free_page(p->page); memset(p, 0, sizeof(*p)); } @@ -626,10 +647,10 @@ static void gen8_initialize_pml4(struct i915_address_space *vm, } static void -gen8_setup_page_directory(struct i915_hw_ppgtt *ppgtt, - struct i915_page_directory_pointer *pdp, - struct i915_page_directory *pd, - int index) +gen8_setup_pdpe(struct i915_hw_ppgtt *ppgtt, + struct i915_page_directory_pointer *pdp, + struct i915_page_directory *pd, + int index) { gen8_ppgtt_pdpe_t *page_directorypo; @@ -642,10 +663,10 @@ gen8_setup_page_directory(struct i915_hw_ppgtt *ppgtt, } static void -gen8_setup_page_directory_pointer(struct i915_hw_ppgtt *ppgtt, - struct i915_pml4 *pml4, - struct i915_page_directory_pointer *pdp, - int index) +gen8_setup_pml4e(struct i915_hw_ppgtt *ppgtt, + struct i915_pml4 *pml4, + struct i915_page_directory_pointer *pdp, + int index) { gen8_ppgtt_pml4e_t *pagemap = kmap_px(pml4); @@ -793,9 +814,6 @@ static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm, struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm); struct i915_page_directory *pd; uint64_t pdpe; - gen8_ppgtt_pdpe_t *pdpe_vaddr; - gen8_ppgtt_pdpe_t scratch_pdpe = - gen8_pdpe_encode(px_dma(vm->scratch_pd), I915_CACHE_LLC); gen8_for_each_pdpe(pd, pdp, start, length, pdpe) { if (WARN_ON(!pdp->page_directory[pdpe])) @@ -803,11 +821,7 @@ static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm, if (gen8_ppgtt_clear_pd(vm, pd, start, length)) { __clear_bit(pdpe, pdp->used_pdpes); - if (USES_FULL_48BIT_PPGTT(dev_priv)) { - pdpe_vaddr = kmap_px(pdp); - pdpe_vaddr[pdpe] = scratch_pdpe; - kunmap_px(ppgtt, pdpe_vaddr); - } + gen8_setup_pdpe(ppgtt, pdp, vm->scratch_pd, pdpe); free_pd(vm->i915, pd); } } @@ -832,9 +846,6 @@ static void gen8_ppgtt_clear_pml4(struct i915_address_space *vm, struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm); struct i915_page_directory_pointer *pdp; uint64_t pml4e; - gen8_ppgtt_pml4e_t *pml4e_vaddr; - gen8_ppgtt_pml4e_t scratch_pml4e = - gen8_pml4e_encode(px_dma(vm->scratch_pdp), I915_CACHE_LLC); GEM_BUG_ON(!USES_FULL_48BIT_PPGTT(vm->i915)); @@ -844,9 +855,7 @@ static void gen8_ppgtt_clear_pml4(struct i915_address_space *vm, if (gen8_ppgtt_clear_pdp(vm, pdp, start, length)) { __clear_bit(pml4e, pml4->used_pml4es); - pml4e_vaddr = kmap_px(pml4); - pml4e_vaddr[pml4e] = scratch_pml4e; - kunmap_px(ppgtt, pml4e_vaddr); + gen8_setup_pml4e(ppgtt, pml4, vm->scratch_pdp, pml4e); free_pdp(vm->i915, pdp); } } @@ -1366,7 +1375,7 @@ static int gen8_alloc_va_range_3lvl(struct i915_address_space *vm, kunmap_px(ppgtt, page_directory); __set_bit(pdpe, pdp->used_pdpes); - gen8_setup_page_directory(ppgtt, pdp, pd, pdpe); + gen8_setup_pdpe(ppgtt, pdp, pd, pdpe); } free_gen8_temp_bitmaps(new_page_dirs, new_page_tables); @@ -1425,7 +1434,7 @@ static int gen8_alloc_va_range_4lvl(struct i915_address_space *vm, if (ret) goto err_out; - gen8_setup_page_directory_pointer(ppgtt, pml4, pdp, pml4e); + gen8_setup_pml4e(ppgtt, pml4, pdp, pml4e); } bitmap_or(pml4->used_pml4es, new_pdps, pml4->used_pml4es, @@ -2044,7 +2053,6 @@ static int gen6_ppgtt_allocate_page_directories(struct i915_hw_ppgtt *ppgtt) struct i915_address_space *vm = &ppgtt->base; struct drm_i915_private *dev_priv = ppgtt->base.i915; struct i915_ggtt *ggtt = &dev_priv->ggtt; - bool retried = false; int ret; /* PPGTT PDEs reside in the GGTT and consists of 512 entries. The @@ -2057,29 +2065,14 @@ static int gen6_ppgtt_allocate_page_directories(struct i915_hw_ppgtt *ppgtt) if (ret) return ret; -alloc: - ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm, &ppgtt->node, - GEN6_PD_SIZE, GEN6_PD_ALIGN, - I915_COLOR_UNEVICTABLE, - 0, ggtt->base.total, - DRM_MM_TOPDOWN); - if (ret == -ENOSPC && !retried) { - ret = i915_gem_evict_something(&ggtt->base, - GEN6_PD_SIZE, GEN6_PD_ALIGN, - I915_COLOR_UNEVICTABLE, - 0, ggtt->base.total, - 0); - if (ret) - goto err_out; - - retried = true; - goto alloc; - } - + ret = i915_gem_gtt_insert(&ggtt->base, &ppgtt->node, + GEN6_PD_SIZE, GEN6_PD_ALIGN, + I915_COLOR_UNEVICTABLE, + 0, ggtt->base.total, + PIN_HIGH); if (ret) goto err_out; - if (ppgtt->node.start < ggtt->mappable_end) DRM_DEBUG("Forced to use aperture for PDEs\n"); @@ -2267,6 +2260,27 @@ i915_ppgtt_create(struct drm_i915_private *dev_priv, return ppgtt; } +void i915_ppgtt_close(struct i915_address_space *vm) +{ + struct list_head *phases[] = { + &vm->active_list, + &vm->inactive_list, + &vm->unbound_list, + NULL, + }, **phase; + + GEM_BUG_ON(vm->closed); + vm->closed = true; + + for (phase = phases; *phase; phase++) { + struct i915_vma *vma, *vn; + + list_for_each_entry_safe(vma, vn, *phase, vm_link) + if (!i915_vma_is_closed(vma)) + i915_vma_close(vma); + } +} + void i915_ppgtt_release(struct kref *kref) { struct i915_hw_ppgtt *ppgtt = @@ -2331,16 +2345,6 @@ void i915_check_and_clear_faults(struct drm_i915_private *dev_priv) POSTING_READ(RING_FAULT_REG(dev_priv->engine[RCS])); } -static void i915_ggtt_flush(struct drm_i915_private *dev_priv) -{ - if (INTEL_INFO(dev_priv)->gen < 6) { - intel_gtt_chipset_flush(); - } else { - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); - } -} - void i915_gem_suspend_gtt_mappings(struct drm_i915_private *dev_priv) { struct i915_ggtt *ggtt = &dev_priv->ggtt; @@ -2355,7 +2359,7 @@ void i915_gem_suspend_gtt_mappings(struct drm_i915_private *dev_priv) ggtt->base.clear_range(&ggtt->base, ggtt->base.start, ggtt->base.total); - i915_ggtt_flush(dev_priv); + i915_ggtt_invalidate(dev_priv); } int i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj, @@ -2394,15 +2398,13 @@ static void gen8_ggtt_insert_page(struct i915_address_space *vm, enum i915_cache_level level, u32 unused) { - struct drm_i915_private *dev_priv = vm->i915; + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); gen8_pte_t __iomem *pte = - (gen8_pte_t __iomem *)dev_priv->ggtt.gsm + - (offset >> PAGE_SHIFT); + (gen8_pte_t __iomem *)ggtt->gsm + (offset >> PAGE_SHIFT); gen8_set_pte(pte, gen8_pte_encode(addr, level)); - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } static void gen8_ggtt_insert_entries(struct i915_address_space *vm, @@ -2410,7 +2412,6 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, uint64_t start, enum i915_cache_level level, u32 unused) { - struct drm_i915_private *dev_priv = vm->i915; struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); struct sgt_iter sgt_iter; gen8_pte_t __iomem *gtt_entries; @@ -2439,8 +2440,7 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, * want to flush the TLBs only after we're certain all the PTE updates * have finished. */ - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } struct insert_entries { @@ -2475,15 +2475,13 @@ static void gen6_ggtt_insert_page(struct i915_address_space *vm, enum i915_cache_level level, u32 flags) { - struct drm_i915_private *dev_priv = vm->i915; + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); gen6_pte_t __iomem *pte = - (gen6_pte_t __iomem *)dev_priv->ggtt.gsm + - (offset >> PAGE_SHIFT); + (gen6_pte_t __iomem *)ggtt->gsm + (offset >> PAGE_SHIFT); iowrite32(vm->pte_encode(addr, level, flags), pte); - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } /* @@ -2497,7 +2495,6 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, uint64_t start, enum i915_cache_level level, u32 flags) { - struct drm_i915_private *dev_priv = vm->i915; struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); struct sgt_iter sgt_iter; gen6_pte_t __iomem *gtt_entries; @@ -2525,8 +2522,7 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, * want to flush the TLBs only after we're certain all the PTE updates * have finished. */ - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } static void nop_clear_range(struct i915_address_space *vm, @@ -2723,11 +2719,11 @@ static void i915_gtt_color_adjust(const struct drm_mm_node *node, u64 *end) { if (node->color != color) - *start += 4096; + *start += I915_GTT_PAGE_SIZE; node = list_next_entry(node, node_list); if (node->allocated && node->color != color) - *end -= 4096; + *end -= I915_GTT_PAGE_SIZE; } int i915_gem_init_ggtt(struct drm_i915_private *dev_priv) @@ -2754,7 +2750,7 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv) /* Reserve a mappable slot for our lockless error capture */ ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm, &ggtt->error_capture, - 4096, 0, + PAGE_SIZE, 0, I915_COLOR_UNEVICTABLE, 0, ggtt->mappable_end, 0, 0); @@ -3086,6 +3082,8 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt) if (IS_CHERRYVIEW(dev_priv)) ggtt->base.insert_entries = gen8_ggtt_insert_entries__BKL; + ggtt->invalidate = gen6_ggtt_invalidate; + return ggtt_probe_common(ggtt, size); } @@ -3123,6 +3121,8 @@ static int gen6_gmch_probe(struct i915_ggtt *ggtt) ggtt->base.unbind_vma = ggtt_unbind_vma; ggtt->base.cleanup = gen6_gmch_remove; + ggtt->invalidate = gen6_ggtt_invalidate; + if (HAS_EDRAM(dev_priv)) ggtt->base.pte_encode = iris_pte_encode; else if (IS_HASWELL(dev_priv)) @@ -3166,6 +3166,8 @@ static int i915_gmch_probe(struct i915_ggtt *ggtt) ggtt->base.unbind_vma = ggtt_unbind_vma; ggtt->base.cleanup = i915_gmch_remove; + ggtt->invalidate = gmch_ggtt_invalidate; + if (unlikely(ggtt->do_idle_maps)) DRM_INFO("applying Ironlake quirks for intel_iommu\n"); @@ -3284,6 +3286,16 @@ int i915_ggtt_enable_hw(struct drm_i915_private *dev_priv) return 0; } +void i915_ggtt_enable_guc(struct drm_i915_private *i915) +{ + i915->ggtt.invalidate = guc_ggtt_invalidate; +} + +void i915_ggtt_disable_guc(struct drm_i915_private *i915) +{ + i915->ggtt.invalidate = gen6_ggtt_invalidate; +} + void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv) { struct i915_ggtt *ggtt = &dev_priv->ggtt; @@ -3347,52 +3359,7 @@ void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv) } } - i915_ggtt_flush(dev_priv); -} - -struct i915_vma * -i915_gem_obj_to_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) -{ - struct rb_node *rb; - - rb = obj->vma_tree.rb_node; - while (rb) { - struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node); - long cmp; - - cmp = i915_vma_compare(vma, vm, view); - if (cmp == 0) - return vma; - - if (cmp < 0) - rb = rb->rb_right; - else - rb = rb->rb_left; - } - - return NULL; -} - -struct i915_vma * -i915_gem_obj_lookup_or_create_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) -{ - struct i915_vma *vma; - - lockdep_assert_held(&obj->base.dev->struct_mutex); - GEM_BUG_ON(view && !i915_is_ggtt(vm)); - - vma = i915_gem_obj_to_vma(obj, vm, view); - if (!vma) { - vma = i915_vma_create(obj, vm, view); - GEM_BUG_ON(vma != i915_gem_obj_to_vma(obj, vm, view)); - } - - GEM_BUG_ON(i915_vma_is_closed(vma)); - return vma; + i915_ggtt_invalidate(dev_priv); } static struct scatterlist * @@ -3492,7 +3459,7 @@ intel_partial_pages(const struct i915_ggtt_view *view, { struct sg_table *st; struct scatterlist *sg, *iter; - unsigned int count = view->params.partial.size; + unsigned int count = view->partial.size; unsigned int offset; int ret = -ENOMEM; @@ -3504,9 +3471,7 @@ intel_partial_pages(const struct i915_ggtt_view *view, if (ret) goto err_sg_alloc; - iter = i915_gem_object_get_sg(obj, - view->params.partial.offset, - &offset); + iter = i915_gem_object_get_sg(obj, view->partial.offset, &offset); GEM_BUG_ON(!iter); sg = st->sgl; @@ -3558,7 +3523,8 @@ i915_get_ggtt_vma_pages(struct i915_vma *vma) vma->pages = vma->obj->mm.pages; else if (vma->ggtt_view.type == I915_GGTT_VIEW_ROTATED) vma->pages = - intel_rotate_fb_obj_pages(&vma->ggtt_view.params.rotated, vma->obj); + intel_rotate_fb_obj_pages(&vma->ggtt_view.rotated, + vma->obj); else if (vma->ggtt_view.type == I915_GGTT_VIEW_PARTIAL) vma->pages = intel_partial_pages(&vma->ggtt_view, vma->obj); else @@ -3579,3 +3545,207 @@ i915_get_ggtt_vma_pages(struct i915_vma *vma) return ret; } +/** + * i915_gem_gtt_reserve - reserve a node in an address_space (GTT) + * @vm: the &struct i915_address_space + * @node: the &struct drm_mm_node (typically i915_vma.mode) + * @size: how much space to allocate inside the GTT, + * must be #I915_GTT_PAGE_SIZE aligned + * @offset: where to insert inside the GTT, + * must be #I915_GTT_MIN_ALIGNMENT aligned, and the node + * (@offset + @size) must fit within the address space + * @color: color to apply to node, if this node is not from a VMA, + * color must be #I915_COLOR_UNEVICTABLE + * @flags: control search and eviction behaviour + * + * i915_gem_gtt_reserve() tries to insert the @node at the exact @offset inside + * the address space (using @size and @color). If the @node does not fit, it + * tries to evict any overlapping nodes from the GTT, including any + * neighbouring nodes if the colors do not match (to ensure guard pages between + * differing domains). See i915_gem_evict_for_node() for the gory details + * on the eviction algorithm. #PIN_NONBLOCK may used to prevent waiting on + * evicting active overlapping objects, and any overlapping node that is pinned + * or marked as unevictable will also result in failure. + * + * Returns: 0 on success, -ENOSPC if no suitable hole is found, -EINTR if + * asked to wait for eviction and interrupted. + */ +int i915_gem_gtt_reserve(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 offset, unsigned long color, + unsigned int flags) +{ + int err; + + GEM_BUG_ON(!size); + GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(!IS_ALIGNED(offset, I915_GTT_MIN_ALIGNMENT)); + GEM_BUG_ON(range_overflows(offset, size, vm->total)); + GEM_BUG_ON(vm == &vm->i915->mm.aliasing_ppgtt->base); + GEM_BUG_ON(drm_mm_node_allocated(node)); + + node->size = size; + node->start = offset; + node->color = color; + + err = drm_mm_reserve_node(&vm->mm, node); + if (err != -ENOSPC) + return err; + + err = i915_gem_evict_for_node(vm, node, flags); + if (err == 0) + err = drm_mm_reserve_node(&vm->mm, node); + + return err; +} + +static u64 random_offset(u64 start, u64 end, u64 len, u64 align) +{ + u64 range, addr; + + GEM_BUG_ON(range_overflows(start, len, end)); + GEM_BUG_ON(round_up(start, align) > round_down(end - len, align)); + + range = round_down(end - len, align) - round_up(start, align); + if (range) { + if (sizeof(unsigned long) == sizeof(u64)) { + addr = get_random_long(); + } else { + addr = get_random_int(); + if (range > U32_MAX) { + addr <<= 32; + addr |= get_random_int(); + } + } + div64_u64_rem(addr, range, &addr); + start += addr; + } + + return round_up(start, align); +} + +/** + * i915_gem_gtt_insert - insert a node into an address_space (GTT) + * @vm: the &struct i915_address_space + * @node: the &struct drm_mm_node (typically i915_vma.node) + * @size: how much space to allocate inside the GTT, + * must be #I915_GTT_PAGE_SIZE aligned + * @alignment: required alignment of starting offset, may be 0 but + * if specified, this must be a power-of-two and at least + * #I915_GTT_MIN_ALIGNMENT + * @color: color to apply to node + * @start: start of any range restriction inside GTT (0 for all), + * must be #I915_GTT_PAGE_SIZE aligned + * @end: end of any range restriction inside GTT (U64_MAX for all), + * must be #I915_GTT_PAGE_SIZE aligned if not U64_MAX + * @flags: control search and eviction behaviour + * + * i915_gem_gtt_insert() first searches for an available hole into which + * is can insert the node. The hole address is aligned to @alignment and + * its @size must then fit entirely within the [@start, @end] bounds. The + * nodes on either side of the hole must match @color, or else a guard page + * will be inserted between the two nodes (or the node evicted). If no + * suitable hole is found, first a victim is randomly selected and tested + * for eviction, otherwise then the LRU list of objects within the GTT + * is scanned to find the first set of replacement nodes to create the hole. + * Those old overlapping nodes are evicted from the GTT (and so must be + * rebound before any future use). Any node that is currently pinned cannot + * be evicted (see i915_vma_pin()). Similar if the node's VMA is currently + * active and #PIN_NONBLOCK is specified, that node is also skipped when + * searching for an eviction candidate. See i915_gem_evict_something() for + * the gory details on the eviction algorithm. + * + * Returns: 0 on success, -ENOSPC if no suitable hole is found, -EINTR if + * asked to wait for eviction and interrupted. + */ +int i915_gem_gtt_insert(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 alignment, unsigned long color, + u64 start, u64 end, unsigned int flags) +{ + u32 search_flag, alloc_flag; + u64 offset; + int err; + + lockdep_assert_held(&vm->i915->drm.struct_mutex); + GEM_BUG_ON(!size); + GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(alignment && !is_power_of_2(alignment)); + GEM_BUG_ON(alignment && !IS_ALIGNED(alignment, I915_GTT_MIN_ALIGNMENT)); + GEM_BUG_ON(start >= end); + GEM_BUG_ON(start > 0 && !IS_ALIGNED(start, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(end < U64_MAX && !IS_ALIGNED(end, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(vm == &vm->i915->mm.aliasing_ppgtt->base); + GEM_BUG_ON(drm_mm_node_allocated(node)); + + if (unlikely(range_overflows(start, size, end))) + return -ENOSPC; + + if (unlikely(round_up(start, alignment) > round_down(end - size, alignment))) + return -ENOSPC; + + if (flags & PIN_HIGH) { + search_flag = DRM_MM_SEARCH_BELOW; + alloc_flag = DRM_MM_CREATE_TOP; + } else { + search_flag = DRM_MM_SEARCH_DEFAULT; + alloc_flag = DRM_MM_CREATE_DEFAULT; + } + + /* We only allocate in PAGE_SIZE/GTT_PAGE_SIZE (4096) chunks, + * so we know that we always have a minimum alignment of 4096. + * The drm_mm range manager is optimised to return results + * with zero alignment, so where possible use the optimal + * path. + */ + BUILD_BUG_ON(I915_GTT_MIN_ALIGNMENT > I915_GTT_PAGE_SIZE); + if (alignment <= I915_GTT_MIN_ALIGNMENT) + alignment = 0; + + err = drm_mm_insert_node_in_range_generic(&vm->mm, node, + size, alignment, color, + start, end, + search_flag, alloc_flag); + if (err != -ENOSPC) + return err; + + /* No free space, pick a slot at random. + * + * There is a pathological case here using a GTT shared between + * mmap and GPU (i.e. ggtt/aliasing_ppgtt but not full-ppgtt): + * + * |<-- 256 MiB aperture -->||<-- 1792 MiB unmappable -->| + * (64k objects) (448k objects) + * + * Now imagine that the eviction LRU is ordered top-down (just because + * pathology meets real life), and that we need to evict an object to + * make room inside the aperture. The eviction scan then has to walk + * the 448k list before it finds one within range. And now imagine that + * it has to search for a new hole between every byte inside the memcpy, + * for several simultaneous clients. + * + * On a full-ppgtt system, if we have run out of available space, there + * will be lots and lots of objects in the eviction list! Again, + * searching that LRU list may be slow if we are also applying any + * range restrictions (e.g. restriction to low 4GiB) and so, for + * simplicity and similarilty between different GTT, try the single + * random replacement first. + */ + offset = random_offset(start, end, + size, alignment ?: I915_GTT_MIN_ALIGNMENT); + err = i915_gem_gtt_reserve(vm, node, size, offset, color, flags); + if (err != -ENOSPC) + return err; + + /* Randomly selected placement is pinned, do a search */ + err = i915_gem_evict_something(vm, size, alignment, color, + start, end, flags); + if (err) + return err; + + search_flag = DRM_MM_SEARCH_DEFAULT; + return drm_mm_insert_node_in_range_generic(&vm->mm, node, + size, alignment, color, + start, end, + search_flag, alloc_flag); +} diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index 9e91d7e6149c..3c5ef5358cef 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -40,6 +40,9 @@ #include "i915_gem_timeline.h" #include "i915_gem_request.h" +#define I915_GTT_PAGE_SIZE 4096UL +#define I915_GTT_MIN_ALIGNMENT I915_GTT_PAGE_SIZE + #define I915_FENCE_REG_NONE -1 #define I915_MAX_NUM_FENCES 32 /* 32 fences + sign bit for FENCE_REG_NONE */ @@ -142,34 +145,57 @@ typedef uint64_t gen8_ppgtt_pml4e_t; struct sg_table; -enum i915_ggtt_view_type { - I915_GGTT_VIEW_NORMAL = 0, - I915_GGTT_VIEW_ROTATED, - I915_GGTT_VIEW_PARTIAL, -}; - struct intel_rotation_info { - struct { + struct intel_rotation_plane_info { /* tiles */ unsigned int width, height, stride, offset; } plane[2]; +} __packed; + +static inline void assert_intel_rotation_info_is_packed(void) +{ + BUILD_BUG_ON(sizeof(struct intel_rotation_info) != 8*sizeof(unsigned int)); +} + +struct intel_partial_info { + u64 offset; + unsigned int size; +} __packed; + +static inline void assert_intel_partial_info_is_packed(void) +{ + BUILD_BUG_ON(sizeof(struct intel_partial_info) != sizeof(u64) + sizeof(unsigned int)); +} + +enum i915_ggtt_view_type { + I915_GGTT_VIEW_NORMAL = 0, + I915_GGTT_VIEW_ROTATED = sizeof(struct intel_rotation_info), + I915_GGTT_VIEW_PARTIAL = sizeof(struct intel_partial_info), }; +static inline void assert_i915_ggtt_view_type_is_unique(void) +{ + /* As we encode the size of each branch inside the union into its type, + * we have to be careful that each branch has a unique size. + */ + switch ((enum i915_ggtt_view_type)0) { + case I915_GGTT_VIEW_NORMAL: + case I915_GGTT_VIEW_PARTIAL: + case I915_GGTT_VIEW_ROTATED: + /* gcc complains if these are identical cases */ + break; + } +} + struct i915_ggtt_view { enum i915_ggtt_view_type type; - union { - struct { - u64 offset; - unsigned int size; - } partial; + /* Members need to contain no holes/padding */ + struct intel_partial_info partial; struct intel_rotation_info rotated; - } params; + }; }; -extern const struct i915_ggtt_view i915_ggtt_view_normal; -extern const struct i915_ggtt_view i915_ggtt_view_rotated; - enum i915_cache_level; struct i915_vma; @@ -333,6 +359,7 @@ struct i915_ggtt { /** "Graphics Stolen Memory" holds the global PTEs */ void __iomem *gsm; + void (*invalidate)(struct drm_i915_private *dev_priv); bool do_idle_maps; @@ -501,6 +528,8 @@ i915_vm_to_ggtt(struct i915_address_space *vm) int i915_ggtt_probe_hw(struct drm_i915_private *dev_priv); int i915_ggtt_init_hw(struct drm_i915_private *dev_priv); int i915_ggtt_enable_hw(struct drm_i915_private *dev_priv); +void i915_ggtt_enable_guc(struct drm_i915_private *i915); +void i915_ggtt_disable_guc(struct drm_i915_private *i915); int i915_gem_init_ggtt(struct drm_i915_private *dev_priv); void i915_ggtt_cleanup_hw(struct drm_i915_private *dev_priv); @@ -509,6 +538,7 @@ void i915_ppgtt_release(struct kref *kref); struct i915_hw_ppgtt *i915_ppgtt_create(struct drm_i915_private *dev_priv, struct drm_i915_file_private *fpriv, const char *name); +void i915_ppgtt_close(struct i915_address_space *vm); static inline void i915_ppgtt_get(struct i915_hw_ppgtt *ppgtt) { if (ppgtt) @@ -529,6 +559,16 @@ int __must_check i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj, void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj, struct sg_table *pages); +int i915_gem_gtt_reserve(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 offset, unsigned long color, + unsigned int flags); + +int i915_gem_gtt_insert(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 alignment, unsigned long color, + u64 start, u64 end, unsigned int flags); + /* Flags used by pin/bind&friends. */ #define PIN_NONBLOCK BIT(0) #define PIN_MAPPABLE BIT(1) @@ -543,6 +583,6 @@ void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj, #define PIN_HIGH BIT(9) #define PIN_OFFSET_BIAS BIT(10) #define PIN_OFFSET_FIXED BIT(11) -#define PIN_OFFSET_MASK (~4095) +#define PIN_OFFSET_MASK (-I915_GTT_PAGE_SIZE) #endif diff --git a/drivers/gpu/drm/i915/i915_gem_internal.c b/drivers/gpu/drm/i915/i915_gem_internal.c index 2222863e505f..17ce53d0d092 100644 --- a/drivers/gpu/drm/i915/i915_gem_internal.c +++ b/drivers/gpu/drm/i915/i915_gem_internal.c @@ -151,10 +151,16 @@ static const struct drm_i915_gem_object_ops i915_gem_object_internal_ops = { */ struct drm_i915_gem_object * i915_gem_object_create_internal(struct drm_i915_private *i915, - unsigned int size) + phys_addr_t size) { struct drm_i915_gem_object *obj; + GEM_BUG_ON(!size); + GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE)); + + if (overflows_type(size, obj->base.size)) + return ERR_PTR(-E2BIG); + obj = i915_gem_object_alloc(i915); if (!obj) return ERR_PTR(-ENOMEM); diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h index 6a368de9d81e..290eaa7fc9eb 100644 --- a/drivers/gpu/drm/i915/i915_gem_object.h +++ b/drivers/gpu/drm/i915/i915_gem_object.h @@ -317,6 +317,29 @@ i915_gem_object_get_stride(struct drm_i915_gem_object *obj) return obj->tiling_and_stride & STRIDE_MASK; } +static inline unsigned int +i915_gem_tile_height(unsigned int tiling) +{ + GEM_BUG_ON(!tiling); + return tiling == I915_TILING_Y ? 32 : 8; +} + +static inline unsigned int +i915_gem_object_get_tile_height(struct drm_i915_gem_object *obj) +{ + return i915_gem_tile_height(i915_gem_object_get_tiling(obj)); +} + +static inline unsigned int +i915_gem_object_get_tile_row_size(struct drm_i915_gem_object *obj) +{ + return (i915_gem_object_get_stride(obj) * + i915_gem_object_get_tile_height(obj)); +} + +int i915_gem_object_set_tiling(struct drm_i915_gem_object *obj, + unsigned int tiling, unsigned int stride); + static inline struct intel_engine_cs * i915_gem_object_last_write_engine(struct drm_i915_gem_object *obj) { diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.c b/drivers/gpu/drm/i915/i915_gem_render_state.c index 5af19b0bf713..b42c81b42487 100644 --- a/drivers/gpu/drm/i915/i915_gem_render_state.c +++ b/drivers/gpu/drm/i915/i915_gem_render_state.c @@ -187,20 +187,20 @@ int i915_gem_render_state_init(struct intel_engine_cs *engine) if (!rodata) return 0; - if (rodata->batch_items * 4 > 4096) + if (rodata->batch_items * 4 > PAGE_SIZE) return -EINVAL; so = kmalloc(sizeof(*so), GFP_KERNEL); if (!so) return -ENOMEM; - obj = i915_gem_object_create_internal(engine->i915, 4096); + obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); if (IS_ERR(obj)) { ret = PTR_ERR(obj); goto err_free; } - so->vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + so->vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(so->vma)) { ret = PTR_ERR(so->vma); goto err_obj; diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c index 99056b948eda..72b7f7d9461d 100644 --- a/drivers/gpu/drm/i915/i915_gem_request.c +++ b/drivers/gpu/drm/i915/i915_gem_request.c @@ -307,26 +307,6 @@ void i915_gem_request_retire_upto(struct drm_i915_gem_request *req) } while (tmp != req); } -static int i915_gem_check_wedge(struct drm_i915_private *dev_priv) -{ - struct i915_gpu_error *error = &dev_priv->gpu_error; - - if (i915_terminally_wedged(error)) - return -EIO; - - if (i915_reset_in_progress(error)) { - /* Non-interruptible callers can't handle -EAGAIN, hence return - * -EIO unconditionally for these. - */ - if (!dev_priv->mm.interruptible) - return -EIO; - - return -EAGAIN; - } - - return 0; -} - static int i915_gem_init_global_seqno(struct drm_i915_private *i915, u32 seqno) { struct i915_gem_timeline *timeline = &i915->gt.global_timeline; @@ -521,12 +501,10 @@ i915_gem_request_alloc(struct intel_engine_cs *engine, lockdep_assert_held(&dev_priv->drm.struct_mutex); /* ABI: Before userspace accesses the GPU (e.g. execbuffer), report - * EIO if the GPU is already wedged, or EAGAIN to drop the struct_mutex - * and restart. + * EIO if the GPU is already wedged. */ - ret = i915_gem_check_wedge(dev_priv); - if (ret) - return ERR_PTR(ret); + if (i915_terminally_wedged(&dev_priv->gpu_error)) + return ERR_PTR(-EIO); /* Pinning the contexts may generate requests in order to acquire * GGTT space, so do this first before we reserve a seqno for @@ -851,6 +829,13 @@ void __i915_add_request(struct drm_i915_gem_request *request, bool flush_caches) lockdep_assert_held(&request->i915->drm.struct_mutex); trace_i915_gem_request_add(request); + /* Make sure that no request gazumped us - if it was allocated after + * our i915_gem_request_alloc() and called __i915_add_request() before + * us, the timeline will hold its seqno which is later than ours. + */ + GEM_BUG_ON(i915_seqno_passed(timeline->last_submitted_seqno, + request->fence.seqno)); + /* * To ensure that this call will not fail, space for its emissions * should already have been reserved in the ring buffer. Let the ring diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index f1a1d33febcd..127d698e7c84 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -647,8 +647,9 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv stolen_offset, gtt_offset, size); /* KISS and expect everything to be page-aligned */ - if (WARN_ON(size == 0) || WARN_ON(size & 4095) || - WARN_ON(stolen_offset & 4095)) + if (WARN_ON(size == 0) || + WARN_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)) || + WARN_ON(!IS_ALIGNED(stolen_offset, I915_GTT_MIN_ALIGNMENT))) return NULL; stolen = kzalloc(sizeof(*stolen), GFP_KERNEL); @@ -682,7 +683,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv if (ret) goto err; - vma = i915_gem_obj_lookup_or_create_vma(obj, &ggtt->base, NULL); + vma = i915_vma_instance(obj, &ggtt->base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto err_pages; @@ -693,15 +694,16 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv * setting up the GTT space. The actual reservation will occur * later. */ - vma->node.start = gtt_offset; - vma->node.size = size; - - ret = drm_mm_reserve_node(&ggtt->base.mm, &vma->node); + ret = i915_gem_gtt_reserve(&ggtt->base, &vma->node, + size, gtt_offset, obj->cache_level, + 0); if (ret) { DRM_DEBUG_KMS("failed to allocate stolen GTT space\n"); goto err_pages; } + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); + vma->pages = obj->mm.pages; vma->flags |= I915_VMA_GLOBAL_BIND; __i915_vma_set_map_and_fenceable(vma); diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c index 62ad375de6ca..b1361cfd4c5c 100644 --- a/drivers/gpu/drm/i915/i915_gem_tiling.c +++ b/drivers/gpu/drm/i915/i915_gem_tiling.c @@ -34,8 +34,8 @@ /** * DOC: buffer object tiling * - * i915_gem_set_tiling() and i915_gem_get_tiling() is the userspace interface to - * declare fence register requirements. + * i915_gem_set_tiling_ioctl() and i915_gem_get_tiling_ioctl() is the userspace + * interface to declare fence register requirements. * * In principle GEM doesn't care at all about the internal data layout of an * object, and hence it also doesn't care about tiling or swizzling. There's two @@ -58,86 +58,147 @@ * invovlement. */ +/** + * i915_gem_fence_size - required global GTT size for a fence + * @i915: i915 device + * @size: object size + * @tiling: tiling mode + * @stride: tiling stride + * + * Return the required global GTT size for a fence (view of a tiled object), + * taking into account potential fence register mapping. + */ +u32 i915_gem_fence_size(struct drm_i915_private *i915, + u32 size, unsigned int tiling, unsigned int stride) +{ + u32 ggtt_size; + + GEM_BUG_ON(!size); + + if (tiling == I915_TILING_NONE) + return size; + + GEM_BUG_ON(!stride); + + if (INTEL_GEN(i915) >= 4) { + stride *= i915_gem_tile_height(tiling); + GEM_BUG_ON(!IS_ALIGNED(stride, I965_FENCE_PAGE)); + return roundup(size, stride); + } + + /* Previous chips need a power-of-two fence region when tiling */ + if (IS_GEN3(i915)) + ggtt_size = 1024*1024; + else + ggtt_size = 512*1024; + + while (ggtt_size < size) + ggtt_size <<= 1; + + return ggtt_size; +} + +/** + * i915_gem_fence_alignment - required global GTT alignment for a fence + * @i915: i915 device + * @size: object size + * @tiling: tiling mode + * @stride: tiling stride + * + * Return the required global GTT alignment for a fence (a view of a tiled + * object), taking into account potential fence register mapping. + */ +u32 i915_gem_fence_alignment(struct drm_i915_private *i915, u32 size, + unsigned int tiling, unsigned int stride) +{ + GEM_BUG_ON(!size); + + /* + * Minimum alignment is 4k (GTT page size), but might be greater + * if a fence register is needed for the object. + */ + if (tiling == I915_TILING_NONE) + return I915_GTT_MIN_ALIGNMENT; + + if (INTEL_GEN(i915) >= 4) + return I965_FENCE_PAGE; + + /* + * Previous chips need to be aligned to the size of the smallest + * fence register that can contain the object. + */ + return i915_gem_fence_size(i915, size, tiling, stride); +} + /* Check pitch constriants for all chips & tiling formats */ static bool -i915_tiling_ok(struct drm_i915_private *dev_priv, - int stride, int size, int tiling_mode) +i915_tiling_ok(struct drm_i915_gem_object *obj, + unsigned int tiling, unsigned int stride) { - int tile_width; + struct drm_i915_private *i915 = to_i915(obj->base.dev); + unsigned int tile_width; /* Linear is always fine */ - if (tiling_mode == I915_TILING_NONE) + if (tiling == I915_TILING_NONE) return true; - if (tiling_mode > I915_TILING_LAST) + if (tiling > I915_TILING_LAST) return false; - if (IS_GEN2(dev_priv) || - (tiling_mode == I915_TILING_Y && HAS_128_BYTE_Y_TILING(dev_priv))) - tile_width = 128; - else - tile_width = 512; - /* check maximum stride & object size */ /* i965+ stores the end address of the gtt mapping in the fence * reg, so dont bother to check the size */ - if (INTEL_GEN(dev_priv) >= 7) { + if (INTEL_GEN(i915) >= 7) { if (stride / 128 > GEN7_FENCE_MAX_PITCH_VAL) return false; - } else if (INTEL_GEN(dev_priv) >= 4) { + } else if (INTEL_GEN(i915) >= 4) { if (stride / 128 > I965_FENCE_MAX_PITCH_VAL) return false; } else { if (stride > 8192) return false; - if (IS_GEN3(dev_priv)) { - if (size > I830_FENCE_MAX_SIZE_VAL << 20) + if (IS_GEN3(i915)) { + if (obj->base.size > I830_FENCE_MAX_SIZE_VAL << 20) return false; } else { - if (size > I830_FENCE_MAX_SIZE_VAL << 19) + if (obj->base.size > I830_FENCE_MAX_SIZE_VAL << 19) return false; } } - if (stride < tile_width) + if (IS_GEN2(i915) || + (tiling == I915_TILING_Y && HAS_128_BYTE_Y_TILING(i915))) + tile_width = 128; + else + tile_width = 512; + + if (!IS_ALIGNED(stride, tile_width)) return false; /* 965+ just needs multiples of tile width */ - if (INTEL_GEN(dev_priv) >= 4) { - if (stride & (tile_width - 1)) - return false; + if (INTEL_GEN(i915) >= 4) return true; - } /* Pre-965 needs power of two tile widths */ - if (stride & (stride - 1)) - return false; - - return true; + return is_power_of_2(stride); } -static bool i915_vma_fence_prepare(struct i915_vma *vma, int tiling_mode) +static bool i915_vma_fence_prepare(struct i915_vma *vma, + int tiling_mode, unsigned int stride) { - struct drm_i915_private *dev_priv = vma->vm->i915; - u32 size; + struct drm_i915_private *i915 = vma->vm->i915; + u32 size, alignment; if (!i915_vma_is_map_and_fenceable(vma)) return true; - if (INTEL_GEN(dev_priv) == 3) { - if (vma->node.start & ~I915_FENCE_START_MASK) - return false; - } else { - if (vma->node.start & ~I830_FENCE_START_MASK) - return false; - } - - size = i915_gem_get_ggtt_size(dev_priv, vma->size, tiling_mode); + size = i915_gem_fence_size(i915, vma->size, tiling_mode, stride); if (vma->node.size < size) return false; - if (vma->node.start & (size - 1)) + alignment = i915_gem_fence_alignment(i915, vma->size, tiling_mode, stride); + if (!IS_ALIGNED(vma->node.start, alignment)) return false; return true; @@ -145,20 +206,20 @@ static bool i915_vma_fence_prepare(struct i915_vma *vma, int tiling_mode) /* Make the current GTT allocation valid for the change in tiling. */ static int -i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, int tiling_mode) +i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, + int tiling_mode, unsigned int stride) { - struct drm_i915_private *dev_priv = to_i915(obj->base.dev); struct i915_vma *vma; int ret; if (tiling_mode == I915_TILING_NONE) return 0; - if (INTEL_GEN(dev_priv) >= 4) - return 0; - list_for_each_entry(vma, &obj->vma_list, obj_link) { - if (i915_vma_fence_prepare(vma, tiling_mode)) + if (!i915_vma_is_ggtt(vma)) + break; + + if (i915_vma_fence_prepare(vma, tiling_mode, stride)) continue; ret = i915_vma_unbind(vma); @@ -169,8 +230,100 @@ i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, int tiling_mode) return 0; } +int +i915_gem_object_set_tiling(struct drm_i915_gem_object *obj, + unsigned int tiling, unsigned int stride) +{ + struct drm_i915_private *i915 = to_i915(obj->base.dev); + struct i915_vma *vma; + int err; + + /* Make sure we don't cross-contaminate obj->tiling_and_stride */ + BUILD_BUG_ON(I915_TILING_LAST & STRIDE_MASK); + + GEM_BUG_ON(!i915_tiling_ok(obj, tiling, stride)); + GEM_BUG_ON(!stride ^ (tiling == I915_TILING_NONE)); + lockdep_assert_held(&i915->drm.struct_mutex); + + if ((tiling | stride) == obj->tiling_and_stride) + return 0; + + if (obj->framebuffer_references) + return -EBUSY; + + /* We need to rebind the object if its current allocation + * no longer meets the alignment restrictions for its new + * tiling mode. Otherwise we can just leave it alone, but + * need to ensure that any fence register is updated before + * the next fenced (either through the GTT or by the BLT unit + * on older GPUs) access. + * + * After updating the tiling parameters, we then flag whether + * we need to update an associated fence register. Note this + * has to also include the unfenced register the GPU uses + * whilst executing a fenced command for an untiled object. + */ + + err = i915_gem_object_fence_prepare(obj, tiling, stride); + if (err) + return err; + + /* If the memory has unknown (i.e. varying) swizzling, we pin the + * pages to prevent them being swapped out and causing corruption + * due to the change in swizzling. + */ + mutex_lock(&obj->mm.lock); + if (obj->mm.pages && + obj->mm.madv == I915_MADV_WILLNEED && + i915->quirks & QUIRK_PIN_SWIZZLED_PAGES) { + if (tiling == I915_TILING_NONE) { + GEM_BUG_ON(!obj->mm.quirked); + __i915_gem_object_unpin_pages(obj); + obj->mm.quirked = false; + } + if (!i915_gem_object_is_tiled(obj)) { + GEM_BUG_ON(!obj->mm.quirked); + __i915_gem_object_pin_pages(obj); + obj->mm.quirked = true; + } + } + mutex_unlock(&obj->mm.lock); + + list_for_each_entry(vma, &obj->vma_list, obj_link) { + if (!i915_vma_is_ggtt(vma)) + break; + + vma->fence_size = + i915_gem_fence_size(i915, vma->size, tiling, stride); + vma->fence_alignment = + i915_gem_fence_alignment(i915, + vma->size, tiling, stride); + + if (vma->fence) + vma->fence->dirty = true; + } + + obj->tiling_and_stride = tiling | stride; + + /* Force the fence to be reacquired for GTT access */ + i915_gem_release_mmap(obj); + + /* Try to preallocate memory required to save swizzling on put-pages */ + if (i915_gem_object_needs_bit17_swizzle(obj)) { + if (!obj->bit_17) { + obj->bit_17 = kcalloc(BITS_TO_LONGS(obj->base.size >> PAGE_SHIFT), + sizeof(long), GFP_KERNEL); + } + } else { + kfree(obj->bit_17); + obj->bit_17 = NULL; + } + + return 0; +} + /** - * i915_gem_set_tiling - IOCTL handler to set tiling mode + * i915_gem_set_tiling_ioctl - IOCTL handler to set tiling mode * @dev: DRM device * @data: data pointer for the ioctl * @file: DRM file for the ioctl call @@ -184,30 +337,19 @@ i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, int tiling_mode) * Zero on success, negative errno on failure. */ int -i915_gem_set_tiling(struct drm_device *dev, void *data, - struct drm_file *file) +i915_gem_set_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file) { struct drm_i915_gem_set_tiling *args = data; - struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_gem_object *obj; - int err = 0; - - /* Make sure we don't cross-contaminate obj->tiling_and_stride */ - BUILD_BUG_ON(I915_TILING_LAST & STRIDE_MASK); + int err; obj = i915_gem_object_lookup(file, args->handle); if (!obj) return -ENOENT; - if (!i915_tiling_ok(dev_priv, - args->stride, obj->base.size, args->tiling_mode)) { - i915_gem_object_put(obj); - return -EINVAL; - } - - mutex_lock(&dev->struct_mutex); - if (obj->pin_display || obj->framebuffer_references) { - err = -EBUSY; + if (!i915_tiling_ok(obj, args->tiling_mode, args->stride)) { + err = -EINVAL; goto err; } @@ -216,9 +358,9 @@ i915_gem_set_tiling(struct drm_device *dev, void *data, args->stride = 0; } else { if (args->tiling_mode == I915_TILING_X) - args->swizzle_mode = dev_priv->mm.bit_6_swizzle_x; + args->swizzle_mode = to_i915(dev)->mm.bit_6_swizzle_x; else - args->swizzle_mode = dev_priv->mm.bit_6_swizzle_y; + args->swizzle_mode = to_i915(dev)->mm.bit_6_swizzle_y; /* Hide bit 17 swizzling from the user. This prevents old Mesa * from aborting the application on sw fallbacks to bit 17, @@ -240,79 +382,24 @@ i915_gem_set_tiling(struct drm_device *dev, void *data, } } - if (args->tiling_mode != i915_gem_object_get_tiling(obj) || - args->stride != i915_gem_object_get_stride(obj)) { - /* We need to rebind the object if its current allocation - * no longer meets the alignment restrictions for its new - * tiling mode. Otherwise we can just leave it alone, but - * need to ensure that any fence register is updated before - * the next fenced (either through the GTT or by the BLT unit - * on older GPUs) access. - * - * After updating the tiling parameters, we then flag whether - * we need to update an associated fence register. Note this - * has to also include the unfenced register the GPU uses - * whilst executing a fenced command for an untiled object. - */ + err = mutex_lock_interruptible(&dev->struct_mutex); + if (err) + goto err; - err = i915_gem_object_fence_prepare(obj, args->tiling_mode); - if (!err) { - struct i915_vma *vma; - - mutex_lock(&obj->mm.lock); - if (obj->mm.pages && - obj->mm.madv == I915_MADV_WILLNEED && - dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES) { - if (args->tiling_mode == I915_TILING_NONE) { - GEM_BUG_ON(!obj->mm.quirked); - __i915_gem_object_unpin_pages(obj); - obj->mm.quirked = false; - } - if (!i915_gem_object_is_tiled(obj)) { - GEM_BUG_ON(!obj->mm.quirked); - __i915_gem_object_pin_pages(obj); - obj->mm.quirked = true; - } - } - mutex_unlock(&obj->mm.lock); - - list_for_each_entry(vma, &obj->vma_list, obj_link) { - if (!vma->fence) - continue; - - vma->fence->dirty = true; - } - obj->tiling_and_stride = - args->stride | args->tiling_mode; - - /* Force the fence to be reacquired for GTT access */ - i915_gem_release_mmap(obj); - } - } - /* we have to maintain this existing ABI... */ + err = i915_gem_object_set_tiling(obj, args->tiling_mode, args->stride); + mutex_unlock(&dev->struct_mutex); + + /* We have to maintain this existing ABI... */ args->stride = i915_gem_object_get_stride(obj); args->tiling_mode = i915_gem_object_get_tiling(obj); - /* Try to preallocate memory required to save swizzling on put-pages */ - if (i915_gem_object_needs_bit17_swizzle(obj)) { - if (obj->bit_17 == NULL) { - obj->bit_17 = kcalloc(BITS_TO_LONGS(obj->base.size >> PAGE_SHIFT), - sizeof(long), GFP_KERNEL); - } - } else { - kfree(obj->bit_17); - obj->bit_17 = NULL; - } - err: i915_gem_object_put(obj); - mutex_unlock(&dev->struct_mutex); - return err; } /** - * i915_gem_get_tiling - IOCTL handler to get tiling mode + * i915_gem_get_tiling_ioctl - IOCTL handler to get tiling mode * @dev: DRM device * @data: data pointer for the ioctl * @file: DRM file for the ioctl call @@ -325,8 +412,8 @@ err: * Zero on success, negative errno on failure. */ int -i915_gem_get_tiling(struct drm_device *dev, void *data, - struct drm_file *file) +i915_gem_get_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file) { struct drm_i915_gem_get_tiling *args = data; struct drm_i915_private *dev_priv = to_i915(dev); diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 396c6f0fd033..9cd22cda17af 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -121,6 +121,7 @@ static void __i915_error_advance(struct drm_i915_error_state_buf *e, e->pos += len; } +__printf(2, 0) static void i915_error_vprintf(struct drm_i915_error_state_buf *e, const char *f, va_list args) { diff --git a/drivers/gpu/drm/i915/i915_guc_reg.h b/drivers/gpu/drm/i915/i915_guc_reg.h index 6a0adafe0523..35cf9918d09a 100644 --- a/drivers/gpu/drm/i915/i915_guc_reg.h +++ b/drivers/gpu/drm/i915/i915_guc_reg.h @@ -61,12 +61,18 @@ #define DMA_ADDRESS_SPACE_GTT (8 << 16) #define DMA_COPY_SIZE _MMIO(0xc310) #define DMA_CTRL _MMIO(0xc314) +#define HUC_UKERNEL (1<<9) #define UOS_MOVE (1<<4) #define START_DMA (1<<0) #define DMA_GUC_WOPCM_OFFSET _MMIO(0xc340) +#define HUC_LOADING_AGENT_VCR (0<<1) +#define HUC_LOADING_AGENT_GUC (1<<1) #define GUC_WOPCM_OFFSET_VALUE 0x80000 /* 512KB */ #define GUC_MAX_IDLE_COUNT _MMIO(0xC3E4) +#define HUC_STATUS2 _MMIO(0xD3B0) +#define HUC_FW_VERIFIED (1<<7) + /* Defines WOPCM space available to GuC firmware */ #define GUC_WOPCM_SIZE _MMIO(0xc050) /* GuC addresses below GUC_WOPCM_TOP don't map through the GTT */ diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index 710fbb9fc63f..8ced9e26f075 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -22,8 +22,6 @@ * */ #include <linux/circ_buf.h> -#include <linux/debugfs.h> -#include <linux/relay.h> #include "i915_drv.h" #include "intel_uc.h" @@ -545,7 +543,7 @@ static void i915_guc_submit(struct drm_i915_gem_request *rq) */ /** - * guc_allocate_vma() - Allocate a GGTT VMA for GuC usage + * intel_guc_allocate_vma() - Allocate a GGTT VMA for GuC usage * @guc: the guc * @size: size of area to allocate (both virtual space and memory) * @@ -557,7 +555,7 @@ static void i915_guc_submit(struct drm_i915_gem_request *rq) * * Return: A i915_vma if successful, otherwise an ERR_PTR. */ -static struct i915_vma *guc_allocate_vma(struct intel_guc *guc, u32 size) +struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size) { struct drm_i915_private *dev_priv = guc_to_i915(guc); struct drm_i915_gem_object *obj; @@ -568,7 +566,7 @@ static struct i915_vma *guc_allocate_vma(struct intel_guc *guc, u32 size) if (IS_ERR(obj)) return ERR_CAST(obj); - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) goto err; @@ -579,9 +577,6 @@ static struct i915_vma *guc_allocate_vma(struct intel_guc *guc, u32 size) goto err; } - /* Invalidate GuC TLB to let GuC take the latest updates to GTT. */ - I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); - return vma; err: @@ -721,7 +716,7 @@ guc_client_alloc(struct drm_i915_private *dev_priv, } /* The first page is doorbell/proc_desc. Two followed pages are wq. */ - vma = guc_allocate_vma(guc, GUC_DB_SIZE + GUC_WQ_SIZE); + vma = intel_guc_allocate_vma(guc, GUC_DB_SIZE + GUC_WQ_SIZE); if (IS_ERR(vma)) goto err; @@ -777,488 +772,7 @@ err: return NULL; } -/* - * Sub buffer switch callback. Called whenever relay has to switch to a new - * sub buffer, relay stays on the same sub buffer if 0 is returned. - */ -static int subbuf_start_callback(struct rchan_buf *buf, - void *subbuf, - void *prev_subbuf, - size_t prev_padding) -{ - /* Use no-overwrite mode by default, where relay will stop accepting - * new data if there are no empty sub buffers left. - * There is no strict synchronization enforced by relay between Consumer - * and Producer. In overwrite mode, there is a possibility of getting - * inconsistent/garbled data, the producer could be writing on to the - * same sub buffer from which Consumer is reading. This can't be avoided - * unless Consumer is fast enough and can always run in tandem with - * Producer. - */ - if (relay_buf_full(buf)) - return 0; - - return 1; -} - -/* - * file_create() callback. Creates relay file in debugfs. - */ -static struct dentry *create_buf_file_callback(const char *filename, - struct dentry *parent, - umode_t mode, - struct rchan_buf *buf, - int *is_global) -{ - struct dentry *buf_file; - - /* This to enable the use of a single buffer for the relay channel and - * correspondingly have a single file exposed to User, through which - * it can collect the logs in order without any post-processing. - * Need to set 'is_global' even if parent is NULL for early logging. - */ - *is_global = 1; - - if (!parent) - return NULL; - - /* Not using the channel filename passed as an argument, since for each - * channel relay appends the corresponding CPU number to the filename - * passed in relay_open(). This should be fine as relay just needs a - * dentry of the file associated with the channel buffer and that file's - * name need not be same as the filename passed as an argument. - */ - buf_file = debugfs_create_file("guc_log", mode, - parent, buf, &relay_file_operations); - return buf_file; -} - -/* - * file_remove() default callback. Removes relay file in debugfs. - */ -static int remove_buf_file_callback(struct dentry *dentry) -{ - debugfs_remove(dentry); - return 0; -} - -/* relay channel callbacks */ -static struct rchan_callbacks relay_callbacks = { - .subbuf_start = subbuf_start_callback, - .create_buf_file = create_buf_file_callback, - .remove_buf_file = remove_buf_file_callback, -}; - -static void guc_log_remove_relay_file(struct intel_guc *guc) -{ - relay_close(guc->log.relay_chan); -} - -static int guc_log_create_relay_channel(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct rchan *guc_log_relay_chan; - size_t n_subbufs, subbuf_size; - - /* Keep the size of sub buffers same as shared log buffer */ - subbuf_size = guc->log.vma->obj->base.size; - - /* Store up to 8 snapshots, which is large enough to buffer sufficient - * boot time logs and provides enough leeway to User, in terms of - * latency, for consuming the logs from relay. Also doesn't take - * up too much memory. - */ - n_subbufs = 8; - - guc_log_relay_chan = relay_open(NULL, NULL, subbuf_size, - n_subbufs, &relay_callbacks, dev_priv); - if (!guc_log_relay_chan) { - DRM_ERROR("Couldn't create relay chan for GuC logging\n"); - return -ENOMEM; - } - - GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); - guc->log.relay_chan = guc_log_relay_chan; - return 0; -} - -static int guc_log_create_relay_file(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct dentry *log_dir; - int ret; - - /* For now create the log file in /sys/kernel/debug/dri/0 dir */ - log_dir = dev_priv->drm.primary->debugfs_root; - - /* If /sys/kernel/debug/dri/0 location do not exist, then debugfs is - * not mounted and so can't create the relay file. - * The relay API seems to fit well with debugfs only, for availing relay - * there are 3 requirements which can be met for debugfs file only in a - * straightforward/clean manner :- - * i) Need the associated dentry pointer of the file, while opening the - * relay channel. - * ii) Should be able to use 'relay_file_operations' fops for the file. - * iii) Set the 'i_private' field of file's inode to the pointer of - * relay channel buffer. - */ - if (!log_dir) { - DRM_ERROR("Debugfs dir not available yet for GuC log file\n"); - return -ENODEV; - } - - ret = relay_late_setup_files(guc->log.relay_chan, "guc_log", log_dir); - if (ret) { - DRM_ERROR("Couldn't associate relay chan with file %d\n", ret); - return ret; - } - - return 0; -} - -static void guc_move_to_next_buf(struct intel_guc *guc) -{ - /* Make sure the updates made in the sub buffer are visible when - * Consumer sees the following update to offset inside the sub buffer. - */ - smp_wmb(); - - /* All data has been written, so now move the offset of sub buffer. */ - relay_reserve(guc->log.relay_chan, guc->log.vma->obj->base.size); - - /* Switch to the next sub buffer */ - relay_flush(guc->log.relay_chan); -} - -static void *guc_get_write_buffer(struct intel_guc *guc) -{ - if (!guc->log.relay_chan) - return NULL; - - /* Just get the base address of a new sub buffer and copy data into it - * ourselves. NULL will be returned in no-overwrite mode, if all sub - * buffers are full. Could have used the relay_write() to indirectly - * copy the data, but that would have been bit convoluted, as we need to - * write to only certain locations inside a sub buffer which cannot be - * done without using relay_reserve() along with relay_write(). So its - * better to use relay_reserve() alone. - */ - return relay_reserve(guc->log.relay_chan, 0); -} - -static bool -guc_check_log_buf_overflow(struct intel_guc *guc, - enum guc_log_buffer_type type, unsigned int full_cnt) -{ - unsigned int prev_full_cnt = guc->log.prev_overflow_count[type]; - bool overflow = false; - - if (full_cnt != prev_full_cnt) { - overflow = true; - - guc->log.prev_overflow_count[type] = full_cnt; - guc->log.total_overflow_count[type] += full_cnt - prev_full_cnt; - - if (full_cnt < prev_full_cnt) { - /* buffer_full_cnt is a 4 bit counter */ - guc->log.total_overflow_count[type] += 16; - } - DRM_ERROR_RATELIMITED("GuC log buffer overflow\n"); - } - - return overflow; -} - -static unsigned int guc_get_log_buffer_size(enum guc_log_buffer_type type) -{ - switch (type) { - case GUC_ISR_LOG_BUFFER: - return (GUC_LOG_ISR_PAGES + 1) * PAGE_SIZE; - case GUC_DPC_LOG_BUFFER: - return (GUC_LOG_DPC_PAGES + 1) * PAGE_SIZE; - case GUC_CRASH_DUMP_LOG_BUFFER: - return (GUC_LOG_CRASH_PAGES + 1) * PAGE_SIZE; - default: - MISSING_CASE(type); - } - - return 0; -} - -static void guc_read_update_log_buffer(struct intel_guc *guc) -{ - unsigned int buffer_size, read_offset, write_offset, bytes_to_copy, full_cnt; - struct guc_log_buffer_state *log_buf_state, *log_buf_snapshot_state; - struct guc_log_buffer_state log_buf_state_local; - enum guc_log_buffer_type type; - void *src_data, *dst_data; - bool new_overflow; - - if (WARN_ON(!guc->log.buf_addr)) - return; - - /* Get the pointer to shared GuC log buffer */ - log_buf_state = src_data = guc->log.buf_addr; - - /* Get the pointer to local buffer to store the logs */ - log_buf_snapshot_state = dst_data = guc_get_write_buffer(guc); - - /* Actual logs are present from the 2nd page */ - src_data += PAGE_SIZE; - dst_data += PAGE_SIZE; - - for (type = GUC_ISR_LOG_BUFFER; type < GUC_MAX_LOG_BUFFER; type++) { - /* Make a copy of the state structure, inside GuC log buffer - * (which is uncached mapped), on the stack to avoid reading - * from it multiple times. - */ - memcpy(&log_buf_state_local, log_buf_state, - sizeof(struct guc_log_buffer_state)); - buffer_size = guc_get_log_buffer_size(type); - read_offset = log_buf_state_local.read_ptr; - write_offset = log_buf_state_local.sampled_write_ptr; - full_cnt = log_buf_state_local.buffer_full_cnt; - - /* Bookkeeping stuff */ - guc->log.flush_count[type] += log_buf_state_local.flush_to_file; - new_overflow = guc_check_log_buf_overflow(guc, type, full_cnt); - - /* Update the state of shared log buffer */ - log_buf_state->read_ptr = write_offset; - log_buf_state->flush_to_file = 0; - log_buf_state++; - - if (unlikely(!log_buf_snapshot_state)) - continue; - - /* First copy the state structure in snapshot buffer */ - memcpy(log_buf_snapshot_state, &log_buf_state_local, - sizeof(struct guc_log_buffer_state)); - - /* The write pointer could have been updated by GuC firmware, - * after sending the flush interrupt to Host, for consistency - * set write pointer value to same value of sampled_write_ptr - * in the snapshot buffer. - */ - log_buf_snapshot_state->write_ptr = write_offset; - log_buf_snapshot_state++; - - /* Now copy the actual logs. */ - if (unlikely(new_overflow)) { - /* copy the whole buffer in case of overflow */ - read_offset = 0; - write_offset = buffer_size; - } else if (unlikely((read_offset > buffer_size) || - (write_offset > buffer_size))) { - DRM_ERROR("invalid log buffer state\n"); - /* copy whole buffer as offsets are unreliable */ - read_offset = 0; - write_offset = buffer_size; - } - - /* Just copy the newly written data */ - if (read_offset > write_offset) { - i915_memcpy_from_wc(dst_data, src_data, write_offset); - bytes_to_copy = buffer_size - read_offset; - } else { - bytes_to_copy = write_offset - read_offset; - } - i915_memcpy_from_wc(dst_data + read_offset, - src_data + read_offset, bytes_to_copy); - - src_data += buffer_size; - dst_data += buffer_size; - } - - if (log_buf_snapshot_state) - guc_move_to_next_buf(guc); - else { - /* Used rate limited to avoid deluge of messages, logs might be - * getting consumed by User at a slow rate. - */ - DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); - guc->log.capture_miss_count++; - } -} - -static void guc_capture_logs_work(struct work_struct *work) -{ - struct drm_i915_private *dev_priv = - container_of(work, struct drm_i915_private, guc.log.flush_work); - - i915_guc_capture_logs(dev_priv); -} - -static void guc_log_cleanup(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - /* First disable the flush interrupt */ - gen9_disable_guc_interrupts(dev_priv); - if (guc->log.flush_wq) - destroy_workqueue(guc->log.flush_wq); - - guc->log.flush_wq = NULL; - - if (guc->log.relay_chan) - guc_log_remove_relay_file(guc); - - guc->log.relay_chan = NULL; - - if (guc->log.buf_addr) - i915_gem_object_unpin_map(guc->log.vma->obj); - - guc->log.buf_addr = NULL; -} - -static int guc_log_create_extras(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - void *vaddr; - int ret; - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - /* Nothing to do */ - if (i915.guc_log_level < 0) - return 0; - - if (!guc->log.buf_addr) { - /* Create a WC (Uncached for read) vmalloc mapping of log - * buffer pages, so that we can directly get the data - * (up-to-date) from memory. - */ - vaddr = i915_gem_object_pin_map(guc->log.vma->obj, I915_MAP_WC); - if (IS_ERR(vaddr)) { - ret = PTR_ERR(vaddr); - DRM_ERROR("Couldn't map log buffer pages %d\n", ret); - return ret; - } - - guc->log.buf_addr = vaddr; - } - - if (!guc->log.relay_chan) { - /* Create a relay channel, so that we have buffers for storing - * the GuC firmware logs, the channel will be linked with a file - * later on when debugfs is registered. - */ - ret = guc_log_create_relay_channel(guc); - if (ret) - return ret; - } - - if (!guc->log.flush_wq) { - INIT_WORK(&guc->log.flush_work, guc_capture_logs_work); - - /* - * GuC log buffer flush work item has to do register access to - * send the ack to GuC and this work item, if not synced before - * suspend, can potentially get executed after the GFX device is - * suspended. - * By marking the WQ as freezable, we don't have to bother about - * flushing of this work item from the suspend hooks, the pending - * work item if any will be either executed before the suspend - * or scheduled later on resume. This way the handling of work - * item can be kept same between system suspend & rpm suspend. - */ - guc->log.flush_wq = alloc_ordered_workqueue("i915-guc_log", - WQ_HIGHPRI | WQ_FREEZABLE); - if (guc->log.flush_wq == NULL) { - DRM_ERROR("Couldn't allocate the wq for GuC logging\n"); - return -ENOMEM; - } - } - - return 0; -} - -static void guc_log_create(struct intel_guc *guc) -{ - struct i915_vma *vma; - unsigned long offset; - uint32_t size, flags; - - if (i915.guc_log_level > GUC_LOG_VERBOSITY_MAX) - i915.guc_log_level = GUC_LOG_VERBOSITY_MAX; - - /* The first page is to save log buffer state. Allocate one - * extra page for others in case for overlap */ - size = (1 + GUC_LOG_DPC_PAGES + 1 + - GUC_LOG_ISR_PAGES + 1 + - GUC_LOG_CRASH_PAGES + 1) << PAGE_SHIFT; - - vma = guc->log.vma; - if (!vma) { - /* We require SSE 4.1 for fast reads from the GuC log buffer and - * it should be present on the chipsets supporting GuC based - * submisssions. - */ - if (WARN_ON(!i915_has_memcpy_from_wc())) { - /* logging will not be enabled */ - i915.guc_log_level = -1; - return; - } - - vma = guc_allocate_vma(guc, size); - if (IS_ERR(vma)) { - /* logging will be off */ - i915.guc_log_level = -1; - return; - } - - guc->log.vma = vma; - - if (guc_log_create_extras(guc)) { - guc_log_cleanup(guc); - i915_vma_unpin_and_release(&guc->log.vma); - i915.guc_log_level = -1; - return; - } - } - - /* each allocated unit is a page */ - flags = GUC_LOG_VALID | GUC_LOG_NOTIFY_ON_HALF_FULL | - (GUC_LOG_DPC_PAGES << GUC_LOG_DPC_SHIFT) | - (GUC_LOG_ISR_PAGES << GUC_LOG_ISR_SHIFT) | - (GUC_LOG_CRASH_PAGES << GUC_LOG_CRASH_SHIFT); - - offset = guc_ggtt_offset(vma) >> PAGE_SHIFT; /* in pages */ - guc->log.flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; -} - -static int guc_log_late_setup(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - int ret; - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - if (i915.guc_log_level < 0) - return -EINVAL; - - /* If log_level was set as -1 at boot time, then setup needed to - * handle log buffer flush interrupts would not have been done yet, - * so do that now. - */ - ret = guc_log_create_extras(guc); - if (ret) - goto err; - - ret = guc_log_create_relay_file(guc); - if (ret) - goto err; - - return 0; -err: - guc_log_cleanup(guc); - /* logging will remain off */ - i915.guc_log_level = -1; - return ret; -} static void guc_policies_init(struct guc_policies *policies) { @@ -1301,7 +815,7 @@ static void guc_addon_create(struct intel_guc *guc) vma = guc->ads_vma; if (!vma) { - vma = guc_allocate_vma(guc, PAGE_ALIGN(size)); + vma = intel_guc_allocate_vma(guc, PAGE_ALIGN(size)); if (IS_ERR(vma)) return; @@ -1376,13 +890,13 @@ int i915_guc_submission_init(struct drm_i915_private *dev_priv) if (guc->ctx_pool_vma) return 0; /* already allocated */ - vma = guc_allocate_vma(guc, gemsize); + vma = intel_guc_allocate_vma(guc, gemsize); if (IS_ERR(vma)) return PTR_ERR(vma); guc->ctx_pool_vma = vma; ida_init(&guc->ctx_ids); - guc_log_create(guc); + intel_guc_log_create(guc); guc_addon_create(guc); guc->execbuf_client = guc_client_alloc(dev_priv, @@ -1484,7 +998,7 @@ int intel_guc_suspend(struct drm_i915_private *dev_priv) struct i915_gem_context *ctx; u32 data[3]; - if (guc->guc_fw.guc_fw_load_status != GUC_FIRMWARE_SUCCESS) + if (guc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) return 0; gen9_disable_guc_interrupts(dev_priv); @@ -1511,7 +1025,7 @@ int intel_guc_resume(struct drm_i915_private *dev_priv) struct i915_gem_context *ctx; u32 data[3]; - if (guc->guc_fw.guc_fw_load_status != GUC_FIRMWARE_SUCCESS) + if (guc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) return 0; if (i915.guc_log_level >= 0) @@ -1527,103 +1041,4 @@ int intel_guc_resume(struct drm_i915_private *dev_priv) return intel_guc_send(guc, data, ARRAY_SIZE(data)); } -void i915_guc_capture_logs(struct drm_i915_private *dev_priv) -{ - guc_read_update_log_buffer(&dev_priv->guc); - - /* Generally device is expected to be active only at this - * time, so get/put should be really quick. - */ - intel_runtime_pm_get(dev_priv); - intel_guc_log_flush_complete(&dev_priv->guc); - intel_runtime_pm_put(dev_priv); -} - -void i915_guc_flush_logs(struct drm_i915_private *dev_priv) -{ - if (!i915.enable_guc_submission || (i915.guc_log_level < 0)) - return; - - /* First disable the interrupts, will be renabled afterwards */ - gen9_disable_guc_interrupts(dev_priv); - - /* Before initiating the forceful flush, wait for any pending/ongoing - * flush to complete otherwise forceful flush may not actually happen. - */ - flush_work(&dev_priv->guc.log.flush_work); - - /* Ask GuC to update the log buffer state */ - intel_guc_log_flush(&dev_priv->guc); - - /* GuC would have updated log buffer by now, so capture it */ - i915_guc_capture_logs(dev_priv); -} - -void i915_guc_unregister(struct drm_i915_private *dev_priv) -{ - if (!i915.enable_guc_submission) - return; - mutex_lock(&dev_priv->drm.struct_mutex); - guc_log_cleanup(&dev_priv->guc); - mutex_unlock(&dev_priv->drm.struct_mutex); -} - -void i915_guc_register(struct drm_i915_private *dev_priv) -{ - if (!i915.enable_guc_submission) - return; - - mutex_lock(&dev_priv->drm.struct_mutex); - guc_log_late_setup(&dev_priv->guc); - mutex_unlock(&dev_priv->drm.struct_mutex); -} - -int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) -{ - union guc_log_control log_param; - int ret; - - log_param.value = control_val; - - if (log_param.verbosity < GUC_LOG_VERBOSITY_MIN || - log_param.verbosity > GUC_LOG_VERBOSITY_MAX) - return -EINVAL; - - /* This combination doesn't make sense & won't have any effect */ - if (!log_param.logging_enabled && (i915.guc_log_level < 0)) - return 0; - - ret = intel_guc_log_control(&dev_priv->guc, log_param.value); - if (ret < 0) { - DRM_DEBUG_DRIVER("guc_logging_control action failed %d\n", ret); - return ret; - } - - i915.guc_log_level = log_param.verbosity; - - /* If log_level was set as -1 at boot time, then the relay channel file - * wouldn't have been created by now and interrupts also would not have - * been enabled. - */ - if (!dev_priv->guc.log.relay_chan) { - ret = guc_log_late_setup(&dev_priv->guc); - if (!ret) - gen9_enable_guc_interrupts(dev_priv); - } else if (!log_param.logging_enabled) { - /* Once logging is disabled, GuC won't generate logs & send an - * interrupt. But there could be some data in the log buffer - * which is yet to be captured. So request GuC to update the log - * buffer state and then collect the left over logs. - */ - i915_guc_flush_logs(dev_priv); - - /* As logging is disabled, update log level to reflect that */ - i915.guc_log_level = -1; - } else { - /* In case interrupts were disabled, enable them now */ - gen9_enable_guc_interrupts(dev_priv); - } - - return ret; -} diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index a0e70f5b3aad..6fefc34ef602 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1170,6 +1170,9 @@ static void gen6_pm_rps_work(struct work_struct *work) adj *= 2; else /* CHV needs even encode values */ adj = IS_CHERRYVIEW(dev_priv) ? 2 : 1; + + if (new_delay >= dev_priv->rps.max_freq_softlimit) + adj = 0; /* * For better performance, jump directly * to RPe if we're below it. @@ -1191,6 +1194,9 @@ static void gen6_pm_rps_work(struct work_struct *work) adj *= 2; else /* CHV needs even encode values */ adj = IS_CHERRYVIEW(dev_priv) ? -2 : -1; + + if (new_delay <= dev_priv->rps.min_freq_softlimit) + adj = 0; } else { /* unknown event */ adj = 0; } @@ -1553,41 +1559,68 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, { struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe]; struct intel_pipe_crc_entry *entry; + struct intel_crtc *crtc = intel_get_crtc_for_pipe(dev_priv, pipe); + struct drm_driver *driver = dev_priv->drm.driver; + uint32_t crcs[5]; int head, tail; spin_lock(&pipe_crc->lock); + if (pipe_crc->source) { + if (!pipe_crc->entries) { + spin_unlock(&pipe_crc->lock); + DRM_DEBUG_KMS("spurious interrupt\n"); + return; + } - if (!pipe_crc->entries) { - spin_unlock(&pipe_crc->lock); - DRM_DEBUG_KMS("spurious interrupt\n"); - return; - } - - head = pipe_crc->head; - tail = pipe_crc->tail; + head = pipe_crc->head; + tail = pipe_crc->tail; - if (CIRC_SPACE(head, tail, INTEL_PIPE_CRC_ENTRIES_NR) < 1) { - spin_unlock(&pipe_crc->lock); - DRM_ERROR("CRC buffer overflowing\n"); - return; - } + if (CIRC_SPACE(head, tail, INTEL_PIPE_CRC_ENTRIES_NR) < 1) { + spin_unlock(&pipe_crc->lock); + DRM_ERROR("CRC buffer overflowing\n"); + return; + } - entry = &pipe_crc->entries[head]; + entry = &pipe_crc->entries[head]; - entry->frame = dev_priv->drm.driver->get_vblank_counter(&dev_priv->drm, - pipe); - entry->crc[0] = crc0; - entry->crc[1] = crc1; - entry->crc[2] = crc2; - entry->crc[3] = crc3; - entry->crc[4] = crc4; + entry->frame = driver->get_vblank_counter(&dev_priv->drm, pipe); + entry->crc[0] = crc0; + entry->crc[1] = crc1; + entry->crc[2] = crc2; + entry->crc[3] = crc3; + entry->crc[4] = crc4; - head = (head + 1) & (INTEL_PIPE_CRC_ENTRIES_NR - 1); - pipe_crc->head = head; + head = (head + 1) & (INTEL_PIPE_CRC_ENTRIES_NR - 1); + pipe_crc->head = head; - spin_unlock(&pipe_crc->lock); + spin_unlock(&pipe_crc->lock); - wake_up_interruptible(&pipe_crc->wq); + wake_up_interruptible(&pipe_crc->wq); + } else { + /* + * For some not yet identified reason, the first CRC is + * bonkers. So let's just wait for the next vblank and read + * out the buggy result. + * + * On CHV sometimes the second CRC is bonkers as well, so + * don't trust that one either. + */ + if (pipe_crc->skipped == 0 || + (IS_CHERRYVIEW(dev_priv) && pipe_crc->skipped == 1)) { + pipe_crc->skipped++; + spin_unlock(&pipe_crc->lock); + return; + } + spin_unlock(&pipe_crc->lock); + crcs[0] = crc0; + crcs[1] = crc1; + crcs[2] = crc2; + crcs[3] = crc3; + crcs[4] = crc4; + drm_crtc_add_crc_entry(&crtc->base, true, + drm_accurate_vblank_count(&crtc->base), + crcs); + } } #else static inline void @@ -2703,12 +2736,13 @@ static void i915_clear_error_registers(struct drm_i915_private *dev_priv) * i915_handle_error - handle a gpu error * @dev_priv: i915 device private * @engine_mask: mask representing engines that are hung + * @fmt: Error message format string + * * Do some basic checking of register state at error time and * dump it to the syslog. Also call i915_capture_error_state() to make * sure we get a record and make it available in debugfs. Fire a uevent * so userspace knows something bad happened (should trigger collection * of a ring dump etc.). - * @fmt: Error message format string */ void i915_handle_error(struct drm_i915_private *dev_priv, u32 engine_mask, diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 00970aa77afa..72f9f36ae5ce 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -3597,9 +3597,12 @@ enum { #define EDP_PSR_PERF_CNT_MASK 0xffffff #define EDP_PSR_DEBUG_CTL _MMIO(dev_priv->psr_mmio_base + 0x60) -#define EDP_PSR_DEBUG_MASK_LPSP (1<<27) -#define EDP_PSR_DEBUG_MASK_MEMUP (1<<26) -#define EDP_PSR_DEBUG_MASK_HPD (1<<25) +#define EDP_PSR_DEBUG_MASK_MAX_SLEEP (1<<28) +#define EDP_PSR_DEBUG_MASK_LPSP (1<<27) +#define EDP_PSR_DEBUG_MASK_MEMUP (1<<26) +#define EDP_PSR_DEBUG_MASK_HPD (1<<25) +#define EDP_PSR_DEBUG_MASK_DISP_REG_WRITE (1<<16) +#define EDP_PSR_DEBUG_EXIT_ON_PIXEL_UNDERRUN (1<<15) #define EDP_PSR2_CTL _MMIO(0x6f900) #define EDP_PSR2_ENABLE (1<<31) @@ -3614,6 +3617,11 @@ enum { #define EDP_PSR2_FRAME_BEFORE_SU_SHIFT 4 #define EDP_PSR2_FRAME_BEFORE_SU_MASK (0xf<<4) #define EDP_PSR2_IDLE_MASK 0xf +#define EDP_FRAMES_BEFORE_SU_ENTRY (1<<4) + +#define EDP_PSR2_STATUS_CTL _MMIO(0x6f940) +#define EDP_PSR2_STATUS_STATE_MASK (0xf<<28) +#define EDP_PSR2_STATUS_STATE_SHIFT 28 /* VGA port control */ #define ADPA _MMIO(0x61100) @@ -6446,6 +6454,12 @@ enum { #define BDW_DPRS_MASK_VBLANK_SRD (1 << 0) #define CHICKEN_PIPESL_1(pipe) _MMIO_PIPE(pipe, _CHICKEN_PIPESL_1_A, _CHICKEN_PIPESL_1_B) +#define CHICKEN_TRANS_A 0x420c0 +#define CHICKEN_TRANS_B 0x420c4 +#define CHICKEN_TRANS(trans) _MMIO_TRANS(trans, CHICKEN_TRANS_A, CHICKEN_TRANS_B) +#define PSR2_VSC_ENABLE_PROG_HEADER (1<<12) +#define PSR2_ADD_VERTICAL_LINE_COUNT (1<<15) + #define DISP_ARB_CTL _MMIO(0x45000) #define DISP_FBC_MEMORY_WAKE (1<<31) #define DISP_TILE_SURFACE_SWIZZLING (1<<13) diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c index f5a88092dacf..40f4e5efaf83 100644 --- a/drivers/gpu/drm/i915/i915_sw_fence.c +++ b/drivers/gpu/drm/i915/i915_sw_fence.c @@ -63,6 +63,7 @@ static inline void debug_fence_destroy(struct i915_sw_fence *fence) static inline void debug_fence_free(struct i915_sw_fence *fence) { debug_object_free(fence, &i915_sw_fence_debug_descr); + smp_wmb(); /* flush the change in state before reallocation */ } static inline void debug_fence_assert(struct i915_sw_fence *fence) diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index 40c0ac70d79d..376ac957cd1c 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -58,7 +58,7 @@ static u32 calc_residency(struct drm_i915_private *dev_priv, if (I915_READ(VLV_COUNTER_CONTROL) & VLV_COUNT_RANGE_HIGH) units <<= 8; - } else if (IS_BROXTON(dev_priv)) { + } else if (IS_GEN9_LP(dev_priv)) { units = 1; div = 1200; /* 833.33ns */ } diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h index 18ae37c411fd..4461df5a94fe 100644 --- a/drivers/gpu/drm/i915/i915_trace.h +++ b/drivers/gpu/drm/i915/i915_trace.h @@ -450,9 +450,9 @@ TRACE_EVENT(i915_gem_evict_vm, TP_printk("dev=%d, vm=%p", __entry->dev, __entry->vm) ); -TRACE_EVENT(i915_gem_evict_vma, - TP_PROTO(struct i915_vma *vma, unsigned int flags), - TP_ARGS(vma, flags), +TRACE_EVENT(i915_gem_evict_node, + TP_PROTO(struct i915_address_space *vm, struct drm_mm_node *node, unsigned int flags), + TP_ARGS(vm, node, flags), TP_STRUCT__entry( __field(u32, dev) @@ -464,11 +464,11 @@ TRACE_EVENT(i915_gem_evict_vma, ), TP_fast_assign( - __entry->dev = vma->vm->i915->drm.primary->index; - __entry->vm = vma->vm; - __entry->start = vma->node.start; - __entry->size = vma->node.size; - __entry->color = vma->node.color; + __entry->dev = vm->i915->drm.primary->index; + __entry->vm = vm; + __entry->start = node->start; + __entry->size = node->size; + __entry->color = node->color; __entry->flags = flags; ), diff --git a/drivers/gpu/drm/i915/i915_vgpu.c b/drivers/gpu/drm/i915/i915_vgpu.c index dae340cfc6c7..d0abfd08a01c 100644 --- a/drivers/gpu/drm/i915/i915_vgpu.c +++ b/drivers/gpu/drm/i915/i915_vgpu.c @@ -116,22 +116,20 @@ void intel_vgt_deballoon(struct drm_i915_private *dev_priv) memset(&bl_info, 0, sizeof(bl_info)); } -static int vgt_balloon_space(struct drm_mm *mm, +static int vgt_balloon_space(struct i915_ggtt *ggtt, struct drm_mm_node *node, unsigned long start, unsigned long end) { unsigned long size = end - start; - if (start == end) + if (start >= end) return -EINVAL; DRM_INFO("balloon space: range [ 0x%lx - 0x%lx ] %lu KiB.\n", start, end, size / 1024); - - node->start = start; - node->size = size; - - return drm_mm_reserve_node(mm, node); + return i915_gem_gtt_reserve(&ggtt->base, node, + size, start, I915_COLOR_UNEVICTABLE, + 0); } /** @@ -214,10 +212,8 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv) /* Unmappable graphic memory ballooning */ if (unmappable_base > ggtt->mappable_end) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[2], - ggtt->mappable_end, - unmappable_base); + ret = vgt_balloon_space(ggtt, &bl_info.space[2], + ggtt->mappable_end, unmappable_base); if (ret) goto err; @@ -228,18 +224,15 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv) * because it is reserved to the guard page. */ if (unmappable_end < ggtt_end - PAGE_SIZE) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[3], - unmappable_end, - ggtt_end - PAGE_SIZE); + ret = vgt_balloon_space(ggtt, &bl_info.space[3], + unmappable_end, ggtt_end - PAGE_SIZE); if (ret) goto err; } /* Mappable graphic memory ballooning */ if (mappable_base > ggtt->base.start) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[0], + ret = vgt_balloon_space(ggtt, &bl_info.space[0], ggtt->base.start, mappable_base); if (ret) @@ -247,10 +240,8 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv) } if (mappable_end < ggtt->mappable_end) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[1], - mappable_end, - ggtt->mappable_end); + ret = vgt_balloon_space(ggtt, &bl_info.space[1], + mappable_end, ggtt->mappable_end); if (ret) goto err; diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index 58f2483362ad..155906e84812 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -45,6 +45,7 @@ i915_vma_retire(struct i915_gem_active *active, if (i915_vma_is_active(vma)) return; + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); list_move_tail(&vma->vm_link, &vma->vm->inactive_list); if (unlikely(i915_vma_is_closed(vma) && !i915_vma_is_pinned(vma))) WARN_ON(i915_vma_unbind(vma)); @@ -69,17 +70,15 @@ i915_vma_retire(struct i915_gem_active *active, } static struct i915_vma * -__i915_vma_create(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) +vma_create(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view) { struct i915_vma *vma; struct rb_node *rb, **p; int i; - GEM_BUG_ON(vm->closed); - - vma = kmem_cache_zalloc(to_i915(obj->base.dev)->vmas, GFP_KERNEL); + vma = kmem_cache_zalloc(vm->i915->vmas, GFP_KERNEL); if (vma == NULL) return ERR_PTR(-ENOMEM); @@ -87,29 +86,50 @@ __i915_vma_create(struct drm_i915_gem_object *obj, for (i = 0; i < ARRAY_SIZE(vma->last_read); i++) init_request_active(&vma->last_read[i], i915_vma_retire); init_request_active(&vma->last_fence, NULL); - list_add(&vma->vm_link, &vm->unbound_list); vma->vm = vm; vma->obj = obj; vma->size = obj->base.size; + vma->display_alignment = I915_GTT_MIN_ALIGNMENT; - if (view) { + if (view && view->type != I915_GGTT_VIEW_NORMAL) { vma->ggtt_view = *view; if (view->type == I915_GGTT_VIEW_PARTIAL) { GEM_BUG_ON(range_overflows_t(u64, - view->params.partial.offset, - view->params.partial.size, + view->partial.offset, + view->partial.size, obj->base.size >> PAGE_SHIFT)); - vma->size = view->params.partial.size; + vma->size = view->partial.size; vma->size <<= PAGE_SHIFT; GEM_BUG_ON(vma->size >= obj->base.size); } else if (view->type == I915_GGTT_VIEW_ROTATED) { - vma->size = - intel_rotation_info_size(&view->params.rotated); + vma->size = intel_rotation_info_size(&view->rotated); vma->size <<= PAGE_SHIFT; } } + if (unlikely(vma->size > vm->total)) + goto err_vma; + + GEM_BUG_ON(!IS_ALIGNED(vma->size, I915_GTT_PAGE_SIZE)); + if (i915_is_ggtt(vm)) { + if (unlikely(overflows_type(vma->size, u32))) + goto err_vma; + + vma->fence_size = i915_gem_fence_size(vm->i915, vma->size, + i915_gem_object_get_tiling(obj), + i915_gem_object_get_stride(obj)); + if (unlikely(vma->fence_size < vma->size || /* overflow */ + vma->fence_size > vm->total)) + goto err_vma; + + GEM_BUG_ON(!IS_ALIGNED(vma->fence_size, I915_GTT_MIN_ALIGNMENT)); + + vma->fence_alignment = i915_gem_fence_alignment(vm->i915, vma->size, + i915_gem_object_get_tiling(obj), + i915_gem_object_get_stride(obj)); + GEM_BUG_ON(!is_power_of_2(vma->fence_alignment)); + vma->flags |= I915_VMA_GGTT; list_add(&vma->obj_link, &obj->vma_list); } else { @@ -131,20 +151,74 @@ __i915_vma_create(struct drm_i915_gem_object *obj, } rb_link_node(&vma->obj_node, rb, p); rb_insert_color(&vma->obj_node, &obj->vma_tree); + list_add(&vma->vm_link, &vm->unbound_list); return vma; + +err_vma: + kmem_cache_free(vm->i915->vmas, vma); + return ERR_PTR(-E2BIG); } +static struct i915_vma * +vma_lookup(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view) +{ + struct rb_node *rb; + + rb = obj->vma_tree.rb_node; + while (rb) { + struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node); + long cmp; + + cmp = i915_vma_compare(vma, vm, view); + if (cmp == 0) + return vma; + + if (cmp < 0) + rb = rb->rb_right; + else + rb = rb->rb_left; + } + + return NULL; +} + +/** + * i915_vma_instance - return the singleton instance of the VMA + * @obj: parent &struct drm_i915_gem_object to be mapped + * @vm: address space in which the mapping is located + * @view: additional mapping requirements + * + * i915_vma_instance() looks up an existing VMA of the @obj in the @vm with + * the same @view characteristics. If a match is not found, one is created. + * Once created, the VMA is kept until either the object is freed, or the + * address space is closed. + * + * Must be called with struct_mutex held. + * + * Returns the vma, or an error pointer. + */ struct i915_vma * -i915_vma_create(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) +i915_vma_instance(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view) { + struct i915_vma *vma; + lockdep_assert_held(&obj->base.dev->struct_mutex); GEM_BUG_ON(view && !i915_is_ggtt(vm)); - GEM_BUG_ON(i915_gem_obj_to_vma(obj, vm, view)); + GEM_BUG_ON(vm->closed); + + vma = vma_lookup(obj, vm, view); + if (!vma) + vma = vma_create(obj, vm, view); - return __i915_vma_create(obj, vm, view); + GEM_BUG_ON(!IS_ERR(vma) && i915_vma_is_closed(vma)); + GEM_BUG_ON(!IS_ERR(vma) && i915_vma_compare(vma, vm, view)); + GEM_BUG_ON(!IS_ERR(vma) && vma_lookup(obj, vm, view) != vma); + return vma; } /** @@ -195,6 +269,7 @@ int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level, return ret; } + trace_i915_vma_bind(vma, bind_flags); ret = vma->vm->bind_vma(vma, cache_level, bind_flags); if (ret) return ret; @@ -258,7 +333,8 @@ i915_vma_misplaced(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) if (vma->node.size < size) return true; - if (alignment && vma->node.start & (alignment - 1)) + GEM_BUG_ON(alignment && !is_power_of_2(alignment)); + if (alignment && !IS_ALIGNED(vma->node.start, alignment)) return true; if (flags & PIN_MAPPABLE && !i915_vma_is_map_and_fenceable(vma)) @@ -277,31 +353,24 @@ i915_vma_misplaced(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) void __i915_vma_set_map_and_fenceable(struct i915_vma *vma) { - struct drm_i915_gem_object *obj = vma->obj; - struct drm_i915_private *dev_priv = to_i915(obj->base.dev); bool mappable, fenceable; - u32 fence_size, fence_alignment; - - fence_size = i915_gem_get_ggtt_size(dev_priv, - vma->size, - i915_gem_object_get_tiling(obj)); - fence_alignment = i915_gem_get_ggtt_alignment(dev_priv, - vma->size, - i915_gem_object_get_tiling(obj), - true); - - fenceable = (vma->node.size == fence_size && - (vma->node.start & (fence_alignment - 1)) == 0); - mappable = (vma->node.start + fence_size <= - dev_priv->ggtt.mappable_end); + GEM_BUG_ON(!i915_vma_is_ggtt(vma)); + GEM_BUG_ON(!vma->fence_size); /* * Explicitly disable for rotated VMA since the display does not * need the fence and the VMA is not accessible to other users. */ - if (mappable && fenceable && - vma->ggtt_view.type != I915_GGTT_VIEW_ROTATED) + if (vma->ggtt_view.type == I915_GGTT_VIEW_ROTATED) + return; + + fenceable = (vma->node.size >= vma->fence_size && + IS_ALIGNED(vma->node.start, vma->fence_alignment)); + + mappable = vma->node.start + vma->fence_size <= i915_vm_to_ggtt(vma->vm)->mappable_end; + + if (mappable && fenceable) vma->flags |= I915_VMA_CAN_FENCE; else vma->flags &= ~I915_VMA_CAN_FENCE; @@ -368,22 +437,26 @@ i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) GEM_BUG_ON(drm_mm_node_allocated(&vma->node)); size = max(size, vma->size); - if (flags & PIN_MAPPABLE) - size = i915_gem_get_ggtt_size(dev_priv, size, - i915_gem_object_get_tiling(obj)); + alignment = max(alignment, vma->display_alignment); + if (flags & PIN_MAPPABLE) { + size = max_t(typeof(size), size, vma->fence_size); + alignment = max_t(typeof(alignment), + alignment, vma->fence_alignment); + } - alignment = max(max(alignment, vma->display_alignment), - i915_gem_get_ggtt_alignment(dev_priv, size, - i915_gem_object_get_tiling(obj), - flags & PIN_MAPPABLE)); + GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(!IS_ALIGNED(alignment, I915_GTT_MIN_ALIGNMENT)); + GEM_BUG_ON(!is_power_of_2(alignment)); start = flags & PIN_OFFSET_BIAS ? flags & PIN_OFFSET_MASK : 0; + GEM_BUG_ON(!IS_ALIGNED(start, I915_GTT_PAGE_SIZE)); end = vma->vm->total; if (flags & PIN_MAPPABLE) end = min_t(u64, end, dev_priv->ggtt.mappable_end); if (flags & PIN_ZONE_4G) - end = min_t(u64, end, (1ULL << 32) - PAGE_SIZE); + end = min_t(u64, end, (1ULL << 32) - I915_GTT_PAGE_SIZE); + GEM_BUG_ON(!IS_ALIGNED(end, I915_GTT_PAGE_SIZE)); /* If binding the object/GGTT view requires more space than the entire * aperture has, reject it early before evicting everything in a vain @@ -403,65 +476,28 @@ i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) if (flags & PIN_OFFSET_FIXED) { u64 offset = flags & PIN_OFFSET_MASK; - if (offset & (alignment - 1) || + if (!IS_ALIGNED(offset, alignment) || range_overflows(offset, size, end)) { ret = -EINVAL; goto err_unpin; } - vma->node.start = offset; - vma->node.size = size; - vma->node.color = obj->cache_level; - ret = drm_mm_reserve_node(&vma->vm->mm, &vma->node); - if (ret) { - ret = i915_gem_evict_for_vma(vma, flags); - if (ret == 0) - ret = drm_mm_reserve_node(&vma->vm->mm, &vma->node); - if (ret) - goto err_unpin; - } + ret = i915_gem_gtt_reserve(vma->vm, &vma->node, + size, offset, obj->cache_level, + flags); + if (ret) + goto err_unpin; } else { - u32 search_flag, alloc_flag; - - if (flags & PIN_HIGH) { - search_flag = DRM_MM_SEARCH_BELOW; - alloc_flag = DRM_MM_CREATE_TOP; - } else { - search_flag = DRM_MM_SEARCH_DEFAULT; - alloc_flag = DRM_MM_CREATE_DEFAULT; - } - - /* We only allocate in PAGE_SIZE/GTT_PAGE_SIZE (4096) chunks, - * so we know that we always have a minimum alignment of 4096. - * The drm_mm range manager is optimised to return results - * with zero alignment, so where possible use the optimal - * path. - */ - if (alignment <= 4096) - alignment = 0; - -search_free: - ret = drm_mm_insert_node_in_range_generic(&vma->vm->mm, - &vma->node, - size, alignment, - obj->cache_level, - start, end, - search_flag, - alloc_flag); - if (ret) { - ret = i915_gem_evict_something(vma->vm, size, alignment, - obj->cache_level, - start, end, - flags); - if (ret == 0) - goto search_free; - + ret = i915_gem_gtt_insert(vma->vm, &vma->node, + size, alignment, obj->cache_level, + start, end, flags); + if (ret) goto err_unpin; - } GEM_BUG_ON(vma->node.start < start); GEM_BUG_ON(vma->node.start + vma->node.size > end); } + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); GEM_BUG_ON(!i915_gem_valid_gtt_space(vma, obj->cache_level)); list_move_tail(&obj->global_link, &dev_priv->mm.bound_list); @@ -504,6 +540,7 @@ int __i915_vma_do_pin(struct i915_vma *vma, if ((bound ^ vma->flags) & I915_VMA_GLOBAL_BIND) __i915_vma_set_map_and_fenceable(vma); + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); GEM_BUG_ON(i915_vma_misplaced(vma, size, alignment, flags)); return 0; diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h index e3b2b3b1e056..e39d922cfb6f 100644 --- a/drivers/gpu/drm/i915/i915_vma.h +++ b/drivers/gpu/drm/i915/i915_vma.h @@ -55,6 +55,9 @@ struct i915_vma { u64 size; u64 display_alignment; + u32 fence_size; + u32 fence_alignment; + unsigned int flags; /** * How many users have pinned this object in GTT space. The following @@ -109,9 +112,9 @@ struct i915_vma { }; struct i915_vma * -i915_vma_create(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view); +i915_vma_instance(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view); void i915_vma_unpin_and_release(struct i915_vma **p_vma); @@ -196,15 +199,30 @@ i915_vma_compare(struct i915_vma *vma, if (cmp) return cmp; + BUILD_BUG_ON(I915_GGTT_VIEW_NORMAL != 0); + cmp = vma->ggtt_view.type; if (!view) - return vma->ggtt_view.type; + return cmp; - if (vma->ggtt_view.type != view->type) - return vma->ggtt_view.type - view->type; + cmp -= view->type; + if (cmp) + return cmp; - return memcmp(&vma->ggtt_view.params, - &view->params, - sizeof(view->params)); + /* ggtt_view.type also encodes its size so that we both distinguish + * different views using it as a "type" and also use a compact (no + * accessing of uninitialised padding bytes) memcmp without storing + * an extra parameter or adding more code. + * + * To ensure that the memcmp is valid for all branches of the union, + * even though the code looks like it is just comparing one branch, + * we assert above that all branches have the same address, and that + * each branch has a unique type/size. + */ + BUILD_BUG_ON(I915_GGTT_VIEW_NORMAL >= I915_GGTT_VIEW_PARTIAL); + BUILD_BUG_ON(I915_GGTT_VIEW_PARTIAL >= I915_GGTT_VIEW_ROTATED); + BUILD_BUG_ON(offsetof(typeof(*view), rotated) != + offsetof(typeof(*view), partial)); + return memcmp(&vma->ggtt_view.partial, &view->partial, view->type); } int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level, @@ -229,8 +247,11 @@ i915_vma_pin(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) /* Pin early to prevent the shrinker/eviction logic from destroying * our vma as we insert and bind. */ - if (likely(((++vma->flags ^ flags) & I915_VMA_BIND_MASK) == 0)) + if (likely(((++vma->flags ^ flags) & I915_VMA_BIND_MASK) == 0)) { + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); + GEM_BUG_ON(i915_vma_misplaced(vma, size, alignment, flags)); return 0; + } return __i915_vma_do_pin(vma, size, alignment, flags); } diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c index 4612ffd555a7..41fd94e62d3c 100644 --- a/drivers/gpu/drm/i915/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/intel_atomic_plane.c @@ -85,6 +85,8 @@ intel_plane_duplicate_state(struct drm_plane *plane) __drm_atomic_helper_plane_duplicate_state(plane, state); + intel_state->vma = NULL; + return state; } @@ -100,6 +102,24 @@ void intel_plane_destroy_state(struct drm_plane *plane, struct drm_plane_state *state) { + struct i915_vma *vma; + + vma = fetch_and_zero(&to_intel_plane_state(state)->vma); + + /* + * FIXME: Normally intel_cleanup_plane_fb handles destruction of vma. + * We currently don't clear all planes during driver unload, so we have + * to be able to unpin vma here for now. + * + * Normally this can only happen during unload when kmscon is disabled + * and userspace doesn't attempt to set a framebuffer at all. + */ + if (vma) { + mutex_lock(&plane->dev->struct_mutex); + intel_unpin_fb_vma(vma); + mutex_unlock(&plane->dev->struct_mutex); + } + drm_atomic_helper_plane_destroy_state(plane, state); } diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index 9cbb8d8363b4..0085bc745f6a 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -34,6 +34,10 @@ * low-power state and comes back to normal. */ +#define I915_CSR_GLK "i915/glk_dmc_ver1_01.bin" +MODULE_FIRMWARE(I915_CSR_GLK); +#define GLK_CSR_VERSION_REQUIRED CSR_VERSION(1, 1) + #define I915_CSR_KBL "i915/kbl_dmc_ver1_01.bin" MODULE_FIRMWARE(I915_CSR_KBL); #define KBL_CSR_VERSION_REQUIRED CSR_VERSION(1, 1) @@ -286,7 +290,9 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, csr->version = css_header->version; - if (IS_KABYLAKE(dev_priv)) { + if (IS_GEMINILAKE(dev_priv)) { + required_version = GLK_CSR_VERSION_REQUIRED; + } else if (IS_KABYLAKE(dev_priv)) { required_version = KBL_CSR_VERSION_REQUIRED; } else if (IS_SKYLAKE(dev_priv)) { required_version = SKL_CSR_VERSION_REQUIRED; @@ -435,7 +441,9 @@ void intel_csr_ucode_init(struct drm_i915_private *dev_priv) if (!HAS_CSR(dev_priv)) return; - if (IS_KABYLAKE(dev_priv)) + if (IS_GEMINILAKE(dev_priv)) + csr->fw_path = I915_CSR_GLK; + else if (IS_KABYLAKE(dev_priv)) csr->fw_path = I915_CSR_KBL; else if (IS_SKYLAKE(dev_priv)) csr->fw_path = I915_CSR_SKL; diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index f642f6ded4ae..fcf81815daff 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -192,7 +192,7 @@ static void gen9_sseu_info_init(struct drm_i915_private *dev_priv) (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) && hweight8(sseu->slice_mask) > 1; sseu->has_subslice_pg = - IS_BROXTON(dev_priv) && sseu_subslice_total(sseu) > 1; + IS_GEN9_LP(dev_priv) && sseu_subslice_total(sseu) > 1; sseu->has_eu_pg = sseu->eu_per_subslice > 2; if (IS_BROXTON(dev_priv)) { diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 1eb63699e115..b3e773c9f872 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2137,11 +2137,10 @@ intel_fill_fb_ggtt_view(struct i915_ggtt_view *view, const struct drm_framebuffer *fb, unsigned int rotation) { + view->type = I915_GGTT_VIEW_NORMAL; if (drm_rotation_90_or_270(rotation)) { - *view = i915_ggtt_view_rotated; - view->params.rotated = to_intel_framebuffer(fb)->rot_info; - } else { - *view = i915_ggtt_view_normal; + view->type = I915_GGTT_VIEW_ROTATED; + view->rotated = to_intel_framebuffer(fb)->rot_info; } } @@ -2235,24 +2234,19 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, unsigned int rotation) i915_vma_pin_fence(vma); } + i915_vma_get(vma); err: intel_runtime_pm_put(dev_priv); return vma; } -void intel_unpin_fb_obj(struct drm_framebuffer *fb, unsigned int rotation) +void intel_unpin_fb_vma(struct i915_vma *vma) { - struct drm_i915_gem_object *obj = intel_fb_obj(fb); - struct i915_ggtt_view view; - struct i915_vma *vma; - - WARN_ON(!mutex_is_locked(&obj->base.dev->struct_mutex)); - - intel_fill_fb_ggtt_view(&view, fb, rotation); - vma = i915_gem_object_to_ggtt(obj, &view); + lockdep_assert_held(&vma->vm->i915->drm.struct_mutex); i915_vma_unpin_fence(vma); i915_gem_object_unpin_from_display_plane(vma); + i915_vma_put(vma); } static int intel_fb_pitch(const struct drm_framebuffer *fb, int plane, @@ -2745,7 +2739,6 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, struct drm_device *dev = intel_crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); struct drm_crtc *c; - struct intel_crtc *i; struct drm_i915_gem_object *obj; struct drm_plane *primary = intel_crtc->base.primary; struct drm_plane_state *plane_state = primary->state; @@ -2770,20 +2763,20 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, * an fb with another CRTC instead */ for_each_crtc(dev, c) { - i = to_intel_crtc(c); + struct intel_plane_state *state; if (c == &intel_crtc->base) continue; - if (!i->active) + if (!to_intel_crtc(c)->active) continue; - fb = c->primary->fb; - if (!fb) + state = to_intel_plane_state(c->primary->state); + if (!state->vma) continue; - obj = intel_fb_obj(fb); - if (i915_gem_object_ggtt_offset(obj, NULL) == plane_config->base) { + if (intel_plane_ggtt_offset(state) == plane_config->base) { + fb = c->primary->fb; drm_framebuffer_reference(fb); goto valid_fb; } @@ -2796,7 +2789,7 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, * simplest solution is to just disable the primary plane now and * pretend the BIOS never had it enabled. */ - to_intel_plane_state(plane_state)->base.visible = false; + plane_state->visible = false; crtc_state->plane_mask &= ~(1 << drm_plane_index(primary)); intel_pre_disable_primary_noatomic(&intel_crtc->base); intel_plane->disable_plane(primary, &intel_crtc->base); @@ -2804,6 +2797,19 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, return; valid_fb: + mutex_lock(&dev->struct_mutex); + intel_state->vma = + intel_pin_and_fence_fb_obj(fb, primary->state->rotation); + mutex_unlock(&dev->struct_mutex); + if (IS_ERR(intel_state->vma)) { + DRM_ERROR("failed to pin boot fb on pipe %d: %li\n", + intel_crtc->pipe, PTR_ERR(intel_state->vma)); + + intel_state->vma = NULL; + drm_framebuffer_unreference(fb); + return; + } + plane_state->src_x = 0; plane_state->src_y = 0; plane_state->src_w = fb->width << 16; @@ -2966,6 +2972,9 @@ int skl_check_plane_surface(struct intel_plane_state *plane_state) unsigned int rotation = plane_state->base.rotation; int ret; + if (!plane_state->base.visible) + return 0; + /* Rotate src coordinates to match rotated GTT view */ if (drm_rotation_90_or_270(rotation)) drm_rect_rotate(&plane_state->base.src, @@ -3096,13 +3105,13 @@ static void i9xx_update_primary_plane(struct drm_plane *primary, I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]); if (INTEL_GEN(dev_priv) >= 4) { I915_WRITE(DSPSURF(plane), - intel_fb_gtt_offset(fb, rotation) + + intel_plane_ggtt_offset(plane_state) + intel_crtc->dspaddr_offset); I915_WRITE(DSPTILEOFF(plane), (y << 16) | x); I915_WRITE(DSPLINOFF(plane), linear_offset); } else { I915_WRITE(DSPADDR(plane), - intel_fb_gtt_offset(fb, rotation) + + intel_plane_ggtt_offset(plane_state) + intel_crtc->dspaddr_offset); } POSTING_READ(reg); @@ -3199,7 +3208,7 @@ static void ironlake_update_primary_plane(struct drm_plane *primary, I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]); I915_WRITE(DSPSURF(plane), - intel_fb_gtt_offset(fb, rotation) + + intel_plane_ggtt_offset(plane_state) + intel_crtc->dspaddr_offset); if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { I915_WRITE(DSPOFFSET(plane), (y << 16) | x); @@ -3222,23 +3231,6 @@ u32 intel_fb_stride_alignment(const struct drm_i915_private *dev_priv, } } -u32 intel_fb_gtt_offset(struct drm_framebuffer *fb, - unsigned int rotation) -{ - struct drm_i915_gem_object *obj = intel_fb_obj(fb); - struct i915_ggtt_view view; - struct i915_vma *vma; - - intel_fill_fb_ggtt_view(&view, fb, rotation); - - vma = i915_gem_object_to_ggtt(obj, &view); - if (WARN(!vma, "ggtt vma for display object not found! (view=%u)\n", - view.type)) - return -1; - - return i915_ggtt_offset(vma); -} - static void skl_detach_scaler(struct intel_crtc *intel_crtc, int id) { struct drm_device *dev = intel_crtc->base.dev; @@ -3434,7 +3426,7 @@ static void skylake_update_primary_plane(struct drm_plane *plane, } I915_WRITE(PLANE_SURF(pipe, plane_id), - intel_fb_gtt_offset(fb, rotation) + surf_addr); + intel_plane_ggtt_offset(plane_state) + surf_addr); POSTING_READ(PLANE_SURF(pipe, plane_id)); } @@ -3558,23 +3550,19 @@ void intel_prepare_reset(struct drm_i915_private *dev_priv) state = drm_atomic_helper_duplicate_state(dev, ctx); if (IS_ERR(state)) { ret = PTR_ERR(state); - state = NULL; DRM_ERROR("Duplicating state failed with %i\n", ret); - goto err; + return; } ret = drm_atomic_helper_disable_all(dev, ctx); if (ret) { DRM_ERROR("Suspending crtc's failed with %i\n", ret); - goto err; + drm_atomic_state_put(state); + return; } dev_priv->modeset_restore_state = state; state->acquire_ctx = ctx; - return; - -err: - drm_atomic_state_put(state); } void intel_finish_reset(struct drm_i915_private *dev_priv) @@ -6883,13 +6871,13 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc) if (!intel_crtc->active) return; - if (to_intel_plane_state(crtc->primary->state)->base.visible) { + if (crtc->primary->state->visible) { WARN_ON(intel_crtc->flip_work); intel_pre_disable_primary_noatomic(crtc); intel_crtc_disable_planes(crtc, 1 << drm_plane_index(crtc->primary)); - to_intel_plane_state(crtc->primary->state)->base.visible = false; + crtc->primary->state->visible = false; } state = drm_atomic_state_alloc(crtc->dev); @@ -8403,7 +8391,6 @@ void intel_mode_from_pipe_config(struct drm_display_mode *mode, mode->type = DRM_MODE_TYPE_DRIVER; mode->clock = pipe_config->base.adjusted_mode.crtc_clock; - mode->flags |= pipe_config->base.adjusted_mode.flags; mode->hsync = drm_mode_hsync(mode); mode->vrefresh = drm_mode_vrefresh(mode); @@ -11288,6 +11275,7 @@ found: } old->restore_state = restore_state; + drm_atomic_state_put(state); /* let the connector get through one full cycle before testing */ intel_wait_for_vblank(dev_priv, intel_crtc->pipe); @@ -11567,7 +11555,7 @@ static void intel_unpin_work_fn(struct work_struct *__work) flush_work(&work->mmio_work); mutex_lock(&dev->struct_mutex); - intel_unpin_fb_obj(work->old_fb, primary->state->rotation); + intel_unpin_fb_vma(work->old_vma); i915_gem_object_put(work->pending_flip_obj); mutex_unlock(&dev->struct_mutex); @@ -12277,8 +12265,10 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, goto cleanup_pending; } - work->gtt_offset = intel_fb_gtt_offset(fb, primary->state->rotation); - work->gtt_offset += intel_crtc->dspaddr_offset; + work->old_vma = to_intel_plane_state(primary->state)->vma; + to_intel_plane_state(primary->state)->vma = vma; + + work->gtt_offset = i915_ggtt_offset(vma) + intel_crtc->dspaddr_offset; work->rotation = crtc->primary->state->rotation; /* @@ -12333,7 +12323,8 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, cleanup_request: i915_add_request_no_flush(request); cleanup_unpin: - intel_unpin_fb_obj(fb, crtc->primary->state->rotation); + to_intel_plane_state(primary->state)->vma = work->old_vma; + intel_unpin_fb_vma(vma); cleanup_pending: atomic_dec(&intel_crtc->unpin_work_count); unlock: @@ -12463,7 +12454,7 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state, } was_visible = old_plane_state->base.visible; - visible = to_intel_plane_state(plane_state)->base.visible; + visible = plane_state->visible; if (!was_crtc_enabled && WARN_ON(was_visible)) was_visible = false; @@ -12479,7 +12470,7 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state, * only combine the results from all planes in the current place? */ if (!is_crtc_enabled) - to_intel_plane_state(plane_state)->base.visible = visible = false; + plane_state->visible = visible = false; if (!was_visible && !visible) return 0; @@ -14737,6 +14728,7 @@ static const struct drm_crtc_funcs intel_crtc_funcs = { .page_flip = intel_crtc_page_flip, .atomic_duplicate_state = intel_crtc_duplicate_state, .atomic_destroy_state = intel_crtc_destroy_state, + .set_crc_source = intel_crtc_set_crc_source, }; /** @@ -14833,6 +14825,8 @@ intel_prepare_plane_fb(struct drm_plane *plane, DRM_DEBUG_KMS("failed to pin object\n"); return PTR_ERR(vma); } + + to_intel_plane_state(new_state)->vma = vma; } return 0; @@ -14851,19 +14845,12 @@ void intel_cleanup_plane_fb(struct drm_plane *plane, struct drm_plane_state *old_state) { - struct drm_i915_private *dev_priv = to_i915(plane->dev); - struct intel_plane_state *old_intel_state; - struct drm_i915_gem_object *old_obj = intel_fb_obj(old_state->fb); - struct drm_i915_gem_object *obj = intel_fb_obj(plane->state->fb); - - old_intel_state = to_intel_plane_state(old_state); - - if (!obj && !old_obj) - return; + struct i915_vma *vma; - if (old_obj && (plane->type != DRM_PLANE_TYPE_CURSOR || - !INTEL_INFO(dev_priv)->cursor_needs_physical)) - intel_unpin_fb_obj(old_state->fb, old_state->rotation); + /* Should only be called after a successful intel_prepare_plane_fb()! */ + vma = fetch_and_zero(&to_intel_plane_state(old_state)->vma); + if (vma) + intel_unpin_fb_vma(vma); } int @@ -15015,6 +15002,7 @@ intel_legacy_cursor_update(struct drm_plane *plane, struct intel_plane *intel_plane = to_intel_plane(plane); struct drm_framebuffer *old_fb; struct drm_crtc_state *crtc_state = crtc->state; + struct i915_vma *old_vma; /* * When crtc is inactive or there is a modeset pending, @@ -15086,9 +15074,12 @@ intel_legacy_cursor_update(struct drm_plane *plane, ret = PTR_ERR(vma); goto out_unlock; } + + to_intel_plane_state(new_plane_state)->vma = vma; } old_fb = old_plane_state->fb; + old_vma = to_intel_plane_state(old_plane_state)->vma; i915_gem_track_fb(intel_fb_obj(old_fb), intel_fb_obj(fb), intel_plane->frontbuffer_bit); @@ -15098,6 +15089,7 @@ intel_legacy_cursor_update(struct drm_plane *plane, *to_intel_plane_state(old_plane_state) = *to_intel_plane_state(new_plane_state); new_plane_state->fence = NULL; new_plane_state->fb = old_fb; + to_intel_plane_state(new_plane_state)->vma = old_vma; intel_plane->update_plane(plane, to_intel_crtc_state(crtc->state), @@ -15336,7 +15328,7 @@ intel_update_cursor_plane(struct drm_plane *plane, if (!obj) addr = 0; else if (!INTEL_INFO(dev_priv)->cursor_needs_physical) - addr = i915_gem_object_ggtt_offset(obj, NULL); + addr = intel_plane_ggtt_offset(state); else addr = obj->phys_handle->busaddr; @@ -16840,7 +16832,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc) * Temporarily change the plane mapping and disable everything * ... */ plane = crtc->plane; - to_intel_plane_state(crtc->base.primary->state)->base.visible = true; + crtc->base.primary->state->visible = true; crtc->plane = !plane; intel_crtc_disable_noatomic(&crtc->base); crtc->plane = plane; @@ -16992,7 +16984,8 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) dev_priv->active_crtcs = 0; for_each_intel_crtc(dev, crtc) { - struct intel_crtc_state *crtc_state = crtc->config; + struct intel_crtc_state *crtc_state = + to_intel_crtc_state(crtc->base.state); __drm_atomic_helper_crtc_destroy_state(&crtc_state->base); memset(crtc_state, 0, sizeof(*crtc_state)); @@ -17011,7 +17004,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) DRM_DEBUG_KMS("[CRTC:%d:%s] hw state readout: %s\n", crtc->base.base.id, crtc->base.name, - enableddisabled(crtc->active)); + enableddisabled(crtc_state->base.active)); } for (i = 0; i < dev_priv->num_shared_dpll; i++) { @@ -17021,7 +17014,11 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) &pll->state.hw_state); pll->state.crtc_mask = 0; for_each_intel_crtc(dev, crtc) { - if (crtc->active && crtc->config->shared_dpll == pll) + struct intel_crtc_state *crtc_state = + to_intel_crtc_state(crtc->base.state); + + if (crtc_state->base.active && + crtc_state->shared_dpll == pll) pll->state.crtc_mask |= 1 << crtc->pipe; } pll->active_mask = pll->state.crtc_mask; @@ -17034,11 +17031,14 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) pipe = 0; if (encoder->get_hw_state(encoder, &pipe)) { + struct intel_crtc_state *crtc_state; + crtc = intel_get_crtc_for_pipe(dev_priv, pipe); + crtc_state = to_intel_crtc_state(crtc->base.state); encoder->base.crtc = &crtc->base; - crtc->config->output_types |= 1 << encoder->type; - encoder->get_config(encoder, crtc->config); + crtc_state->output_types |= 1 << encoder->type; + encoder->get_config(encoder, crtc_state); } else { encoder->base.crtc = NULL; } @@ -17079,14 +17079,16 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) } for_each_intel_crtc(dev, crtc) { + struct intel_crtc_state *crtc_state = + to_intel_crtc_state(crtc->base.state); int pixclk = 0; - crtc->base.hwmode = crtc->config->base.adjusted_mode; + crtc->base.hwmode = crtc_state->base.adjusted_mode; memset(&crtc->base.mode, 0, sizeof(crtc->base.mode)); - if (crtc->base.state->active) { - intel_mode_from_pipe_config(&crtc->base.mode, crtc->config); - intel_mode_from_pipe_config(&crtc->base.state->adjusted_mode, crtc->config); + if (crtc_state->base.active) { + intel_mode_from_pipe_config(&crtc->base.mode, crtc_state); + intel_mode_from_pipe_config(&crtc_state->base.adjusted_mode, crtc_state); WARN_ON(drm_atomic_set_mode_for_crtc(crtc->base.state, &crtc->base.mode)); /* @@ -17098,17 +17100,17 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) * set a flag to indicate that a full recalculation is * needed on the next commit. */ - crtc->base.state->mode.private_flags = I915_MODE_FLAG_INHERITED; + crtc_state->base.mode.private_flags = I915_MODE_FLAG_INHERITED; if (INTEL_GEN(dev_priv) >= 9 || IS_BROADWELL(dev_priv)) - pixclk = ilk_pipe_pixel_rate(crtc->config); + pixclk = ilk_pipe_pixel_rate(crtc_state); else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - pixclk = crtc->config->base.adjusted_mode.crtc_clock; + pixclk = crtc_state->base.adjusted_mode.crtc_clock; else WARN_ON(dev_priv->display.modeset_calc_cdclk); /* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */ - if (IS_BROADWELL(dev_priv) && crtc->config->ips_enabled) + if (IS_BROADWELL(dev_priv) && crtc_state->ips_enabled) pixclk = DIV_ROUND_UP(pixclk * 100, 95); drm_calc_timestamping_constants(&crtc->base, &crtc->base.hwmode); @@ -17117,7 +17119,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) dev_priv->min_pixclk[crtc->pipe] = pixclk; - intel_pipe_config_sanity_check(dev_priv, crtc->config); + intel_pipe_config_sanity_check(dev_priv, crtc_state); } } @@ -17218,47 +17220,19 @@ void intel_display_resume(struct drm_device *dev) if (ret) DRM_ERROR("Restoring old state failed with %i\n", ret); - drm_atomic_state_put(state); + if (state) + drm_atomic_state_put(state); } void intel_modeset_gem_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = to_i915(dev); - struct drm_crtc *c; - struct drm_i915_gem_object *obj; intel_init_gt_powersave(dev_priv); intel_modeset_init_hw(dev); intel_setup_overlay(dev_priv); - - /* - * Make sure any fbs we allocated at startup are properly - * pinned & fenced. When we do the allocation it's too early - * for this. - */ - for_each_crtc(dev, c) { - struct i915_vma *vma; - - obj = intel_fb_obj(c->primary->fb); - if (obj == NULL) - continue; - - mutex_lock(&dev->struct_mutex); - vma = intel_pin_and_fence_fb_obj(c->primary->fb, - c->primary->state->rotation); - mutex_unlock(&dev->struct_mutex); - if (IS_ERR(vma)) { - DRM_ERROR("failed to pin boot fb on pipe %d\n", - to_intel_crtc(c)->pipe); - drm_framebuffer_unreference(c->primary->fb); - c->primary->fb = NULL; - c->primary->crtc = c->primary->state->crtc = NULL; - update_state_fb(c->primary); - c->state->plane_mask &= ~(1 << drm_plane_index(c->primary)); - } - } } int intel_connector_register(struct drm_connector *connector) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index be89aca0dbe8..3d8ac8aa7214 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -3044,6 +3044,32 @@ intel_dp_get_link_status(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_ DP_LINK_STATUS_SIZE) == DP_LINK_STATUS_SIZE; } +static bool intel_dp_get_y_cord_status(struct intel_dp *intel_dp) +{ + uint8_t psr_caps = 0; + + drm_dp_dpcd_readb(&intel_dp->aux, DP_PSR_CAPS, &psr_caps); + return psr_caps & DP_PSR2_SU_Y_COORDINATE_REQUIRED; +} + +static bool intel_dp_get_colorimetry_status(struct intel_dp *intel_dp) +{ + uint8_t dprx = 0; + + drm_dp_dpcd_readb(&intel_dp->aux, + DP_DPRX_FEATURE_ENUMERATION_LIST, + &dprx); + return dprx & DP_VSC_SDP_EXT_FOR_COLORIMETRY_SUPPORTED; +} + +static bool intel_dp_get_alpm_status(struct intel_dp *intel_dp) +{ + uint8_t alpm_caps = 0; + + drm_dp_dpcd_readb(&intel_dp->aux, DP_RECEIVER_ALPM_CAP, &alpm_caps); + return alpm_caps & DP_ALPM_CAP; +} + /* These are source-specific values. */ uint8_t intel_dp_voltage_max(struct intel_dp *intel_dp) @@ -3414,7 +3440,7 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp) if (HAS_DDI(dev_priv)) { signal_levels = ddi_signal_levels(intel_dp); - if (IS_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) signal_levels = 0; else mask = DDI_BUF_EMP_MASK; @@ -3622,6 +3648,16 @@ intel_edp_init_dpcd(struct intel_dp *intel_dp) dev_priv->psr.psr2_support = dev_priv->psr.aux_frame_sync; DRM_DEBUG_KMS("PSR2 %s on sink", dev_priv->psr.psr2_support ? "supported" : "not supported"); + + if (dev_priv->psr.psr2_support) { + dev_priv->psr.y_cord_support = + intel_dp_get_y_cord_status(intel_dp); + dev_priv->psr.colorimetry_support = + intel_dp_get_colorimetry_status(intel_dp); + dev_priv->psr.alpm = + intel_dp_get_alpm_status(intel_dp); + } + } /* Read the eDP Display control capabilities registers */ diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 6b02dac6ea26..0cec0013ace0 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -376,6 +376,7 @@ struct intel_atomic_state { struct intel_plane_state { struct drm_plane_state base; struct drm_rect clip; + struct i915_vma *vma; struct { u32 offset; @@ -1067,6 +1068,7 @@ struct intel_flip_work { struct work_struct mmio_work; struct drm_crtc *crtc; + struct i915_vma *old_vma; struct drm_framebuffer *old_fb; struct drm_i915_gem_object *pending_flip_obj; struct drm_pending_vblank_event *event; @@ -1302,7 +1304,7 @@ void intel_release_load_detect_pipe(struct drm_connector *connector, struct drm_modeset_acquire_ctx *ctx); struct i915_vma * intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, unsigned int rotation); -void intel_unpin_fb_obj(struct drm_framebuffer *fb, unsigned int rotation); +void intel_unpin_fb_vma(struct i915_vma *vma); struct drm_framebuffer * __intel_framebuffer_create(struct drm_device *dev, struct drm_mode_fb_cmd2 *mode_cmd, @@ -1391,7 +1393,10 @@ void intel_mode_from_pipe_config(struct drm_display_mode *mode, int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state); int skl_max_scale(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state); -u32 intel_fb_gtt_offset(struct drm_framebuffer *fb, unsigned int rotation); +static inline u32 intel_plane_ggtt_offset(const struct intel_plane_state *state) +{ + return i915_ggtt_offset(state->vma); +} u32 skl_plane_ctl_format(uint32_t pixel_format); u32 skl_plane_ctl_tiling(uint64_t fb_modifier); @@ -1880,5 +1885,11 @@ void lspcon_wait_pcon_mode(struct intel_lspcon *lspcon); /* intel_pipe_crc.c */ int intel_pipe_crc_create(struct drm_minor *minor); void intel_pipe_crc_cleanup(struct drm_minor *minor); +#ifdef CONFIG_DEBUG_FS +int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, + size_t *values_cnt); +#else +#define intel_crtc_set_crc_source NULL +#endif extern const struct file_operations i915_display_crc_ctl_fops; #endif /* __INTEL_DRV_H__ */ diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 97bbbc3d6aa8..371acf109e34 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -264,7 +264,7 @@ int intel_engine_create_scratch(struct intel_engine_cs *engine, int size) return PTR_ERR(obj); } - vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto err_unref; diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c index 26a81a9e9c1d..89fe5c8464df 100644 --- a/drivers/gpu/drm/i915/intel_fbc.c +++ b/drivers/gpu/drm/i915/intel_fbc.c @@ -173,7 +173,7 @@ static void i8xx_fbc_activate(struct drm_i915_private *dev_priv) if (IS_I945GM(dev_priv)) fbc_ctl |= FBC_CTL_C3_IDLE; /* 945 needs special SR handling */ fbc_ctl |= (cfb_pitch & 0xff) << FBC_CTL_STRIDE_SHIFT; - fbc_ctl |= params->fb.fence_reg; + fbc_ctl |= params->vma->fence->id; I915_WRITE(FBC_CONTROL, fbc_ctl); } @@ -193,8 +193,8 @@ static void g4x_fbc_activate(struct drm_i915_private *dev_priv) else dpfc_ctl |= DPFC_CTL_LIMIT_1X; - if (params->fb.fence_reg != I915_FENCE_REG_NONE) { - dpfc_ctl |= DPFC_CTL_FENCE_EN | params->fb.fence_reg; + if (params->vma->fence) { + dpfc_ctl |= DPFC_CTL_FENCE_EN | params->vma->fence->id; I915_WRITE(DPFC_FENCE_YOFF, params->crtc.fence_y_offset); } else { I915_WRITE(DPFC_FENCE_YOFF, 0); @@ -251,13 +251,14 @@ static void ilk_fbc_activate(struct drm_i915_private *dev_priv) break; } - if (params->fb.fence_reg != I915_FENCE_REG_NONE) { + if (params->vma->fence) { dpfc_ctl |= DPFC_CTL_FENCE_EN; if (IS_GEN5(dev_priv)) - dpfc_ctl |= params->fb.fence_reg; + dpfc_ctl |= params->vma->fence->id; if (IS_GEN6(dev_priv)) { I915_WRITE(SNB_DPFC_CTL_SA, - SNB_CPU_FENCE_ENABLE | params->fb.fence_reg); + SNB_CPU_FENCE_ENABLE | + params->vma->fence->id); I915_WRITE(DPFC_CPU_FENCE_OFFSET, params->crtc.fence_y_offset); } @@ -269,7 +270,8 @@ static void ilk_fbc_activate(struct drm_i915_private *dev_priv) } I915_WRITE(ILK_DPFC_FENCE_YOFF, params->crtc.fence_y_offset); - I915_WRITE(ILK_FBC_RT_BASE, params->fb.ggtt_offset | ILK_FBC_RT_VALID); + I915_WRITE(ILK_FBC_RT_BASE, + i915_ggtt_offset(params->vma) | ILK_FBC_RT_VALID); /* enable it... */ I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN); @@ -319,10 +321,11 @@ static void gen7_fbc_activate(struct drm_i915_private *dev_priv) break; } - if (params->fb.fence_reg != I915_FENCE_REG_NONE) { + if (params->vma->fence) { dpfc_ctl |= IVB_DPFC_CTL_FENCE_EN; I915_WRITE(SNB_DPFC_CTL_SA, - SNB_CPU_FENCE_ENABLE | params->fb.fence_reg); + SNB_CPU_FENCE_ENABLE | + params->vma->fence->id); I915_WRITE(DPFC_CPU_FENCE_OFFSET, params->crtc.fence_y_offset); } else { I915_WRITE(SNB_DPFC_CTL_SA,0); @@ -727,14 +730,6 @@ static bool intel_fbc_hw_tracking_covers_screen(struct intel_crtc *crtc) return effective_w <= max_w && effective_h <= max_h; } -/* XXX replace me when we have VMA tracking for intel_plane_state */ -static int get_fence_id(struct drm_framebuffer *fb) -{ - struct i915_vma *vma = i915_gem_object_to_ggtt(intel_fb_obj(fb), NULL); - - return vma && vma->fence ? vma->fence->id : I915_FENCE_REG_NONE; -} - static void intel_fbc_update_state_cache(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, struct intel_plane_state *plane_state) @@ -743,7 +738,8 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc, struct intel_fbc *fbc = &dev_priv->fbc; struct intel_fbc_state_cache *cache = &fbc->state_cache; struct drm_framebuffer *fb = plane_state->base.fb; - struct drm_i915_gem_object *obj; + + cache->vma = NULL; cache->crtc.mode_flags = crtc_state->base.adjusted_mode.flags; if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) @@ -758,16 +754,10 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc, if (!cache->plane.visible) return; - obj = intel_fb_obj(fb); - - /* FIXME: We lack the proper locking here, so only run this on the - * platforms that need. */ - if (IS_GEN(dev_priv, 5, 6)) - cache->fb.ilk_ggtt_offset = i915_gem_object_ggtt_offset(obj, NULL); cache->fb.format = fb->format; cache->fb.stride = fb->pitches[0]; - cache->fb.fence_reg = get_fence_id(fb); - cache->fb.tiling_mode = i915_gem_object_get_tiling(obj); + + cache->vma = plane_state->vma; } static bool intel_fbc_can_activate(struct intel_crtc *crtc) @@ -784,7 +774,7 @@ static bool intel_fbc_can_activate(struct intel_crtc *crtc) return false; } - if (!cache->plane.visible) { + if (!cache->vma) { fbc->no_fbc_reason = "primary plane not visible"; return false; } @@ -807,8 +797,7 @@ static bool intel_fbc_can_activate(struct intel_crtc *crtc) * so have no fence associated with it) due to aperture constaints * at the time of pinning. */ - if (cache->fb.tiling_mode != I915_TILING_X || - cache->fb.fence_reg == I915_FENCE_REG_NONE) { + if (!cache->vma->fence) { fbc->no_fbc_reason = "framebuffer not tiled or fenced"; return false; } @@ -888,17 +877,16 @@ static void intel_fbc_get_reg_params(struct intel_crtc *crtc, * zero. */ memset(params, 0, sizeof(*params)); + params->vma = cache->vma; + params->crtc.pipe = crtc->pipe; params->crtc.plane = crtc->plane; params->crtc.fence_y_offset = get_crtc_fence_y_offset(crtc); params->fb.format = cache->fb.format; params->fb.stride = cache->fb.stride; - params->fb.fence_reg = cache->fb.fence_reg; params->cfb_size = intel_fbc_calculate_cfb_size(dev_priv, cache); - - params->fb.ggtt_offset = cache->fb.ilk_ggtt_offset; } static bool intel_fbc_reg_params_equal(struct intel_fbc_reg_params *params1, @@ -1296,7 +1284,7 @@ void intel_fbc_init_pipe_state(struct drm_i915_private *dev_priv) for_each_intel_crtc(&dev_priv->drm, crtc) if (intel_crtc_active(crtc) && - to_intel_plane_state(crtc->base.primary->state)->base.visible) + crtc->base.primary->state->visible) dev_priv->fbc.visible_pipes_mask |= (1 << crtc->pipe); } diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index bdefa61f2e60..e0d9e72cf3d1 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -284,7 +284,7 @@ static int intelfb_create(struct drm_fb_helper *helper, out_destroy_fbi: drm_fb_helper_release_fbi(helper); out_unpin: - intel_unpin_fb_obj(&ifbdev->fb->base, DRM_ROTATE_0); + intel_unpin_fb_vma(vma); out_unlock: mutex_unlock(&dev->struct_mutex); return ret; @@ -549,7 +549,7 @@ static void intel_fbdev_destroy(struct intel_fbdev *ifbdev) if (ifbdev->fb) { mutex_lock(&ifbdev->helper.dev->struct_mutex); - intel_unpin_fb_obj(&ifbdev->fb->base, DRM_ROTATE_0); + intel_unpin_fb_vma(ifbdev->vma); mutex_unlock(&ifbdev->helper.dev->struct_mutex); drm_framebuffer_remove(&ifbdev->fb->base); @@ -742,6 +742,9 @@ void intel_fbdev_initial_config_async(struct drm_device *dev) { struct intel_fbdev *ifbdev = to_i915(dev)->fbdev; + if (!ifbdev) + return; + ifbdev->cookie = async_schedule(intel_fbdev_initial_config, ifbdev); } diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 3202b32b5638..25691f0e4c50 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -145,7 +145,7 @@ * The GuC firmware layout looks like this: * * +-------------------------------+ - * | guc_css_header | + * | uc_css_header | * | | * | contains major/minor version | * +-------------------------------+ @@ -172,9 +172,16 @@ * 3. Length info of each component can be found in header, in dwords. * 4. Modulus and exponent key are not required by driver. They may not appear * in fw. So driver will load a truncated firmware in this case. + * + * HuC firmware layout is same as GuC firmware. + * + * HuC firmware css header is different. However, the only difference is where + * the version information is saved. The uc_css_header is unified to support + * both. Driver should get HuC version from uc_css_header.huc_sw_version, while + * uc_css_header.guc_sw_version for GuC. */ -struct guc_css_header { +struct uc_css_header { uint32_t module_type; /* header_size includes all non-uCode bits, including css_header, rsa * key, modulus key and exponent data. */ @@ -205,8 +212,16 @@ struct guc_css_header { char username[8]; char buildnumber[12]; - uint32_t device_id; - uint32_t guc_sw_version; + union { + struct { + uint32_t branch_client_version; + uint32_t sw_version; + } guc; + struct { + uint32_t sw_version; + uint32_t reserved; + } huc; + }; uint32_t prod_preprod_fw; uint32_t reserved[12]; uint32_t header_info; @@ -490,6 +505,7 @@ enum intel_guc_action { INTEL_GUC_ACTION_ENTER_S_STATE = 0x501, INTEL_GUC_ACTION_EXIT_S_STATE = 0x502, INTEL_GUC_ACTION_SLPC_REQUEST = 0x3003, + INTEL_GUC_ACTION_AUTHENTICATE_HUC = 0x4000, INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING = 0x0E000, INTEL_GUC_ACTION_LIMIT }; diff --git a/drivers/gpu/drm/i915/intel_guc_loader.c b/drivers/gpu/drm/i915/intel_guc_loader.c index 35d5690f47a2..2f1cf9aea04e 100644 --- a/drivers/gpu/drm/i915/intel_guc_loader.c +++ b/drivers/gpu/drm/i915/intel_guc_loader.c @@ -51,12 +51,6 @@ * 512K. In order to exclude 0-512K address space from GGTT, all gfx objects * used by GuC is pinned with PIN_OFFSET_BIAS along with size of WOPCM. * - * Firmware log: - * Firmware log is enabled by setting i915.guc_log_level to non-negative level. - * Log data is printed out via reading debugfs i915_guc_log_dump. Reading from - * i915_guc_load_status will print out firmware loading status and scratch - * registers value. - * */ #define SKL_FW_MAJOR 6 @@ -81,16 +75,16 @@ MODULE_FIRMWARE(I915_BXT_GUC_UCODE); MODULE_FIRMWARE(I915_KBL_GUC_UCODE); /* User-friendly representation of an enum */ -const char *intel_guc_fw_status_repr(enum intel_guc_fw_status status) +const char *intel_uc_fw_status_repr(enum intel_uc_fw_status status) { switch (status) { - case GUC_FIRMWARE_FAIL: + case INTEL_UC_FIRMWARE_FAIL: return "FAIL"; - case GUC_FIRMWARE_NONE: + case INTEL_UC_FIRMWARE_NONE: return "NONE"; - case GUC_FIRMWARE_PENDING: + case INTEL_UC_FIRMWARE_PENDING: return "PENDING"; - case GUC_FIRMWARE_SUCCESS: + case INTEL_UC_FIRMWARE_SUCCESS: return "SUCCESS"; default: return "UNKNOWN!"; @@ -278,7 +272,7 @@ static inline bool guc_ucode_response(struct drm_i915_private *dev_priv, static int guc_ucode_xfer_dma(struct drm_i915_private *dev_priv, struct i915_vma *vma) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; unsigned long offset; struct sg_table *sg = vma->pages; u32 status, rsa[UOS_RSA_SCRATCH_MAX_COUNT]; @@ -334,12 +328,12 @@ static int guc_ucode_xfer_dma(struct drm_i915_private *dev_priv, return ret; } -static u32 guc_wopcm_size(struct drm_i915_private *dev_priv) +u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv) { u32 wopcm_size = GUC_WOPCM_TOP; /* On BXT, the top of WOPCM is reserved for RC6 context */ - if (IS_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) wopcm_size -= BXT_GUC_WOPCM_RC6_RESERVED; return wopcm_size; @@ -350,29 +344,27 @@ static u32 guc_wopcm_size(struct drm_i915_private *dev_priv) */ static int guc_ucode_xfer(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; struct i915_vma *vma; int ret; - ret = i915_gem_object_set_to_gtt_domain(guc_fw->guc_fw_obj, false); + ret = i915_gem_object_set_to_gtt_domain(guc_fw->obj, false); if (ret) { DRM_DEBUG_DRIVER("set-domain failed %d\n", ret); return ret; } - vma = i915_gem_object_ggtt_pin(guc_fw->guc_fw_obj, NULL, 0, 0, 0); + vma = i915_gem_object_ggtt_pin(guc_fw->obj, NULL, 0, 0, + PIN_OFFSET_BIAS | GUC_WOPCM_TOP); if (IS_ERR(vma)) { DRM_DEBUG_DRIVER("pin failed %d\n", (int)PTR_ERR(vma)); return PTR_ERR(vma); } - /* Invalidate GuC TLB to let GuC take the latest updates to GTT. */ - I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); - intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); /* init WOPCM */ - I915_WRITE(GUC_WOPCM_SIZE, guc_wopcm_size(dev_priv)); + I915_WRITE(GUC_WOPCM_SIZE, intel_guc_wopcm_size(dev_priv)); I915_WRITE(DMA_GUC_WOPCM_OFFSET, GUC_WOPCM_OFFSET_VALUE); /* Enable MIA caching. GuC clock gating is disabled. */ @@ -388,7 +380,7 @@ static int guc_ucode_xfer(struct drm_i915_private *dev_priv) if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_B0)) I915_WRITE(GEN6_GFXPAUSE, 0x30FFF); - if (IS_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) I915_WRITE(GEN9LP_GT_PM_CONFIG, GT_DOORBELL_ENABLE); else I915_WRITE(GEN9_GT_PM_CONFIG, GT_DOORBELL_ENABLE); @@ -450,14 +442,14 @@ static int guc_hw_reset(struct drm_i915_private *dev_priv) */ int intel_guc_setup(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; - const char *fw_path = guc_fw->guc_fw_path; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; + const char *fw_path = guc_fw->path; int retries, ret, err; DRM_DEBUG_DRIVER("GuC fw status: path %s, fetch %s, load %s\n", fw_path, - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status), - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status), + intel_uc_fw_status_repr(guc_fw->load_status)); /* Loading forbidden, or no firmware to load? */ if (!i915.enable_guc_loading) { @@ -475,10 +467,10 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) } /* Fetch failed, or already fetched but failed to load? */ - if (guc_fw->guc_fw_fetch_status != GUC_FIRMWARE_SUCCESS) { + if (guc_fw->fetch_status != INTEL_UC_FIRMWARE_SUCCESS) { err = -EIO; goto fail; - } else if (guc_fw->guc_fw_load_status == GUC_FIRMWARE_FAIL) { + } else if (guc_fw->load_status == INTEL_UC_FIRMWARE_FAIL) { err = -ENOEXEC; goto fail; } @@ -486,11 +478,14 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) guc_interrupts_release(dev_priv); gen9_reset_guc_interrupts(dev_priv); - guc_fw->guc_fw_load_status = GUC_FIRMWARE_PENDING; + /* We need to notify the guc whenever we change the GGTT */ + i915_ggtt_enable_guc(dev_priv); + + guc_fw->load_status = INTEL_UC_FIRMWARE_PENDING; DRM_DEBUG_DRIVER("GuC fw status: fetch %s, load %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status), - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status), + intel_uc_fw_status_repr(guc_fw->load_status)); err = i915_guc_submission_init(dev_priv); if (err) @@ -511,6 +506,7 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) if (err) goto fail; + intel_huc_load(dev_priv); err = guc_ucode_xfer(dev_priv); if (!err) break; @@ -522,11 +518,13 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) "retry %d more time(s)\n", err, retries); } - guc_fw->guc_fw_load_status = GUC_FIRMWARE_SUCCESS; + guc_fw->load_status = INTEL_UC_FIRMWARE_SUCCESS; DRM_DEBUG_DRIVER("GuC fw status: fetch %s, load %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status), - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status), + intel_uc_fw_status_repr(guc_fw->load_status)); + + intel_guc_auth_huc(dev_priv); if (i915.enable_guc_submission) { if (i915.guc_log_level >= 0) @@ -541,12 +539,13 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) return 0; fail: - if (guc_fw->guc_fw_load_status == GUC_FIRMWARE_PENDING) - guc_fw->guc_fw_load_status = GUC_FIRMWARE_FAIL; + if (guc_fw->load_status == INTEL_UC_FIRMWARE_PENDING) + guc_fw->load_status = INTEL_UC_FIRMWARE_FAIL; guc_interrupts_release(dev_priv); i915_guc_submission_disable(dev_priv); i915_guc_submission_fini(dev_priv); + i915_ggtt_disable_guc(dev_priv); /* * We've failed to load the firmware :( @@ -587,93 +586,108 @@ fail: return ret; } -static void guc_fw_fetch(struct drm_i915_private *dev_priv, - struct intel_guc_fw *guc_fw) +void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, + struct intel_uc_fw *uc_fw) { struct pci_dev *pdev = dev_priv->drm.pdev; struct drm_i915_gem_object *obj; const struct firmware *fw = NULL; - struct guc_css_header *css; + struct uc_css_header *css; size_t size; int err; - DRM_DEBUG_DRIVER("before requesting firmware: GuC fw fetch status %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status)); + DRM_DEBUG_DRIVER("before requesting firmware: uC fw fetch status %s\n", + intel_uc_fw_status_repr(uc_fw->fetch_status)); - err = request_firmware(&fw, guc_fw->guc_fw_path, &pdev->dev); + err = request_firmware(&fw, uc_fw->path, &pdev->dev); if (err) goto fail; if (!fw) goto fail; - DRM_DEBUG_DRIVER("fetch GuC fw from %s succeeded, fw %p\n", - guc_fw->guc_fw_path, fw); + DRM_DEBUG_DRIVER("fetch uC fw from %s succeeded, fw %p\n", + uc_fw->path, fw); /* Check the size of the blob before examining buffer contents */ - if (fw->size < sizeof(struct guc_css_header)) { + if (fw->size < sizeof(struct uc_css_header)) { DRM_NOTE("Firmware header is missing\n"); goto fail; } - css = (struct guc_css_header *)fw->data; + css = (struct uc_css_header *)fw->data; /* Firmware bits always start from header */ - guc_fw->header_offset = 0; - guc_fw->header_size = (css->header_size_dw - css->modulus_size_dw - + uc_fw->header_offset = 0; + uc_fw->header_size = (css->header_size_dw - css->modulus_size_dw - css->key_size_dw - css->exponent_size_dw) * sizeof(u32); - if (guc_fw->header_size != sizeof(struct guc_css_header)) { + if (uc_fw->header_size != sizeof(struct uc_css_header)) { DRM_NOTE("CSS header definition mismatch\n"); goto fail; } /* then, uCode */ - guc_fw->ucode_offset = guc_fw->header_offset + guc_fw->header_size; - guc_fw->ucode_size = (css->size_dw - css->header_size_dw) * sizeof(u32); + uc_fw->ucode_offset = uc_fw->header_offset + uc_fw->header_size; + uc_fw->ucode_size = (css->size_dw - css->header_size_dw) * sizeof(u32); /* now RSA */ if (css->key_size_dw != UOS_RSA_SCRATCH_MAX_COUNT) { DRM_NOTE("RSA key size is bad\n"); goto fail; } - guc_fw->rsa_offset = guc_fw->ucode_offset + guc_fw->ucode_size; - guc_fw->rsa_size = css->key_size_dw * sizeof(u32); + uc_fw->rsa_offset = uc_fw->ucode_offset + uc_fw->ucode_size; + uc_fw->rsa_size = css->key_size_dw * sizeof(u32); /* At least, it should have header, uCode and RSA. Size of all three. */ - size = guc_fw->header_size + guc_fw->ucode_size + guc_fw->rsa_size; + size = uc_fw->header_size + uc_fw->ucode_size + uc_fw->rsa_size; if (fw->size < size) { DRM_NOTE("Missing firmware components\n"); goto fail; } - /* Header and uCode will be loaded to WOPCM. Size of the two. */ - size = guc_fw->header_size + guc_fw->ucode_size; - if (size > guc_wopcm_size(dev_priv)) { - DRM_NOTE("Firmware is too large to fit in WOPCM\n"); - goto fail; - } - /* * The GuC firmware image has the version number embedded at a well-known * offset within the firmware blob; note that major / minor version are * TWO bytes each (i.e. u16), although all pointers and offsets are defined * in terms of bytes (u8). */ - guc_fw->guc_fw_major_found = css->guc_sw_version >> 16; - guc_fw->guc_fw_minor_found = css->guc_sw_version & 0xFFFF; - - if (guc_fw->guc_fw_major_found != guc_fw->guc_fw_major_wanted || - guc_fw->guc_fw_minor_found < guc_fw->guc_fw_minor_wanted) { - DRM_NOTE("GuC firmware version %d.%d, required %d.%d\n", - guc_fw->guc_fw_major_found, guc_fw->guc_fw_minor_found, - guc_fw->guc_fw_major_wanted, guc_fw->guc_fw_minor_wanted); + switch (uc_fw->fw) { + case INTEL_UC_FW_TYPE_GUC: + /* Header and uCode will be loaded to WOPCM. Size of the two. */ + size = uc_fw->header_size + uc_fw->ucode_size; + + /* Top 32k of WOPCM is reserved (8K stack + 24k RC6 context). */ + if (size > intel_guc_wopcm_size(dev_priv)) { + DRM_ERROR("Firmware is too large to fit in WOPCM\n"); + goto fail; + } + uc_fw->major_ver_found = css->guc.sw_version >> 16; + uc_fw->minor_ver_found = css->guc.sw_version & 0xFFFF; + break; + + case INTEL_UC_FW_TYPE_HUC: + uc_fw->major_ver_found = css->huc.sw_version >> 16; + uc_fw->minor_ver_found = css->huc.sw_version & 0xFFFF; + break; + + default: + DRM_ERROR("Unknown firmware type %d\n", uc_fw->fw); + err = -ENOEXEC; + goto fail; + } + + if (uc_fw->major_ver_found != uc_fw->major_ver_wanted || + uc_fw->minor_ver_found < uc_fw->minor_ver_wanted) { + DRM_NOTE("uC firmware version %d.%d, required %d.%d\n", + uc_fw->major_ver_found, uc_fw->minor_ver_found, + uc_fw->major_ver_wanted, uc_fw->minor_ver_wanted); err = -ENOEXEC; goto fail; } DRM_DEBUG_DRIVER("firmware version %d.%d OK (minimum %d.%d)\n", - guc_fw->guc_fw_major_found, guc_fw->guc_fw_minor_found, - guc_fw->guc_fw_major_wanted, guc_fw->guc_fw_minor_wanted); + uc_fw->major_ver_found, uc_fw->minor_ver_found, + uc_fw->major_ver_wanted, uc_fw->minor_ver_wanted); mutex_lock(&dev_priv->drm.struct_mutex); obj = i915_gem_object_create_from_data(dev_priv, fw->data, fw->size); @@ -683,31 +697,31 @@ static void guc_fw_fetch(struct drm_i915_private *dev_priv, goto fail; } - guc_fw->guc_fw_obj = obj; - guc_fw->guc_fw_size = fw->size; + uc_fw->obj = obj; + uc_fw->size = fw->size; - DRM_DEBUG_DRIVER("GuC fw fetch status SUCCESS, obj %p\n", - guc_fw->guc_fw_obj); + DRM_DEBUG_DRIVER("uC fw fetch status SUCCESS, obj %p\n", + uc_fw->obj); release_firmware(fw); - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_SUCCESS; + uc_fw->fetch_status = INTEL_UC_FIRMWARE_SUCCESS; return; fail: - DRM_WARN("Failed to fetch valid GuC firmware from %s (error %d)\n", - guc_fw->guc_fw_path, err); - DRM_DEBUG_DRIVER("GuC fw fetch status FAIL; err %d, fw %p, obj %p\n", - err, fw, guc_fw->guc_fw_obj); + DRM_WARN("Failed to fetch valid uC firmware from %s (error %d)\n", + uc_fw->path, err); + DRM_DEBUG_DRIVER("uC fw fetch status FAIL; err %d, fw %p, obj %p\n", + err, fw, uc_fw->obj); mutex_lock(&dev_priv->drm.struct_mutex); - obj = guc_fw->guc_fw_obj; + obj = uc_fw->obj; if (obj) i915_gem_object_put(obj); - guc_fw->guc_fw_obj = NULL; + uc_fw->obj = NULL; mutex_unlock(&dev_priv->drm.struct_mutex); release_firmware(fw); /* OK even if fw is NULL */ - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_FAIL; + uc_fw->fetch_status = INTEL_UC_FIRMWARE_FAIL; } /** @@ -721,7 +735,7 @@ fail: */ void intel_guc_init(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; const char *fw_path; if (!HAS_GUC(dev_priv)) { @@ -739,23 +753,23 @@ void intel_guc_init(struct drm_i915_private *dev_priv) fw_path = NULL; } else if (IS_SKYLAKE(dev_priv)) { fw_path = I915_SKL_GUC_UCODE; - guc_fw->guc_fw_major_wanted = SKL_FW_MAJOR; - guc_fw->guc_fw_minor_wanted = SKL_FW_MINOR; + guc_fw->major_ver_wanted = SKL_FW_MAJOR; + guc_fw->minor_ver_wanted = SKL_FW_MINOR; } else if (IS_BROXTON(dev_priv)) { fw_path = I915_BXT_GUC_UCODE; - guc_fw->guc_fw_major_wanted = BXT_FW_MAJOR; - guc_fw->guc_fw_minor_wanted = BXT_FW_MINOR; + guc_fw->major_ver_wanted = BXT_FW_MAJOR; + guc_fw->minor_ver_wanted = BXT_FW_MINOR; } else if (IS_KABYLAKE(dev_priv)) { fw_path = I915_KBL_GUC_UCODE; - guc_fw->guc_fw_major_wanted = KBL_FW_MAJOR; - guc_fw->guc_fw_minor_wanted = KBL_FW_MINOR; + guc_fw->major_ver_wanted = KBL_FW_MAJOR; + guc_fw->minor_ver_wanted = KBL_FW_MINOR; } else { fw_path = ""; /* unknown device */ } - guc_fw->guc_fw_path = fw_path; - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_NONE; - guc_fw->guc_fw_load_status = GUC_FIRMWARE_NONE; + guc_fw->path = fw_path; + guc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; + guc_fw->load_status = INTEL_UC_FIRMWARE_NONE; /* Early (and silent) return if GuC loading is disabled */ if (!i915.enable_guc_loading) @@ -765,9 +779,9 @@ void intel_guc_init(struct drm_i915_private *dev_priv) if (*fw_path == '\0') return; - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_PENDING; + guc_fw->fetch_status = INTEL_UC_FIRMWARE_PENDING; DRM_DEBUG_DRIVER("GuC firmware pending, path %s\n", fw_path); - guc_fw_fetch(dev_priv, guc_fw); + intel_uc_fw_fetch(dev_priv, guc_fw); /* status must now be FAIL or SUCCESS */ } @@ -777,17 +791,17 @@ void intel_guc_init(struct drm_i915_private *dev_priv) */ void intel_guc_fini(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; mutex_lock(&dev_priv->drm.struct_mutex); guc_interrupts_release(dev_priv); i915_guc_submission_disable(dev_priv); i915_guc_submission_fini(dev_priv); - if (guc_fw->guc_fw_obj) - i915_gem_object_put(guc_fw->guc_fw_obj); - guc_fw->guc_fw_obj = NULL; + if (guc_fw->obj) + i915_gem_object_put(guc_fw->obj); + guc_fw->obj = NULL; mutex_unlock(&dev_priv->drm.struct_mutex); - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_NONE; + guc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; } diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c new file mode 100644 index 000000000000..5c0f9a49da0e --- /dev/null +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -0,0 +1,658 @@ +/* + * Copyright © 2014-2017 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + */ +#include <linux/debugfs.h> +#include <linux/relay.h> +#include "i915_drv.h" + +static void guc_log_capture_logs(struct intel_guc *guc); + +/** + * DOC: GuC firmware log + * + * Firmware log is enabled by setting i915.guc_log_level to non-negative level. + * Log data is printed out via reading debugfs i915_guc_log_dump. Reading from + * i915_guc_load_status will print out firmware loading status and scratch + * registers value. + * + */ + +static int guc_log_flush_complete(struct intel_guc *guc) +{ + u32 action[] = { + INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE + }; + + return intel_guc_send(guc, action, ARRAY_SIZE(action)); +} + +static int guc_log_flush(struct intel_guc *guc) +{ + u32 action[] = { + INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH, + 0 + }; + + return intel_guc_send(guc, action, ARRAY_SIZE(action)); +} + +static int guc_log_control(struct intel_guc *guc, u32 control_val) +{ + u32 action[] = { + INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING, + control_val + }; + + return intel_guc_send(guc, action, ARRAY_SIZE(action)); +} + + +/* + * Sub buffer switch callback. Called whenever relay has to switch to a new + * sub buffer, relay stays on the same sub buffer if 0 is returned. + */ +static int subbuf_start_callback(struct rchan_buf *buf, + void *subbuf, + void *prev_subbuf, + size_t prev_padding) +{ + /* Use no-overwrite mode by default, where relay will stop accepting + * new data if there are no empty sub buffers left. + * There is no strict synchronization enforced by relay between Consumer + * and Producer. In overwrite mode, there is a possibility of getting + * inconsistent/garbled data, the producer could be writing on to the + * same sub buffer from which Consumer is reading. This can't be avoided + * unless Consumer is fast enough and can always run in tandem with + * Producer. + */ + if (relay_buf_full(buf)) + return 0; + + return 1; +} + +/* + * file_create() callback. Creates relay file in debugfs. + */ +static struct dentry *create_buf_file_callback(const char *filename, + struct dentry *parent, + umode_t mode, + struct rchan_buf *buf, + int *is_global) +{ + struct dentry *buf_file; + + /* This to enable the use of a single buffer for the relay channel and + * correspondingly have a single file exposed to User, through which + * it can collect the logs in order without any post-processing. + * Need to set 'is_global' even if parent is NULL for early logging. + */ + *is_global = 1; + + if (!parent) + return NULL; + + /* Not using the channel filename passed as an argument, since for each + * channel relay appends the corresponding CPU number to the filename + * passed in relay_open(). This should be fine as relay just needs a + * dentry of the file associated with the channel buffer and that file's + * name need not be same as the filename passed as an argument. + */ + buf_file = debugfs_create_file("guc_log", mode, + parent, buf, &relay_file_operations); + return buf_file; +} + +/* + * file_remove() default callback. Removes relay file in debugfs. + */ +static int remove_buf_file_callback(struct dentry *dentry) +{ + debugfs_remove(dentry); + return 0; +} + +/* relay channel callbacks */ +static struct rchan_callbacks relay_callbacks = { + .subbuf_start = subbuf_start_callback, + .create_buf_file = create_buf_file_callback, + .remove_buf_file = remove_buf_file_callback, +}; + +static void guc_log_remove_relay_file(struct intel_guc *guc) +{ + relay_close(guc->log.relay_chan); +} + +static int guc_log_create_relay_channel(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + struct rchan *guc_log_relay_chan; + size_t n_subbufs, subbuf_size; + + /* Keep the size of sub buffers same as shared log buffer */ + subbuf_size = guc->log.vma->obj->base.size; + + /* Store up to 8 snapshots, which is large enough to buffer sufficient + * boot time logs and provides enough leeway to User, in terms of + * latency, for consuming the logs from relay. Also doesn't take + * up too much memory. + */ + n_subbufs = 8; + + guc_log_relay_chan = relay_open(NULL, NULL, subbuf_size, + n_subbufs, &relay_callbacks, dev_priv); + if (!guc_log_relay_chan) { + DRM_ERROR("Couldn't create relay chan for GuC logging\n"); + return -ENOMEM; + } + + GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); + guc->log.relay_chan = guc_log_relay_chan; + return 0; +} + +static int guc_log_create_relay_file(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + struct dentry *log_dir; + int ret; + + /* For now create the log file in /sys/kernel/debug/dri/0 dir */ + log_dir = dev_priv->drm.primary->debugfs_root; + + /* If /sys/kernel/debug/dri/0 location do not exist, then debugfs is + * not mounted and so can't create the relay file. + * The relay API seems to fit well with debugfs only, for availing relay + * there are 3 requirements which can be met for debugfs file only in a + * straightforward/clean manner :- + * i) Need the associated dentry pointer of the file, while opening the + * relay channel. + * ii) Should be able to use 'relay_file_operations' fops for the file. + * iii) Set the 'i_private' field of file's inode to the pointer of + * relay channel buffer. + */ + if (!log_dir) { + DRM_ERROR("Debugfs dir not available yet for GuC log file\n"); + return -ENODEV; + } + + ret = relay_late_setup_files(guc->log.relay_chan, "guc_log", log_dir); + if (ret) { + DRM_ERROR("Couldn't associate relay chan with file %d\n", ret); + return ret; + } + + return 0; +} + +static void guc_move_to_next_buf(struct intel_guc *guc) +{ + /* Make sure the updates made in the sub buffer are visible when + * Consumer sees the following update to offset inside the sub buffer. + */ + smp_wmb(); + + /* All data has been written, so now move the offset of sub buffer. */ + relay_reserve(guc->log.relay_chan, guc->log.vma->obj->base.size); + + /* Switch to the next sub buffer */ + relay_flush(guc->log.relay_chan); +} + +static void *guc_get_write_buffer(struct intel_guc *guc) +{ + if (!guc->log.relay_chan) + return NULL; + + /* Just get the base address of a new sub buffer and copy data into it + * ourselves. NULL will be returned in no-overwrite mode, if all sub + * buffers are full. Could have used the relay_write() to indirectly + * copy the data, but that would have been bit convoluted, as we need to + * write to only certain locations inside a sub buffer which cannot be + * done without using relay_reserve() along with relay_write(). So its + * better to use relay_reserve() alone. + */ + return relay_reserve(guc->log.relay_chan, 0); +} + +static bool guc_check_log_buf_overflow(struct intel_guc *guc, + enum guc_log_buffer_type type, + unsigned int full_cnt) +{ + unsigned int prev_full_cnt = guc->log.prev_overflow_count[type]; + bool overflow = false; + + if (full_cnt != prev_full_cnt) { + overflow = true; + + guc->log.prev_overflow_count[type] = full_cnt; + guc->log.total_overflow_count[type] += full_cnt - prev_full_cnt; + + if (full_cnt < prev_full_cnt) { + /* buffer_full_cnt is a 4 bit counter */ + guc->log.total_overflow_count[type] += 16; + } + DRM_ERROR_RATELIMITED("GuC log buffer overflow\n"); + } + + return overflow; +} + +static unsigned int guc_get_log_buffer_size(enum guc_log_buffer_type type) +{ + switch (type) { + case GUC_ISR_LOG_BUFFER: + return (GUC_LOG_ISR_PAGES + 1) * PAGE_SIZE; + case GUC_DPC_LOG_BUFFER: + return (GUC_LOG_DPC_PAGES + 1) * PAGE_SIZE; + case GUC_CRASH_DUMP_LOG_BUFFER: + return (GUC_LOG_CRASH_PAGES + 1) * PAGE_SIZE; + default: + MISSING_CASE(type); + } + + return 0; +} + +static void guc_read_update_log_buffer(struct intel_guc *guc) +{ + unsigned int buffer_size, read_offset, write_offset, bytes_to_copy, full_cnt; + struct guc_log_buffer_state *log_buf_state, *log_buf_snapshot_state; + struct guc_log_buffer_state log_buf_state_local; + enum guc_log_buffer_type type; + void *src_data, *dst_data; + bool new_overflow; + + if (WARN_ON(!guc->log.buf_addr)) + return; + + /* Get the pointer to shared GuC log buffer */ + log_buf_state = src_data = guc->log.buf_addr; + + /* Get the pointer to local buffer to store the logs */ + log_buf_snapshot_state = dst_data = guc_get_write_buffer(guc); + + /* Actual logs are present from the 2nd page */ + src_data += PAGE_SIZE; + dst_data += PAGE_SIZE; + + for (type = GUC_ISR_LOG_BUFFER; type < GUC_MAX_LOG_BUFFER; type++) { + /* Make a copy of the state structure, inside GuC log buffer + * (which is uncached mapped), on the stack to avoid reading + * from it multiple times. + */ + memcpy(&log_buf_state_local, log_buf_state, + sizeof(struct guc_log_buffer_state)); + buffer_size = guc_get_log_buffer_size(type); + read_offset = log_buf_state_local.read_ptr; + write_offset = log_buf_state_local.sampled_write_ptr; + full_cnt = log_buf_state_local.buffer_full_cnt; + + /* Bookkeeping stuff */ + guc->log.flush_count[type] += log_buf_state_local.flush_to_file; + new_overflow = guc_check_log_buf_overflow(guc, type, full_cnt); + + /* Update the state of shared log buffer */ + log_buf_state->read_ptr = write_offset; + log_buf_state->flush_to_file = 0; + log_buf_state++; + + if (unlikely(!log_buf_snapshot_state)) + continue; + + /* First copy the state structure in snapshot buffer */ + memcpy(log_buf_snapshot_state, &log_buf_state_local, + sizeof(struct guc_log_buffer_state)); + + /* The write pointer could have been updated by GuC firmware, + * after sending the flush interrupt to Host, for consistency + * set write pointer value to same value of sampled_write_ptr + * in the snapshot buffer. + */ + log_buf_snapshot_state->write_ptr = write_offset; + log_buf_snapshot_state++; + + /* Now copy the actual logs. */ + if (unlikely(new_overflow)) { + /* copy the whole buffer in case of overflow */ + read_offset = 0; + write_offset = buffer_size; + } else if (unlikely((read_offset > buffer_size) || + (write_offset > buffer_size))) { + DRM_ERROR("invalid log buffer state\n"); + /* copy whole buffer as offsets are unreliable */ + read_offset = 0; + write_offset = buffer_size; + } + + /* Just copy the newly written data */ + if (read_offset > write_offset) { + i915_memcpy_from_wc(dst_data, src_data, write_offset); + bytes_to_copy = buffer_size - read_offset; + } else { + bytes_to_copy = write_offset - read_offset; + } + i915_memcpy_from_wc(dst_data + read_offset, + src_data + read_offset, bytes_to_copy); + + src_data += buffer_size; + dst_data += buffer_size; + } + + if (log_buf_snapshot_state) + guc_move_to_next_buf(guc); + else { + /* Used rate limited to avoid deluge of messages, logs might be + * getting consumed by User at a slow rate. + */ + DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); + guc->log.capture_miss_count++; + } +} + +static void guc_log_cleanup(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + + lockdep_assert_held(&dev_priv->drm.struct_mutex); + + /* First disable the flush interrupt */ + gen9_disable_guc_interrupts(dev_priv); + + if (guc->log.flush_wq) + destroy_workqueue(guc->log.flush_wq); + + guc->log.flush_wq = NULL; + + if (guc->log.relay_chan) + guc_log_remove_relay_file(guc); + + guc->log.relay_chan = NULL; + + if (guc->log.buf_addr) + i915_gem_object_unpin_map(guc->log.vma->obj); + + guc->log.buf_addr = NULL; +} + +static void capture_logs_work(struct work_struct *work) +{ + struct intel_guc *guc = + container_of(work, struct intel_guc, log.flush_work); + + guc_log_capture_logs(guc); +} + +static int guc_log_create_extras(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + void *vaddr; + int ret; + + lockdep_assert_held(&dev_priv->drm.struct_mutex); + + /* Nothing to do */ + if (i915.guc_log_level < 0) + return 0; + + if (!guc->log.buf_addr) { + /* Create a WC (Uncached for read) vmalloc mapping of log + * buffer pages, so that we can directly get the data + * (up-to-date) from memory. + */ + vaddr = i915_gem_object_pin_map(guc->log.vma->obj, I915_MAP_WC); + if (IS_ERR(vaddr)) { + ret = PTR_ERR(vaddr); + DRM_ERROR("Couldn't map log buffer pages %d\n", ret); + return ret; + } + + guc->log.buf_addr = vaddr; + } + + if (!guc->log.relay_chan) { + /* Create a relay channel, so that we have buffers for storing + * the GuC firmware logs, the channel will be linked with a file + * later on when debugfs is registered. + */ + ret = guc_log_create_relay_channel(guc); + if (ret) + return ret; + } + + if (!guc->log.flush_wq) { + INIT_WORK(&guc->log.flush_work, capture_logs_work); + + /* + * GuC log buffer flush work item has to do register access to + * send the ack to GuC and this work item, if not synced before + * suspend, can potentially get executed after the GFX device is + * suspended. + * By marking the WQ as freezable, we don't have to bother about + * flushing of this work item from the suspend hooks, the pending + * work item if any will be either executed before the suspend + * or scheduled later on resume. This way the handling of work + * item can be kept same between system suspend & rpm suspend. + */ + guc->log.flush_wq = alloc_ordered_workqueue("i915-guc_log", + WQ_HIGHPRI | WQ_FREEZABLE); + if (guc->log.flush_wq == NULL) { + DRM_ERROR("Couldn't allocate the wq for GuC logging\n"); + return -ENOMEM; + } + } + + return 0; +} + +void intel_guc_log_create(struct intel_guc *guc) +{ + struct i915_vma *vma; + unsigned long offset; + uint32_t size, flags; + + if (i915.guc_log_level > GUC_LOG_VERBOSITY_MAX) + i915.guc_log_level = GUC_LOG_VERBOSITY_MAX; + + /* The first page is to save log buffer state. Allocate one + * extra page for others in case for overlap */ + size = (1 + GUC_LOG_DPC_PAGES + 1 + + GUC_LOG_ISR_PAGES + 1 + + GUC_LOG_CRASH_PAGES + 1) << PAGE_SHIFT; + + vma = guc->log.vma; + if (!vma) { + /* We require SSE 4.1 for fast reads from the GuC log buffer and + * it should be present on the chipsets supporting GuC based + * submisssions. + */ + if (WARN_ON(!i915_has_memcpy_from_wc())) { + /* logging will not be enabled */ + i915.guc_log_level = -1; + return; + } + + vma = intel_guc_allocate_vma(guc, size); + if (IS_ERR(vma)) { + /* logging will be off */ + i915.guc_log_level = -1; + return; + } + + guc->log.vma = vma; + + if (guc_log_create_extras(guc)) { + guc_log_cleanup(guc); + i915_vma_unpin_and_release(&guc->log.vma); + i915.guc_log_level = -1; + return; + } + } + + /* each allocated unit is a page */ + flags = GUC_LOG_VALID | GUC_LOG_NOTIFY_ON_HALF_FULL | + (GUC_LOG_DPC_PAGES << GUC_LOG_DPC_SHIFT) | + (GUC_LOG_ISR_PAGES << GUC_LOG_ISR_SHIFT) | + (GUC_LOG_CRASH_PAGES << GUC_LOG_CRASH_SHIFT); + + offset = guc_ggtt_offset(vma) >> PAGE_SHIFT; /* in pages */ + guc->log.flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; +} + +static int guc_log_late_setup(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + int ret; + + lockdep_assert_held(&dev_priv->drm.struct_mutex); + + if (i915.guc_log_level < 0) + return -EINVAL; + + /* If log_level was set as -1 at boot time, then setup needed to + * handle log buffer flush interrupts would not have been done yet, + * so do that now. + */ + ret = guc_log_create_extras(guc); + if (ret) + goto err; + + ret = guc_log_create_relay_file(guc); + if (ret) + goto err; + + return 0; +err: + guc_log_cleanup(guc); + /* logging will remain off */ + i915.guc_log_level = -1; + return ret; +} + +static void guc_log_capture_logs(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + + guc_read_update_log_buffer(guc); + + /* Generally device is expected to be active only at this + * time, so get/put should be really quick. + */ + intel_runtime_pm_get(dev_priv); + guc_log_flush_complete(guc); + intel_runtime_pm_put(dev_priv); +} + +static void guc_flush_logs(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + + if (!i915.enable_guc_submission || (i915.guc_log_level < 0)) + return; + + /* First disable the interrupts, will be renabled afterwards */ + gen9_disable_guc_interrupts(dev_priv); + + /* Before initiating the forceful flush, wait for any pending/ongoing + * flush to complete otherwise forceful flush may not actually happen. + */ + flush_work(&guc->log.flush_work); + + /* Ask GuC to update the log buffer state */ + guc_log_flush(guc); + + /* GuC would have updated log buffer by now, so capture it */ + guc_log_capture_logs(guc); +} + +int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) +{ + struct intel_guc *guc = &dev_priv->guc; + + union guc_log_control log_param; + int ret; + + log_param.value = control_val; + + if (log_param.verbosity < GUC_LOG_VERBOSITY_MIN || + log_param.verbosity > GUC_LOG_VERBOSITY_MAX) + return -EINVAL; + + /* This combination doesn't make sense & won't have any effect */ + if (!log_param.logging_enabled && (i915.guc_log_level < 0)) + return 0; + + ret = guc_log_control(guc, log_param.value); + if (ret < 0) { + DRM_DEBUG_DRIVER("guc_logging_control action failed %d\n", ret); + return ret; + } + + i915.guc_log_level = log_param.verbosity; + + /* If log_level was set as -1 at boot time, then the relay channel file + * wouldn't have been created by now and interrupts also would not have + * been enabled. + */ + if (!dev_priv->guc.log.relay_chan) { + ret = guc_log_late_setup(guc); + if (!ret) + gen9_enable_guc_interrupts(dev_priv); + } else if (!log_param.logging_enabled) { + /* Once logging is disabled, GuC won't generate logs & send an + * interrupt. But there could be some data in the log buffer + * which is yet to be captured. So request GuC to update the log + * buffer state and then collect the left over logs. + */ + guc_flush_logs(guc); + + /* As logging is disabled, update log level to reflect that */ + i915.guc_log_level = -1; + } else { + /* In case interrupts were disabled, enable them now */ + gen9_enable_guc_interrupts(dev_priv); + } + + return ret; +} + +void i915_guc_log_register(struct drm_i915_private *dev_priv) +{ + if (!i915.enable_guc_submission) + return; + + mutex_lock(&dev_priv->drm.struct_mutex); + guc_log_late_setup(&dev_priv->guc); + mutex_unlock(&dev_priv->drm.struct_mutex); +} + +void i915_guc_log_unregister(struct drm_i915_private *dev_priv) +{ + if (!i915.enable_guc_submission) + return; + + mutex_lock(&dev_priv->drm.struct_mutex); + guc_log_cleanup(&dev_priv->guc); + mutex_unlock(&dev_priv->drm.struct_mutex); +} diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c index 3d546c019de0..b62e3f8ad415 100644 --- a/drivers/gpu/drm/i915/intel_hotplug.c +++ b/drivers/gpu/drm/i915/intel_hotplug.c @@ -180,7 +180,7 @@ static void intel_hpd_irq_storm_disable(struct drm_i915_private *dev_priv) /* Enable polling and queue hotplug re-enabling. */ if (hpd_disabled) { - drm_kms_helper_poll_enable_locked(dev); + drm_kms_helper_poll_enable(dev); mod_delayed_work(system_wq, &dev_priv->hotplug.reenable_work, msecs_to_jiffies(HPD_STORM_REENABLE_DELAY)); } @@ -511,7 +511,7 @@ static void i915_hpd_poll_init_work(struct work_struct *work) } if (enabled) - drm_kms_helper_poll_enable_locked(dev); + drm_kms_helper_poll_enable(dev); mutex_unlock(&dev->mode_config.mutex); diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c new file mode 100644 index 000000000000..c144609425f6 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -0,0 +1,338 @@ +/* + * Copyright © 2016-2017 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + */ +#include <linux/firmware.h> +#include "i915_drv.h" +#include "intel_uc.h" + +/** + * DOC: HuC Firmware + * + * Motivation: + * GEN9 introduces a new dedicated firmware for usage in media HEVC (High + * Efficiency Video Coding) operations. Userspace can use the firmware + * capabilities by adding HuC specific commands to batch buffers. + * + * Implementation: + * The same firmware loader is used as the GuC. However, the actual + * loading to HW is deferred until GEM initialization is done. + * + * Note that HuC firmware loading must be done before GuC loading. + */ + +#define BXT_HUC_FW_MAJOR 01 +#define BXT_HUC_FW_MINOR 07 +#define BXT_BLD_NUM 1398 + +#define SKL_HUC_FW_MAJOR 01 +#define SKL_HUC_FW_MINOR 07 +#define SKL_BLD_NUM 1398 + +#define KBL_HUC_FW_MAJOR 02 +#define KBL_HUC_FW_MINOR 00 +#define KBL_BLD_NUM 1810 + +#define HUC_FW_PATH(platform, major, minor, bld_num) \ + "i915/" __stringify(platform) "_huc_ver" __stringify(major) "_" \ + __stringify(minor) "_" __stringify(bld_num) ".bin" + +#define I915_SKL_HUC_UCODE HUC_FW_PATH(skl, SKL_HUC_FW_MAJOR, \ + SKL_HUC_FW_MINOR, SKL_BLD_NUM) +MODULE_FIRMWARE(I915_SKL_HUC_UCODE); + +#define I915_BXT_HUC_UCODE HUC_FW_PATH(bxt, BXT_HUC_FW_MAJOR, \ + BXT_HUC_FW_MINOR, BXT_BLD_NUM) +MODULE_FIRMWARE(I915_BXT_HUC_UCODE); + +#define I915_KBL_HUC_UCODE HUC_FW_PATH(kbl, KBL_HUC_FW_MAJOR, \ + KBL_HUC_FW_MINOR, KBL_BLD_NUM) +MODULE_FIRMWARE(I915_KBL_HUC_UCODE); + +/** + * huc_ucode_xfer() - DMA's the firmware + * @dev_priv: the drm_i915_private device + * + * Transfer the firmware image to RAM for execution by the microcontroller. + * + * Return: 0 on success, non-zero on failure + */ +static int huc_ucode_xfer(struct drm_i915_private *dev_priv) +{ + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + struct i915_vma *vma; + unsigned long offset = 0; + u32 size; + int ret; + + ret = i915_gem_object_set_to_gtt_domain(huc_fw->obj, false); + if (ret) { + DRM_DEBUG_DRIVER("set-domain failed %d\n", ret); + return ret; + } + + vma = i915_gem_object_ggtt_pin(huc_fw->obj, NULL, 0, 0, + PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + if (IS_ERR(vma)) { + DRM_DEBUG_DRIVER("pin failed %d\n", (int)PTR_ERR(vma)); + return PTR_ERR(vma); + } + + intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); + + /* init WOPCM */ + I915_WRITE(GUC_WOPCM_SIZE, intel_guc_wopcm_size(dev_priv)); + I915_WRITE(DMA_GUC_WOPCM_OFFSET, GUC_WOPCM_OFFSET_VALUE | + HUC_LOADING_AGENT_GUC); + + /* Set the source address for the uCode */ + offset = guc_ggtt_offset(vma) + huc_fw->header_offset; + I915_WRITE(DMA_ADDR_0_LOW, lower_32_bits(offset)); + I915_WRITE(DMA_ADDR_0_HIGH, upper_32_bits(offset) & 0xFFFF); + + /* Hardware doesn't look at destination address for HuC. Set it to 0, + * but still program the correct address space. + */ + I915_WRITE(DMA_ADDR_1_LOW, 0); + I915_WRITE(DMA_ADDR_1_HIGH, DMA_ADDRESS_SPACE_WOPCM); + + size = huc_fw->header_size + huc_fw->ucode_size; + I915_WRITE(DMA_COPY_SIZE, size); + + /* Start the DMA */ + I915_WRITE(DMA_CTRL, _MASKED_BIT_ENABLE(HUC_UKERNEL | START_DMA)); + + /* Wait for DMA to finish */ + ret = wait_for((I915_READ(DMA_CTRL) & START_DMA) == 0, 100); + + DRM_DEBUG_DRIVER("HuC DMA transfer wait over with ret %d\n", ret); + + /* Disable the bits once DMA is over */ + I915_WRITE(DMA_CTRL, _MASKED_BIT_DISABLE(HUC_UKERNEL)); + + intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); + + /* + * We keep the object pages for reuse during resume. But we can unpin it + * now that DMA has completed, so it doesn't continue to take up space. + */ + i915_vma_unpin(vma); + + return ret; +} + +/** + * intel_huc_init() - initiate HuC firmware loading request + * @dev_priv: the drm_i915_private device + * + * Called early during driver load, but after GEM is initialised. The loading + * will continue only when driver explicitly specify firmware name and version. + * All other cases are considered as INTEL_UC_FIRMWARE_NONE either because HW + * is not capable or driver yet support it. And there will be no error message + * for INTEL_UC_FIRMWARE_NONE cases. + * + * The DMA-copying to HW is done later when intel_huc_load() is called. + */ +void intel_huc_init(struct drm_i915_private *dev_priv) +{ + struct intel_huc *huc = &dev_priv->huc; + struct intel_uc_fw *huc_fw = &huc->fw; + const char *fw_path = NULL; + + huc_fw->path = NULL; + huc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; + huc_fw->load_status = INTEL_UC_FIRMWARE_NONE; + huc_fw->fw = INTEL_UC_FW_TYPE_HUC; + + if (!HAS_HUC_UCODE(dev_priv)) + return; + + if (IS_SKYLAKE(dev_priv)) { + fw_path = I915_SKL_HUC_UCODE; + huc_fw->major_ver_wanted = SKL_HUC_FW_MAJOR; + huc_fw->minor_ver_wanted = SKL_HUC_FW_MINOR; + } else if (IS_BROXTON(dev_priv)) { + fw_path = I915_BXT_HUC_UCODE; + huc_fw->major_ver_wanted = BXT_HUC_FW_MAJOR; + huc_fw->minor_ver_wanted = BXT_HUC_FW_MINOR; + } else if (IS_KABYLAKE(dev_priv)) { + fw_path = I915_KBL_HUC_UCODE; + huc_fw->major_ver_wanted = KBL_HUC_FW_MAJOR; + huc_fw->minor_ver_wanted = KBL_HUC_FW_MINOR; + } + + huc_fw->path = fw_path; + huc_fw->fetch_status = INTEL_UC_FIRMWARE_PENDING; + + DRM_DEBUG_DRIVER("HuC firmware pending, path %s\n", fw_path); + + WARN(huc_fw->path == NULL, "HuC present but no fw path\n"); + + intel_uc_fw_fetch(dev_priv, huc_fw); +} + +/** + * intel_huc_load() - load HuC uCode to device + * @dev_priv: the drm_i915_private device + * + * Called from guc_setup() during driver loading and also after a GPU reset. + * Be note that HuC loading must be done before GuC loading. + * + * The firmware image should have already been fetched into memory by the + * earlier call to intel_huc_init(), so here we need only check that + * is succeeded, and then transfer the image to the h/w. + * + * Return: non-zero code on error + */ +int intel_huc_load(struct drm_i915_private *dev_priv) +{ + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + int err; + + if (huc_fw->fetch_status == INTEL_UC_FIRMWARE_NONE) + return 0; + + DRM_DEBUG_DRIVER("%s fw status: fetch %s, load %s\n", + huc_fw->path, + intel_uc_fw_status_repr(huc_fw->fetch_status), + intel_uc_fw_status_repr(huc_fw->load_status)); + + if (huc_fw->fetch_status == INTEL_UC_FIRMWARE_SUCCESS && + huc_fw->load_status == INTEL_UC_FIRMWARE_FAIL) + return -ENOEXEC; + + huc_fw->load_status = INTEL_UC_FIRMWARE_PENDING; + + switch (huc_fw->fetch_status) { + case INTEL_UC_FIRMWARE_FAIL: + /* something went wrong :( */ + err = -EIO; + goto fail; + + case INTEL_UC_FIRMWARE_NONE: + case INTEL_UC_FIRMWARE_PENDING: + default: + /* "can't happen" */ + WARN_ONCE(1, "HuC fw %s invalid fetch_status %s [%d]\n", + huc_fw->path, + intel_uc_fw_status_repr(huc_fw->fetch_status), + huc_fw->fetch_status); + err = -ENXIO; + goto fail; + + case INTEL_UC_FIRMWARE_SUCCESS: + break; + } + + err = huc_ucode_xfer(dev_priv); + if (err) + goto fail; + + huc_fw->load_status = INTEL_UC_FIRMWARE_SUCCESS; + + DRM_DEBUG_DRIVER("%s fw status: fetch %s, load %s\n", + huc_fw->path, + intel_uc_fw_status_repr(huc_fw->fetch_status), + intel_uc_fw_status_repr(huc_fw->load_status)); + + return 0; + +fail: + if (huc_fw->load_status == INTEL_UC_FIRMWARE_PENDING) + huc_fw->load_status = INTEL_UC_FIRMWARE_FAIL; + + DRM_ERROR("Failed to complete HuC uCode load with ret %d\n", err); + + return err; +} + +/** + * intel_huc_fini() - clean up resources allocated for HuC + * @dev_priv: the drm_i915_private device + * + * Cleans up by releasing the huc firmware GEM obj. + */ +void intel_huc_fini(struct drm_i915_private *dev_priv) +{ + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + + mutex_lock(&dev_priv->drm.struct_mutex); + if (huc_fw->obj) + i915_gem_object_put(huc_fw->obj); + huc_fw->obj = NULL; + mutex_unlock(&dev_priv->drm.struct_mutex); + + huc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; +} + +/** + * intel_guc_auth_huc() - authenticate ucode + * @dev_priv: the drm_i915_device + * + * Triggers a HuC fw authentication request to the GuC via intel_guc_action_ + * authenticate_huc interface. + */ +void intel_guc_auth_huc(struct drm_i915_private *dev_priv) +{ + struct intel_guc *guc = &dev_priv->guc; + struct intel_huc *huc = &dev_priv->huc; + struct i915_vma *vma; + int ret; + u32 data[2]; + + if (huc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) + return; + + vma = i915_gem_object_ggtt_pin(huc->fw.obj, NULL, 0, 0, + PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + if (IS_ERR(vma)) { + DRM_ERROR("failed to pin huc fw object %d\n", + (int)PTR_ERR(vma)); + return; + } + + /* Specify auth action and where public signature is. */ + data[0] = INTEL_GUC_ACTION_AUTHENTICATE_HUC; + data[1] = guc_ggtt_offset(vma) + huc->fw.rsa_offset; + + ret = intel_guc_send(guc, data, ARRAY_SIZE(data)); + if (ret) { + DRM_ERROR("HuC: GuC did not ack Auth request %d\n", ret); + goto out; + } + + /* Check authentication status, it should be done by now */ + ret = intel_wait_for_register(dev_priv, + HUC_STATUS2, + HUC_FW_VERIFIED, + HUC_FW_VERIFIED, + 50); + + if (ret) { + DRM_ERROR("HuC: Authentication failed %d\n", ret); + goto out; + } + +out: + i915_vma_unpin(vma); +} + diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 6db246ad2f13..432ee495dec2 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -811,12 +811,6 @@ static int execlists_context_pin(struct intel_engine_cs *engine, ce->state->obj->mm.dirty = true; - /* Invalidate GuC TLB. */ - if (i915.enable_guc_submission) { - struct drm_i915_private *dev_priv = ctx->i915; - I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); - } - i915_gem_context_get(ctx); return 0; @@ -970,18 +964,8 @@ static inline int gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine, uint32_t *batch, uint32_t index) { - struct drm_i915_private *dev_priv = engine->i915; uint32_t l3sqc4_flush = (0x40400000 | GEN8_LQSC_FLUSH_COHERENT_LINES); - /* - * WaDisableLSQCROPERFforOCL:kbl - * This WA is implemented in skl_init_clock_gating() but since - * this batch updates GEN8_L3SQCREG4 with default value we need to - * set this bit here to retain the WA during flush. - */ - if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_E0)) - l3sqc4_flush |= GEN8_LQSC_RO_PERF_DIS; - wa_ctx_emit(batch, index, (MI_STORE_REGISTER_MEM_GEN8 | MI_SRM_LRM_GLOBAL_GTT)); wa_ctx_emit_reg(batch, index, GEN8_L3SQCREG4); @@ -1241,7 +1225,7 @@ static int lrc_setup_wa_ctx_obj(struct intel_engine_cs *engine, u32 size) if (IS_ERR(obj)) return PTR_ERR(obj); - vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto err; @@ -1927,7 +1911,7 @@ int logical_render_ring_init(struct intel_engine_cs *engine) engine->emit_breadcrumb = gen8_emit_breadcrumb_render; engine->emit_breadcrumb_sz = gen8_emit_breadcrumb_render_sz; - ret = intel_engine_create_scratch(engine, 4096); + ret = intel_engine_create_scratch(engine, PAGE_SIZE); if (ret) return ret; @@ -2103,19 +2087,12 @@ static void execlists_init_reg_state(u32 *reg_state, ASSIGN_CTX_REG(reg_state, CTX_PDP0_LDW, GEN8_RING_PDP_LDW(engine, 0), 0); - if (USES_FULL_48BIT_PPGTT(ppgtt->base.dev)) { + if (ppgtt && USES_FULL_48BIT_PPGTT(ppgtt->base.dev)) { /* 64b PPGTT (48bit canonical) * PDP0_DESCRIPTOR contains the base address to PML4 and * other PDP Descriptors are ignored. */ ASSIGN_CTX_PML4(ppgtt, reg_state); - } else { - /* 32b PPGTT - * PDP*_DESCRIPTOR contains the base address of space supported. - * With dynamic page allocation, PDPs may not be allocated at - * this point. Point the unallocated PDPs to the scratch page - */ - execlists_update_context_pdps(ppgtt, reg_state); } if (engine->id == RCS) { @@ -2209,7 +2186,8 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, WARN_ON(ce->state); - context_size = round_up(intel_lr_context_size(engine), 4096); + context_size = round_up(intel_lr_context_size(engine), + I915_GTT_PAGE_SIZE); /* One extra page as the sharing data between driver and GuC */ context_size += PAGE_SIZE * LRC_PPHWSP_PN; @@ -2220,7 +2198,7 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, return PTR_ERR(ctx_obj); } - vma = i915_vma_create(ctx_obj, &ctx->i915->ggtt.base, NULL); + vma = i915_vma_instance(ctx_obj, &ctx->i915->ggtt.base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto error_deref_obj; diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h index 01ba36ea125e..0c852c024227 100644 --- a/drivers/gpu/drm/i915/intel_lrc.h +++ b/drivers/gpu/drm/i915/intel_lrc.h @@ -26,7 +26,7 @@ #include "intel_ringbuffer.h" -#define GEN8_LR_CONTEXT_ALIGN 4096 +#define GEN8_LR_CONTEXT_ALIGN I915_GTT_MIN_ALIGNMENT /* Execlists regs */ #define RING_ELSP(engine) _MMIO((engine)->mmio_base + 0x230) diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 4473a611c664..0608fad7f593 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -811,8 +811,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay, if (ret != 0) return ret; - vma = i915_gem_object_pin_to_display_plane(new_bo, 0, - &i915_ggtt_view_normal); + vma = i915_gem_object_pin_to_display_plane(new_bo, 0, NULL); if (IS_ERR(vma)) return PTR_ERR(vma); diff --git a/drivers/gpu/drm/i915/intel_pipe_crc.c b/drivers/gpu/drm/i915/intel_pipe_crc.c index ef0c0e195164..c0b1f99da37b 100644 --- a/drivers/gpu/drm/i915/intel_pipe_crc.c +++ b/drivers/gpu/drm/i915/intel_pipe_crc.c @@ -560,14 +560,14 @@ static void hsw_trans_edp_pipe_A_crc_wa(struct drm_i915_private *dev_priv, state = drm_atomic_state_alloc(dev); if (!state) { ret = -ENOMEM; - goto out; + goto unlock; } state->acquire_ctx = drm_modeset_legacy_acquire_ctx(&crtc->base); pipe_config = intel_atomic_get_crtc_state(state, crtc); if (IS_ERR(pipe_config)) { ret = PTR_ERR(pipe_config); - goto out; + goto put_state; } pipe_config->pch_pfit.force_thru = enable; @@ -576,10 +576,12 @@ static void hsw_trans_edp_pipe_A_crc_wa(struct drm_i915_private *dev_priv, pipe_config->base.connectors_changed = true; ret = drm_atomic_commit(state); -out: + +put_state: + drm_atomic_state_put(state); +unlock: WARN(ret, "Toggling workaround to %i returns %i\n", enable, ret); drm_modeset_unlock_all(dev); - drm_atomic_state_put(state); } static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, @@ -613,6 +615,22 @@ static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, return 0; } +static int get_new_crc_ctl_reg(struct drm_i915_private *dev_priv, + enum pipe pipe, + enum intel_pipe_crc_source *source, u32 *val) +{ + if (IS_GEN2(dev_priv)) + return i8xx_pipe_crc_ctl_reg(source, val); + else if (INTEL_GEN(dev_priv) < 5) + return i9xx_pipe_crc_ctl_reg(dev_priv, pipe, source, val); + else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + return vlv_pipe_crc_ctl_reg(dev_priv, pipe, source, val); + else if (IS_GEN5(dev_priv) || IS_GEN6(dev_priv)) + return ilk_pipe_crc_ctl_reg(source, val); + else + return ivb_pipe_crc_ctl_reg(dev_priv, pipe, source, val); +} + static int pipe_crc_set_source(struct drm_i915_private *dev_priv, enum pipe pipe, enum intel_pipe_crc_source source) @@ -636,17 +654,7 @@ static int pipe_crc_set_source(struct drm_i915_private *dev_priv, return -EIO; } - if (IS_GEN2(dev_priv)) - ret = i8xx_pipe_crc_ctl_reg(&source, &val); - else if (INTEL_GEN(dev_priv) < 5) - ret = i9xx_pipe_crc_ctl_reg(dev_priv, pipe, &source, &val); - else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - ret = vlv_pipe_crc_ctl_reg(dev_priv, pipe, &source, &val); - else if (IS_GEN5(dev_priv) || IS_GEN6(dev_priv)) - ret = ilk_pipe_crc_ctl_reg(&source, &val); - else - ret = ivb_pipe_crc_ctl_reg(dev_priv, pipe, &source, &val); - + ret = get_new_crc_ctl_reg(dev_priv, pipe, &source, &val); if (ret != 0) goto out; @@ -687,7 +695,7 @@ static int pipe_crc_set_source(struct drm_i915_private *dev_priv, POSTING_READ(PIPE_CRC_CTL(pipe)); /* real source -> none transition */ - if (source == INTEL_PIPE_CRC_SOURCE_NONE) { + if (!source) { struct intel_pipe_crc_entry *entries; struct intel_crtc *crtc = intel_get_crtc_for_pipe(dev_priv, pipe); @@ -809,6 +817,11 @@ display_crc_ctl_parse_source(const char *buf, enum intel_pipe_crc_source *s) { int i; + if (!buf) { + *s = INTEL_PIPE_CRC_SOURCE_NONE; + return 0; + } + for (i = 0; i < ARRAY_SIZE(pipe_crc_sources); i++) if (!strcmp(buf, pipe_crc_sources[i])) { *s = i; @@ -937,3 +950,62 @@ void intel_pipe_crc_cleanup(struct drm_minor *minor) drm_debugfs_remove_files(info_list, 1, minor); } } + +int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, + size_t *values_cnt) +{ + struct drm_i915_private *dev_priv = crtc->dev->dev_private; + struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + enum intel_display_power_domain power_domain; + enum intel_pipe_crc_source source; + u32 val = 0; /* shut up gcc */ + int ret = 0; + + if (display_crc_ctl_parse_source(source_name, &source) < 0) { + DRM_DEBUG_DRIVER("unknown source %s\n", source_name); + return -EINVAL; + } + + power_domain = POWER_DOMAIN_PIPE(crtc->index); + if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) { + DRM_DEBUG_KMS("Trying to capture CRC while pipe is off\n"); + return -EIO; + } + + ret = get_new_crc_ctl_reg(dev_priv, crtc->index, &source, &val); + if (ret != 0) + goto out; + + if (source) { + /* + * When IPS gets enabled, the pipe CRC changes. Since IPS gets + * enabled and disabled dynamically based on package C states, + * user space can't make reliable use of the CRCs, so let's just + * completely disable it. + */ + hsw_disable_ips(intel_crtc); + } + + I915_WRITE(PIPE_CRC_CTL(crtc->index), val); + POSTING_READ(PIPE_CRC_CTL(crtc->index)); + + if (!source) { + if (IS_G4X(dev_priv)) + g4x_undo_pipe_scramble_reset(dev_priv, crtc->index); + else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + vlv_undo_pipe_scramble_reset(dev_priv, crtc->index); + else if (IS_HASWELL(dev_priv) && crtc->index == PIPE_A) + hsw_trans_edp_pipe_A_crc_wa(dev_priv, false); + + hsw_enable_ips(intel_crtc); + } + + pipe_crc->skipped = 0; + *values_cnt = 5; + +out: + intel_display_power_put(dev_priv, power_domain); + + return ret; +} diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 6aca8ff14989..c3780d0d2baf 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -122,13 +122,26 @@ static void vlv_psr_setup_vsc(struct intel_dp *intel_dp) static void skl_psr_setup_su_vsc(struct intel_dp *intel_dp) { struct edp_vsc_psr psr_vsc; + struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); + struct drm_device *dev = intel_dig_port->base.base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); /* Prepare VSC Header for SU as per EDP 1.4 spec, Table 6.11 */ memset(&psr_vsc, 0, sizeof(psr_vsc)); psr_vsc.sdp_header.HB0 = 0; psr_vsc.sdp_header.HB1 = 0x7; - psr_vsc.sdp_header.HB2 = 0x3; - psr_vsc.sdp_header.HB3 = 0xb; + if (dev_priv->psr.colorimetry_support && + dev_priv->psr.y_cord_support) { + psr_vsc.sdp_header.HB2 = 0x5; + psr_vsc.sdp_header.HB3 = 0x13; + } else if (dev_priv->psr.y_cord_support) { + psr_vsc.sdp_header.HB2 = 0x4; + psr_vsc.sdp_header.HB3 = 0xe; + } else { + psr_vsc.sdp_header.HB2 = 0x3; + psr_vsc.sdp_header.HB3 = 0xc; + } + intel_psr_write_vsc(intel_dp, &psr_vsc); } @@ -196,7 +209,11 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) drm_dp_dpcd_writeb(&intel_dp->aux, DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, DP_AUX_FRAME_SYNC_ENABLE); - + /* Enable ALPM at sink for psr2 */ + if (dev_priv->psr.psr2_support && dev_priv->psr.alpm) + drm_dp_dpcd_writeb(&intel_dp->aux, + DP_RECEIVER_ALPM_CONFIG, + DP_ALPM_ENABLE); if (dev_priv->psr.link_standby) drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE); @@ -248,7 +265,7 @@ static void vlv_psr_activate(struct intel_dp *intel_dp) VLV_EDP_PSR_ACTIVE_ENTRY); } -static void hsw_psr_enable_source(struct intel_dp *intel_dp) +static void intel_enable_source_psr1(struct intel_dp *intel_dp) { struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct drm_device *dev = dig_port->base.base.dev; @@ -299,14 +316,31 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp) val |= EDP_PSR_TP1_TP2_SEL; I915_WRITE(EDP_PSR_CTL, val); +} - if (!dev_priv->psr.psr2_support) - return; +static void intel_enable_source_psr2(struct intel_dp *intel_dp) +{ + struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); + struct drm_device *dev = dig_port->base.base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); + /* + * Let's respect VBT in case VBT asks a higher idle_frame value. + * Let's use 6 as the minimum to cover all known cases including + * the off-by-one issue that HW has in some cases. Also there are + * cases where sink should be able to train + * with the 5 or 6 idle patterns. + */ + uint32_t idle_frames = max(6, dev_priv->vbt.psr.idle_frames); + uint32_t val; + + val = idle_frames << EDP_PSR_IDLE_FRAME_SHIFT; /* FIXME: selective update is probably totally broken because it doesn't * mesh at all with our frontbuffer tracking. And the hw alone isn't * good enough. */ - val = EDP_PSR2_ENABLE | EDP_SU_TRACK_ENABLE; + val |= EDP_PSR2_ENABLE | + EDP_SU_TRACK_ENABLE | + EDP_FRAMES_BEFORE_SU_ENTRY; if (dev_priv->vbt.psr.tp2_tp3_wakeup_time > 5) val |= EDP_PSR2_TP2_TIME_2500; @@ -320,6 +354,19 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp) I915_WRITE(EDP_PSR2_CTL, val); } +static void hsw_psr_enable_source(struct intel_dp *intel_dp) +{ + struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); + struct drm_device *dev = dig_port->base.base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); + + /* psr1 and psr2 are mutually exclusive.*/ + if (dev_priv->psr.psr2_support) + intel_enable_source_psr2(intel_dp); + else + intel_enable_source_psr1(intel_dp); +} + static bool intel_psr_match_conditions(struct intel_dp *intel_dp) { struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); @@ -387,6 +434,22 @@ static bool intel_psr_match_conditions(struct intel_dp *intel_dp) return false; } + /* PSR2 is restricted to work with panel resolutions upto 3200x2000 */ + if (intel_crtc->config->pipe_src_w > 3200 || + intel_crtc->config->pipe_src_h > 2000) { + dev_priv->psr.psr2_support = false; + return false; + } + + /* + * FIXME:enable psr2 only for y-cordinate psr2 panels + * After gtc implementation , remove this restriction. + */ + if (!dev_priv->psr.y_cord_support && dev_priv->psr.psr2_support) { + DRM_DEBUG_KMS("PSR2 disabled, panel does not support Y coordinate\n"); + return false; + } + dev_priv->psr.source_ok = true; return true; } @@ -397,7 +460,10 @@ static void intel_psr_activate(struct intel_dp *intel_dp) struct drm_device *dev = intel_dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); + if (dev_priv->psr.psr2_support) + WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); + else + WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); WARN_ON(dev_priv->psr.active); lockdep_assert_held(&dev_priv->psr.lock); @@ -426,6 +492,8 @@ void intel_psr_enable(struct intel_dp *intel_dp) struct drm_device *dev = intel_dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); struct intel_crtc *crtc = to_intel_crtc(intel_dig_port->base.base.crtc); + enum transcoder cpu_transcoder = crtc->config->cpu_transcoder; + u32 chicken; if (!HAS_PSR(dev_priv)) { DRM_DEBUG_KMS("PSR not supported on this platform\n"); @@ -449,26 +517,34 @@ void intel_psr_enable(struct intel_dp *intel_dp) dev_priv->psr.busy_frontbuffer_bits = 0; if (HAS_DDI(dev_priv)) { - hsw_psr_setup_vsc(intel_dp); - if (dev_priv->psr.psr2_support) { - /* PSR2 is restricted to work with panel resolutions upto 3200x2000 */ - if (crtc->config->pipe_src_w > 3200 || - crtc->config->pipe_src_h > 2000) - dev_priv->psr.psr2_support = false; - else - skl_psr_setup_su_vsc(intel_dp); + skl_psr_setup_su_vsc(intel_dp); + chicken = PSR2_VSC_ENABLE_PROG_HEADER; + if (dev_priv->psr.y_cord_support) + chicken |= PSR2_ADD_VERTICAL_LINE_COUNT; + I915_WRITE(CHICKEN_TRANS(cpu_transcoder), chicken); + I915_WRITE(EDP_PSR_DEBUG_CTL, + EDP_PSR_DEBUG_MASK_MEMUP | + EDP_PSR_DEBUG_MASK_HPD | + EDP_PSR_DEBUG_MASK_LPSP | + EDP_PSR_DEBUG_MASK_MAX_SLEEP | + EDP_PSR_DEBUG_MASK_DISP_REG_WRITE); + } else { + /* set up vsc header for psr1 */ + hsw_psr_setup_vsc(intel_dp); + /* + * Per Spec: Avoid continuous PSR exit by masking MEMUP + * and HPD. also mask LPSP to avoid dependency on other + * drivers that might block runtime_pm besides + * preventing other hw tracking issues now we can rely + * on frontbuffer tracking. + */ + I915_WRITE(EDP_PSR_DEBUG_CTL, + EDP_PSR_DEBUG_MASK_MEMUP | + EDP_PSR_DEBUG_MASK_HPD | + EDP_PSR_DEBUG_MASK_LPSP); } - /* - * Per Spec: Avoid continuous PSR exit by masking MEMUP and HPD. - * Also mask LPSP to avoid dependency on other drivers that - * might block runtime_pm besides preventing other hw tracking - * issues now we can rely on frontbuffer tracking. - */ - I915_WRITE(EDP_PSR_DEBUG_CTL, EDP_PSR_DEBUG_MASK_MEMUP | - EDP_PSR_DEBUG_MASK_HPD | EDP_PSR_DEBUG_MASK_LPSP); - /* Enable PSR on the panel */ hsw_psr_enable_sink(intel_dp); @@ -544,20 +620,42 @@ static void hsw_psr_disable(struct intel_dp *intel_dp) struct drm_i915_private *dev_priv = to_i915(dev); if (dev_priv->psr.active) { - I915_WRITE(EDP_PSR_CTL, - I915_READ(EDP_PSR_CTL) & ~EDP_PSR_ENABLE); + i915_reg_t psr_ctl; + u32 psr_status_mask; + + if (dev_priv->psr.aux_frame_sync) + drm_dp_dpcd_writeb(&intel_dp->aux, + DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, + 0); + + if (dev_priv->psr.psr2_support) { + psr_ctl = EDP_PSR2_CTL; + psr_status_mask = EDP_PSR2_STATUS_STATE_MASK; + + I915_WRITE(psr_ctl, + I915_READ(psr_ctl) & + ~(EDP_PSR2_ENABLE | EDP_SU_TRACK_ENABLE)); + + } else { + psr_ctl = EDP_PSR_STATUS_CTL; + psr_status_mask = EDP_PSR_STATUS_STATE_MASK; + + I915_WRITE(psr_ctl, + I915_READ(psr_ctl) & ~EDP_PSR_ENABLE); + } /* Wait till PSR is idle */ if (intel_wait_for_register(dev_priv, - EDP_PSR_STATUS_CTL, - EDP_PSR_STATUS_STATE_MASK, - 0, + psr_ctl, psr_status_mask, 0, 2000)) DRM_ERROR("Timed out waiting for PSR Idle State\n"); dev_priv->psr.active = false; } else { - WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); + if (dev_priv->psr.psr2_support) + WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); + else + WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); } } @@ -608,13 +706,24 @@ static void intel_psr_work(struct work_struct *work) * and be ready for re-enable. */ if (HAS_DDI(dev_priv)) { - if (intel_wait_for_register(dev_priv, - EDP_PSR_STATUS_CTL, - EDP_PSR_STATUS_STATE_MASK, - 0, - 50)) { - DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n"); - return; + if (dev_priv->psr.psr2_support) { + if (intel_wait_for_register(dev_priv, + EDP_PSR2_STATUS_CTL, + EDP_PSR2_STATUS_STATE_MASK, + 0, + 50)) { + DRM_ERROR("Timed out waiting for PSR2 Idle for re-enable\n"); + return; + } + } else { + if (intel_wait_for_register(dev_priv, + EDP_PSR_STATUS_CTL, + EDP_PSR_STATUS_STATE_MASK, + 0, + 50)) { + DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n"); + return; + } } } else { if (intel_wait_for_register(dev_priv, @@ -656,11 +765,19 @@ static void intel_psr_exit(struct drm_i915_private *dev_priv) return; if (HAS_DDI(dev_priv)) { - val = I915_READ(EDP_PSR_CTL); - - WARN_ON(!(val & EDP_PSR_ENABLE)); - - I915_WRITE(EDP_PSR_CTL, val & ~EDP_PSR_ENABLE); + if (dev_priv->psr.aux_frame_sync) + drm_dp_dpcd_writeb(&intel_dp->aux, + DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, + 0); + if (dev_priv->psr.psr2_support) { + val = I915_READ(EDP_PSR2_CTL); + WARN_ON(!(val & EDP_PSR2_ENABLE)); + I915_WRITE(EDP_PSR2_CTL, val & ~EDP_PSR2_ENABLE); + } else { + val = I915_READ(EDP_PSR_CTL); + WARN_ON(!(val & EDP_PSR_ENABLE)); + I915_WRITE(EDP_PSR_CTL, val & ~EDP_PSR_ENABLE); + } } else { val = I915_READ(VLV_PSRCTL(pipe)); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 0971ac396b60..69035e4f9b3b 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1095,14 +1095,6 @@ static int kbl_init_workarounds(struct intel_engine_cs *engine) WA_SET_BIT_MASKED(HDC_CHICKEN0, HDC_FENCE_DEST_SLM_DISABLE); - /* GEN8_L3SQCREG4 has a dependency with WA batch so any new changes - * involving this register should also be added to WA batch as required. - */ - if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_E0)) - /* WaDisableLSQCROPERFforOCL:kbl */ - I915_WRITE(GEN8_L3SQCREG4, I915_READ(GEN8_L3SQCREG4) | - GEN8_LQSC_RO_PERF_DIS); - /* WaToEnableHwFixForPushConstHWBug:kbl */ if (IS_KBL_REVID(dev_priv, KBL_REVID_C0, REVID_FOREVER)) WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, @@ -1736,7 +1728,7 @@ static int init_status_page(struct intel_engine_cs *engine) void *vaddr; int ret; - obj = i915_gem_object_create_internal(engine->i915, 4096); + obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); if (IS_ERR(obj)) { DRM_ERROR("Failed to allocate status page\n"); return PTR_ERR(obj); @@ -1746,7 +1738,7 @@ static int init_status_page(struct intel_engine_cs *engine) if (ret) goto err; - vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto err; @@ -1777,7 +1769,7 @@ static int init_status_page(struct intel_engine_cs *engine) engine->status_page.vma = vma; engine->status_page.ggtt_offset = i915_ggtt_offset(vma); - engine->status_page.page_addr = memset(vaddr, 0, 4096); + engine->status_page.page_addr = memset(vaddr, 0, PAGE_SIZE); DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n", engine->name, i915_ggtt_offset(vma)); @@ -1880,7 +1872,7 @@ intel_ring_create_vma(struct drm_i915_private *dev_priv, int size) /* mark ring buffers as read-only from GPU side by default */ obj->gt_ro = 1; - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) goto err; @@ -2049,7 +2041,7 @@ static int intel_init_ring_buffer(struct intel_engine_cs *engine) } /* Ring wraparound at offset 0 sometimes hangs. No idea why. */ - ret = intel_ring_pin(ring, 4096); + ret = intel_ring_pin(ring, I915_GTT_PAGE_SIZE); if (ret) { intel_ring_free(ring); goto error; @@ -2466,11 +2458,11 @@ static void intel_ring_init_semaphores(struct drm_i915_private *dev_priv, if (INTEL_GEN(dev_priv) >= 8 && !dev_priv->semaphore) { struct i915_vma *vma; - obj = i915_gem_object_create(dev_priv, 4096); + obj = i915_gem_object_create(dev_priv, PAGE_SIZE); if (IS_ERR(obj)) goto err; - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) goto err_obj; @@ -2683,7 +2675,7 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine) return ret; if (INTEL_GEN(dev_priv) >= 6) { - ret = intel_engine_create_scratch(engine, 4096); + ret = intel_engine_create_scratch(engine, PAGE_SIZE); if (ret) return ret; } else if (HAS_BROKEN_CS_TLB(dev_priv)) { diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c index 7031bc733d97..9ef54688872a 100644 --- a/drivers/gpu/drm/i915/intel_sprite.c +++ b/drivers/gpu/drm/i915/intel_sprite.c @@ -273,7 +273,7 @@ skl_update_plane(struct drm_plane *drm_plane, I915_WRITE(PLANE_CTL(pipe, plane_id), plane_ctl); I915_WRITE(PLANE_SURF(pipe, plane_id), - intel_fb_gtt_offset(fb, rotation) + surf_addr); + intel_plane_ggtt_offset(plane_state) + surf_addr); POSTING_READ(PLANE_SURF(pipe, plane_id)); } @@ -458,7 +458,7 @@ vlv_update_plane(struct drm_plane *dplane, I915_WRITE(SPSIZE(pipe, plane_id), (crtc_h << 16) | crtc_w); I915_WRITE(SPCNTR(pipe, plane_id), sprctl); I915_WRITE(SPSURF(pipe, plane_id), - intel_fb_gtt_offset(fb, rotation) + sprsurf_offset); + intel_plane_ggtt_offset(plane_state) + sprsurf_offset); POSTING_READ(SPSURF(pipe, plane_id)); } @@ -594,7 +594,7 @@ ivb_update_plane(struct drm_plane *plane, I915_WRITE(SPRSCALE(pipe), sprscale); I915_WRITE(SPRCTL(pipe), sprctl); I915_WRITE(SPRSURF(pipe), - intel_fb_gtt_offset(fb, rotation) + sprsurf_offset); + intel_plane_ggtt_offset(plane_state) + sprsurf_offset); POSTING_READ(SPRSURF(pipe)); } @@ -721,7 +721,7 @@ ilk_update_plane(struct drm_plane *plane, I915_WRITE(DVSSCALE(pipe), dvsscale); I915_WRITE(DVSCNTR(pipe), dvscntr); I915_WRITE(DVSSURF(pipe), - intel_fb_gtt_offset(fb, rotation) + dvssurf_offset); + intel_plane_ggtt_offset(plane_state) + dvssurf_offset); POSTING_READ(DVSSURF(pipe)); } diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index c6be35220955..c46bc8594f22 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -114,29 +114,3 @@ int intel_guc_sample_forcewake(struct intel_guc *guc) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -int intel_guc_log_flush_complete(struct intel_guc *guc) -{ - u32 action[] = { INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE }; - - return intel_guc_send(guc, action, ARRAY_SIZE(action)); -} - -int intel_guc_log_flush(struct intel_guc *guc) -{ - u32 action[] = { - INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH, - 0 - }; - - return intel_guc_send(guc, action, ARRAY_SIZE(action)); -} - -int intel_guc_log_control(struct intel_guc *guc, u32 control_val) -{ - u32 action[] = { - INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING, - control_val - }; - - return intel_guc_send(guc, action, ARRAY_SIZE(action)); -} diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index 9490a8e049c3..d74f4d3ad8dc 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -93,29 +93,35 @@ struct i915_guc_client { uint64_t submissions[I915_NUM_ENGINES]; }; -enum intel_guc_fw_status { - GUC_FIRMWARE_FAIL = -1, - GUC_FIRMWARE_NONE = 0, - GUC_FIRMWARE_PENDING, - GUC_FIRMWARE_SUCCESS +enum intel_uc_fw_status { + INTEL_UC_FIRMWARE_FAIL = -1, + INTEL_UC_FIRMWARE_NONE = 0, + INTEL_UC_FIRMWARE_PENDING, + INTEL_UC_FIRMWARE_SUCCESS +}; + +enum intel_uc_fw_type { + INTEL_UC_FW_TYPE_GUC, + INTEL_UC_FW_TYPE_HUC }; /* * This structure encapsulates all the data needed during the process * of fetching, caching, and loading the firmware image into the GuC. */ -struct intel_guc_fw { - const char * guc_fw_path; - size_t guc_fw_size; - struct drm_i915_gem_object * guc_fw_obj; - enum intel_guc_fw_status guc_fw_fetch_status; - enum intel_guc_fw_status guc_fw_load_status; - - uint16_t guc_fw_major_wanted; - uint16_t guc_fw_minor_wanted; - uint16_t guc_fw_major_found; - uint16_t guc_fw_minor_found; - +struct intel_uc_fw { + const char *path; + size_t size; + struct drm_i915_gem_object *obj; + enum intel_uc_fw_status fetch_status; + enum intel_uc_fw_status load_status; + + uint16_t major_ver_wanted; + uint16_t minor_ver_wanted; + uint16_t major_ver_found; + uint16_t minor_ver_found; + + enum intel_uc_fw_type fw; uint32_t header_size; uint32_t header_offset; uint32_t rsa_size; @@ -141,7 +147,7 @@ struct intel_guc_log { }; struct intel_guc { - struct intel_guc_fw guc_fw; + struct intel_uc_fw fw; struct intel_guc_log log; /* intel_guc_recv interrupt related state */ @@ -170,21 +176,28 @@ struct intel_guc { struct mutex send_mutex; }; +struct intel_huc { + /* Generic uC firmware management */ + struct intel_uc_fw fw; + + /* HuC-specific additions */ +}; + /* intel_uc.c */ void intel_uc_init_early(struct drm_i915_private *dev_priv); int intel_guc_send(struct intel_guc *guc, const u32 *action, u32 len); int intel_guc_sample_forcewake(struct intel_guc *guc); -int intel_guc_log_flush_complete(struct intel_guc *guc); -int intel_guc_log_flush(struct intel_guc *guc); -int intel_guc_log_control(struct intel_guc *guc, u32 control_val); /* intel_guc_loader.c */ extern void intel_guc_init(struct drm_i915_private *dev_priv); extern int intel_guc_setup(struct drm_i915_private *dev_priv); extern void intel_guc_fini(struct drm_i915_private *dev_priv); -extern const char *intel_guc_fw_status_repr(enum intel_guc_fw_status status); +extern const char *intel_uc_fw_status_repr(enum intel_uc_fw_status status); extern int intel_guc_suspend(struct drm_i915_private *dev_priv); extern int intel_guc_resume(struct drm_i915_private *dev_priv); +void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, + struct intel_uc_fw *uc_fw); +u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv); /* i915_guc_submission.c */ int i915_guc_submission_init(struct drm_i915_private *dev_priv); @@ -193,10 +206,12 @@ int i915_guc_wq_reserve(struct drm_i915_gem_request *rq); void i915_guc_wq_unreserve(struct drm_i915_gem_request *request); void i915_guc_submission_disable(struct drm_i915_private *dev_priv); void i915_guc_submission_fini(struct drm_i915_private *dev_priv); -void i915_guc_capture_logs(struct drm_i915_private *dev_priv); -void i915_guc_flush_logs(struct drm_i915_private *dev_priv); -void i915_guc_register(struct drm_i915_private *dev_priv); -void i915_guc_unregister(struct drm_i915_private *dev_priv); +struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size); + +/* intel_guc_log.c */ +void intel_guc_log_create(struct intel_guc *guc); +void i915_guc_log_register(struct drm_i915_private *dev_priv); +void i915_guc_log_unregister(struct drm_i915_private *dev_priv); int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val); static inline u32 guc_ggtt_offset(struct i915_vma *vma) @@ -207,4 +222,10 @@ static inline u32 guc_ggtt_offset(struct i915_vma *vma) return offset; } +/* intel_huc.c */ +void intel_huc_init(struct drm_i915_private *dev_priv); +void intel_huc_fini(struct drm_i915_private *dev_priv); +int intel_huc_load(struct drm_i915_private *dev_priv); +void intel_guc_auth_huc(struct drm_i915_private *dev_priv); + #endif diff --git a/drivers/gpu/drm/meson/meson_plane.c b/drivers/gpu/drm/meson/meson_plane.c index 642b2fab42ff..a32d3b6e2e12 100644 --- a/drivers/gpu/drm/meson/meson_plane.c +++ b/drivers/gpu/drm/meson/meson_plane.c @@ -51,6 +51,9 @@ static int meson_plane_atomic_check(struct drm_plane *plane, struct drm_crtc_state *crtc_state; struct drm_rect clip = { 0, }; + if (!state->crtc) + return 0; + crtc_state = drm_atomic_get_crtc_state(state->state, state->crtc); if (IS_ERR(crtc_state)) return PTR_ERR(crtc_state); diff --git a/drivers/gpu/drm/meson/meson_venc.c b/drivers/gpu/drm/meson/meson_venc.c index d836b2274531..f7c870172220 100644 --- a/drivers/gpu/drm/meson/meson_venc.c +++ b/drivers/gpu/drm/meson/meson_venc.c @@ -38,6 +38,11 @@ * - TV Panel encoding via ENCT */ +/* HHI Registers */ +#define HHI_VDAC_CNTL0 0x2F4 /* 0xbd offset in data sheet */ +#define HHI_VDAC_CNTL1 0x2F8 /* 0xbe offset in data sheet */ +#define HHI_HDMI_PHY_CNTL0 0x3a0 /* 0xe8 offset in data sheet */ + struct meson_cvbs_enci_mode meson_cvbs_enci_pal = { .mode_tag = MESON_VENC_MODE_CVBS_PAL, .hso_begin = 3, @@ -242,6 +247,20 @@ void meson_venc_disable_vsync(struct meson_drm *priv) void meson_venc_init(struct meson_drm *priv) { + /* Disable CVBS VDAC */ + regmap_write(priv->hhi, HHI_VDAC_CNTL0, 0); + regmap_write(priv->hhi, HHI_VDAC_CNTL1, 8); + + /* Power Down Dacs */ + writel_relaxed(0xff, priv->io_base + _REG(VENC_VDAC_SETTING)); + + /* Disable HDMI PHY */ + regmap_write(priv->hhi, HHI_HDMI_PHY_CNTL0, 0); + + /* Disable HDMI */ + writel_bits_relaxed(0x3, 0, + priv->io_base + _REG(VPU_HDMI_SETTING)); + /* Disable all encoders */ writel_relaxed(0, priv->io_base + _REG(ENCI_VIDEO_EN)); writel_relaxed(0, priv->io_base + _REG(ENCP_VIDEO_EN)); diff --git a/drivers/gpu/drm/meson/meson_venc_cvbs.c b/drivers/gpu/drm/meson/meson_venc_cvbs.c index c809c085fd78..a2bcc70a03ef 100644 --- a/drivers/gpu/drm/meson/meson_venc_cvbs.c +++ b/drivers/gpu/drm/meson/meson_venc_cvbs.c @@ -167,7 +167,7 @@ static void meson_venc_cvbs_encoder_disable(struct drm_encoder *encoder) /* Disable CVBS VDAC */ regmap_write(priv->hhi, HHI_VDAC_CNTL0, 0); - regmap_write(priv->hhi, HHI_VDAC_CNTL1, 0); + regmap_write(priv->hhi, HHI_VDAC_CNTL1, 8); } static void meson_venc_cvbs_encoder_enable(struct drm_encoder *encoder) diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.c b/drivers/gpu/drm/mgag200/mgag200_drv.c index b0b874264f9d..9ac007880328 100644 --- a/drivers/gpu/drm/mgag200/mgag200_drv.c +++ b/drivers/gpu/drm/mgag200/mgag200_drv.c @@ -36,6 +36,7 @@ static const struct pci_device_id pciidlist[] = { { PCI_VENDOR_ID_MATROX, 0x533, PCI_ANY_ID, PCI_ANY_ID, 0, 0, G200_EH }, { PCI_VENDOR_ID_MATROX, 0x534, PCI_ANY_ID, PCI_ANY_ID, 0, 0, G200_ER }, { PCI_VENDOR_ID_MATROX, 0x536, PCI_ANY_ID, PCI_ANY_ID, 0, 0, G200_EW3 }, + { PCI_VENDOR_ID_MATROX, 0x538, PCI_ANY_ID, PCI_ANY_ID, 0, 0, G200_EH3 }, {0,} }; diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h b/drivers/gpu/drm/mgag200/mgag200_drv.h index 0d6e998d63e6..c88b6ec88dd2 100644 --- a/drivers/gpu/drm/mgag200/mgag200_drv.h +++ b/drivers/gpu/drm/mgag200/mgag200_drv.h @@ -180,6 +180,7 @@ enum mga_type { G200_WB, G200_EV, G200_EH, + G200_EH3, G200_ER, G200_EW3, }; diff --git a/drivers/gpu/drm/mgag200/mgag200_i2c.c b/drivers/gpu/drm/mgag200/mgag200_i2c.c index 10535e3b75f2..77d1c4771786 100644 --- a/drivers/gpu/drm/mgag200/mgag200_i2c.c +++ b/drivers/gpu/drm/mgag200/mgag200_i2c.c @@ -106,6 +106,7 @@ struct mga_i2c_chan *mgag200_i2c_create(struct drm_device *dev) clock = 2; break; case G200_EH: + case G200_EH3: case G200_ER: data = 2; clock = 1; diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c index 067dfbc91b1c..3938120e5051 100644 --- a/drivers/gpu/drm/mgag200/mgag200_mode.c +++ b/drivers/gpu/drm/mgag200/mgag200_mode.c @@ -497,34 +497,70 @@ static int mga_g200eh_set_plls(struct mga_device *mdev, long clock) bool pll_locked = false; m = n = p = 0; - vcomax = 800000; - vcomin = 400000; - pllreffreq = 33333; - delta = 0xffffffff; + if (mdev->type == G200_EH3) { + vcomax = 3000000; + vcomin = 1500000; + pllreffreq = 25000; - for (testp = 16; testp > 0; testp >>= 1) { - if (clock * testp > vcomax) - continue; - if (clock * testp < vcomin) - continue; + delta = 0xffffffff; - for (testm = 1; testm < 33; testm++) { - for (testn = 17; testn < 257; testn++) { - computed = (pllreffreq * testn) / - (testm * testp); + testp = 0; + + for (testm = 150; testm >= 6; testm--) { + if (clock * testm > vcomax) + continue; + if (clock * testm < vcomin) + continue; + for (testn = 120; testn >= 60; testn--) { + computed = (pllreffreq * testn) / testm; if (computed > clock) tmpdelta = computed - clock; else tmpdelta = clock - computed; if (tmpdelta < delta) { delta = tmpdelta; - n = testn - 1; - m = (testm - 1); - p = testp - 1; + n = testn; + m = testm; + p = testp; + } + if (delta == 0) + break; + } + if (delta == 0) + break; + } + } else { + + vcomax = 800000; + vcomin = 400000; + pllreffreq = 33333; + + delta = 0xffffffff; + + for (testp = 16; testp > 0; testp >>= 1) { + if (clock * testp > vcomax) + continue; + if (clock * testp < vcomin) + continue; + + for (testm = 1; testm < 33; testm++) { + for (testn = 17; testn < 257; testn++) { + computed = (pllreffreq * testn) / + (testm * testp); + if (computed > clock) + tmpdelta = computed - clock; + else + tmpdelta = clock - computed; + if (tmpdelta < delta) { + delta = tmpdelta; + n = testn - 1; + m = (testm - 1); + p = testp - 1; + } + if ((clock * testp) >= 600000) + p |= 0x80; } - if ((clock * testp) >= 600000) - p |= 0x80; } } } @@ -674,6 +710,7 @@ static int mga_crtc_set_plls(struct mga_device *mdev, long clock) return mga_g200ev_set_plls(mdev, clock); break; case G200_EH: + case G200_EH3: return mga_g200eh_set_plls(mdev, clock); break; case G200_ER: @@ -933,6 +970,7 @@ static int mga_crtc_mode_set(struct drm_crtc *crtc, option2 = 0x0000b000; break; case G200_EH: + case G200_EH3: dacvalue[MGA1064_MISC_CTL] = MGA1064_MISC_CTL_VGA8 | MGA1064_MISC_CTL_DAC_RAM_CS; option = 0x00000120; @@ -979,7 +1017,8 @@ static int mga_crtc_mode_set(struct drm_crtc *crtc, if ((mdev->type == G200_EV || mdev->type == G200_WB || mdev->type == G200_EH || - mdev->type == G200_EW3) && + mdev->type == G200_EW3 || + mdev->type == G200_EH3) && (i >= 0x44) && (i <= 0x4e)) continue; diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index a18126150e11..686a580c711a 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -213,7 +213,14 @@ void adreno_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit, void adreno_flush(struct msm_gpu *gpu) { struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); - uint32_t wptr = get_wptr(gpu->rb); + uint32_t wptr; + + /* + * Mask wptr value that we calculate to fit in the HW range. This is + * to account for the possibility that the last command fit exactly into + * the ringbuffer and rb->next hasn't wrapped to zero yet + */ + wptr = get_wptr(gpu->rb) & ((gpu->rb->size / 4) - 1); /* ensure writes to ringbuffer have hit system memory: */ mb(); @@ -338,7 +345,6 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, { struct adreno_platform_config *config = pdev->dev.platform_data; struct msm_gpu *gpu = &adreno_gpu->base; - struct msm_mmu *mmu; int ret; adreno_gpu->funcs = funcs; @@ -378,8 +384,8 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, return ret; } - mmu = gpu->aspace->mmu; - if (mmu) { + if (gpu->aspace && gpu->aspace->mmu) { + struct msm_mmu *mmu = gpu->aspace->mmu; ret = mmu->funcs->attach(mmu, iommu_ports, ARRAY_SIZE(iommu_ports)); if (ret) diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c index 5f6cd8745dbc..c396d459a9d0 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c @@ -119,13 +119,7 @@ static void mdp5_prepare_commit(struct msm_kms *kms, struct drm_atomic_state *st static void mdp5_complete_commit(struct msm_kms *kms, struct drm_atomic_state *state) { - int i; struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); - struct drm_plane *plane; - struct drm_plane_state *plane_state; - - for_each_plane_in_state(state, plane, plane_state, i) - mdp5_plane_complete_commit(plane, plane_state); if (mdp5_kms->smp) mdp5_smp_complete_commit(mdp5_kms->smp, &mdp5_kms->state->smp); diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h index 17b0cc101171..cdfc63d90c7b 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h @@ -104,8 +104,6 @@ struct mdp5_plane_state { /* assigned by crtc blender */ enum mdp_mixer_stage_id stage; - - bool pending : 1; }; #define to_mdp5_plane_state(x) \ container_of(x, struct mdp5_plane_state, base) @@ -232,8 +230,6 @@ int mdp5_irq_domain_init(struct mdp5_kms *mdp5_kms); void mdp5_irq_domain_fini(struct mdp5_kms *mdp5_kms); uint32_t mdp5_plane_get_flush(struct drm_plane *plane); -void mdp5_plane_complete_commit(struct drm_plane *plane, - struct drm_plane_state *state); enum mdp5_pipe mdp5_plane_pipe(struct drm_plane *plane); struct drm_plane *mdp5_plane_init(struct drm_device *dev, bool primary); diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c index 75247ea4335b..b9fb111d3428 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c @@ -179,7 +179,6 @@ mdp5_plane_atomic_print_state(struct drm_printer *p, drm_printf(p, "\tzpos=%u\n", pstate->zpos); drm_printf(p, "\talpha=%u\n", pstate->alpha); drm_printf(p, "\tstage=%s\n", stage2name(pstate->stage)); - drm_printf(p, "\tpending=%u\n", pstate->pending); } static void mdp5_plane_reset(struct drm_plane *plane) @@ -220,8 +219,6 @@ mdp5_plane_duplicate_state(struct drm_plane *plane) if (mdp5_state && mdp5_state->base.fb) drm_framebuffer_reference(mdp5_state->base.fb); - mdp5_state->pending = false; - return &mdp5_state->base; } @@ -288,13 +285,6 @@ static int mdp5_plane_atomic_check(struct drm_plane *plane, DBG("%s: check (%d -> %d)", plane->name, plane_enabled(old_state), plane_enabled(state)); - /* We don't allow faster-than-vblank updates.. if we did add this - * some day, we would need to disallow in cases where hwpipe - * changes - */ - if (WARN_ON(to_mdp5_plane_state(old_state)->pending)) - return -EBUSY; - max_width = config->hw->lm.max_width << 16; max_height = config->hw->lm.max_height << 16; @@ -370,12 +360,9 @@ static void mdp5_plane_atomic_update(struct drm_plane *plane, struct drm_plane_state *old_state) { struct drm_plane_state *state = plane->state; - struct mdp5_plane_state *mdp5_state = to_mdp5_plane_state(state); DBG("%s: update", plane->name); - mdp5_state->pending = true; - if (plane_enabled(state)) { int ret; @@ -851,15 +838,6 @@ uint32_t mdp5_plane_get_flush(struct drm_plane *plane) return pstate->hwpipe->flush_mask; } -/* called after vsync in thread context */ -void mdp5_plane_complete_commit(struct drm_plane *plane, - struct drm_plane_state *state) -{ - struct mdp5_plane_state *pstate = to_mdp5_plane_state(plane->state); - - pstate->pending = false; -} - /* initialize plane */ struct drm_plane *mdp5_plane_init(struct drm_device *dev, bool primary) { diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c index d8bc59c7e261..8098677a3916 100644 --- a/drivers/gpu/drm/msm/msm_gem.c +++ b/drivers/gpu/drm/msm/msm_gem.c @@ -294,6 +294,8 @@ put_iova(struct drm_gem_object *obj) WARN_ON(!mutex_is_locked(&dev->struct_mutex)); for (id = 0; id < ARRAY_SIZE(msm_obj->domain); id++) { + if (!priv->aspace[id]) + continue; msm_gem_unmap_vma(priv->aspace[id], &msm_obj->domain[id], msm_obj->sgt); } diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c index 166e84e4f0d4..489676568a10 100644 --- a/drivers/gpu/drm/msm/msm_gem_submit.c +++ b/drivers/gpu/drm/msm/msm_gem_submit.c @@ -106,7 +106,8 @@ static int submit_lookup_objects(struct msm_gem_submit *submit, pagefault_disable(); } - if (submit_bo.flags & ~MSM_SUBMIT_BO_FLAGS) { + if ((submit_bo.flags & ~MSM_SUBMIT_BO_FLAGS) || + !(submit_bo.flags & MSM_SUBMIT_BO_FLAGS)) { DRM_ERROR("invalid flags: %x\n", submit_bo.flags); ret = -EINVAL; goto out_unlock; @@ -290,7 +291,7 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob { uint32_t i, last_offset = 0; uint32_t *ptr; - int ret; + int ret = 0; if (offset % 4) { DRM_ERROR("non-aligned cmdstream buffer: %u\n", offset); @@ -318,12 +319,13 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob ret = copy_from_user(&submit_reloc, userptr, sizeof(submit_reloc)); if (ret) - return -EFAULT; + goto out; if (submit_reloc.submit_offset % 4) { DRM_ERROR("non-aligned reloc offset: %u\n", submit_reloc.submit_offset); - return -EINVAL; + ret = -EINVAL; + goto out; } /* offset in dwords: */ @@ -332,12 +334,13 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob if ((off >= (obj->base.size / 4)) || (off < last_offset)) { DRM_ERROR("invalid offset %u at reloc %u\n", off, i); - return -EINVAL; + ret = -EINVAL; + goto out; } ret = submit_bo(submit, submit_reloc.reloc_idx, NULL, &iova, &valid); if (ret) - return ret; + goto out; if (valid) continue; @@ -354,9 +357,10 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob last_offset = off; } +out: msm_gem_put_vaddr_locked(&obj->base); - return 0; + return ret; } static void submit_cleanup(struct msm_gem_submit *submit) diff --git a/drivers/gpu/drm/msm/msm_ringbuffer.c b/drivers/gpu/drm/msm/msm_ringbuffer.c index f326cf6a32e6..67b34e069abf 100644 --- a/drivers/gpu/drm/msm/msm_ringbuffer.c +++ b/drivers/gpu/drm/msm/msm_ringbuffer.c @@ -23,7 +23,8 @@ struct msm_ringbuffer *msm_ringbuffer_new(struct msm_gpu *gpu, int size) struct msm_ringbuffer *ring; int ret; - size = ALIGN(size, 4); /* size should be dword aligned */ + if (WARN_ON(!is_power_of_2(size))) + return ERR_PTR(-EINVAL); ring = kzalloc(sizeof(*ring), GFP_KERNEL); if (!ring) { diff --git a/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c b/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c index dc026a843712..a2bb855a2851 100644 --- a/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c +++ b/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c @@ -1253,7 +1253,7 @@ static int dsicm_probe(struct platform_device *pdev) dsicm_hw_reset(ddata); if (ddata->use_dsi_backlight) { - memset(&props, 0, sizeof(struct backlight_properties)); + memset(&props, 0, sizeof(props)); props.max_brightness = 255; props.type = BACKLIGHT_RAW; diff --git a/drivers/gpu/drm/omapdrm/displays/panel-sony-acx565akm.c b/drivers/gpu/drm/omapdrm/displays/panel-sony-acx565akm.c index 746cb8d9cba1..5ab39e0060f2 100644 --- a/drivers/gpu/drm/omapdrm/displays/panel-sony-acx565akm.c +++ b/drivers/gpu/drm/omapdrm/displays/panel-sony-acx565akm.c @@ -909,6 +909,7 @@ static struct spi_driver acx565akm_driver = { module_spi_driver(acx565akm_driver); +MODULE_ALIAS("spi:sony,acx565akm"); MODULE_AUTHOR("Nokia Corporation"); MODULE_DESCRIPTION("acx565akm LCD Driver"); MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/omapdrm/dss/dispc.c b/drivers/gpu/drm/omapdrm/dss/dispc.c index c839f6456db2..5554b72cf56a 100644 --- a/drivers/gpu/drm/omapdrm/dss/dispc.c +++ b/drivers/gpu/drm/omapdrm/dss/dispc.c @@ -620,6 +620,19 @@ u32 dispc_wb_get_framedone_irq(void) return DISPC_IRQ_FRAMEDONEWB; } +void dispc_mgr_enable(enum omap_channel channel, bool enable) +{ + mgr_fld_write(channel, DISPC_MGR_FLD_ENABLE, enable); + /* flush posted write */ + mgr_fld_read(channel, DISPC_MGR_FLD_ENABLE); +} +EXPORT_SYMBOL(dispc_mgr_enable); + +static bool dispc_mgr_is_enabled(enum omap_channel channel) +{ + return !!mgr_fld_read(channel, DISPC_MGR_FLD_ENABLE); +} + bool dispc_mgr_go_busy(enum omap_channel channel) { return mgr_fld_read(channel, DISPC_MGR_FLD_GO) == 1; @@ -2901,20 +2914,6 @@ enum omap_dss_output_id dispc_mgr_get_supported_outputs(enum omap_channel channe } EXPORT_SYMBOL(dispc_mgr_get_supported_outputs); -void dispc_mgr_enable(enum omap_channel channel, bool enable) -{ - mgr_fld_write(channel, DISPC_MGR_FLD_ENABLE, enable); - /* flush posted write */ - mgr_fld_read(channel, DISPC_MGR_FLD_ENABLE); -} -EXPORT_SYMBOL(dispc_mgr_enable); - -bool dispc_mgr_is_enabled(enum omap_channel channel) -{ - return !!mgr_fld_read(channel, DISPC_MGR_FLD_ENABLE); -} -EXPORT_SYMBOL(dispc_mgr_is_enabled); - void dispc_wb_enable(bool enable) { dispc_ovl_enable(OMAP_DSS_WB, enable); diff --git a/drivers/gpu/drm/omapdrm/dss/dsi.c b/drivers/gpu/drm/omapdrm/dss/dsi.c index f060bda31235..f74615d005a8 100644 --- a/drivers/gpu/drm/omapdrm/dss/dsi.c +++ b/drivers/gpu/drm/omapdrm/dss/dsi.c @@ -4336,7 +4336,7 @@ static void print_dsi_vm(const char *str, wc = DIV_ROUND_UP(t->hact * t->bitspp, 8); pps = DIV_ROUND_UP(wc + 6, t->ndl); /* pixel packet size */ - bl = t->hss + t->hsa + t->hse + t->hbp + t->hfront_porch; + bl = t->hss + t->hsa + t->hse + t->hbp + t->hfp; tot = bl + pps; #define TO_DSI_T(x) ((u32)div64_u64((u64)x * 1000000000llu, byteclk)) @@ -4345,14 +4345,14 @@ static void print_dsi_vm(const char *str, "%u/%u/%u/%u/%u/%u = %u + %u = %u\n", str, byteclk, - t->hss, t->hsa, t->hse, t->hbp, pps, t->hfront_porch, + t->hss, t->hsa, t->hse, t->hbp, pps, t->hfp, bl, pps, tot, TO_DSI_T(t->hss), TO_DSI_T(t->hsa), TO_DSI_T(t->hse), TO_DSI_T(t->hbp), TO_DSI_T(pps), - TO_DSI_T(t->hfront_porch), + TO_DSI_T(t->hfp), TO_DSI_T(bl), TO_DSI_T(pps), @@ -4367,7 +4367,7 @@ static void print_dispc_vm(const char *str, const struct videomode *vm) int hact, bl, tot; hact = vm->hactive; - bl = vm->hsync_len + vm->hbp + vm->hfront_porch; + bl = vm->hsync_len + vm->hback_porch + vm->hfront_porch; tot = hact + bl; #define TO_DISPC_T(x) ((u32)div64_u64((u64)x * 1000000000llu, pck)) @@ -4376,10 +4376,10 @@ static void print_dispc_vm(const char *str, const struct videomode *vm) "%u/%u/%u/%u = %u + %u = %u\n", str, pck, - vm->hsync_len, vm->hbp, hact, vm->hfront_porch, + vm->hsync_len, vm->hback_porch, hact, vm->hfront_porch, bl, hact, tot, TO_DISPC_T(vm->hsync_len), - TO_DISPC_T(vm->hbp), + TO_DISPC_T(vm->hback_porch), TO_DISPC_T(hact), TO_DISPC_T(vm->hfront_porch), TO_DISPC_T(bl), @@ -4401,12 +4401,12 @@ static void print_dsi_dispc_vm(const char *str, dsi_tput = (u64)byteclk * t->ndl * 8; pck = (u32)div64_u64(dsi_tput, t->bitspp); dsi_hact = DIV_ROUND_UP(DIV_ROUND_UP(t->hact * t->bitspp, 8) + 6, t->ndl); - dsi_htot = t->hss + t->hsa + t->hse + t->hbp + dsi_hact + t->hfront_porch; + dsi_htot = t->hss + t->hsa + t->hse + t->hbp + dsi_hact + t->hfp; vm.pixelclock = pck; vm.hsync_len = div64_u64((u64)(t->hsa + t->hse) * pck, byteclk); - vm.hbp = div64_u64((u64)t->hbp * pck, byteclk); - vm.hfront_porch = div64_u64((u64)t->hfront_porch * pck, byteclk); + vm.hback_porch = div64_u64((u64)t->hbp * pck, byteclk); + vm.hfront_porch = div64_u64((u64)t->hfp * pck, byteclk); vm.hactive = t->hact; print_dispc_vm(str, &vm); diff --git a/drivers/gpu/drm/omapdrm/dss/omapdss-boot-init.c b/drivers/gpu/drm/omapdrm/dss/omapdss-boot-init.c index 136d30484d02..bf626acae271 100644 --- a/drivers/gpu/drm/omapdrm/dss/omapdss-boot-init.c +++ b/drivers/gpu/drm/omapdrm/dss/omapdss-boot-init.c @@ -119,8 +119,7 @@ static void __init omapdss_omapify_node(struct device_node *node) static void __init omapdss_add_to_list(struct device_node *node, bool root) { - struct dss_conv_node *n = kmalloc(sizeof(struct dss_conv_node), - GFP_KERNEL); + struct dss_conv_node *n = kmalloc(sizeof(*n), GFP_KERNEL); if (n) { n->node = node; n->root = root; diff --git a/drivers/gpu/drm/omapdrm/dss/omapdss.h b/drivers/gpu/drm/omapdrm/dss/omapdss.h index b420dde8c0fb..5b3b961127bd 100644 --- a/drivers/gpu/drm/omapdrm/dss/omapdss.h +++ b/drivers/gpu/drm/omapdrm/dss/omapdss.h @@ -856,7 +856,6 @@ int dispc_runtime_get(void); void dispc_runtime_put(void); void dispc_mgr_enable(enum omap_channel channel, bool enable); -bool dispc_mgr_is_enabled(enum omap_channel channel); u32 dispc_mgr_get_vsync_irq(enum omap_channel channel); u32 dispc_mgr_get_framedone_irq(enum omap_channel channel); u32 dispc_mgr_get_sync_lost_irq(enum omap_channel channel); diff --git a/drivers/gpu/drm/omapdrm/omap_connector.c b/drivers/gpu/drm/omapdrm/omap_connector.c index 2580e8673908..f90e2d22c5ec 100644 --- a/drivers/gpu/drm/omapdrm/omap_connector.c +++ b/drivers/gpu/drm/omapdrm/omap_connector.c @@ -162,7 +162,7 @@ static int omap_connector_mode_valid(struct drm_connector *connector, dssdrv->get_timings(dssdev, &t); - if (memcmp(&vm, &t, sizeof(struct videomode))) + if (memcmp(&vm, &t, sizeof(vm))) r = -EINVAL; else r = 0; @@ -217,7 +217,7 @@ struct drm_connector *omap_connector_init(struct drm_device *dev, omap_dss_get_device(dssdev); - omap_connector = kzalloc(sizeof(struct omap_connector), GFP_KERNEL); + omap_connector = kzalloc(sizeof(*omap_connector), GFP_KERNEL); if (!omap_connector) goto fail; @@ -240,8 +240,6 @@ struct drm_connector *omap_connector_init(struct drm_device *dev, connector->interlace_allowed = 1; connector->doublescan_allowed = 0; - drm_connector_register(connector); - return connector; fail: diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c index 8dea89030e66..dd47dc191e6b 100644 --- a/drivers/gpu/drm/omapdrm/omap_crtc.c +++ b/drivers/gpu/drm/omapdrm/omap_crtc.c @@ -36,26 +36,18 @@ struct omap_crtc { struct videomode vm; - struct omap_drm_irq vblank_irq; - struct omap_drm_irq error_irq; - bool ignore_digit_sync_lost; + bool enabled; bool pending; wait_queue_head_t pending_wait; + struct drm_pending_vblank_event *event; }; /* ----------------------------------------------------------------------------- * Helper Functions */ -uint32_t pipe2vbl(struct drm_crtc *crtc) -{ - struct omap_crtc *omap_crtc = to_omap_crtc(crtc); - - return dispc_mgr_get_vsync_irq(omap_crtc->channel); -} - struct videomode *omap_crtc_timings(struct drm_crtc *crtc) { struct omap_crtc *omap_crtc = to_omap_crtc(crtc); @@ -68,6 +60,19 @@ enum omap_channel omap_crtc_channel(struct drm_crtc *crtc) return omap_crtc->channel; } +static bool omap_crtc_is_pending(struct drm_crtc *crtc) +{ + struct omap_crtc *omap_crtc = to_omap_crtc(crtc); + unsigned long flags; + bool pending; + + spin_lock_irqsave(&crtc->dev->event_lock, flags); + pending = omap_crtc->pending; + spin_unlock_irqrestore(&crtc->dev->event_lock, flags); + + return pending; +} + int omap_crtc_wait_pending(struct drm_crtc *crtc) { struct omap_crtc *omap_crtc = to_omap_crtc(crtc); @@ -77,7 +82,7 @@ int omap_crtc_wait_pending(struct drm_crtc *crtc) * a single frame refresh even on slower displays. */ return wait_event_timeout(omap_crtc->pending_wait, - !omap_crtc->pending, + !omap_crtc_is_pending(crtc), msecs_to_jiffies(250)); } @@ -135,14 +140,15 @@ static void omap_crtc_set_enabled(struct drm_crtc *crtc, bool enable) u32 framedone_irq, vsync_irq; int ret; + if (WARN_ON(omap_crtc->enabled == enable)) + return; + if (omap_crtc_output[channel]->output_type == OMAP_DISPLAY_TYPE_HDMI) { dispc_mgr_enable(channel, enable); + omap_crtc->enabled = enable; return; } - if (dispc_mgr_is_enabled(channel) == enable) - return; - if (omap_crtc->channel == OMAP_DSS_CHANNEL_DIGIT) { /* * Digit output produces some sync lost interrupts during the @@ -173,6 +179,7 @@ static void omap_crtc_set_enabled(struct drm_crtc *crtc, bool enable) } dispc_mgr_enable(channel, enable); + omap_crtc->enabled = enable; ret = omap_irq_wait(dev, wait, msecs_to_jiffies(100)); if (ret) { @@ -259,26 +266,9 @@ static const struct dss_mgr_ops mgr_ops = { * Setup, Flush and Page Flip */ -static void omap_crtc_complete_page_flip(struct drm_crtc *crtc) +void omap_crtc_error_irq(struct drm_crtc *crtc, uint32_t irqstatus) { - struct drm_pending_vblank_event *event; - struct drm_device *dev = crtc->dev; - unsigned long flags; - - event = crtc->state->event; - - if (!event) - return; - - spin_lock_irqsave(&dev->event_lock, flags); - drm_crtc_send_vblank_event(crtc, event); - spin_unlock_irqrestore(&dev->event_lock, flags); -} - -static void omap_crtc_error_irq(struct omap_drm_irq *irq, uint32_t irqstatus) -{ - struct omap_crtc *omap_crtc = - container_of(irq, struct omap_crtc, error_irq); + struct omap_crtc *omap_crtc = to_omap_crtc(crtc); if (omap_crtc->ignore_digit_sync_lost) { irqstatus &= ~DISPC_IRQ_SYNC_LOST_DIGIT; @@ -289,29 +279,38 @@ static void omap_crtc_error_irq(struct omap_drm_irq *irq, uint32_t irqstatus) DRM_ERROR_RATELIMITED("%s: errors: %08x\n", omap_crtc->name, irqstatus); } -static void omap_crtc_vblank_irq(struct omap_drm_irq *irq, uint32_t irqstatus) +void omap_crtc_vblank_irq(struct drm_crtc *crtc) { - struct omap_crtc *omap_crtc = - container_of(irq, struct omap_crtc, vblank_irq); - struct drm_device *dev = omap_crtc->base.dev; + struct omap_crtc *omap_crtc = to_omap_crtc(crtc); + bool pending; - if (dispc_mgr_go_busy(omap_crtc->channel)) + spin_lock(&crtc->dev->event_lock); + /* + * If the dispc is busy we're racing the flush operation. Try again on + * the next vblank interrupt. + */ + if (dispc_mgr_go_busy(omap_crtc->channel)) { + spin_unlock(&crtc->dev->event_lock); return; + } - DBG("%s: apply done", omap_crtc->name); - - __omap_irq_unregister(dev, &omap_crtc->vblank_irq); + /* Send the vblank event if one has been requested. */ + if (omap_crtc->event) { + drm_crtc_send_vblank_event(crtc, omap_crtc->event); + omap_crtc->event = NULL; + } - rmb(); - WARN_ON(!omap_crtc->pending); + pending = omap_crtc->pending; omap_crtc->pending = false; - wmb(); + spin_unlock(&crtc->dev->event_lock); - /* wake up userspace */ - omap_crtc_complete_page_flip(&omap_crtc->base); + if (pending) + drm_crtc_vblank_put(crtc); - /* wake up omap_atomic_complete */ + /* Wake up omap_atomic_complete. */ wake_up(&omap_crtc->pending_wait); + + DBG("%s: apply done", omap_crtc->name); } /* ----------------------------------------------------------------------------- @@ -324,9 +323,6 @@ static void omap_crtc_destroy(struct drm_crtc *crtc) DBG("%s", omap_crtc->name); - WARN_ON(omap_crtc->vblank_irq.registered); - omap_irq_unregister(crtc->dev, &omap_crtc->error_irq); - drm_crtc_cleanup(crtc); kfree(omap_crtc); @@ -335,17 +331,18 @@ static void omap_crtc_destroy(struct drm_crtc *crtc) static void omap_crtc_enable(struct drm_crtc *crtc) { struct omap_crtc *omap_crtc = to_omap_crtc(crtc); + int ret; DBG("%s", omap_crtc->name); - rmb(); + spin_lock_irq(&crtc->dev->event_lock); + drm_crtc_vblank_on(crtc); + ret = drm_crtc_vblank_get(crtc); + WARN_ON(ret != 0); + WARN_ON(omap_crtc->pending); omap_crtc->pending = true; - wmb(); - - omap_irq_register(crtc->dev, &omap_crtc->vblank_irq); - - drm_crtc_vblank_on(crtc); + spin_unlock_irq(&crtc->dev->event_lock); } static void omap_crtc_disable(struct drm_crtc *crtc) @@ -390,16 +387,15 @@ static int omap_crtc_atomic_check(struct drm_crtc *crtc, } static void omap_crtc_atomic_begin(struct drm_crtc *crtc, - struct drm_crtc_state *old_crtc_state) + struct drm_crtc_state *old_crtc_state) { } static void omap_crtc_atomic_flush(struct drm_crtc *crtc, - struct drm_crtc_state *old_crtc_state) + struct drm_crtc_state *old_crtc_state) { struct omap_crtc *omap_crtc = to_omap_crtc(crtc); - - WARN_ON(omap_crtc->vblank_irq.registered); + int ret; if (crtc->state->color_mgmt_changed) { struct drm_color_lut *lut = NULL; @@ -414,18 +410,30 @@ static void omap_crtc_atomic_flush(struct drm_crtc *crtc, dispc_mgr_set_gamma(omap_crtc->channel, lut, length); } - if (dispc_mgr_is_enabled(omap_crtc->channel)) { + /* + * Only flush the CRTC if it is currently enabled. CRTCs that require a + * mode set are disabled prior plane updates and enabled afterwards. + * They are thus not active (regardless of what their CRTC core state + * reports) and the DRM core could thus call this function even though + * the CRTC is currently disabled. Do nothing in that case. + */ + if (!omap_crtc->enabled) + return; + + DBG("%s: GO", omap_crtc->name); - DBG("%s: GO", omap_crtc->name); + ret = drm_crtc_vblank_get(crtc); + WARN_ON(ret != 0); - rmb(); - WARN_ON(omap_crtc->pending); - omap_crtc->pending = true; - wmb(); + spin_lock_irq(&crtc->dev->event_lock); + dispc_mgr_go(omap_crtc->channel); - dispc_mgr_go(omap_crtc->channel); - omap_irq_register(crtc->dev, &omap_crtc->vblank_irq); - } + WARN_ON(omap_crtc->pending); + omap_crtc->pending = true; + + if (crtc->state->event) + omap_crtc->event = crtc->state->event; + spin_unlock_irq(&crtc->dev->event_lock); } static bool omap_crtc_is_plane_prop(struct drm_crtc *crtc, @@ -546,14 +554,6 @@ struct drm_crtc *omap_crtc_init(struct drm_device *dev, omap_crtc->channel = channel; omap_crtc->name = channel_names[channel]; - omap_crtc->vblank_irq.irqmask = pipe2vbl(crtc); - omap_crtc->vblank_irq.irq = omap_crtc_vblank_irq; - - omap_crtc->error_irq.irqmask = - dispc_mgr_get_sync_lost_irq(channel); - omap_crtc->error_irq.irq = omap_crtc_error_irq; - omap_irq_register(dev, &omap_crtc->error_irq); - ret = drm_crtc_init_with_planes(dev, crtc, plane, NULL, &omap_crtc_funcs, NULL); if (ret < 0) { diff --git a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c index 4ceed7a9762f..3cab06661a08 100644 --- a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c +++ b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c @@ -224,7 +224,7 @@ static void dmm_txn_append(struct dmm_txn *txn, struct pat_area *area, int rows = (1 + area->y1 - area->y0); int i = columns*rows; - pat = alloc_dma(txn, sizeof(struct pat), &pat_pa); + pat = alloc_dma(txn, sizeof(*pat), &pat_pa); if (txn->last_pat) txn->last_pat->next_pa = (uint32_t)pat_pa; @@ -735,7 +735,7 @@ static int omap_dmm_probe(struct platform_device *dev) /* alloc engines */ omap_dmm->engines = kcalloc(omap_dmm->num_engines, - sizeof(struct refill_engine), GFP_KERNEL); + sizeof(*omap_dmm->engines), GFP_KERNEL); if (!omap_dmm->engines) { ret = -ENOMEM; goto fail; diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c index 42330e0c3324..afe8f05b927b 100644 --- a/drivers/gpu/drm/omapdrm/omap_drv.c +++ b/drivers/gpu/drm/omapdrm/omap_drv.c @@ -96,7 +96,8 @@ static void omap_atomic_complete(struct omap_atomic_state_commit *commit) dispc_runtime_get(); drm_atomic_helper_commit_modeset_disables(dev, old_state); - drm_atomic_helper_commit_planes(dev, old_state, 0); + drm_atomic_helper_commit_planes(dev, old_state, + DRM_PLANE_COMMIT_ACTIVE_ONLY); drm_atomic_helper_commit_modeset_enables(dev, old_state); omap_atomic_wait_for_completion(dev, old_state); @@ -315,8 +316,6 @@ static int omap_modeset_init(struct drm_device *dev) drm_mode_config_init(dev); - omap_drm_irq_install(dev); - ret = omap_modeset_init_properties(dev); if (ret < 0) return ret; @@ -489,12 +488,9 @@ static int omap_modeset_init(struct drm_device *dev) drm_mode_config_reset(dev); - return 0; -} + omap_drm_irq_install(dev); -static void omap_modeset_free(struct drm_device *dev) -{ - drm_mode_config_cleanup(dev); + return 0; } /* @@ -632,93 +628,6 @@ static const struct drm_ioctl_desc ioctls[DRM_COMMAND_END - DRM_COMMAND_BASE] = * drm driver funcs */ -/** - * load - setup chip and create an initial config - * @dev: DRM device - * @flags: startup flags - * - * The driver load routine has to do several things: - * - initialize the memory manager - * - allocate initial config memory - * - setup the DRM framebuffer with the allocated memory - */ -static int dev_load(struct drm_device *dev, unsigned long flags) -{ - struct omap_drm_platform_data *pdata = dev->dev->platform_data; - struct omap_drm_private *priv; - unsigned int i; - int ret; - - DBG("load: dev=%p", dev); - - priv = kzalloc(sizeof(*priv), GFP_KERNEL); - if (!priv) - return -ENOMEM; - - priv->omaprev = pdata->omaprev; - - dev->dev_private = priv; - - priv->wq = alloc_ordered_workqueue("omapdrm", 0); - init_waitqueue_head(&priv->commit.wait); - spin_lock_init(&priv->commit.lock); - - spin_lock_init(&priv->list_lock); - INIT_LIST_HEAD(&priv->obj_list); - - omap_gem_init(dev); - - ret = omap_modeset_init(dev); - if (ret) { - dev_err(dev->dev, "omap_modeset_init failed: ret=%d\n", ret); - dev->dev_private = NULL; - kfree(priv); - return ret; - } - - /* Initialize vblank handling, start with all CRTCs disabled. */ - ret = drm_vblank_init(dev, priv->num_crtcs); - if (ret) - dev_warn(dev->dev, "could not init vblank\n"); - - for (i = 0; i < priv->num_crtcs; i++) - drm_crtc_vblank_off(priv->crtcs[i]); - - priv->fbdev = omap_fbdev_init(dev); - - /* store off drm_device for use in pm ops */ - dev_set_drvdata(dev->dev, dev); - - drm_kms_helper_poll_init(dev); - - return 0; -} - -static void dev_unload(struct drm_device *dev) -{ - struct omap_drm_private *priv = dev->dev_private; - - DBG("unload: dev=%p", dev); - - drm_kms_helper_poll_fini(dev); - - if (priv->fbdev) - omap_fbdev_free(dev); - - omap_modeset_free(dev); - omap_gem_deinit(dev); - - destroy_workqueue(priv->wq); - - drm_vblank_cleanup(dev); - omap_drm_irq_uninstall(dev); - - kfree(dev->dev_private); - dev->dev_private = NULL; - - dev_set_drvdata(dev->dev, NULL); -} - static int dev_open(struct drm_device *dev, struct drm_file *file) { file->driver_priv = NULL; @@ -803,8 +712,6 @@ static const struct file_operations omapdriver_fops = { static struct drm_driver omap_drm_driver = { .driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME | DRIVER_ATOMIC, - .load = dev_load, - .unload = dev_unload, .open = dev_open, .lastclose = dev_lastclose, .get_vblank_counter = drm_vblank_no_hw_counter, @@ -833,30 +740,125 @@ static struct drm_driver omap_drm_driver = { .patchlevel = DRIVER_PATCHLEVEL, }; -static int pdev_probe(struct platform_device *device) +static int pdev_probe(struct platform_device *pdev) { - int r; + struct omap_drm_platform_data *pdata = pdev->dev.platform_data; + struct omap_drm_private *priv; + struct drm_device *ddev; + unsigned int i; + int ret; + + DBG("%s", pdev->name); if (omapdss_is_initialized() == false) return -EPROBE_DEFER; omap_crtc_pre_init(); - r = omap_connect_dssdevs(); - if (r) { - omap_crtc_pre_uninit(); - return r; + ret = omap_connect_dssdevs(); + if (ret) + goto err_crtc_uninit; + + /* Allocate and initialize the driver private structure. */ + priv = kzalloc(sizeof(*priv), GFP_KERNEL); + if (!priv) { + ret = -ENOMEM; + goto err_disconnect_dssdevs; } - DBG("%s", device->name); - return drm_platform_init(&omap_drm_driver, device); + priv->omaprev = pdata->omaprev; + priv->wq = alloc_ordered_workqueue("omapdrm", 0); + + init_waitqueue_head(&priv->commit.wait); + spin_lock_init(&priv->commit.lock); + spin_lock_init(&priv->list_lock); + INIT_LIST_HEAD(&priv->obj_list); + + /* Allocate and initialize the DRM device. */ + ddev = drm_dev_alloc(&omap_drm_driver, &pdev->dev); + if (IS_ERR(ddev)) { + ret = PTR_ERR(ddev); + goto err_free_priv; + } + + ddev->dev_private = priv; + platform_set_drvdata(pdev, ddev); + + omap_gem_init(ddev); + + ret = omap_modeset_init(ddev); + if (ret) { + dev_err(&pdev->dev, "omap_modeset_init failed: ret=%d\n", ret); + goto err_free_drm_dev; + } + + /* Initialize vblank handling, start with all CRTCs disabled. */ + ret = drm_vblank_init(ddev, priv->num_crtcs); + if (ret) { + dev_err(&pdev->dev, "could not init vblank\n"); + goto err_cleanup_modeset; + } + + for (i = 0; i < priv->num_crtcs; i++) + drm_crtc_vblank_off(priv->crtcs[i]); + + priv->fbdev = omap_fbdev_init(ddev); + + drm_kms_helper_poll_init(ddev); + + /* + * Register the DRM device with the core and the connectors with + * sysfs. + */ + ret = drm_dev_register(ddev, 0); + if (ret) + goto err_cleanup_helpers; + + return 0; + +err_cleanup_helpers: + drm_kms_helper_poll_fini(ddev); + if (priv->fbdev) + omap_fbdev_free(ddev); +err_cleanup_modeset: + drm_mode_config_cleanup(ddev); + omap_drm_irq_uninstall(ddev); +err_free_drm_dev: + omap_gem_deinit(ddev); + drm_dev_unref(ddev); +err_free_priv: + destroy_workqueue(priv->wq); + kfree(priv); +err_disconnect_dssdevs: + omap_disconnect_dssdevs(); +err_crtc_uninit: + omap_crtc_pre_uninit(); + return ret; } -static int pdev_remove(struct platform_device *device) +static int pdev_remove(struct platform_device *pdev) { + struct drm_device *ddev = platform_get_drvdata(pdev); + struct omap_drm_private *priv = ddev->dev_private; + DBG(""); - drm_put_dev(platform_get_drvdata(device)); + drm_dev_unregister(ddev); + + drm_kms_helper_poll_fini(ddev); + + if (priv->fbdev) + omap_fbdev_free(ddev); + + drm_mode_config_cleanup(ddev); + + omap_drm_irq_uninstall(ddev); + omap_gem_deinit(ddev); + + drm_dev_unref(ddev); + + destroy_workqueue(priv->wq); + kfree(priv); omap_disconnect_dssdevs(); omap_crtc_pre_uninit(); diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h index 8b113ba39da0..36d93ce84a29 100644 --- a/drivers/gpu/drm/omapdrm/omap_drv.h +++ b/drivers/gpu/drm/omapdrm/omap_drv.h @@ -48,19 +48,6 @@ struct omap_drm_window { uint32_t src_w, src_h; }; -/* For transiently registering for different DSS irqs that various parts - * of the KMS code need during setup/configuration. We these are not - * necessarily the same as what drm_vblank_get/put() are requesting, and - * the hysteresis in drm_vblank_put() is not necessarily desirable for - * internal housekeeping related irq usage. - */ -struct omap_drm_irq { - struct list_head node; - uint32_t irqmask; - bool registered; - void (*irq)(struct omap_drm_irq *irq, uint32_t irqstatus); -}; - /* For KMS code that needs to wait for a certain # of IRQs: */ struct omap_irq_wait; @@ -101,9 +88,9 @@ struct omap_drm_private { struct drm_property *zorder_prop; /* irq handling: */ - struct list_head irq_list; /* list of omap_drm_irq */ - uint32_t vblank_mask; /* irq bits set for userspace vblank */ - struct omap_drm_irq error_handler; + spinlock_t wait_lock; /* protects the wait_list */ + struct list_head wait_list; /* list of omap_irq_wait */ + uint32_t irq_mask; /* enabled irqs in addition to wait_list */ /* atomic commit */ struct { @@ -127,10 +114,6 @@ int omap_gem_resume(struct device *dev); int omap_irq_enable_vblank(struct drm_device *dev, unsigned int pipe); void omap_irq_disable_vblank(struct drm_device *dev, unsigned int pipe); -void __omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq); -void __omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq); -void omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq); -void omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq); void omap_drm_irq_uninstall(struct drm_device *dev); int omap_drm_irq_install(struct drm_device *dev); @@ -154,6 +137,8 @@ void omap_crtc_pre_uninit(void); struct drm_crtc *omap_crtc_init(struct drm_device *dev, struct drm_plane *plane, enum omap_channel channel, int id); int omap_crtc_wait_pending(struct drm_crtc *crtc); +void omap_crtc_error_irq(struct drm_crtc *crtc, uint32_t irqstatus); +void omap_crtc_vblank_irq(struct drm_crtc *crtc); struct drm_plane *omap_plane_init(struct drm_device *dev, int id, enum drm_plane_type type, @@ -232,32 +217,6 @@ struct drm_gem_object *omap_gem_prime_import(struct drm_device *dev, struct dma_buf *buffer); /* map crtc to vblank mask */ -uint32_t pipe2vbl(struct drm_crtc *crtc); struct omap_dss_device *omap_encoder_get_dssdev(struct drm_encoder *encoder); -/* should these be made into common util helpers? - */ - -static inline int objects_lookup( - struct drm_file *filp, uint32_t pixel_format, - struct drm_gem_object **bos, const uint32_t *handles) -{ - int i, n = drm_format_num_planes(pixel_format); - - for (i = 0; i < n; i++) { - bos[i] = drm_gem_object_lookup(filp, handles[i]); - if (!bos[i]) - goto fail; - - } - - return 0; - -fail: - while (--i > 0) - drm_gem_object_unreference_unlocked(bos[i]); - - return -ENOENT; -} - #endif /* __OMAP_DRV_H__ */ diff --git a/drivers/gpu/drm/omapdrm/omap_encoder.c b/drivers/gpu/drm/omapdrm/omap_encoder.c index a20f30039aee..86c977b7189a 100644 --- a/drivers/gpu/drm/omapdrm/omap_encoder.c +++ b/drivers/gpu/drm/omapdrm/omap_encoder.c @@ -117,7 +117,7 @@ static int omap_encoder_update(struct drm_encoder *encoder, dssdrv->get_timings(dssdev, &t); - if (memcmp(vm, &t, sizeof(struct videomode))) + if (memcmp(vm, &t, sizeof(*vm))) ret = -EINVAL; else ret = 0; diff --git a/drivers/gpu/drm/omapdrm/omap_fb.c b/drivers/gpu/drm/omapdrm/omap_fb.c index bd6b94c38613..29dc677dd4d3 100644 --- a/drivers/gpu/drm/omapdrm/omap_fb.c +++ b/drivers/gpu/drm/omapdrm/omap_fb.c @@ -29,37 +29,30 @@ * framebuffer funcs */ -/* per-format info: */ -struct format { +/* DSS to DRM formats mapping */ +static const struct { enum omap_color_mode dss_format; uint32_t pixel_format; - struct { - int stride_bpp; /* this times width is stride */ - int sub_y; /* sub-sample in y dimension */ - } planes[4]; - bool yuv; -}; - -static const struct format formats[] = { +} formats[] = { /* 16bpp [A]RGB: */ - { OMAP_DSS_COLOR_RGB16, DRM_FORMAT_RGB565, {{2, 1}}, false }, /* RGB16-565 */ - { OMAP_DSS_COLOR_RGB12U, DRM_FORMAT_RGBX4444, {{2, 1}}, false }, /* RGB12x-4444 */ - { OMAP_DSS_COLOR_RGBX16, DRM_FORMAT_XRGB4444, {{2, 1}}, false }, /* xRGB12-4444 */ - { OMAP_DSS_COLOR_RGBA16, DRM_FORMAT_RGBA4444, {{2, 1}}, false }, /* RGBA12-4444 */ - { OMAP_DSS_COLOR_ARGB16, DRM_FORMAT_ARGB4444, {{2, 1}}, false }, /* ARGB16-4444 */ - { OMAP_DSS_COLOR_XRGB16_1555, DRM_FORMAT_XRGB1555, {{2, 1}}, false }, /* xRGB15-1555 */ - { OMAP_DSS_COLOR_ARGB16_1555, DRM_FORMAT_ARGB1555, {{2, 1}}, false }, /* ARGB16-1555 */ + { OMAP_DSS_COLOR_RGB16, DRM_FORMAT_RGB565 }, /* RGB16-565 */ + { OMAP_DSS_COLOR_RGB12U, DRM_FORMAT_RGBX4444 }, /* RGB12x-4444 */ + { OMAP_DSS_COLOR_RGBX16, DRM_FORMAT_XRGB4444 }, /* xRGB12-4444 */ + { OMAP_DSS_COLOR_RGBA16, DRM_FORMAT_RGBA4444 }, /* RGBA12-4444 */ + { OMAP_DSS_COLOR_ARGB16, DRM_FORMAT_ARGB4444 }, /* ARGB16-4444 */ + { OMAP_DSS_COLOR_XRGB16_1555, DRM_FORMAT_XRGB1555 }, /* xRGB15-1555 */ + { OMAP_DSS_COLOR_ARGB16_1555, DRM_FORMAT_ARGB1555 }, /* ARGB16-1555 */ /* 24bpp RGB: */ - { OMAP_DSS_COLOR_RGB24P, DRM_FORMAT_RGB888, {{3, 1}}, false }, /* RGB24-888 */ + { OMAP_DSS_COLOR_RGB24P, DRM_FORMAT_RGB888 }, /* RGB24-888 */ /* 32bpp [A]RGB: */ - { OMAP_DSS_COLOR_RGBX32, DRM_FORMAT_RGBX8888, {{4, 1}}, false }, /* RGBx24-8888 */ - { OMAP_DSS_COLOR_RGB24U, DRM_FORMAT_XRGB8888, {{4, 1}}, false }, /* xRGB24-8888 */ - { OMAP_DSS_COLOR_RGBA32, DRM_FORMAT_RGBA8888, {{4, 1}}, false }, /* RGBA32-8888 */ - { OMAP_DSS_COLOR_ARGB32, DRM_FORMAT_ARGB8888, {{4, 1}}, false }, /* ARGB32-8888 */ + { OMAP_DSS_COLOR_RGBX32, DRM_FORMAT_RGBX8888 }, /* RGBx24-8888 */ + { OMAP_DSS_COLOR_RGB24U, DRM_FORMAT_XRGB8888 }, /* xRGB24-8888 */ + { OMAP_DSS_COLOR_RGBA32, DRM_FORMAT_RGBA8888 }, /* RGBA32-8888 */ + { OMAP_DSS_COLOR_ARGB32, DRM_FORMAT_ARGB8888 }, /* ARGB32-8888 */ /* YUV: */ - { OMAP_DSS_COLOR_NV12, DRM_FORMAT_NV12, {{1, 1}, {1, 2}}, true }, - { OMAP_DSS_COLOR_YUV2, DRM_FORMAT_YUYV, {{2, 1}}, true }, - { OMAP_DSS_COLOR_UYVY, DRM_FORMAT_UYVY, {{2, 1}}, true }, + { OMAP_DSS_COLOR_NV12, DRM_FORMAT_NV12 }, + { OMAP_DSS_COLOR_YUV2, DRM_FORMAT_YUYV }, + { OMAP_DSS_COLOR_UYVY, DRM_FORMAT_UYVY }, }; /* convert from overlay's pixel formats bitmask to an array of fourcc's */ @@ -89,8 +82,9 @@ struct plane { struct omap_framebuffer { struct drm_framebuffer base; int pin_count; - const struct format *format; - struct plane planes[4]; + const struct drm_format_info *format; + enum omap_color_mode dss_format; + struct plane planes[2]; /* lock for pinning (pin_count and planes.paddr) */ struct mutex lock; }; @@ -128,13 +122,13 @@ static const struct drm_framebuffer_funcs omap_framebuffer_funcs = { }; static uint32_t get_linear_addr(struct plane *plane, - const struct format *format, int n, int x, int y) + const struct drm_format_info *format, int n, int x, int y) { uint32_t offset; - offset = plane->offset + - (x * format->planes[n].stride_bpp) + - (y * plane->pitch / format->planes[n].sub_y); + offset = plane->offset + + (x * format->cpp[n] / (n == 0 ? 1 : format->hsub)) + + (y * plane->pitch / (n == 0 ? 1 : format->vsub)); return plane->paddr + offset; } @@ -153,11 +147,11 @@ void omap_framebuffer_update_scanout(struct drm_framebuffer *fb, struct omap_drm_window *win, struct omap_overlay_info *info) { struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb); - const struct format *format = omap_fb->format; + const struct drm_format_info *format = omap_fb->format; struct plane *plane = &omap_fb->planes[0]; uint32_t x, y, orient = 0; - info->color_mode = format->dss_format; + info->color_mode = omap_fb->dss_format; info->pos_x = win->crtc_x; info->pos_y = win->crtc_y; @@ -231,9 +225,9 @@ void omap_framebuffer_update_scanout(struct drm_framebuffer *fb, } /* convert to pixels: */ - info->screen_width /= format->planes[0].stride_bpp; + info->screen_width /= format->cpp[0]; - if (format->dss_format == OMAP_DSS_COLOR_NV12) { + if (omap_fb->dss_format == OMAP_DSS_COLOR_NV12) { plane = &omap_fb->planes[1]; if (info->rotation_type == OMAP_DSS_ROT_TILER) { @@ -360,47 +354,58 @@ void omap_framebuffer_describe(struct drm_framebuffer *fb, struct seq_file *m) struct drm_framebuffer *omap_framebuffer_create(struct drm_device *dev, struct drm_file *file, const struct drm_mode_fb_cmd2 *mode_cmd) { + unsigned int num_planes = drm_format_num_planes(mode_cmd->pixel_format); struct drm_gem_object *bos[4]; struct drm_framebuffer *fb; - int ret; + int i; - ret = objects_lookup(file, mode_cmd->pixel_format, - bos, mode_cmd->handles); - if (ret) - return ERR_PTR(ret); + for (i = 0; i < num_planes; i++) { + bos[i] = drm_gem_object_lookup(file, mode_cmd->handles[i]); + if (!bos[i]) { + fb = ERR_PTR(-ENOENT); + goto error; + } + } fb = omap_framebuffer_init(dev, mode_cmd, bos); - if (IS_ERR(fb)) { - int i, n = drm_format_num_planes(mode_cmd->pixel_format); - for (i = 0; i < n; i++) - drm_gem_object_unreference_unlocked(bos[i]); - return fb; - } + if (IS_ERR(fb)) + goto error; + + return fb; + +error: + while (--i > 0) + drm_gem_object_unreference_unlocked(bos[i]); + return fb; } struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev, const struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object **bos) { + const struct drm_format_info *format = NULL; struct omap_framebuffer *omap_fb = NULL; struct drm_framebuffer *fb = NULL; - const struct format *format = NULL; - int ret, i, n = drm_format_num_planes(mode_cmd->pixel_format); + enum omap_color_mode dss_format = 0; + unsigned int pitch = mode_cmd->pitches[0]; + int ret, i; DBG("create framebuffer: dev=%p, mode_cmd=%p (%dx%d@%4.4s)", dev, mode_cmd, mode_cmd->width, mode_cmd->height, (char *)&mode_cmd->pixel_format); + format = drm_format_info(mode_cmd->pixel_format); + for (i = 0; i < ARRAY_SIZE(formats); i++) { if (formats[i].pixel_format == mode_cmd->pixel_format) { - format = &formats[i]; + dss_format = formats[i].dss_format; break; } } - if (!format) { - dev_err(dev->dev, "unsupported pixel format: %4.4s\n", - (char *)&mode_cmd->pixel_format); + if (!format || !dss_format) { + dev_dbg(dev->dev, "unsupported pixel format: %4.4s\n", + (char *)&mode_cmd->pixel_format); ret = -EINVAL; goto fail; } @@ -413,40 +418,39 @@ struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev, fb = &omap_fb->base; omap_fb->format = format; + omap_fb->dss_format = dss_format; mutex_init(&omap_fb->lock); - for (i = 0; i < n; i++) { - struct plane *plane = &omap_fb->planes[i]; - int size, pitch = mode_cmd->pitches[i]; - - if (pitch < (mode_cmd->width * format->planes[i].stride_bpp)) { - dev_err(dev->dev, "provided buffer pitch is too small! %d < %d\n", - pitch, mode_cmd->width * format->planes[i].stride_bpp); - ret = -EINVAL; - goto fail; - } + /* + * The code below assumes that no format use more than two planes, and + * that the two planes of multiplane formats need the same number of + * bytes per pixel. + */ + if (format->num_planes == 2 && pitch != mode_cmd->pitches[1]) { + dev_dbg(dev->dev, "pitches differ between planes 0 and 1\n"); + ret = -EINVAL; + goto fail; + } - if (pitch % format->planes[i].stride_bpp != 0) { - dev_err(dev->dev, - "buffer pitch (%d bytes) is not a multiple of pixel size (%d bytes)\n", - pitch, format->planes[i].stride_bpp); - ret = -EINVAL; - goto fail; - } + if (pitch % format->cpp[0]) { + dev_dbg(dev->dev, + "buffer pitch (%u bytes) is not a multiple of pixel size (%u bytes)\n", + pitch, format->cpp[0]); + ret = -EINVAL; + goto fail; + } - size = pitch * mode_cmd->height / format->planes[i].sub_y; + for (i = 0; i < format->num_planes; i++) { + struct plane *plane = &omap_fb->planes[i]; + unsigned int vsub = i == 0 ? 1 : format->vsub; + unsigned int size; - if (size > (omap_gem_mmap_size(bos[i]) - mode_cmd->offsets[i])) { - dev_err(dev->dev, "provided buffer object is too small! %d < %d\n", - bos[i]->size - mode_cmd->offsets[i], size); - ret = -EINVAL; - goto fail; - } + size = pitch * mode_cmd->height / vsub; - if (i > 0 && pitch != mode_cmd->pitches[i - 1]) { - dev_err(dev->dev, - "pitches are not the same between framebuffer planes %d != %d\n", - pitch, mode_cmd->pitches[i - 1]); + if (size > omap_gem_mmap_size(bos[i]) - mode_cmd->offsets[i]) { + dev_dbg(dev->dev, + "provided buffer object is too small! %d < %d\n", + bos[i]->size - mode_cmd->offsets[i], size); ret = -EINVAL; goto fail; } diff --git a/drivers/gpu/drm/omapdrm/omap_irq.c b/drivers/gpu/drm/omapdrm/omap_irq.c index 60e1e8016708..9adfa7c99695 100644 --- a/drivers/gpu/drm/omapdrm/omap_irq.c +++ b/drivers/gpu/drm/omapdrm/omap_irq.c @@ -19,25 +19,24 @@ #include "omap_drv.h" -static DEFINE_SPINLOCK(list_lock); - -static void omap_irq_error_handler(struct omap_drm_irq *irq, - uint32_t irqstatus) -{ - DRM_ERROR("errors: %08x\n", irqstatus); -} +struct omap_irq_wait { + struct list_head node; + wait_queue_head_t wq; + uint32_t irqmask; + int count; +}; -/* call with list_lock and dispc runtime held */ +/* call with wait_lock and dispc runtime held */ static void omap_irq_update(struct drm_device *dev) { struct omap_drm_private *priv = dev->dev_private; - struct omap_drm_irq *irq; - uint32_t irqmask = priv->vblank_mask; + struct omap_irq_wait *wait; + uint32_t irqmask = priv->irq_mask; - assert_spin_locked(&list_lock); + assert_spin_locked(&priv->wait_lock); - list_for_each_entry(irq, &priv->irq_list, node) - irqmask |= irq->irqmask; + list_for_each_entry(wait, &priv->wait_list, node) + irqmask |= wait->irqmask; DBG("irqmask=%08x", irqmask); @@ -45,90 +44,48 @@ static void omap_irq_update(struct drm_device *dev) dispc_read_irqenable(); /* flush posted write */ } -void __omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq) -{ - struct omap_drm_private *priv = dev->dev_private; - unsigned long flags; - - spin_lock_irqsave(&list_lock, flags); - - if (!WARN_ON(irq->registered)) { - irq->registered = true; - list_add(&irq->node, &priv->irq_list); - omap_irq_update(dev); - } - - spin_unlock_irqrestore(&list_lock, flags); -} - -void omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq) -{ - dispc_runtime_get(); - - __omap_irq_register(dev, irq); - - dispc_runtime_put(); -} - -void __omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq) +static void omap_irq_wait_handler(struct omap_irq_wait *wait) { - unsigned long flags; - - spin_lock_irqsave(&list_lock, flags); - - if (!WARN_ON(!irq->registered)) { - irq->registered = false; - list_del(&irq->node); - omap_irq_update(dev); - } - - spin_unlock_irqrestore(&list_lock, flags); -} - -void omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq) -{ - dispc_runtime_get(); - - __omap_irq_unregister(dev, irq); - - dispc_runtime_put(); -} - -struct omap_irq_wait { - struct omap_drm_irq irq; - int count; -}; - -static DECLARE_WAIT_QUEUE_HEAD(wait_event); - -static void wait_irq(struct omap_drm_irq *irq, uint32_t irqstatus) -{ - struct omap_irq_wait *wait = - container_of(irq, struct omap_irq_wait, irq); wait->count--; - wake_up_all(&wait_event); + wake_up(&wait->wq); } struct omap_irq_wait * omap_irq_wait_init(struct drm_device *dev, uint32_t irqmask, int count) { + struct omap_drm_private *priv = dev->dev_private; struct omap_irq_wait *wait = kzalloc(sizeof(*wait), GFP_KERNEL); - wait->irq.irq = wait_irq; - wait->irq.irqmask = irqmask; + unsigned long flags; + + init_waitqueue_head(&wait->wq); + wait->irqmask = irqmask; wait->count = count; - omap_irq_register(dev, &wait->irq); + + spin_lock_irqsave(&priv->wait_lock, flags); + list_add(&wait->node, &priv->wait_list); + omap_irq_update(dev); + spin_unlock_irqrestore(&priv->wait_lock, flags); + return wait; } int omap_irq_wait(struct drm_device *dev, struct omap_irq_wait *wait, unsigned long timeout) { - int ret = wait_event_timeout(wait_event, (wait->count <= 0), timeout); - omap_irq_unregister(dev, &wait->irq); + struct omap_drm_private *priv = dev->dev_private; + unsigned long flags; + int ret; + + ret = wait_event_timeout(wait->wq, (wait->count <= 0), timeout); + + spin_lock_irqsave(&priv->wait_lock, flags); + list_del(&wait->node); + omap_irq_update(dev); + spin_unlock_irqrestore(&priv->wait_lock, flags); + kfree(wait); - if (ret == 0) - return -1; - return 0; + + return ret == 0 ? -1 : 0; } /** @@ -152,10 +109,10 @@ int omap_irq_enable_vblank(struct drm_device *dev, unsigned int pipe) DBG("dev=%p, crtc=%u", dev, pipe); - spin_lock_irqsave(&list_lock, flags); - priv->vblank_mask |= pipe2vbl(crtc); + spin_lock_irqsave(&priv->wait_lock, flags); + priv->irq_mask |= dispc_mgr_get_vsync_irq(omap_crtc_channel(crtc)); omap_irq_update(dev); - spin_unlock_irqrestore(&list_lock, flags); + spin_unlock_irqrestore(&priv->wait_lock, flags); return 0; } @@ -177,17 +134,66 @@ void omap_irq_disable_vblank(struct drm_device *dev, unsigned int pipe) DBG("dev=%p, crtc=%u", dev, pipe); - spin_lock_irqsave(&list_lock, flags); - priv->vblank_mask &= ~pipe2vbl(crtc); + spin_lock_irqsave(&priv->wait_lock, flags); + priv->irq_mask &= ~dispc_mgr_get_vsync_irq(omap_crtc_channel(crtc)); omap_irq_update(dev); - spin_unlock_irqrestore(&list_lock, flags); + spin_unlock_irqrestore(&priv->wait_lock, flags); +} + +static void omap_irq_fifo_underflow(struct omap_drm_private *priv, + u32 irqstatus) +{ + static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, + DEFAULT_RATELIMIT_BURST); + static const struct { + const char *name; + u32 mask; + } sources[] = { + { "gfx", DISPC_IRQ_GFX_FIFO_UNDERFLOW }, + { "vid1", DISPC_IRQ_VID1_FIFO_UNDERFLOW }, + { "vid2", DISPC_IRQ_VID2_FIFO_UNDERFLOW }, + { "vid3", DISPC_IRQ_VID3_FIFO_UNDERFLOW }, + }; + + const u32 mask = DISPC_IRQ_GFX_FIFO_UNDERFLOW + | DISPC_IRQ_VID1_FIFO_UNDERFLOW + | DISPC_IRQ_VID2_FIFO_UNDERFLOW + | DISPC_IRQ_VID3_FIFO_UNDERFLOW; + unsigned int i; + + spin_lock(&priv->wait_lock); + irqstatus &= priv->irq_mask & mask; + spin_unlock(&priv->wait_lock); + + if (!irqstatus) + return; + + if (!__ratelimit(&_rs)) + return; + + DRM_ERROR("FIFO underflow on "); + + for (i = 0; i < ARRAY_SIZE(sources); ++i) { + if (sources[i].mask & irqstatus) + pr_cont("%s ", sources[i].name); + } + + pr_cont("(0x%08x)\n", irqstatus); +} + +static void omap_irq_ocp_error_handler(u32 irqstatus) +{ + if (!(irqstatus & DISPC_IRQ_OCP_ERR)) + return; + + DRM_ERROR("OCP error\n"); } static irqreturn_t omap_irq_handler(int irq, void *arg) { struct drm_device *dev = (struct drm_device *) arg; struct omap_drm_private *priv = dev->dev_private; - struct omap_drm_irq *handler, *n; + struct omap_irq_wait *wait, *n; unsigned long flags; unsigned int id; u32 irqstatus; @@ -200,24 +206,37 @@ static irqreturn_t omap_irq_handler(int irq, void *arg) for (id = 0; id < priv->num_crtcs; id++) { struct drm_crtc *crtc = priv->crtcs[id]; + enum omap_channel channel = omap_crtc_channel(crtc); - if (irqstatus & pipe2vbl(crtc)) + if (irqstatus & dispc_mgr_get_vsync_irq(channel)) { drm_handle_vblank(dev, id); + omap_crtc_vblank_irq(crtc); + } + + if (irqstatus & dispc_mgr_get_sync_lost_irq(channel)) + omap_crtc_error_irq(crtc, irqstatus); } - spin_lock_irqsave(&list_lock, flags); - list_for_each_entry_safe(handler, n, &priv->irq_list, node) { - if (handler->irqmask & irqstatus) { - spin_unlock_irqrestore(&list_lock, flags); - handler->irq(handler, handler->irqmask & irqstatus); - spin_lock_irqsave(&list_lock, flags); - } + omap_irq_ocp_error_handler(irqstatus); + omap_irq_fifo_underflow(priv, irqstatus); + + spin_lock_irqsave(&priv->wait_lock, flags); + list_for_each_entry_safe(wait, n, &priv->wait_list, node) { + if (wait->irqmask & irqstatus) + omap_irq_wait_handler(wait); } - spin_unlock_irqrestore(&list_lock, flags); + spin_unlock_irqrestore(&priv->wait_lock, flags); return IRQ_HANDLED; } +static const u32 omap_underflow_irqs[] = { + [OMAP_DSS_GFX] = DISPC_IRQ_GFX_FIFO_UNDERFLOW, + [OMAP_DSS_VIDEO1] = DISPC_IRQ_VID1_FIFO_UNDERFLOW, + [OMAP_DSS_VIDEO2] = DISPC_IRQ_VID2_FIFO_UNDERFLOW, + [OMAP_DSS_VIDEO3] = DISPC_IRQ_VID3_FIFO_UNDERFLOW, +}; + /* * We need a special version, instead of just using drm_irq_install(), * because we need to register the irq via omapdss. Once omapdss and @@ -228,10 +247,25 @@ static irqreturn_t omap_irq_handler(int irq, void *arg) int omap_drm_irq_install(struct drm_device *dev) { struct omap_drm_private *priv = dev->dev_private; - struct omap_drm_irq *error_handler = &priv->error_handler; + unsigned int num_mgrs = dss_feat_get_num_mgrs(); + unsigned int max_planes; + unsigned int i; int ret; - INIT_LIST_HEAD(&priv->irq_list); + spin_lock_init(&priv->wait_lock); + INIT_LIST_HEAD(&priv->wait_list); + + priv->irq_mask = DISPC_IRQ_OCP_ERR; + + max_planes = min(ARRAY_SIZE(priv->planes), + ARRAY_SIZE(omap_underflow_irqs)); + for (i = 0; i < max_planes; ++i) { + if (priv->planes[i]) + priv->irq_mask |= omap_underflow_irqs[i]; + } + + for (i = 0; i < num_mgrs; ++i) + priv->irq_mask |= dispc_mgr_get_sync_lost_irq(i); dispc_runtime_get(); dispc_clear_irqstatus(0xffffffff); @@ -241,16 +275,6 @@ int omap_drm_irq_install(struct drm_device *dev) if (ret < 0) return ret; - error_handler->irq = omap_irq_error_handler; - error_handler->irqmask = DISPC_IRQ_OCP_ERR; - - /* for now ignore DISPC_IRQ_SYNC_LOST_DIGIT.. really I think - * we just need to ignore it while enabling tv-out - */ - error_handler->irqmask &= ~DISPC_IRQ_SYNC_LOST_DIGIT; - - omap_irq_register(dev, error_handler); - dev->irq_enabled = true; return 0; diff --git a/drivers/gpu/drm/omapdrm/omap_plane.c b/drivers/gpu/drm/omapdrm/omap_plane.c index 82b2c23d6769..386d90af70f7 100644 --- a/drivers/gpu/drm/omapdrm/omap_plane.c +++ b/drivers/gpu/drm/omapdrm/omap_plane.c @@ -43,8 +43,6 @@ struct omap_plane { uint32_t nformats; uint32_t formats[32]; - - struct omap_drm_irq error_irq; }; struct omap_plane_state { @@ -204,8 +202,6 @@ static void omap_plane_destroy(struct drm_plane *plane) DBG("%s", omap_plane->name); - omap_irq_unregister(plane->dev, &omap_plane->error_irq); - drm_plane_cleanup(plane); kfree(omap_plane); @@ -332,14 +328,6 @@ static const struct drm_plane_funcs omap_plane_funcs = { .atomic_get_property = omap_plane_atomic_get_property, }; -static void omap_plane_error_irq(struct omap_drm_irq *irq, uint32_t irqstatus) -{ - struct omap_plane *omap_plane = - container_of(irq, struct omap_plane, error_irq); - DRM_ERROR_RATELIMITED("%s: errors: %08x\n", omap_plane->name, - irqstatus); -} - static const char *plane_names[] = { [OMAP_DSS_GFX] = "gfx", [OMAP_DSS_VIDEO1] = "vid1", @@ -347,13 +335,6 @@ static const char *plane_names[] = { [OMAP_DSS_VIDEO3] = "vid3", }; -static const uint32_t error_irqs[] = { - [OMAP_DSS_GFX] = DISPC_IRQ_GFX_FIFO_UNDERFLOW, - [OMAP_DSS_VIDEO1] = DISPC_IRQ_VID1_FIFO_UNDERFLOW, - [OMAP_DSS_VIDEO2] = DISPC_IRQ_VID2_FIFO_UNDERFLOW, - [OMAP_DSS_VIDEO3] = DISPC_IRQ_VID3_FIFO_UNDERFLOW, -}; - /* initialize plane */ struct drm_plane *omap_plane_init(struct drm_device *dev, int id, enum drm_plane_type type, @@ -377,10 +358,6 @@ struct drm_plane *omap_plane_init(struct drm_device *dev, plane = &omap_plane->base; - omap_plane->error_irq.irqmask = error_irqs[id]; - omap_plane->error_irq.irq = omap_plane_error_irq; - omap_irq_register(dev, &omap_plane->error_irq); - ret = drm_universal_plane_init(dev, plane, possible_crtcs, &omap_plane_funcs, omap_plane->formats, omap_plane->nformats, type, NULL); @@ -394,7 +371,6 @@ struct drm_plane *omap_plane_init(struct drm_device *dev, return plane; error: - omap_irq_unregister(plane->dev, &omap_plane->error_irq); kfree(omap_plane); return NULL; } diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index ad4d7b8b8322..414776811e71 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -50,7 +50,6 @@ MODULE_FIRMWARE("radeon/tahiti_ce.bin"); MODULE_FIRMWARE("radeon/tahiti_mc.bin"); MODULE_FIRMWARE("radeon/tahiti_rlc.bin"); MODULE_FIRMWARE("radeon/tahiti_smc.bin"); -MODULE_FIRMWARE("radeon/tahiti_k_smc.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_me.bin"); @@ -115,6 +114,9 @@ MODULE_FIRMWARE("radeon/hainan_mc.bin"); MODULE_FIRMWARE("radeon/hainan_rlc.bin"); MODULE_FIRMWARE("radeon/hainan_smc.bin"); MODULE_FIRMWARE("radeon/hainan_k_smc.bin"); +MODULE_FIRMWARE("radeon/banks_k_2_smc.bin"); + +MODULE_FIRMWARE("radeon/si58_mc.bin"); static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh); static void si_pcie_gen3_enable(struct radeon_device *rdev); @@ -1651,15 +1653,14 @@ static int si_init_microcode(struct radeon_device *rdev) int err; int new_fw = 0; bool new_smc = false; + bool si58_fw = false; + bool banks2_fw = false; DRM_DEBUG("\n"); switch (rdev->family) { case CHIP_TAHITI: chip_name = "TAHITI"; - /* XXX: figure out which Tahitis need the new ucode */ - if (0) - new_smc = true; new_chip_name = "tahiti"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1671,12 +1672,9 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_PITCAIRN: chip_name = "PITCAIRN"; - if ((rdev->pdev->revision == 0x81) || - (rdev->pdev->device == 0x6810) || - (rdev->pdev->device == 0x6811) || - (rdev->pdev->device == 0x6816) || - (rdev->pdev->device == 0x6817) || - (rdev->pdev->device == 0x6806)) + if ((rdev->pdev->revision == 0x81) && + ((rdev->pdev->device == 0x6810) || + (rdev->pdev->device == 0x6811))) new_smc = true; new_chip_name = "pitcairn"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; @@ -1689,15 +1687,15 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_VERDE: chip_name = "VERDE"; - if ((rdev->pdev->revision == 0x81) || - (rdev->pdev->revision == 0x83) || - (rdev->pdev->revision == 0x87) || - (rdev->pdev->device == 0x6820) || - (rdev->pdev->device == 0x6821) || - (rdev->pdev->device == 0x6822) || - (rdev->pdev->device == 0x6823) || - (rdev->pdev->device == 0x682A) || - (rdev->pdev->device == 0x682B)) + if (((rdev->pdev->device == 0x6820) && + ((rdev->pdev->revision == 0x81) || + (rdev->pdev->revision == 0x83))) || + ((rdev->pdev->device == 0x6821) && + ((rdev->pdev->revision == 0x83) || + (rdev->pdev->revision == 0x87))) || + ((rdev->pdev->revision == 0x87) && + ((rdev->pdev->device == 0x6823) || + (rdev->pdev->device == 0x682b)))) new_smc = true; new_chip_name = "verde"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; @@ -1710,13 +1708,13 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_OLAND: chip_name = "OLAND"; - if ((rdev->pdev->revision == 0xC7) || - (rdev->pdev->revision == 0x80) || - (rdev->pdev->revision == 0x81) || - (rdev->pdev->revision == 0x83) || - (rdev->pdev->revision == 0x87) || - (rdev->pdev->device == 0x6604) || - (rdev->pdev->device == 0x6605)) + if (((rdev->pdev->revision == 0x81) && + ((rdev->pdev->device == 0x6600) || + (rdev->pdev->device == 0x6604) || + (rdev->pdev->device == 0x6605) || + (rdev->pdev->device == 0x6610))) || + ((rdev->pdev->revision == 0x83) && + (rdev->pdev->device == 0x6610))) new_smc = true; new_chip_name = "oland"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; @@ -1728,13 +1726,17 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_HAINAN: chip_name = "HAINAN"; - if ((rdev->pdev->revision == 0x81) || - (rdev->pdev->revision == 0x83) || - (rdev->pdev->revision == 0xC3) || - (rdev->pdev->device == 0x6664) || - (rdev->pdev->device == 0x6665) || - (rdev->pdev->device == 0x6667)) + if (((rdev->pdev->revision == 0x81) && + (rdev->pdev->device == 0x6660)) || + ((rdev->pdev->revision == 0x83) && + ((rdev->pdev->device == 0x6660) || + (rdev->pdev->device == 0x6663) || + (rdev->pdev->device == 0x6665) || + (rdev->pdev->device == 0x6667)))) new_smc = true; + else if ((rdev->pdev->revision == 0xc3) && + (rdev->pdev->device == 0x6665)) + banks2_fw = true; new_chip_name = "hainan"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1746,6 +1748,10 @@ static int si_init_microcode(struct radeon_device *rdev) default: BUG(); } + /* this memory configuration requires special firmware */ + if (((RREG32(MC_SEQ_MISC0) & 0xff000000) >> 24) == 0x58) + si58_fw = true; + DRM_INFO("Loading %s Microcode\n", new_chip_name); snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", new_chip_name); @@ -1849,7 +1855,10 @@ static int si_init_microcode(struct radeon_device *rdev) } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name); + if (si58_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/si58_mc.bin"); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name); err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); if (err) { snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name); @@ -1880,7 +1889,9 @@ static int si_init_microcode(struct radeon_device *rdev) } } - if (new_smc) + if (banks2_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/banks_k_2_smc.bin"); + else if (new_smc) snprintf(fw_name, sizeof(fw_name), "radeon/%s_k_smc.bin", new_chip_name); else snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name); diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c index 8b5e697f2549..2944916f7102 100644 --- a/drivers/gpu/drm/radeon/si_dpm.c +++ b/drivers/gpu/drm/radeon/si_dpm.c @@ -3008,30 +3008,6 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev, (rdev->pdev->device == 0x6817) || (rdev->pdev->device == 0x6806)) max_mclk = 120000; - } else if (rdev->family == CHIP_VERDE) { - if ((rdev->pdev->revision == 0x81) || - (rdev->pdev->revision == 0x83) || - (rdev->pdev->revision == 0x87) || - (rdev->pdev->device == 0x6820) || - (rdev->pdev->device == 0x6821) || - (rdev->pdev->device == 0x6822) || - (rdev->pdev->device == 0x6823) || - (rdev->pdev->device == 0x682A) || - (rdev->pdev->device == 0x682B)) { - max_sclk = 75000; - max_mclk = 80000; - } - } else if (rdev->family == CHIP_OLAND) { - if ((rdev->pdev->revision == 0xC7) || - (rdev->pdev->revision == 0x80) || - (rdev->pdev->revision == 0x81) || - (rdev->pdev->revision == 0x83) || - (rdev->pdev->revision == 0x87) || - (rdev->pdev->device == 0x6604) || - (rdev->pdev->device == 0x6605)) { - max_sclk = 75000; - max_mclk = 80000; - } } else if (rdev->family == CHIP_HAINAN) { if ((rdev->pdev->revision == 0x81) || (rdev->pdev->revision == 0x83) || @@ -3040,7 +3016,6 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev, (rdev->pdev->device == 0x6665) || (rdev->pdev->device == 0x6667)) { max_sclk = 75000; - max_mclk = 80000; } } /* Apply dpm quirks */ diff --git a/drivers/gpu/drm/sti/Makefile b/drivers/gpu/drm/sti/Makefile index d20f7c0b4eac..c35db12435c3 100644 --- a/drivers/gpu/drm/sti/Makefile +++ b/drivers/gpu/drm/sti/Makefile @@ -13,7 +13,6 @@ sti-drm-y := \ sti_dvo.o \ sti_awg_utils.o \ sti_vtg.o \ - sti_vtac.o \ sti_hda.o \ sti_tvout.o \ sti_hqvdp.o \ diff --git a/drivers/gpu/drm/sti/sti_drv.c b/drivers/gpu/drm/sti/sti_drv.c index d3db22488f96..acc056644cd0 100644 --- a/drivers/gpu/drm/sti/sti_drv.c +++ b/drivers/gpu/drm/sti/sti_drv.c @@ -217,19 +217,7 @@ static void sti_output_poll_changed(struct drm_device *ddev) { struct sti_private *private = ddev->dev_private; - if (!ddev->mode_config.num_connector) - return; - - if (private->fbdev) { - drm_fbdev_cma_hotplug_event(private->fbdev); - return; - } - - private->fbdev = drm_fbdev_cma_init(ddev, 32, - ddev->mode_config.num_crtc, - ddev->mode_config.num_connector); - if (IS_ERR(private->fbdev)) - private->fbdev = NULL; + drm_fbdev_cma_hotplug_event(private->fbdev); } static const struct drm_mode_config_funcs sti_mode_config_funcs = { @@ -346,6 +334,8 @@ static void sti_cleanup(struct drm_device *ddev) static int sti_bind(struct device *dev) { struct drm_device *ddev; + struct sti_private *private; + struct drm_fbdev_cma *fbdev; int ret; ddev = drm_dev_alloc(&sti_driver, dev); @@ -368,6 +358,17 @@ static int sti_bind(struct device *dev) drm_mode_config_reset(ddev); + private = ddev->dev_private; + if (ddev->mode_config.num_connector) { + fbdev = drm_fbdev_cma_init(ddev, 32, ddev->mode_config.num_crtc, + ddev->mode_config.num_connector); + if (IS_ERR(fbdev)) { + DRM_DEBUG_DRIVER("Warning: fails to create fbdev\n"); + fbdev = NULL; + } + private->fbdev = fbdev; + } + return 0; err_register: @@ -440,7 +441,6 @@ static struct platform_driver sti_platform_driver = { static struct platform_driver * const drivers[] = { &sti_tvout_driver, - &sti_vtac_driver, &sti_hqvdp_driver, &sti_hdmi_driver, &sti_hda_driver, diff --git a/drivers/gpu/drm/sti/sti_drv.h b/drivers/gpu/drm/sti/sti_drv.h index 78ebe5e30f53..4c75845cc9ab 100644 --- a/drivers/gpu/drm/sti/sti_drv.h +++ b/drivers/gpu/drm/sti/sti_drv.h @@ -34,7 +34,6 @@ struct sti_private { }; extern struct platform_driver sti_tvout_driver; -extern struct platform_driver sti_vtac_driver; extern struct platform_driver sti_hqvdp_driver; extern struct platform_driver sti_hdmi_driver; extern struct platform_driver sti_hda_driver; diff --git a/drivers/gpu/drm/sti/sti_hdmi.c b/drivers/gpu/drm/sti/sti_hdmi.c index 4a8bd620b90c..c9151849d604 100644 --- a/drivers/gpu/drm/sti/sti_hdmi.c +++ b/drivers/gpu/drm/sti/sti_hdmi.c @@ -781,6 +781,95 @@ static void sti_hdmi_disable(struct drm_bridge *bridge) hdmi->enabled = false; } +/** + * sti_hdmi_audio_get_non_coherent_n() - get N parameter for non-coherent + * clocks. None-coherent clocks means that audio and TMDS clocks have not the + * same source (drifts between clocks). In this case assumption is that CTS is + * automatically calculated by hardware. + * + * @audio_fs: audio frame clock frequency in Hz + * + * Values computed are based on table described in HDMI specification 1.4b + * + * Returns n value. + */ +static int sti_hdmi_audio_get_non_coherent_n(unsigned int audio_fs) +{ + unsigned int n; + + switch (audio_fs) { + case 32000: + n = 4096; + break; + case 44100: + n = 6272; + break; + case 48000: + n = 6144; + break; + case 88200: + n = 6272 * 2; + break; + case 96000: + n = 6144 * 2; + break; + case 176400: + n = 6272 * 4; + break; + case 192000: + n = 6144 * 4; + break; + default: + /* Not pre-defined, recommended value: 128 * fs / 1000 */ + n = (audio_fs * 128) / 1000; + } + + return n; +} + +static int hdmi_audio_configure(struct sti_hdmi *hdmi) +{ + int audio_cfg, n; + struct hdmi_audio_params *params = &hdmi->audio; + struct hdmi_audio_infoframe *info = ¶ms->cea; + + DRM_DEBUG_DRIVER("\n"); + + if (!hdmi->enabled) + return 0; + + /* update N parameter */ + n = sti_hdmi_audio_get_non_coherent_n(params->sample_rate); + + DRM_DEBUG_DRIVER("Audio rate = %d Hz, TMDS clock = %d Hz, n = %d\n", + params->sample_rate, hdmi->mode.clock * 1000, n); + hdmi_write(hdmi, n, HDMI_AUDN); + + /* update HDMI registers according to configuration */ + audio_cfg = HDMI_AUD_CFG_SPDIF_DIV_2 | HDMI_AUD_CFG_DTS_INVALID | + HDMI_AUD_CFG_ONE_BIT_INVALID; + + switch (info->channels) { + case 8: + audio_cfg |= HDMI_AUD_CFG_CH78_VALID; + case 6: + audio_cfg |= HDMI_AUD_CFG_CH56_VALID; + case 4: + audio_cfg |= HDMI_AUD_CFG_CH34_VALID | HDMI_AUD_CFG_8CH; + case 2: + audio_cfg |= HDMI_AUD_CFG_CH12_VALID; + break; + default: + DRM_ERROR("ERROR: Unsupported number of channels (%d)!\n", + info->channels); + return -EINVAL; + } + + hdmi_write(hdmi, audio_cfg, HDMI_AUDIO_CFG); + + return hdmi_audio_infoframe_config(hdmi); +} + static void sti_hdmi_pre_enable(struct drm_bridge *bridge) { struct sti_hdmi *hdmi = bridge->driver_private; @@ -819,9 +908,12 @@ static void sti_hdmi_pre_enable(struct drm_bridge *bridge) if (hdmi_avi_infoframe_config(hdmi)) DRM_ERROR("Unable to configure AVI infoframe\n"); - /* Program AUDIO infoframe */ - if (hdmi_audio_infoframe_config(hdmi)) - DRM_ERROR("Unable to configure AUDIO infoframe\n"); + if (hdmi->audio.enabled) { + if (hdmi_audio_configure(hdmi)) + DRM_ERROR("Unable to configure audio\n"); + } else { + hdmi_audio_infoframe_config(hdmi); + } /* Program VS infoframe */ if (hdmi_vendor_infoframe_config(hdmi)) @@ -1071,97 +1163,6 @@ static struct drm_encoder *sti_hdmi_find_encoder(struct drm_device *dev) return NULL; } -/** - * sti_hdmi_audio_get_non_coherent_n() - get N parameter for non-coherent - * clocks. None-coherent clocks means that audio and TMDS clocks have not the - * same source (drifts between clocks). In this case assumption is that CTS is - * automatically calculated by hardware. - * - * @audio_fs: audio frame clock frequency in Hz - * - * Values computed are based on table described in HDMI specification 1.4b - * - * Returns n value. - */ -static int sti_hdmi_audio_get_non_coherent_n(unsigned int audio_fs) -{ - unsigned int n; - - switch (audio_fs) { - case 32000: - n = 4096; - break; - case 44100: - n = 6272; - break; - case 48000: - n = 6144; - break; - case 88200: - n = 6272 * 2; - break; - case 96000: - n = 6144 * 2; - break; - case 176400: - n = 6272 * 4; - break; - case 192000: - n = 6144 * 4; - break; - default: - /* Not pre-defined, recommended value: 128 * fs / 1000 */ - n = (audio_fs * 128) / 1000; - } - - return n; -} - -static int hdmi_audio_configure(struct sti_hdmi *hdmi, - struct hdmi_audio_params *params) -{ - int audio_cfg, n; - struct hdmi_audio_infoframe *info = ¶ms->cea; - - DRM_DEBUG_DRIVER("\n"); - - if (!hdmi->enabled) - return 0; - - /* update N parameter */ - n = sti_hdmi_audio_get_non_coherent_n(params->sample_rate); - - DRM_DEBUG_DRIVER("Audio rate = %d Hz, TMDS clock = %d Hz, n = %d\n", - params->sample_rate, hdmi->mode.clock * 1000, n); - hdmi_write(hdmi, n, HDMI_AUDN); - - /* update HDMI registers according to configuration */ - audio_cfg = HDMI_AUD_CFG_SPDIF_DIV_2 | HDMI_AUD_CFG_DTS_INVALID | - HDMI_AUD_CFG_ONE_BIT_INVALID; - - switch (info->channels) { - case 8: - audio_cfg |= HDMI_AUD_CFG_CH78_VALID; - case 6: - audio_cfg |= HDMI_AUD_CFG_CH56_VALID; - case 4: - audio_cfg |= HDMI_AUD_CFG_CH34_VALID | HDMI_AUD_CFG_8CH; - case 2: - audio_cfg |= HDMI_AUD_CFG_CH12_VALID; - break; - default: - DRM_ERROR("ERROR: Unsupported number of channels (%d)!\n", - info->channels); - return -EINVAL; - } - - hdmi_write(hdmi, audio_cfg, HDMI_AUDIO_CFG); - - hdmi->audio = *params; - - return hdmi_audio_infoframe_config(hdmi); -} - static void hdmi_audio_shutdown(struct device *dev, void *data) { struct sti_hdmi *hdmi = dev_get_drvdata(dev); @@ -1185,17 +1186,9 @@ static int hdmi_audio_hw_params(struct device *dev, { struct sti_hdmi *hdmi = dev_get_drvdata(dev); int ret; - struct hdmi_audio_params audio = { - .sample_width = params->sample_width, - .sample_rate = params->sample_rate, - .cea = params->cea, - }; DRM_DEBUG_DRIVER("\n"); - if (!hdmi->enabled) - return 0; - if ((daifmt->fmt != HDMI_I2S) || daifmt->bit_clk_inv || daifmt->frame_clk_inv || daifmt->bit_clk_master || daifmt->frame_clk_master) { @@ -1206,9 +1199,13 @@ static int hdmi_audio_hw_params(struct device *dev, return -EINVAL; } - audio.enabled = true; + hdmi->audio.sample_width = params->sample_width; + hdmi->audio.sample_rate = params->sample_rate; + hdmi->audio.cea = params->cea; + + hdmi->audio.enabled = true; - ret = hdmi_audio_configure(hdmi, &audio); + ret = hdmi_audio_configure(hdmi); if (ret < 0) return ret; diff --git a/drivers/gpu/drm/sti/sti_hqvdp.c b/drivers/gpu/drm/sti/sti_hqvdp.c index becf10d255c4..4376fd8a8e52 100644 --- a/drivers/gpu/drm/sti/sti_hqvdp.c +++ b/drivers/gpu/drm/sti/sti_hqvdp.c @@ -332,6 +332,7 @@ struct sti_hqvdp_cmd { * @hqvdp_cmd_paddr: physical address of hqvdp_cmd * @vtg: vtg for main data path * @xp70_initialized: true if xp70 is already initialized + * @vtg_registered: true if registered to VTG */ struct sti_hqvdp { struct device *dev; @@ -347,6 +348,7 @@ struct sti_hqvdp { u32 hqvdp_cmd_paddr; struct sti_vtg *vtg; bool xp70_initialized; + bool vtg_registered; }; #define to_sti_hqvdp(x) container_of(x, struct sti_hqvdp, plane) @@ -771,7 +773,7 @@ static void sti_hqvdp_disable(struct sti_hqvdp *hqvdp) DRM_ERROR("XP70 could not revert to idle\n"); hqvdp->plane.status = STI_PLANE_DISABLED; - hqvdp->xp70_initialized = false; + hqvdp->vtg_registered = false; } /** @@ -1064,10 +1066,11 @@ static int sti_hqvdp_atomic_check(struct drm_plane *drm_plane, return -EINVAL; } - if (!hqvdp->xp70_initialized) { + if (!hqvdp->xp70_initialized) /* Start HQVDP XP70 coprocessor */ sti_hqvdp_start_xp70(hqvdp); + if (!hqvdp->vtg_registered) { /* Prevent VTG shutdown */ if (clk_prepare_enable(hqvdp->clk_pix_main)) { DRM_ERROR("Failed to prepare/enable pix main clk\n"); @@ -1081,6 +1084,7 @@ static int sti_hqvdp_atomic_check(struct drm_plane *drm_plane, DRM_ERROR("Cannot register VTG notifier\n"); return -EINVAL; } + hqvdp->vtg_registered = true; } DRM_DEBUG_KMS("CRTC:%d (%s) drm plane:%d (%s)\n", @@ -1113,6 +1117,21 @@ static void sti_hqvdp_atomic_update(struct drm_plane *drm_plane, if (!crtc || !fb) return; + if ((oldstate->fb == state->fb) && + (oldstate->crtc_x == state->crtc_x) && + (oldstate->crtc_y == state->crtc_y) && + (oldstate->crtc_w == state->crtc_w) && + (oldstate->crtc_h == state->crtc_h) && + (oldstate->src_x == state->src_x) && + (oldstate->src_y == state->src_y) && + (oldstate->src_w == state->src_w) && + (oldstate->src_h == state->src_h)) { + /* No change since last update, do not post cmd */ + DRM_DEBUG_DRIVER("No change, not posting cmd\n"); + plane->status = STI_PLANE_UPDATED; + return; + } + mode = &crtc->mode; dst_x = state->crtc_x; dst_y = state->crtc_y; diff --git a/drivers/gpu/drm/sti/sti_plane.c b/drivers/gpu/drm/sti/sti_plane.c index ca4b3719a64a..427d8f58c6b1 100644 --- a/drivers/gpu/drm/sti/sti_plane.c +++ b/drivers/gpu/drm/sti/sti_plane.c @@ -65,9 +65,18 @@ void sti_plane_update_fps(struct sti_plane *plane, fps->last_timestamp = now; fps->last_frame_counter = fps->curr_frame_counter; - fpks = (num_frames * 1000000) / ms_since_last; - snprintf(plane->fps_info.fps_str, FPS_LENGTH, "%-6s @ %d.%.3d fps", - sti_plane_to_str(plane), fpks / 1000, fpks % 1000); + + if (plane->drm_plane.fb) { + fpks = (num_frames * 1000000) / ms_since_last; + snprintf(plane->fps_info.fps_str, FPS_LENGTH, + "%-8s %4dx%-4d %.4s @ %3d.%-3.3d fps (%s)", + plane->drm_plane.name, + plane->drm_plane.fb->width, + plane->drm_plane.fb->height, + (char *)&plane->drm_plane.fb->format->format, + fpks / 1000, fpks % 1000, + sti_plane_to_str(plane)); + } if (fps->curr_field_counter) { /* Compute number of field updates */ @@ -75,7 +84,7 @@ void sti_plane_update_fps(struct sti_plane *plane, fps->last_field_counter = fps->curr_field_counter; fipks = (num_fields * 1000000) / ms_since_last; snprintf(plane->fps_info.fips_str, - FPS_LENGTH, " - %d.%.3d field/sec", + FPS_LENGTH, " - %3d.%-3.3d field/sec", fipks / 1000, fipks % 1000); } else { plane->fps_info.fips_str[0] = '\0'; diff --git a/drivers/gpu/drm/sti/sti_plane.h b/drivers/gpu/drm/sti/sti_plane.h index ce3e8d6c88bb..c36c13faaa18 100644 --- a/drivers/gpu/drm/sti/sti_plane.h +++ b/drivers/gpu/drm/sti/sti_plane.h @@ -48,7 +48,7 @@ enum sti_plane_status { STI_PLANE_DISABLED, }; -#define FPS_LENGTH 64 +#define FPS_LENGTH 128 struct sti_fps_info { bool output; unsigned int curr_frame_counter; diff --git a/drivers/gpu/drm/sti/sti_vtac.c b/drivers/gpu/drm/sti/sti_vtac.c deleted file mode 100644 index cf7fe8a1db42..000000000000 --- a/drivers/gpu/drm/sti/sti_vtac.c +++ /dev/null @@ -1,223 +0,0 @@ -/* - * Copyright (C) STMicroelectronics SA 2014 - * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics. - * License terms: GNU General Public License (GPL), version 2 - */ - -#include <linux/clk.h> -#include <linux/io.h> -#include <linux/module.h> -#include <linux/of.h> -#include <linux/platform_device.h> - -#include <drm/drmP.h> - -#include "sti_drv.h" - -/* registers offset */ -#define VTAC_CONFIG 0x00 -#define VTAC_RX_FIFO_CONFIG 0x04 -#define VTAC_FIFO_CONFIG_VAL 0x04 - -#define VTAC_SYS_CFG8521 0x824 -#define VTAC_SYS_CFG8522 0x828 - -/* Number of phyts per pixel */ -#define VTAC_2_5_PPP 0x0005 -#define VTAC_3_PPP 0x0006 -#define VTAC_4_PPP 0x0008 -#define VTAC_5_PPP 0x000A -#define VTAC_6_PPP 0x000C -#define VTAC_13_PPP 0x001A -#define VTAC_14_PPP 0x001C -#define VTAC_15_PPP 0x001E -#define VTAC_16_PPP 0x0020 -#define VTAC_17_PPP 0x0022 -#define VTAC_18_PPP 0x0024 - -/* enable bits */ -#define VTAC_ENABLE 0x3003 - -#define VTAC_TX_PHY_ENABLE_CLK_PHY BIT(0) -#define VTAC_TX_PHY_ENABLE_CLK_DLL BIT(1) -#define VTAC_TX_PHY_PLL_NOT_OSC_MODE BIT(3) -#define VTAC_TX_PHY_RST_N_DLL_SWITCH BIT(4) -#define VTAC_TX_PHY_PROG_N3 BIT(9) - - -/** - * VTAC mode structure - * - * @vid_in_width: Video Data Resolution - * @phyts_width: Width of phyt buses(phyt low and phyt high). - * @phyts_per_pixel: Number of phyts sent per pixel - */ -struct sti_vtac_mode { - u32 vid_in_width; - u32 phyts_width; - u32 phyts_per_pixel; -}; - -static const struct sti_vtac_mode vtac_mode_main = { - .vid_in_width = 0x2, - .phyts_width = 0x2, - .phyts_per_pixel = VTAC_5_PPP, -}; -static const struct sti_vtac_mode vtac_mode_aux = { - .vid_in_width = 0x1, - .phyts_width = 0x0, - .phyts_per_pixel = VTAC_17_PPP, -}; - -/** - * VTAC structure - * - * @dev: pointer to device structure - * @regs: ioremapped registers for RX and TX devices - * @phy_regs: phy registers for TX device - * @clk: clock - * @mode: main or auxillary configuration mode - */ -struct sti_vtac { - struct device *dev; - void __iomem *regs; - void __iomem *phy_regs; - struct clk *clk; - const struct sti_vtac_mode *mode; -}; - -static void sti_vtac_rx_set_config(struct sti_vtac *vtac) -{ - u32 config; - - /* Enable VTAC clock */ - if (clk_prepare_enable(vtac->clk)) - DRM_ERROR("Failed to prepare/enable vtac_rx clock.\n"); - - writel(VTAC_FIFO_CONFIG_VAL, vtac->regs + VTAC_RX_FIFO_CONFIG); - - config = VTAC_ENABLE; - config |= vtac->mode->vid_in_width << 4; - config |= vtac->mode->phyts_width << 16; - config |= vtac->mode->phyts_per_pixel << 23; - writel(config, vtac->regs + VTAC_CONFIG); -} - -static void sti_vtac_tx_set_config(struct sti_vtac *vtac) -{ - u32 phy_config; - u32 config; - - /* Enable VTAC clock */ - if (clk_prepare_enable(vtac->clk)) - DRM_ERROR("Failed to prepare/enable vtac_tx clock.\n"); - - /* Configure vtac phy */ - phy_config = 0x00000000; - writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8522); - phy_config = VTAC_TX_PHY_ENABLE_CLK_PHY; - writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config |= VTAC_TX_PHY_PROG_N3; - writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config |= VTAC_TX_PHY_ENABLE_CLK_DLL; - writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config |= VTAC_TX_PHY_RST_N_DLL_SWITCH; - writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521); - phy_config |= VTAC_TX_PHY_PLL_NOT_OSC_MODE; - writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521); - - /* Configure vtac tx */ - config = VTAC_ENABLE; - config |= vtac->mode->vid_in_width << 4; - config |= vtac->mode->phyts_width << 16; - config |= vtac->mode->phyts_per_pixel << 23; - writel(config, vtac->regs + VTAC_CONFIG); -} - -static const struct of_device_id vtac_of_match[] = { - { - .compatible = "st,vtac-main", - .data = &vtac_mode_main, - }, { - .compatible = "st,vtac-aux", - .data = &vtac_mode_aux, - }, { - /* end node */ - } -}; -MODULE_DEVICE_TABLE(of, vtac_of_match); - -static int sti_vtac_probe(struct platform_device *pdev) -{ - struct device *dev = &pdev->dev; - struct device_node *np = dev->of_node; - const struct of_device_id *id; - struct sti_vtac *vtac; - struct resource *res; - - vtac = devm_kzalloc(dev, sizeof(*vtac), GFP_KERNEL); - if (!vtac) - return -ENOMEM; - - vtac->dev = dev; - - id = of_match_node(vtac_of_match, np); - if (!id) - return -ENOMEM; - - vtac->mode = id->data; - - res = platform_get_resource(pdev, IORESOURCE_MEM, 0); - if (!res) { - DRM_ERROR("Invalid resource\n"); - return -ENOMEM; - } - vtac->regs = devm_ioremap_resource(dev, res); - if (IS_ERR(vtac->regs)) - return PTR_ERR(vtac->regs); - - - vtac->clk = devm_clk_get(dev, "vtac"); - if (IS_ERR(vtac->clk)) { - DRM_ERROR("Cannot get vtac clock\n"); - return PTR_ERR(vtac->clk); - } - - res = platform_get_resource(pdev, IORESOURCE_MEM, 1); - if (res) { - vtac->phy_regs = devm_ioremap_nocache(dev, res->start, - resource_size(res)); - sti_vtac_tx_set_config(vtac); - } else { - - sti_vtac_rx_set_config(vtac); - } - - platform_set_drvdata(pdev, vtac); - DRM_INFO("%s %s\n", __func__, dev_name(vtac->dev)); - - return 0; -} - -static int sti_vtac_remove(struct platform_device *pdev) -{ - return 0; -} - -struct platform_driver sti_vtac_driver = { - .driver = { - .name = "sti-vtac", - .owner = THIS_MODULE, - .of_match_table = vtac_of_match, - }, - .probe = sti_vtac_probe, - .remove = sti_vtac_remove, -}; - -MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>"); -MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver"); -MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/sti/sti_vtg.c b/drivers/gpu/drm/sti/sti_vtg.c index a8882bdd0f8b..c3d9c8ae14af 100644 --- a/drivers/gpu/drm/sti/sti_vtg.c +++ b/drivers/gpu/drm/sti/sti_vtg.c @@ -429,6 +429,10 @@ static int vtg_probe(struct platform_device *pdev) return -ENOMEM; } vtg->regs = devm_ioremap_nocache(dev, res->start, resource_size(res)); + if (!vtg->regs) { + DRM_ERROR("failed to remap I/O memory\n"); + return -ENOMEM; + } np = of_parse_phandle(pdev->dev.of_node, "st,slave", 0); if (np) { diff --git a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c index 3a763f7cb743..f80bf9385e41 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c @@ -856,7 +856,7 @@ irqreturn_t tilcdc_crtc_irq(struct drm_crtc *crtc) struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc); struct drm_device *dev = crtc->dev; struct tilcdc_drm_private *priv = dev->dev_private; - uint32_t stat; + uint32_t stat, reg; stat = tilcdc_read_irqstatus(dev); tilcdc_clear_irqstatus(dev, stat); @@ -921,17 +921,26 @@ irqreturn_t tilcdc_crtc_irq(struct drm_crtc *crtc) dev_err_ratelimited(dev->dev, "%s(0x%08x): Sync lost", __func__, stat); tilcdc_crtc->frame_intact = false; - if (tilcdc_crtc->sync_lost_count++ > - SYNC_LOST_COUNT_LIMIT) { - dev_err(dev->dev, "%s(0x%08x): Sync lost flood detected, recovering", __func__, stat); - queue_work(system_wq, &tilcdc_crtc->recover_work); - if (priv->rev == 1) + if (priv->rev == 1) { + reg = tilcdc_read(dev, LCDC_RASTER_CTRL_REG); + if (reg & LCDC_RASTER_ENABLE) { tilcdc_clear(dev, LCDC_RASTER_CTRL_REG, - LCDC_V1_SYNC_LOST_INT_ENA); - else + LCDC_RASTER_ENABLE); + tilcdc_set(dev, LCDC_RASTER_CTRL_REG, + LCDC_RASTER_ENABLE); + } + } else { + if (tilcdc_crtc->sync_lost_count++ > + SYNC_LOST_COUNT_LIMIT) { + dev_err(dev->dev, + "%s(0x%08x): Sync lost flood detected, recovering", + __func__, stat); + queue_work(system_wq, + &tilcdc_crtc->recover_work); tilcdc_write(dev, LCDC_INT_ENABLE_CLR_REG, LCDC_SYNC_LOST); - tilcdc_crtc->sync_lost_count = 0; + tilcdc_crtc->sync_lost_count = 0; + } } } diff --git a/drivers/gpu/drm/virtio/virtgpu_fb.c b/drivers/gpu/drm/virtio/virtgpu_fb.c index 61254b991265..24f99fc9d8a4 100644 --- a/drivers/gpu/drm/virtio/virtgpu_fb.c +++ b/drivers/gpu/drm/virtio/virtgpu_fb.c @@ -331,7 +331,7 @@ static int virtio_gpufb_create(struct drm_fb_helper *helper, info->fbops = &virtio_gpufb_ops; info->pixmap.flags = FB_PIXMAP_SYSTEM; - info->screen_base = obj->vmap; + info->screen_buffer = obj->vmap; info->screen_size = obj->gem_base.size; drm_fb_helper_fill_fix(info, fb->pitches[0], fb->format->depth); drm_fb_helper_fill_var(info, &vfbdev->helper, diff --git a/drivers/gpu/drm/zte/Kconfig b/drivers/gpu/drm/zte/Kconfig index 4065b2840f1c..5b36421ef3e5 100644 --- a/drivers/gpu/drm/zte/Kconfig +++ b/drivers/gpu/drm/zte/Kconfig @@ -4,5 +4,7 @@ config DRM_ZTE select DRM_KMS_CMA_HELPER select DRM_KMS_FB_HELPER select DRM_KMS_HELPER + select SND_SOC_HDMI_CODEC if SND_SOC + select VIDEOMODE_HELPERS help Choose this option to enable DRM on ZTE ZX SoCs. diff --git a/drivers/gpu/drm/zte/Makefile b/drivers/gpu/drm/zte/Makefile index 699180bfd57c..01352b56c418 100644 --- a/drivers/gpu/drm/zte/Makefile +++ b/drivers/gpu/drm/zte/Makefile @@ -2,6 +2,7 @@ zxdrm-y := \ zx_drm_drv.o \ zx_hdmi.o \ zx_plane.o \ + zx_tvenc.o \ zx_vou.o obj-$(CONFIG_DRM_ZTE) += zxdrm.o diff --git a/drivers/gpu/drm/zte/zx_drm_drv.c b/drivers/gpu/drm/zte/zx_drm_drv.c index 3e76f72c92ff..13081fed902d 100644 --- a/drivers/gpu/drm/zte/zx_drm_drv.c +++ b/drivers/gpu/drm/zte/zx_drm_drv.c @@ -247,6 +247,7 @@ static struct platform_driver zx_drm_platform_driver = { static struct platform_driver *drivers[] = { &zx_crtc_driver, &zx_hdmi_driver, + &zx_tvenc_driver, &zx_drm_platform_driver, }; diff --git a/drivers/gpu/drm/zte/zx_drm_drv.h b/drivers/gpu/drm/zte/zx_drm_drv.h index e65cd18a6cba..5ca035b079c7 100644 --- a/drivers/gpu/drm/zte/zx_drm_drv.h +++ b/drivers/gpu/drm/zte/zx_drm_drv.h @@ -13,6 +13,7 @@ extern struct platform_driver zx_crtc_driver; extern struct platform_driver zx_hdmi_driver; +extern struct platform_driver zx_tvenc_driver; static inline u32 zx_readl(void __iomem *reg) { diff --git a/drivers/gpu/drm/zte/zx_hdmi.c b/drivers/gpu/drm/zte/zx_hdmi.c index 6bf6c364811e..c47b9cbfe270 100644 --- a/drivers/gpu/drm/zte/zx_hdmi.c +++ b/drivers/gpu/drm/zte/zx_hdmi.c @@ -25,6 +25,8 @@ #include <drm/drm_of.h> #include <drm/drmP.h> +#include <sound/hdmi-codec.h> + #include "zx_hdmi_regs.h" #include "zx_vou.h" @@ -49,17 +51,11 @@ struct zx_hdmi { bool sink_is_hdmi; bool sink_has_audio; const struct vou_inf *inf; + struct platform_device *audio_pdev; }; #define to_zx_hdmi(x) container_of(x, struct zx_hdmi, x) -static const struct vou_inf vou_inf_hdmi = { - .id = VOU_HDMI, - .data_sel = VOU_YUV444, - .clocks_en_bits = BIT(24) | BIT(18) | BIT(6), - .clocks_sel_bits = BIT(13) | BIT(2), -}; - static inline u8 hdmi_readb(struct zx_hdmi *hdmi, u16 offset) { return readl_relaxed(hdmi->mmio + offset * 4); @@ -238,14 +234,14 @@ static void zx_hdmi_encoder_enable(struct drm_encoder *encoder) zx_hdmi_hw_enable(hdmi); - vou_inf_enable(hdmi->inf, encoder->crtc); + vou_inf_enable(VOU_HDMI, encoder->crtc); } static void zx_hdmi_encoder_disable(struct drm_encoder *encoder) { struct zx_hdmi *hdmi = to_zx_hdmi(encoder); - vou_inf_disable(hdmi->inf, encoder->crtc); + vou_inf_disable(VOU_HDMI, encoder->crtc); zx_hdmi_hw_disable(hdmi); @@ -366,6 +362,142 @@ static irqreturn_t zx_hdmi_irq_handler(int irq, void *dev_id) return IRQ_NONE; } +static int zx_hdmi_audio_startup(struct device *dev, void *data) +{ + struct zx_hdmi *hdmi = dev_get_drvdata(dev); + struct drm_encoder *encoder = &hdmi->encoder; + + vou_inf_hdmi_audio_sel(encoder->crtc, VOU_HDMI_AUD_SPDIF); + + return 0; +} + +static void zx_hdmi_audio_shutdown(struct device *dev, void *data) +{ + struct zx_hdmi *hdmi = dev_get_drvdata(dev); + + /* Disable audio input */ + hdmi_writeb_mask(hdmi, AUD_EN, AUD_IN_EN, 0); +} + +static inline int zx_hdmi_audio_get_n(unsigned int fs) +{ + unsigned int n; + + if (fs && (fs % 44100) == 0) + n = 6272 * (fs / 44100); + else + n = fs * 128 / 1000; + + return n; +} + +static int zx_hdmi_audio_hw_params(struct device *dev, + void *data, + struct hdmi_codec_daifmt *daifmt, + struct hdmi_codec_params *params) +{ + struct zx_hdmi *hdmi = dev_get_drvdata(dev); + struct hdmi_audio_infoframe *cea = ¶ms->cea; + union hdmi_infoframe frame; + int n; + + /* We only support spdif for now */ + if (daifmt->fmt != HDMI_SPDIF) { + DRM_DEV_ERROR(hdmi->dev, "invalid daifmt %d\n", daifmt->fmt); + return -EINVAL; + } + + switch (params->sample_width) { + case 16: + hdmi_writeb_mask(hdmi, TPI_AUD_CONFIG, SPDIF_SAMPLE_SIZE_MASK, + SPDIF_SAMPLE_SIZE_16BIT); + break; + case 20: + hdmi_writeb_mask(hdmi, TPI_AUD_CONFIG, SPDIF_SAMPLE_SIZE_MASK, + SPDIF_SAMPLE_SIZE_20BIT); + break; + case 24: + hdmi_writeb_mask(hdmi, TPI_AUD_CONFIG, SPDIF_SAMPLE_SIZE_MASK, + SPDIF_SAMPLE_SIZE_24BIT); + break; + default: + DRM_DEV_ERROR(hdmi->dev, "invalid sample width %d\n", + params->sample_width); + return -EINVAL; + } + + /* CTS is calculated by hardware, and we only need to take care of N */ + n = zx_hdmi_audio_get_n(params->sample_rate); + hdmi_writeb(hdmi, N_SVAL1, n & 0xff); + hdmi_writeb(hdmi, N_SVAL2, (n >> 8) & 0xff); + hdmi_writeb(hdmi, N_SVAL3, (n >> 16) & 0xf); + + /* Enable spdif mode */ + hdmi_writeb_mask(hdmi, AUD_MODE, SPDIF_EN, SPDIF_EN); + + /* Enable audio input */ + hdmi_writeb_mask(hdmi, AUD_EN, AUD_IN_EN, AUD_IN_EN); + + memcpy(&frame.audio, cea, sizeof(*cea)); + + return zx_hdmi_infoframe_trans(hdmi, &frame, FSEL_AUDIO); +} + +static int zx_hdmi_audio_digital_mute(struct device *dev, void *data, + bool enable) +{ + struct zx_hdmi *hdmi = dev_get_drvdata(dev); + + if (enable) + hdmi_writeb_mask(hdmi, TPI_AUD_CONFIG, TPI_AUD_MUTE, + TPI_AUD_MUTE); + else + hdmi_writeb_mask(hdmi, TPI_AUD_CONFIG, TPI_AUD_MUTE, 0); + + return 0; +} + +static int zx_hdmi_audio_get_eld(struct device *dev, void *data, + uint8_t *buf, size_t len) +{ + struct zx_hdmi *hdmi = dev_get_drvdata(dev); + struct drm_connector *connector = &hdmi->connector; + + memcpy(buf, connector->eld, min(sizeof(connector->eld), len)); + + return 0; +} + +static const struct hdmi_codec_ops zx_hdmi_codec_ops = { + .audio_startup = zx_hdmi_audio_startup, + .hw_params = zx_hdmi_audio_hw_params, + .audio_shutdown = zx_hdmi_audio_shutdown, + .digital_mute = zx_hdmi_audio_digital_mute, + .get_eld = zx_hdmi_audio_get_eld, +}; + +static struct hdmi_codec_pdata zx_hdmi_codec_pdata = { + .ops = &zx_hdmi_codec_ops, + .spdif = 1, +}; + +static int zx_hdmi_audio_register(struct zx_hdmi *hdmi) +{ + struct platform_device *pdev; + + pdev = platform_device_register_data(hdmi->dev, HDMI_CODEC_DRV_NAME, + PLATFORM_DEVID_AUTO, + &zx_hdmi_codec_pdata, + sizeof(zx_hdmi_codec_pdata)); + if (IS_ERR(pdev)) + return PTR_ERR(pdev); + + hdmi->audio_pdev = pdev; + + return 0; +} + static int zx_hdmi_i2c_read(struct zx_hdmi *hdmi, struct i2c_msg *msg) { int len = msg->len; @@ -523,7 +655,6 @@ static int zx_hdmi_bind(struct device *dev, struct device *master, void *data) hdmi->dev = dev; hdmi->drm = drm; - hdmi->inf = &vou_inf_hdmi; dev_set_drvdata(dev, hdmi); @@ -566,6 +697,12 @@ static int zx_hdmi_bind(struct device *dev, struct device *master, void *data) return ret; } + ret = zx_hdmi_audio_register(hdmi); + if (ret) { + DRM_DEV_ERROR(dev, "failed to register audio: %d\n", ret); + return ret; + } + ret = zx_hdmi_register(drm, hdmi); if (ret) { DRM_DEV_ERROR(dev, "failed to register hdmi: %d\n", ret); @@ -590,6 +727,9 @@ static void zx_hdmi_unbind(struct device *dev, struct device *master, hdmi->connector.funcs->destroy(&hdmi->connector); hdmi->encoder.funcs->destroy(&hdmi->encoder); + + if (hdmi->audio_pdev) + platform_device_unregister(hdmi->audio_pdev); } static const struct component_ops zx_hdmi_component_ops = { diff --git a/drivers/gpu/drm/zte/zx_hdmi_regs.h b/drivers/gpu/drm/zte/zx_hdmi_regs.h index de911f66b658..c6d5d8211725 100644 --- a/drivers/gpu/drm/zte/zx_hdmi_regs.h +++ b/drivers/gpu/drm/zte/zx_hdmi_regs.h @@ -52,5 +52,19 @@ #define TPI_INFO_TRANS_RPT BIT(6) #define TPI_DDC_MASTER_EN 0x06f8 #define HW_DDC_MASTER BIT(7) +#define N_SVAL1 0xa03 +#define N_SVAL2 0xa04 +#define N_SVAL3 0xa05 +#define AUD_EN 0xa13 +#define AUD_IN_EN BIT(0) +#define AUD_MODE 0xa14 +#define SPDIF_EN BIT(1) +#define TPI_AUD_CONFIG 0xa62 +#define SPDIF_SAMPLE_SIZE_SHIFT 6 +#define SPDIF_SAMPLE_SIZE_MASK (0x3 << SPDIF_SAMPLE_SIZE_SHIFT) +#define SPDIF_SAMPLE_SIZE_16BIT (0x1 << SPDIF_SAMPLE_SIZE_SHIFT) +#define SPDIF_SAMPLE_SIZE_20BIT (0x2 << SPDIF_SAMPLE_SIZE_SHIFT) +#define SPDIF_SAMPLE_SIZE_24BIT (0x3 << SPDIF_SAMPLE_SIZE_SHIFT) +#define TPI_AUD_MUTE BIT(4) #endif /* __ZX_HDMI_REGS_H__ */ diff --git a/drivers/gpu/drm/zte/zx_plane.c b/drivers/gpu/drm/zte/zx_plane.c index b634b090cdc1..1d08ba381098 100644 --- a/drivers/gpu/drm/zte/zx_plane.c +++ b/drivers/gpu/drm/zte/zx_plane.c @@ -21,16 +21,6 @@ #include "zx_plane_regs.h" #include "zx_vou.h" -struct zx_plane { - struct drm_plane plane; - void __iomem *layer; - void __iomem *csc; - void __iomem *hbsc; - void __iomem *rsz; -}; - -#define to_zx_plane(plane) container_of(plane, struct zx_plane, plane) - static const uint32_t gl_formats[] = { DRM_FORMAT_ARGB8888, DRM_FORMAT_XRGB8888, @@ -40,6 +30,261 @@ static const uint32_t gl_formats[] = { DRM_FORMAT_ARGB4444, }; +static const uint32_t vl_formats[] = { + DRM_FORMAT_NV12, /* Semi-planar YUV420 */ + DRM_FORMAT_YUV420, /* Planar YUV420 */ + DRM_FORMAT_YUYV, /* Packed YUV422 */ + DRM_FORMAT_YVYU, + DRM_FORMAT_UYVY, + DRM_FORMAT_VYUY, + DRM_FORMAT_YUV444, /* YUV444 8bit */ + /* + * TODO: add formats below that HW supports: + * - YUV420 P010 + * - YUV420 Hantro + * - YUV444 10bit + */ +}; + +#define FRAC_16_16(mult, div) (((mult) << 16) / (div)) + +static int zx_vl_plane_atomic_check(struct drm_plane *plane, + struct drm_plane_state *plane_state) +{ + struct drm_framebuffer *fb = plane_state->fb; + struct drm_crtc *crtc = plane_state->crtc; + struct drm_crtc_state *crtc_state; + struct drm_rect clip; + int min_scale = FRAC_16_16(1, 8); + int max_scale = FRAC_16_16(8, 1); + + if (!crtc || !fb) + return 0; + + crtc_state = drm_atomic_get_existing_crtc_state(plane_state->state, + crtc); + if (WARN_ON(!crtc_state)) + return -EINVAL; + + /* nothing to check when disabling or disabled */ + if (!crtc_state->enable) + return 0; + + /* plane must be enabled */ + if (!plane_state->crtc) + return -EINVAL; + + clip.x1 = 0; + clip.y1 = 0; + clip.x2 = crtc_state->adjusted_mode.hdisplay; + clip.y2 = crtc_state->adjusted_mode.vdisplay; + + return drm_plane_helper_check_state(plane_state, &clip, + min_scale, max_scale, + true, true); +} + +static int zx_vl_get_fmt(uint32_t format) +{ + switch (format) { + case DRM_FORMAT_NV12: + return VL_FMT_YUV420; + case DRM_FORMAT_YUV420: + return VL_YUV420_PLANAR | VL_FMT_YUV420; + case DRM_FORMAT_YUYV: + return VL_YUV422_YUYV | VL_FMT_YUV422; + case DRM_FORMAT_YVYU: + return VL_YUV422_YVYU | VL_FMT_YUV422; + case DRM_FORMAT_UYVY: + return VL_YUV422_UYVY | VL_FMT_YUV422; + case DRM_FORMAT_VYUY: + return VL_YUV422_VYUY | VL_FMT_YUV422; + case DRM_FORMAT_YUV444: + return VL_FMT_YUV444_8BIT; + default: + WARN_ONCE(1, "invalid pixel format %d\n", format); + return -EINVAL; + } +} + +static inline void zx_vl_set_update(struct zx_plane *zplane) +{ + void __iomem *layer = zplane->layer; + + zx_writel_mask(layer + VL_CTRL0, VL_UPDATE, VL_UPDATE); +} + +static inline void zx_vl_rsz_set_update(struct zx_plane *zplane) +{ + zx_writel(zplane->rsz + RSZ_VL_ENABLE_CFG, 1); +} + +static int zx_vl_rsz_get_fmt(uint32_t format) +{ + switch (format) { + case DRM_FORMAT_NV12: + case DRM_FORMAT_YUV420: + return RSZ_VL_FMT_YCBCR420; + case DRM_FORMAT_YUYV: + case DRM_FORMAT_YVYU: + case DRM_FORMAT_UYVY: + case DRM_FORMAT_VYUY: + return RSZ_VL_FMT_YCBCR422; + case DRM_FORMAT_YUV444: + return RSZ_VL_FMT_YCBCR444; + default: + WARN_ONCE(1, "invalid pixel format %d\n", format); + return -EINVAL; + } +} + +static inline u32 rsz_step_value(u32 src, u32 dst) +{ + u32 val = 0; + + if (src == dst) + val = 0; + else if (src < dst) + val = RSZ_PARA_STEP((src << 16) / dst); + else if (src > dst) + val = RSZ_DATA_STEP(src / dst) | + RSZ_PARA_STEP(((src << 16) / dst) & 0xffff); + + return val; +} + +static void zx_vl_rsz_setup(struct zx_plane *zplane, uint32_t format, + u32 src_w, u32 src_h, u32 dst_w, u32 dst_h) +{ + void __iomem *rsz = zplane->rsz; + u32 src_chroma_w = src_w; + u32 src_chroma_h = src_h; + u32 fmt; + + /* Set up source and destination resolution */ + zx_writel(rsz + RSZ_SRC_CFG, RSZ_VER(src_h - 1) | RSZ_HOR(src_w - 1)); + zx_writel(rsz + RSZ_DEST_CFG, RSZ_VER(dst_h - 1) | RSZ_HOR(dst_w - 1)); + + /* Configure data format for VL RSZ */ + fmt = zx_vl_rsz_get_fmt(format); + if (fmt >= 0) + zx_writel_mask(rsz + RSZ_VL_CTRL_CFG, RSZ_VL_FMT_MASK, fmt); + + /* Calculate Chroma height and width */ + if (fmt == RSZ_VL_FMT_YCBCR420) { + src_chroma_w = src_w >> 1; + src_chroma_h = src_h >> 1; + } else if (fmt == RSZ_VL_FMT_YCBCR422) { + src_chroma_w = src_w >> 1; + } + + /* Set up Luma and Chroma step registers */ + zx_writel(rsz + RSZ_VL_LUMA_HOR, rsz_step_value(src_w, dst_w)); + zx_writel(rsz + RSZ_VL_LUMA_VER, rsz_step_value(src_h, dst_h)); + zx_writel(rsz + RSZ_VL_CHROMA_HOR, rsz_step_value(src_chroma_w, dst_w)); + zx_writel(rsz + RSZ_VL_CHROMA_VER, rsz_step_value(src_chroma_h, dst_h)); + + zx_vl_rsz_set_update(zplane); +} + +static void zx_vl_plane_atomic_update(struct drm_plane *plane, + struct drm_plane_state *old_state) +{ + struct zx_plane *zplane = to_zx_plane(plane); + struct drm_plane_state *state = plane->state; + struct drm_framebuffer *fb = state->fb; + struct drm_rect *src = &state->src; + struct drm_rect *dst = &state->dst; + struct drm_gem_cma_object *cma_obj; + void __iomem *layer = zplane->layer; + void __iomem *hbsc = zplane->hbsc; + void __iomem *paddr_reg; + dma_addr_t paddr; + u32 src_x, src_y, src_w, src_h; + u32 dst_x, dst_y, dst_w, dst_h; + uint32_t format; + u32 fmt; + int num_planes; + int i; + + if (!fb) + return; + + format = fb->format->format; + + src_x = src->x1 >> 16; + src_y = src->y1 >> 16; + src_w = drm_rect_width(src) >> 16; + src_h = drm_rect_height(src) >> 16; + + dst_x = dst->x1; + dst_y = dst->y1; + dst_w = drm_rect_width(dst); + dst_h = drm_rect_height(dst); + + /* Set up data address registers for Y, Cb and Cr planes */ + num_planes = drm_format_num_planes(format); + paddr_reg = layer + VL_Y; + for (i = 0; i < num_planes; i++) { + cma_obj = drm_fb_cma_get_gem_obj(fb, i); + paddr = cma_obj->paddr + fb->offsets[i]; + paddr += src_y * fb->pitches[i]; + paddr += src_x * drm_format_plane_cpp(format, i); + zx_writel(paddr_reg, paddr); + paddr_reg += 4; + } + + /* Set up source height/width register */ + zx_writel(layer + VL_SRC_SIZE, GL_SRC_W(src_w) | GL_SRC_H(src_h)); + + /* Set up start position register */ + zx_writel(layer + VL_POS_START, GL_POS_X(dst_x) | GL_POS_Y(dst_y)); + + /* Set up end position register */ + zx_writel(layer + VL_POS_END, + GL_POS_X(dst_x + dst_w) | GL_POS_Y(dst_y + dst_h)); + + /* Strides of Cb and Cr planes should be identical */ + zx_writel(layer + VL_STRIDE, LUMA_STRIDE(fb->pitches[0]) | + CHROMA_STRIDE(fb->pitches[1])); + + /* Set up video layer data format */ + fmt = zx_vl_get_fmt(format); + if (fmt >= 0) + zx_writel(layer + VL_CTRL1, fmt); + + /* Always use scaler since it exists (set for not bypass) */ + zx_writel_mask(layer + VL_CTRL2, VL_SCALER_BYPASS_MODE, + VL_SCALER_BYPASS_MODE); + + zx_vl_rsz_setup(zplane, format, src_w, src_h, dst_w, dst_h); + + /* Enable HBSC block */ + zx_writel_mask(hbsc + HBSC_CTRL0, HBSC_CTRL_EN, HBSC_CTRL_EN); + + zx_vou_layer_enable(plane); + + zx_vl_set_update(zplane); +} + +static void zx_plane_atomic_disable(struct drm_plane *plane, + struct drm_plane_state *old_state) +{ + struct zx_plane *zplane = to_zx_plane(plane); + void __iomem *hbsc = zplane->hbsc; + + zx_vou_layer_disable(plane); + + /* Disable HBSC block */ + zx_writel_mask(hbsc + HBSC_CTRL0, HBSC_CTRL_EN, 0); +} + +static const struct drm_plane_helper_funcs zx_vl_plane_helper_funcs = { + .atomic_check = zx_vl_plane_atomic_check, + .atomic_update = zx_vl_plane_atomic_update, + .atomic_disable = zx_plane_atomic_disable, +}; + static int zx_gl_plane_atomic_check(struct drm_plane *plane, struct drm_plane_state *plane_state) { @@ -107,14 +352,6 @@ static inline void zx_gl_rsz_set_update(struct zx_plane *zplane) zx_writel(zplane->rsz + RSZ_ENABLE_CFG, 1); } -void zx_plane_set_update(struct drm_plane *plane) -{ - struct zx_plane *zplane = to_zx_plane(plane); - - zx_gl_rsz_set_update(zplane); - zx_gl_set_update(zplane); -} - static void zx_gl_rsz_setup(struct zx_plane *zplane, u32 src_w, u32 src_h, u32 dst_w, u32 dst_h) { @@ -207,12 +444,15 @@ static void zx_gl_plane_atomic_update(struct drm_plane *plane, /* Enable HBSC block */ zx_writel_mask(hbsc + HBSC_CTRL0, HBSC_CTRL_EN, HBSC_CTRL_EN); + zx_vou_layer_enable(plane); + zx_gl_set_update(zplane); } static const struct drm_plane_helper_funcs zx_gl_plane_helper_funcs = { .atomic_check = zx_gl_plane_atomic_check, .atomic_update = zx_gl_plane_atomic_update, + .atomic_disable = zx_plane_atomic_disable, }; static void zx_plane_destroy(struct drm_plane *plane) @@ -230,6 +470,28 @@ static const struct drm_plane_funcs zx_plane_funcs = { .atomic_destroy_state = drm_atomic_helper_plane_destroy_state, }; +void zx_plane_set_update(struct drm_plane *plane) +{ + struct zx_plane *zplane = to_zx_plane(plane); + + /* Do nothing if the plane is not enabled */ + if (!plane->state->crtc) + return; + + switch (plane->type) { + case DRM_PLANE_TYPE_PRIMARY: + zx_gl_rsz_set_update(zplane); + zx_gl_set_update(zplane); + break; + case DRM_PLANE_TYPE_OVERLAY: + zx_vl_rsz_set_update(zplane); + zx_vl_set_update(zplane); + break; + default: + WARN_ONCE(1, "unsupported plane type %d\n", plane->type); + } +} + static void zx_plane_hbsc_init(struct zx_plane *zplane) { void __iomem *hbsc = zplane->hbsc; @@ -248,28 +510,16 @@ static void zx_plane_hbsc_init(struct zx_plane *zplane) zx_writel(hbsc + HBSC_THRESHOLD_COL3, (0x3c0 << 16) | 0x40); } -struct drm_plane *zx_plane_init(struct drm_device *drm, struct device *dev, - struct zx_layer_data *data, - enum drm_plane_type type) +int zx_plane_init(struct drm_device *drm, struct zx_plane *zplane, + enum drm_plane_type type) { const struct drm_plane_helper_funcs *helper; - struct zx_plane *zplane; - struct drm_plane *plane; + struct drm_plane *plane = &zplane->plane; + struct device *dev = zplane->dev; const uint32_t *formats; unsigned int format_count; int ret; - zplane = devm_kzalloc(dev, sizeof(*zplane), GFP_KERNEL); - if (!zplane) - return ERR_PTR(-ENOMEM); - - plane = &zplane->plane; - - zplane->layer = data->layer; - zplane->hbsc = data->hbsc; - zplane->csc = data->csc; - zplane->rsz = data->rsz; - zx_plane_hbsc_init(zplane); switch (type) { @@ -279,10 +529,12 @@ struct drm_plane *zx_plane_init(struct drm_device *drm, struct device *dev, format_count = ARRAY_SIZE(gl_formats); break; case DRM_PLANE_TYPE_OVERLAY: - /* TODO: add video layer (vl) support */ + helper = &zx_vl_plane_helper_funcs; + formats = vl_formats; + format_count = ARRAY_SIZE(vl_formats); break; default: - return ERR_PTR(-ENODEV); + return -ENODEV; } ret = drm_universal_plane_init(drm, plane, VOU_CRTC_MASK, @@ -290,10 +542,10 @@ struct drm_plane *zx_plane_init(struct drm_device *drm, struct device *dev, type, NULL); if (ret) { DRM_DEV_ERROR(dev, "failed to init universal plane: %d\n", ret); - return ERR_PTR(ret); + return ret; } drm_plane_helper_add(plane, helper); - return plane; + return 0; } diff --git a/drivers/gpu/drm/zte/zx_plane.h b/drivers/gpu/drm/zte/zx_plane.h index 2b82cd558d9d..933611ddffd0 100644 --- a/drivers/gpu/drm/zte/zx_plane.h +++ b/drivers/gpu/drm/zte/zx_plane.h @@ -11,16 +11,20 @@ #ifndef __ZX_PLANE_H__ #define __ZX_PLANE_H__ -struct zx_layer_data { +struct zx_plane { + struct drm_plane plane; + struct device *dev; void __iomem *layer; void __iomem *csc; void __iomem *hbsc; void __iomem *rsz; + const struct vou_layer_bits *bits; }; -struct drm_plane *zx_plane_init(struct drm_device *drm, struct device *dev, - struct zx_layer_data *data, - enum drm_plane_type type); +#define to_zx_plane(plane) container_of(plane, struct zx_plane, plane) + +int zx_plane_init(struct drm_device *drm, struct zx_plane *zplane, + enum drm_plane_type type); void zx_plane_set_update(struct drm_plane *plane); #endif /* __ZX_PLANE_H__ */ diff --git a/drivers/gpu/drm/zte/zx_plane_regs.h b/drivers/gpu/drm/zte/zx_plane_regs.h index 3dde6716a558..65f271aeabed 100644 --- a/drivers/gpu/drm/zte/zx_plane_regs.h +++ b/drivers/gpu/drm/zte/zx_plane_regs.h @@ -46,6 +46,37 @@ #define GL_POS_X(x) (((x) << GL_POS_X_SHIFT) & GL_POS_X_MASK) #define GL_POS_Y(x) (((x) << GL_POS_Y_SHIFT) & GL_POS_Y_MASK) +/* VL registers */ +#define VL_CTRL0 0x00 +#define VL_UPDATE BIT(3) +#define VL_CTRL1 0x04 +#define VL_YUV420_PLANAR BIT(5) +#define VL_YUV422_SHIFT 3 +#define VL_YUV422_YUYV (0 << VL_YUV422_SHIFT) +#define VL_YUV422_YVYU (1 << VL_YUV422_SHIFT) +#define VL_YUV422_UYVY (2 << VL_YUV422_SHIFT) +#define VL_YUV422_VYUY (3 << VL_YUV422_SHIFT) +#define VL_FMT_YUV420 0 +#define VL_FMT_YUV422 1 +#define VL_FMT_YUV420_P010 2 +#define VL_FMT_YUV420_HANTRO 3 +#define VL_FMT_YUV444_8BIT 4 +#define VL_FMT_YUV444_10BIT 5 +#define VL_CTRL2 0x08 +#define VL_SCALER_BYPASS_MODE BIT(0) +#define VL_STRIDE 0x0c +#define LUMA_STRIDE_SHIFT 16 +#define LUMA_STRIDE_MASK (0xffff << LUMA_STRIDE_SHIFT) +#define CHROMA_STRIDE_SHIFT 0 +#define CHROMA_STRIDE_MASK (0xffff << CHROMA_STRIDE_SHIFT) +#define VL_SRC_SIZE 0x10 +#define VL_Y 0x14 +#define VL_POS_START 0x30 +#define VL_POS_END 0x34 + +#define LUMA_STRIDE(x) (((x) << LUMA_STRIDE_SHIFT) & LUMA_STRIDE_MASK) +#define CHROMA_STRIDE(x) (((x) << CHROMA_STRIDE_SHIFT) & CHROMA_STRIDE_MASK) + /* CSC registers */ #define CSC_CTRL0 0x30 #define CSC_COV_MODE_SHIFT 16 @@ -69,6 +100,18 @@ #define RSZ_DEST_CFG 0x04 #define RSZ_ENABLE_CFG 0x14 +#define RSZ_VL_LUMA_HOR 0x08 +#define RSZ_VL_LUMA_VER 0x0c +#define RSZ_VL_CHROMA_HOR 0x10 +#define RSZ_VL_CHROMA_VER 0x14 +#define RSZ_VL_CTRL_CFG 0x18 +#define RSZ_VL_FMT_SHIFT 3 +#define RSZ_VL_FMT_MASK (0x3 << RSZ_VL_FMT_SHIFT) +#define RSZ_VL_FMT_YCBCR420 (0x0 << RSZ_VL_FMT_SHIFT) +#define RSZ_VL_FMT_YCBCR422 (0x1 << RSZ_VL_FMT_SHIFT) +#define RSZ_VL_FMT_YCBCR444 (0x2 << RSZ_VL_FMT_SHIFT) +#define RSZ_VL_ENABLE_CFG 0x1c + #define RSZ_VER_SHIFT 16 #define RSZ_VER_MASK (0xffff << RSZ_VER_SHIFT) #define RSZ_HOR_SHIFT 0 @@ -77,6 +120,14 @@ #define RSZ_VER(x) (((x) << RSZ_VER_SHIFT) & RSZ_VER_MASK) #define RSZ_HOR(x) (((x) << RSZ_HOR_SHIFT) & RSZ_HOR_MASK) +#define RSZ_DATA_STEP_SHIFT 16 +#define RSZ_DATA_STEP_MASK (0xffff << RSZ_DATA_STEP_SHIFT) +#define RSZ_PARA_STEP_SHIFT 0 +#define RSZ_PARA_STEP_MASK (0xffff << RSZ_PARA_STEP_SHIFT) + +#define RSZ_DATA_STEP(x) (((x) << RSZ_DATA_STEP_SHIFT) & RSZ_DATA_STEP_MASK) +#define RSZ_PARA_STEP(x) (((x) << RSZ_PARA_STEP_SHIFT) & RSZ_PARA_STEP_MASK) + /* HBSC registers */ #define HBSC_SATURATION 0x00 #define HBSC_HUE 0x04 diff --git a/drivers/gpu/drm/zte/zx_tvenc.c b/drivers/gpu/drm/zte/zx_tvenc.c new file mode 100644 index 000000000000..b56dc69843fc --- /dev/null +++ b/drivers/gpu/drm/zte/zx_tvenc.c @@ -0,0 +1,407 @@ +/* + * Copyright 2017 Linaro Ltd. + * Copyright 2017 ZTE Corporation. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include <linux/clk.h> +#include <linux/component.h> +#include <linux/mfd/syscon.h> +#include <linux/regmap.h> + +#include <drm/drm_atomic_helper.h> +#include <drm/drm_crtc_helper.h> +#include <drm/drmP.h> + +#include "zx_drm_drv.h" +#include "zx_tvenc_regs.h" +#include "zx_vou.h" + +struct zx_tvenc_pwrctrl { + struct regmap *regmap; + u32 reg; + u32 mask; +}; + +struct zx_tvenc { + struct drm_connector connector; + struct drm_encoder encoder; + struct device *dev; + void __iomem *mmio; + const struct vou_inf *inf; + struct zx_tvenc_pwrctrl pwrctrl; +}; + +#define to_zx_tvenc(x) container_of(x, struct zx_tvenc, x) + +struct zx_tvenc_mode { + struct drm_display_mode mode; + u32 video_info; + u32 video_res; + u32 field1_param; + u32 field2_param; + u32 burst_line_odd1; + u32 burst_line_even1; + u32 burst_line_odd2; + u32 burst_line_even2; + u32 line_timing_param; + u32 weight_value; + u32 blank_black_level; + u32 burst_level; + u32 control_param; + u32 sub_carrier_phase1; + u32 phase_line_incr_cvbs; +}; + +/* + * The CRM cannot directly provide a suitable frequency, and we have to + * ask a multiplied rate from CRM and use the divider in VOU to get the + * desired one. + */ +#define TVENC_CLOCK_MULTIPLIER 4 + +static const struct zx_tvenc_mode tvenc_mode_pal = { + .mode = { + .clock = 13500 * TVENC_CLOCK_MULTIPLIER, + .hdisplay = 720, + .hsync_start = 720 + 12, + .hsync_end = 720 + 12 + 2, + .htotal = 720 + 12 + 2 + 130, + .vdisplay = 576, + .vsync_start = 576 + 2, + .vsync_end = 576 + 2 + 2, + .vtotal = 576 + 2 + 2 + 20, + .flags = DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE, + }, + .video_info = 0x00040040, + .video_res = 0x05a9c760, + .field1_param = 0x0004d416, + .field2_param = 0x0009b94f, + .burst_line_odd1 = 0x0004d406, + .burst_line_even1 = 0x0009b53e, + .burst_line_odd2 = 0x0004d805, + .burst_line_even2 = 0x0009b93f, + .line_timing_param = 0x06a96fdf, + .weight_value = 0x00c188a0, + .blank_black_level = 0x0000fcfc, + .burst_level = 0x00001595, + .control_param = 0x00000001, + .sub_carrier_phase1 = 0x1504c566, + .phase_line_incr_cvbs = 0xc068db8c, +}; + +static const struct zx_tvenc_mode tvenc_mode_ntsc = { + .mode = { + .clock = 13500 * TVENC_CLOCK_MULTIPLIER, + .hdisplay = 720, + .hsync_start = 720 + 16, + .hsync_end = 720 + 16 + 2, + .htotal = 720 + 16 + 2 + 120, + .vdisplay = 480, + .vsync_start = 480 + 3, + .vsync_end = 480 + 3 + 2, + .vtotal = 480 + 3 + 2 + 17, + .flags = DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE, + }, + .video_info = 0x00040080, + .video_res = 0x05a8375a, + .field1_param = 0x00041817, + .field2_param = 0x0008351e, + .burst_line_odd1 = 0x00041006, + .burst_line_even1 = 0x0008290d, + .burst_line_odd2 = 0x00000000, + .burst_line_even2 = 0x00000000, + .line_timing_param = 0x06a8ef9e, + .weight_value = 0x00b68197, + .blank_black_level = 0x0000f0f0, + .burst_level = 0x0000009c, + .control_param = 0x00000001, + .sub_carrier_phase1 = 0x10f83e10, + .phase_line_incr_cvbs = 0x80000000, +}; + +static const struct zx_tvenc_mode *tvenc_modes[] = { + &tvenc_mode_pal, + &tvenc_mode_ntsc, +}; + +static const struct zx_tvenc_mode * +zx_tvenc_find_zmode(struct drm_display_mode *mode) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(tvenc_modes); i++) { + const struct zx_tvenc_mode *zmode = tvenc_modes[i]; + + if (drm_mode_equal(mode, &zmode->mode)) + return zmode; + } + + return NULL; +} + +static void zx_tvenc_encoder_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adj_mode) +{ + struct zx_tvenc *tvenc = to_zx_tvenc(encoder); + const struct zx_tvenc_mode *zmode; + struct vou_div_config configs[] = { + { VOU_DIV_INF, VOU_DIV_4 }, + { VOU_DIV_TVENC, VOU_DIV_1 }, + { VOU_DIV_LAYER, VOU_DIV_2 }, + }; + + zx_vou_config_dividers(encoder->crtc, configs, ARRAY_SIZE(configs)); + + zmode = zx_tvenc_find_zmode(mode); + if (!zmode) { + DRM_DEV_ERROR(tvenc->dev, "failed to find zmode\n"); + return; + } + + zx_writel(tvenc->mmio + VENC_VIDEO_INFO, zmode->video_info); + zx_writel(tvenc->mmio + VENC_VIDEO_RES, zmode->video_res); + zx_writel(tvenc->mmio + VENC_FIELD1_PARAM, zmode->field1_param); + zx_writel(tvenc->mmio + VENC_FIELD2_PARAM, zmode->field2_param); + zx_writel(tvenc->mmio + VENC_LINE_O_1, zmode->burst_line_odd1); + zx_writel(tvenc->mmio + VENC_LINE_E_1, zmode->burst_line_even1); + zx_writel(tvenc->mmio + VENC_LINE_O_2, zmode->burst_line_odd2); + zx_writel(tvenc->mmio + VENC_LINE_E_2, zmode->burst_line_even2); + zx_writel(tvenc->mmio + VENC_LINE_TIMING_PARAM, + zmode->line_timing_param); + zx_writel(tvenc->mmio + VENC_WEIGHT_VALUE, zmode->weight_value); + zx_writel(tvenc->mmio + VENC_BLANK_BLACK_LEVEL, + zmode->blank_black_level); + zx_writel(tvenc->mmio + VENC_BURST_LEVEL, zmode->burst_level); + zx_writel(tvenc->mmio + VENC_CONTROL_PARAM, zmode->control_param); + zx_writel(tvenc->mmio + VENC_SUB_CARRIER_PHASE1, + zmode->sub_carrier_phase1); + zx_writel(tvenc->mmio + VENC_PHASE_LINE_INCR_CVBS, + zmode->phase_line_incr_cvbs); +} + +static void zx_tvenc_encoder_enable(struct drm_encoder *encoder) +{ + struct zx_tvenc *tvenc = to_zx_tvenc(encoder); + struct zx_tvenc_pwrctrl *pwrctrl = &tvenc->pwrctrl; + + /* Set bit to power up TVENC DAC */ + regmap_update_bits(pwrctrl->regmap, pwrctrl->reg, pwrctrl->mask, + pwrctrl->mask); + + vou_inf_enable(VOU_TV_ENC, encoder->crtc); + + zx_writel(tvenc->mmio + VENC_ENABLE, 1); +} + +static void zx_tvenc_encoder_disable(struct drm_encoder *encoder) +{ + struct zx_tvenc *tvenc = to_zx_tvenc(encoder); + struct zx_tvenc_pwrctrl *pwrctrl = &tvenc->pwrctrl; + + zx_writel(tvenc->mmio + VENC_ENABLE, 0); + + vou_inf_disable(VOU_TV_ENC, encoder->crtc); + + /* Clear bit to power down TVENC DAC */ + regmap_update_bits(pwrctrl->regmap, pwrctrl->reg, pwrctrl->mask, 0); +} + +static const struct drm_encoder_helper_funcs zx_tvenc_encoder_helper_funcs = { + .enable = zx_tvenc_encoder_enable, + .disable = zx_tvenc_encoder_disable, + .mode_set = zx_tvenc_encoder_mode_set, +}; + +static const struct drm_encoder_funcs zx_tvenc_encoder_funcs = { + .destroy = drm_encoder_cleanup, +}; + +static int zx_tvenc_connector_get_modes(struct drm_connector *connector) +{ + struct zx_tvenc *tvenc = to_zx_tvenc(connector); + struct device *dev = tvenc->dev; + int i; + + for (i = 0; i < ARRAY_SIZE(tvenc_modes); i++) { + const struct zx_tvenc_mode *zmode = tvenc_modes[i]; + struct drm_display_mode *mode; + + mode = drm_mode_duplicate(connector->dev, &zmode->mode); + if (!mode) { + DRM_DEV_ERROR(dev, "failed to duplicate drm mode\n"); + continue; + } + + drm_mode_set_name(mode); + drm_mode_probed_add(connector, mode); + } + + return i; +} + +static enum drm_mode_status +zx_tvenc_connector_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct zx_tvenc *tvenc = to_zx_tvenc(connector); + const struct zx_tvenc_mode *zmode; + + zmode = zx_tvenc_find_zmode(mode); + if (!zmode) { + DRM_DEV_ERROR(tvenc->dev, "unsupported mode: %s\n", mode->name); + return MODE_NOMODE; + } + + return MODE_OK; +} + +static struct drm_connector_helper_funcs zx_tvenc_connector_helper_funcs = { + .get_modes = zx_tvenc_connector_get_modes, + .mode_valid = zx_tvenc_connector_mode_valid, +}; + +static const struct drm_connector_funcs zx_tvenc_connector_funcs = { + .dpms = drm_atomic_helper_connector_dpms, + .fill_modes = drm_helper_probe_single_connector_modes, + .destroy = drm_connector_cleanup, + .reset = drm_atomic_helper_connector_reset, + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, +}; + +static int zx_tvenc_register(struct drm_device *drm, struct zx_tvenc *tvenc) +{ + struct drm_encoder *encoder = &tvenc->encoder; + struct drm_connector *connector = &tvenc->connector; + + /* + * The tvenc is designed to use aux channel, as there is a deflicker + * block for the channel. + */ + encoder->possible_crtcs = BIT(1); + + drm_encoder_init(drm, encoder, &zx_tvenc_encoder_funcs, + DRM_MODE_ENCODER_TVDAC, NULL); + drm_encoder_helper_add(encoder, &zx_tvenc_encoder_helper_funcs); + + connector->interlace_allowed = true; + + drm_connector_init(drm, connector, &zx_tvenc_connector_funcs, + DRM_MODE_CONNECTOR_Composite); + drm_connector_helper_add(connector, &zx_tvenc_connector_helper_funcs); + + drm_mode_connector_attach_encoder(connector, encoder); + + return 0; +} + +static int zx_tvenc_pwrctrl_init(struct zx_tvenc *tvenc) +{ + struct zx_tvenc_pwrctrl *pwrctrl = &tvenc->pwrctrl; + struct device *dev = tvenc->dev; + struct of_phandle_args out_args; + struct regmap *regmap; + int ret; + + ret = of_parse_phandle_with_fixed_args(dev->of_node, + "zte,tvenc-power-control", 2, 0, &out_args); + if (ret) + return ret; + + regmap = syscon_node_to_regmap(out_args.np); + if (IS_ERR(regmap)) { + ret = PTR_ERR(regmap); + goto out; + } + + pwrctrl->regmap = regmap; + pwrctrl->reg = out_args.args[0]; + pwrctrl->mask = out_args.args[1]; + +out: + of_node_put(out_args.np); + return ret; +} + +static int zx_tvenc_bind(struct device *dev, struct device *master, void *data) +{ + struct platform_device *pdev = to_platform_device(dev); + struct drm_device *drm = data; + struct resource *res; + struct zx_tvenc *tvenc; + int ret; + + tvenc = devm_kzalloc(dev, sizeof(*tvenc), GFP_KERNEL); + if (!tvenc) + return -ENOMEM; + + tvenc->dev = dev; + dev_set_drvdata(dev, tvenc); + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + tvenc->mmio = devm_ioremap_resource(dev, res); + if (IS_ERR(tvenc->mmio)) { + ret = PTR_ERR(tvenc->mmio); + DRM_DEV_ERROR(dev, "failed to remap tvenc region: %d\n", ret); + return ret; + } + + ret = zx_tvenc_pwrctrl_init(tvenc); + if (ret) { + DRM_DEV_ERROR(dev, "failed to init power control: %d\n", ret); + return ret; + } + + ret = zx_tvenc_register(drm, tvenc); + if (ret) { + DRM_DEV_ERROR(dev, "failed to register tvenc: %d\n", ret); + return ret; + } + + return 0; +} + +static void zx_tvenc_unbind(struct device *dev, struct device *master, + void *data) +{ + /* Nothing to do */ +} + +static const struct component_ops zx_tvenc_component_ops = { + .bind = zx_tvenc_bind, + .unbind = zx_tvenc_unbind, +}; + +static int zx_tvenc_probe(struct platform_device *pdev) +{ + return component_add(&pdev->dev, &zx_tvenc_component_ops); +} + +static int zx_tvenc_remove(struct platform_device *pdev) +{ + component_del(&pdev->dev, &zx_tvenc_component_ops); + return 0; +} + +static const struct of_device_id zx_tvenc_of_match[] = { + { .compatible = "zte,zx296718-tvenc", }, + { /* end */ }, +}; +MODULE_DEVICE_TABLE(of, zx_tvenc_of_match); + +struct platform_driver zx_tvenc_driver = { + .probe = zx_tvenc_probe, + .remove = zx_tvenc_remove, + .driver = { + .name = "zx-tvenc", + .of_match_table = zx_tvenc_of_match, + }, +}; diff --git a/drivers/gpu/drm/zte/zx_tvenc_regs.h b/drivers/gpu/drm/zte/zx_tvenc_regs.h new file mode 100644 index 000000000000..bd91f5dcc1f3 --- /dev/null +++ b/drivers/gpu/drm/zte/zx_tvenc_regs.h @@ -0,0 +1,31 @@ +/* + * Copyright 2017 Linaro Ltd. + * Copyright 2017 ZTE Corporation. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#ifndef __ZX_TVENC_REGS_H__ +#define __ZX_TVENC_REGS_H__ + +#define VENC_VIDEO_INFO 0x04 +#define VENC_VIDEO_RES 0x08 +#define VENC_FIELD1_PARAM 0x10 +#define VENC_FIELD2_PARAM 0x14 +#define VENC_LINE_O_1 0x18 +#define VENC_LINE_E_1 0x1c +#define VENC_LINE_O_2 0x20 +#define VENC_LINE_E_2 0x24 +#define VENC_LINE_TIMING_PARAM 0x28 +#define VENC_WEIGHT_VALUE 0x2c +#define VENC_BLANK_BLACK_LEVEL 0x30 +#define VENC_BURST_LEVEL 0x34 +#define VENC_CONTROL_PARAM 0x3c +#define VENC_SUB_CARRIER_PHASE1 0x40 +#define VENC_PHASE_LINE_INCR_CVBS 0x48 +#define VENC_ENABLE 0xa8 + +#endif /* __ZX_TVENC_REGS_H__ */ diff --git a/drivers/gpu/drm/zte/zx_vou.c b/drivers/gpu/drm/zte/zx_vou.c index a86e3a5852a2..cf92d675feaa 100644 --- a/drivers/gpu/drm/zte/zx_vou.c +++ b/drivers/gpu/drm/zte/zx_vou.c @@ -40,6 +40,7 @@ struct zx_crtc_regs { u32 fir_active; u32 fir_htiming; u32 fir_vtiming; + u32 sec_vtiming; u32 timing_shift; u32 timing_pi_shift; }; @@ -48,6 +49,7 @@ static const struct zx_crtc_regs main_crtc_regs = { .fir_active = FIR_MAIN_ACTIVE, .fir_htiming = FIR_MAIN_H_TIMING, .fir_vtiming = FIR_MAIN_V_TIMING, + .sec_vtiming = SEC_MAIN_V_TIMING, .timing_shift = TIMING_MAIN_SHIFT, .timing_pi_shift = TIMING_MAIN_PI_SHIFT, }; @@ -56,6 +58,7 @@ static const struct zx_crtc_regs aux_crtc_regs = { .fir_active = FIR_AUX_ACTIVE, .fir_htiming = FIR_AUX_H_TIMING, .fir_vtiming = FIR_AUX_V_TIMING, + .sec_vtiming = SEC_AUX_V_TIMING, .timing_shift = TIMING_AUX_SHIFT, .timing_pi_shift = TIMING_AUX_PI_SHIFT, }; @@ -65,7 +68,17 @@ struct zx_crtc_bits { u32 polarity_shift; u32 int_frame_mask; u32 tc_enable; - u32 gl_enable; + u32 sec_vactive_shift; + u32 sec_vactive_mask; + u32 interlace_select; + u32 pi_enable; + u32 div_vga_shift; + u32 div_pic_shift; + u32 div_tvenc_shift; + u32 div_hdmi_pnx_shift; + u32 div_hdmi_shift; + u32 div_inf_shift; + u32 div_layer_shift; }; static const struct zx_crtc_bits main_crtc_bits = { @@ -73,7 +86,17 @@ static const struct zx_crtc_bits main_crtc_bits = { .polarity_shift = MAIN_POL_SHIFT, .int_frame_mask = TIMING_INT_MAIN_FRAME, .tc_enable = MAIN_TC_EN, - .gl_enable = OSD_CTRL0_GL0_EN, + .sec_vactive_shift = SEC_VACT_MAIN_SHIFT, + .sec_vactive_mask = SEC_VACT_MAIN_MASK, + .interlace_select = MAIN_INTERLACE_SEL, + .pi_enable = MAIN_PI_EN, + .div_vga_shift = VGA_MAIN_DIV_SHIFT, + .div_pic_shift = PIC_MAIN_DIV_SHIFT, + .div_tvenc_shift = TVENC_MAIN_DIV_SHIFT, + .div_hdmi_pnx_shift = HDMI_MAIN_PNX_DIV_SHIFT, + .div_hdmi_shift = HDMI_MAIN_DIV_SHIFT, + .div_inf_shift = INF_MAIN_DIV_SHIFT, + .div_layer_shift = LAYER_MAIN_DIV_SHIFT, }; static const struct zx_crtc_bits aux_crtc_bits = { @@ -81,7 +104,17 @@ static const struct zx_crtc_bits aux_crtc_bits = { .polarity_shift = AUX_POL_SHIFT, .int_frame_mask = TIMING_INT_AUX_FRAME, .tc_enable = AUX_TC_EN, - .gl_enable = OSD_CTRL0_GL1_EN, + .sec_vactive_shift = SEC_VACT_AUX_SHIFT, + .sec_vactive_mask = SEC_VACT_AUX_MASK, + .interlace_select = AUX_INTERLACE_SEL, + .pi_enable = AUX_PI_EN, + .div_vga_shift = VGA_AUX_DIV_SHIFT, + .div_pic_shift = PIC_AUX_DIV_SHIFT, + .div_tvenc_shift = TVENC_AUX_DIV_SHIFT, + .div_hdmi_pnx_shift = HDMI_AUX_PNX_DIV_SHIFT, + .div_hdmi_shift = HDMI_AUX_DIV_SHIFT, + .div_inf_shift = INF_AUX_DIV_SHIFT, + .div_layer_shift = LAYER_AUX_DIV_SHIFT, }; struct zx_crtc { @@ -97,6 +130,40 @@ struct zx_crtc { #define to_zx_crtc(x) container_of(x, struct zx_crtc, crtc) +struct vou_layer_bits { + u32 enable; + u32 chnsel; + u32 clksel; +}; + +static const struct vou_layer_bits zx_gl_bits[GL_NUM] = { + { + .enable = OSD_CTRL0_GL0_EN, + .chnsel = OSD_CTRL0_GL0_SEL, + .clksel = VOU_CLK_GL0_SEL, + }, { + .enable = OSD_CTRL0_GL1_EN, + .chnsel = OSD_CTRL0_GL1_SEL, + .clksel = VOU_CLK_GL1_SEL, + }, +}; + +static const struct vou_layer_bits zx_vl_bits[VL_NUM] = { + { + .enable = OSD_CTRL0_VL0_EN, + .chnsel = OSD_CTRL0_VL0_SEL, + .clksel = VOU_CLK_VL0_SEL, + }, { + .enable = OSD_CTRL0_VL1_EN, + .chnsel = OSD_CTRL0_VL1_SEL, + .clksel = VOU_CLK_VL1_SEL, + }, { + .enable = OSD_CTRL0_VL2_EN, + .chnsel = OSD_CTRL0_VL2_SEL, + .clksel = VOU_CLK_VL2_SEL, + }, +}; + struct zx_vou_hw { struct device *dev; void __iomem *osd; @@ -112,6 +179,33 @@ struct zx_vou_hw { struct zx_crtc *aux_crtc; }; +enum vou_inf_data_sel { + VOU_YUV444 = 0, + VOU_RGB_101010 = 1, + VOU_RGB_888 = 2, + VOU_RGB_666 = 3, +}; + +struct vou_inf { + enum vou_inf_id id; + enum vou_inf_data_sel data_sel; + u32 clocks_en_bits; + u32 clocks_sel_bits; +}; + +static struct vou_inf vou_infs[] = { + [VOU_HDMI] = { + .data_sel = VOU_YUV444, + .clocks_en_bits = BIT(24) | BIT(18) | BIT(6), + .clocks_sel_bits = BIT(13) | BIT(2), + }, + [VOU_TV_ENC] = { + .data_sel = VOU_YUV444, + .clocks_en_bits = BIT(15), + .clocks_sel_bits = BIT(11) | BIT(0), + }, +}; + static inline struct zx_vou_hw *crtc_to_vou(struct drm_crtc *crtc) { struct zx_crtc *zcrtc = to_zx_crtc(crtc); @@ -119,20 +213,30 @@ static inline struct zx_vou_hw *crtc_to_vou(struct drm_crtc *crtc) return zcrtc->vou; } -void vou_inf_enable(const struct vou_inf *inf, struct drm_crtc *crtc) +void vou_inf_hdmi_audio_sel(struct drm_crtc *crtc, + enum vou_inf_hdmi_audio aud) { struct zx_crtc *zcrtc = to_zx_crtc(crtc); struct zx_vou_hw *vou = zcrtc->vou; + + zx_writel_mask(vou->vouctl + VOU_INF_HDMI_CTRL, VOU_HDMI_AUD_MASK, aud); +} + +void vou_inf_enable(enum vou_inf_id id, struct drm_crtc *crtc) +{ + struct zx_crtc *zcrtc = to_zx_crtc(crtc); + struct zx_vou_hw *vou = zcrtc->vou; + struct vou_inf *inf = &vou_infs[id]; bool is_main = zcrtc->chn_type == VOU_CHN_MAIN; - u32 data_sel_shift = inf->id << 1; + u32 data_sel_shift = id << 1; /* Select data format */ zx_writel_mask(vou->vouctl + VOU_INF_DATA_SEL, 0x3 << data_sel_shift, inf->data_sel << data_sel_shift); /* Select channel */ - zx_writel_mask(vou->vouctl + VOU_INF_CH_SEL, 0x1 << inf->id, - zcrtc->chn_type << inf->id); + zx_writel_mask(vou->vouctl + VOU_INF_CH_SEL, 0x1 << id, + zcrtc->chn_type << id); /* Select interface clocks */ zx_writel_mask(vou->vouctl + VOU_CLK_SEL, inf->clocks_sel_bits, @@ -143,20 +247,79 @@ void vou_inf_enable(const struct vou_inf *inf, struct drm_crtc *crtc) inf->clocks_en_bits); /* Enable the device */ - zx_writel_mask(vou->vouctl + VOU_INF_EN, 1 << inf->id, 1 << inf->id); + zx_writel_mask(vou->vouctl + VOU_INF_EN, 1 << id, 1 << id); } -void vou_inf_disable(const struct vou_inf *inf, struct drm_crtc *crtc) +void vou_inf_disable(enum vou_inf_id id, struct drm_crtc *crtc) { struct zx_vou_hw *vou = crtc_to_vou(crtc); + struct vou_inf *inf = &vou_infs[id]; /* Disable the device */ - zx_writel_mask(vou->vouctl + VOU_INF_EN, 1 << inf->id, 0); + zx_writel_mask(vou->vouctl + VOU_INF_EN, 1 << id, 0); /* Disable interface clocks */ zx_writel_mask(vou->vouctl + VOU_CLK_EN, inf->clocks_en_bits, 0); } +void zx_vou_config_dividers(struct drm_crtc *crtc, + struct vou_div_config *configs, int num) +{ + struct zx_crtc *zcrtc = to_zx_crtc(crtc); + struct zx_vou_hw *vou = zcrtc->vou; + const struct zx_crtc_bits *bits = zcrtc->bits; + int i; + + /* Clear update flag bit */ + zx_writel_mask(vou->vouctl + VOU_DIV_PARA, DIV_PARA_UPDATE, 0); + + for (i = 0; i < num; i++) { + struct vou_div_config *cfg = configs + i; + u32 reg, shift; + + switch (cfg->id) { + case VOU_DIV_VGA: + reg = VOU_CLK_SEL; + shift = bits->div_vga_shift; + break; + case VOU_DIV_PIC: + reg = VOU_CLK_SEL; + shift = bits->div_pic_shift; + break; + case VOU_DIV_TVENC: + reg = VOU_DIV_PARA; + shift = bits->div_tvenc_shift; + break; + case VOU_DIV_HDMI_PNX: + reg = VOU_DIV_PARA; + shift = bits->div_hdmi_pnx_shift; + break; + case VOU_DIV_HDMI: + reg = VOU_DIV_PARA; + shift = bits->div_hdmi_shift; + break; + case VOU_DIV_INF: + reg = VOU_DIV_PARA; + shift = bits->div_inf_shift; + break; + case VOU_DIV_LAYER: + reg = VOU_DIV_PARA; + shift = bits->div_layer_shift; + break; + default: + continue; + } + + /* Each divider occupies 3 bits */ + zx_writel_mask(vou->vouctl + reg, 0x7 << shift, + cfg->val << shift); + } + + /* Set update flag bit to get dividers effected */ + zx_writel_mask(vou->vouctl + VOU_DIV_PARA, DIV_PARA_UPDATE, + DIV_PARA_UPDATE); +} + static inline void vou_chn_set_update(struct zx_crtc *zcrtc) { zx_writel(zcrtc->chnreg + CHN_UPDATE, 1); @@ -165,11 +328,13 @@ static inline void vou_chn_set_update(struct zx_crtc *zcrtc) static void zx_crtc_enable(struct drm_crtc *crtc) { struct drm_display_mode *mode = &crtc->state->adjusted_mode; + bool interlaced = mode->flags & DRM_MODE_FLAG_INTERLACE; struct zx_crtc *zcrtc = to_zx_crtc(crtc); struct zx_vou_hw *vou = zcrtc->vou; const struct zx_crtc_regs *regs = zcrtc->regs; const struct zx_crtc_bits *bits = zcrtc->bits; struct videomode vm; + u32 scan_mask; u32 pol = 0; u32 val; int ret; @@ -177,7 +342,7 @@ static void zx_crtc_enable(struct drm_crtc *crtc) drm_display_mode_to_videomode(mode, &vm); /* Set up timing parameters */ - val = V_ACTIVE(vm.vactive - 1); + val = V_ACTIVE((interlaced ? vm.vactive / 2 : vm.vactive) - 1); val |= H_ACTIVE(vm.hactive - 1); zx_writel(vou->timing + regs->fir_active, val); @@ -191,6 +356,25 @@ static void zx_crtc_enable(struct drm_crtc *crtc) val |= FRONT_PORCH(vm.vfront_porch - 1); zx_writel(vou->timing + regs->fir_vtiming, val); + if (interlaced) { + u32 shift = bits->sec_vactive_shift; + u32 mask = bits->sec_vactive_mask; + + val = zx_readl(vou->timing + SEC_V_ACTIVE); + val &= ~mask; + val |= ((vm.vactive / 2 - 1) << shift) & mask; + zx_writel(vou->timing + SEC_V_ACTIVE, val); + + val = SYNC_WIDE(vm.vsync_len - 1); + /* + * The vback_porch for the second field needs to shift one on + * the value for the first field. + */ + val |= BACK_PORCH(vm.vback_porch); + val |= FRONT_PORCH(vm.vfront_porch - 1); + zx_writel(vou->timing + regs->sec_vtiming, val); + } + /* Set up polarities */ if (vm.flags & DISPLAY_FLAGS_VSYNC_LOW) pol |= 1 << POL_VSYNC_SHIFT; @@ -201,9 +385,17 @@ static void zx_crtc_enable(struct drm_crtc *crtc) pol << bits->polarity_shift); /* Setup SHIFT register by following what ZTE BSP does */ - zx_writel(vou->timing + regs->timing_shift, H_SHIFT_VAL); + val = H_SHIFT_VAL; + if (interlaced) + val |= V_SHIFT_VAL << 16; + zx_writel(vou->timing + regs->timing_shift, val); zx_writel(vou->timing + regs->timing_pi_shift, H_PI_SHIFT_VAL); + /* Progressive or interlace scan select */ + scan_mask = bits->interlace_select | bits->pi_enable; + zx_writel_mask(vou->timing + SCAN_CTRL, scan_mask, + interlaced ? scan_mask : 0); + /* Enable TIMING_CTRL */ zx_writel_mask(vou->timing + TIMING_TC_ENABLE, bits->tc_enable, bits->tc_enable); @@ -214,16 +406,16 @@ static void zx_crtc_enable(struct drm_crtc *crtc) zx_writel_mask(zcrtc->chnreg + CHN_CTRL1, CHN_SCREEN_H_MASK, vm.vactive << CHN_SCREEN_H_SHIFT); + /* Configure channel interlace buffer control */ + zx_writel_mask(zcrtc->chnreg + CHN_INTERLACE_BUF_CTRL, CHN_INTERLACE_EN, + interlaced ? CHN_INTERLACE_EN : 0); + /* Update channel */ vou_chn_set_update(zcrtc); /* Enable channel */ zx_writel_mask(zcrtc->chnreg + CHN_CTRL0, CHN_ENABLE, CHN_ENABLE); - /* Enable Graphic Layer */ - zx_writel_mask(vou->osd + OSD_CTRL0, bits->gl_enable, - bits->gl_enable); - drm_crtc_vblank_on(crtc); ret = clk_set_rate(zcrtc->pixclk, mode->clock * 1000); @@ -247,9 +439,6 @@ static void zx_crtc_disable(struct drm_crtc *crtc) drm_crtc_vblank_off(crtc); - /* Disable Graphic Layer */ - zx_writel_mask(vou->osd + OSD_CTRL0, bits->gl_enable, 0); - /* Disable channel */ zx_writel_mask(zcrtc->chnreg + CHN_CTRL0, CHN_ENABLE, 0); @@ -294,7 +483,7 @@ static int zx_crtc_init(struct drm_device *drm, struct zx_vou_hw *vou, enum vou_chn_type chn_type) { struct device *dev = vou->dev; - struct zx_layer_data data; + struct zx_plane *zplane; struct zx_crtc *zcrtc; int ret; @@ -305,19 +494,27 @@ static int zx_crtc_init(struct drm_device *drm, struct zx_vou_hw *vou, zcrtc->vou = vou; zcrtc->chn_type = chn_type; + zplane = devm_kzalloc(dev, sizeof(*zplane), GFP_KERNEL); + if (!zplane) + return -ENOMEM; + + zplane->dev = dev; + if (chn_type == VOU_CHN_MAIN) { - data.layer = vou->osd + MAIN_GL_OFFSET; - data.csc = vou->osd + MAIN_CSC_OFFSET; - data.hbsc = vou->osd + MAIN_HBSC_OFFSET; - data.rsz = vou->otfppu + MAIN_RSZ_OFFSET; + zplane->layer = vou->osd + MAIN_GL_OFFSET; + zplane->csc = vou->osd + MAIN_CSC_OFFSET; + zplane->hbsc = vou->osd + MAIN_HBSC_OFFSET; + zplane->rsz = vou->otfppu + MAIN_RSZ_OFFSET; + zplane->bits = &zx_gl_bits[0]; zcrtc->chnreg = vou->osd + OSD_MAIN_CHN; zcrtc->regs = &main_crtc_regs; zcrtc->bits = &main_crtc_bits; } else { - data.layer = vou->osd + AUX_GL_OFFSET; - data.csc = vou->osd + AUX_CSC_OFFSET; - data.hbsc = vou->osd + AUX_HBSC_OFFSET; - data.rsz = vou->otfppu + AUX_RSZ_OFFSET; + zplane->layer = vou->osd + AUX_GL_OFFSET; + zplane->csc = vou->osd + AUX_CSC_OFFSET; + zplane->hbsc = vou->osd + AUX_HBSC_OFFSET; + zplane->rsz = vou->otfppu + AUX_RSZ_OFFSET; + zplane->bits = &zx_gl_bits[1]; zcrtc->chnreg = vou->osd + OSD_AUX_CHN; zcrtc->regs = &aux_crtc_regs; zcrtc->bits = &aux_crtc_bits; @@ -331,13 +528,14 @@ static int zx_crtc_init(struct drm_device *drm, struct zx_vou_hw *vou, return ret; } - zcrtc->primary = zx_plane_init(drm, dev, &data, DRM_PLANE_TYPE_PRIMARY); - if (IS_ERR(zcrtc->primary)) { - ret = PTR_ERR(zcrtc->primary); + ret = zx_plane_init(drm, zplane, DRM_PLANE_TYPE_PRIMARY); + if (ret) { DRM_DEV_ERROR(dev, "failed to init primary plane: %d\n", ret); return ret; } + zcrtc->primary = &zplane->plane; + ret = drm_crtc_init_with_planes(drm, &zcrtc->crtc, zcrtc->primary, NULL, &zx_crtc_funcs, NULL); if (ret) { @@ -393,6 +591,78 @@ void zx_vou_disable_vblank(struct drm_device *drm, unsigned int pipe) zcrtc->bits->int_frame_mask, 0); } +void zx_vou_layer_enable(struct drm_plane *plane) +{ + struct zx_crtc *zcrtc = to_zx_crtc(plane->state->crtc); + struct zx_vou_hw *vou = zcrtc->vou; + struct zx_plane *zplane = to_zx_plane(plane); + const struct vou_layer_bits *bits = zplane->bits; + + if (zcrtc->chn_type == VOU_CHN_MAIN) { + zx_writel_mask(vou->osd + OSD_CTRL0, bits->chnsel, 0); + zx_writel_mask(vou->vouctl + VOU_CLK_SEL, bits->clksel, 0); + } else { + zx_writel_mask(vou->osd + OSD_CTRL0, bits->chnsel, + bits->chnsel); + zx_writel_mask(vou->vouctl + VOU_CLK_SEL, bits->clksel, + bits->clksel); + } + + zx_writel_mask(vou->osd + OSD_CTRL0, bits->enable, bits->enable); +} + +void zx_vou_layer_disable(struct drm_plane *plane) +{ + struct zx_crtc *zcrtc = to_zx_crtc(plane->crtc); + struct zx_vou_hw *vou = zcrtc->vou; + struct zx_plane *zplane = to_zx_plane(plane); + const struct vou_layer_bits *bits = zplane->bits; + + zx_writel_mask(vou->osd + OSD_CTRL0, bits->enable, 0); +} + +static void zx_overlay_init(struct drm_device *drm, struct zx_vou_hw *vou) +{ + struct device *dev = vou->dev; + struct zx_plane *zplane; + int i; + int ret; + + /* + * VL0 has some quirks on scaling support which need special handling. + * Let's leave it out for now. + */ + for (i = 1; i < VL_NUM; i++) { + zplane = devm_kzalloc(dev, sizeof(*zplane), GFP_KERNEL); + if (!zplane) { + DRM_DEV_ERROR(dev, "failed to allocate zplane %d\n", i); + return; + } + + zplane->layer = vou->osd + OSD_VL_OFFSET(i); + zplane->hbsc = vou->osd + HBSC_VL_OFFSET(i); + zplane->rsz = vou->otfppu + RSZ_VL_OFFSET(i); + zplane->bits = &zx_vl_bits[i]; + + ret = zx_plane_init(drm, zplane, DRM_PLANE_TYPE_OVERLAY); + if (ret) { + DRM_DEV_ERROR(dev, "failed to init overlay %d\n", i); + continue; + } + } +} + +static inline void zx_osd_int_update(struct zx_crtc *zcrtc) +{ + struct drm_crtc *crtc = &zcrtc->crtc; + struct drm_plane *plane; + + vou_chn_set_update(zcrtc); + + drm_for_each_plane_mask(plane, crtc->dev, crtc->state->plane_mask) + zx_plane_set_update(plane); +} + static irqreturn_t vou_irq_handler(int irq, void *dev_id) { struct zx_vou_hw *vou = dev_id; @@ -412,15 +682,11 @@ static irqreturn_t vou_irq_handler(int irq, void *dev_id) state = zx_readl(vou->osd + OSD_INT_STA); zx_writel(vou->osd + OSD_INT_CLRSTA, state); - if (state & OSD_INT_MAIN_UPT) { - vou_chn_set_update(vou->main_crtc); - zx_plane_set_update(vou->main_crtc->primary); - } + if (state & OSD_INT_MAIN_UPT) + zx_osd_int_update(vou->main_crtc); - if (state & OSD_INT_AUX_UPT) { - vou_chn_set_update(vou->aux_crtc); - zx_plane_set_update(vou->aux_crtc->primary); - } + if (state & OSD_INT_AUX_UPT) + zx_osd_int_update(vou->aux_crtc); if (state & OSD_INT_ERROR) DRM_DEV_ERROR(vou->dev, "OSD ERROR: 0x%08x!\n", state); @@ -451,19 +717,9 @@ static void vou_dtrc_init(struct zx_vou_hw *vou) static void vou_hw_init(struct zx_vou_hw *vou) { - /* Set GL0 to main channel and GL1 to aux channel */ - zx_writel_mask(vou->osd + OSD_CTRL0, OSD_CTRL0_GL0_SEL, 0); - zx_writel_mask(vou->osd + OSD_CTRL0, OSD_CTRL0_GL1_SEL, - OSD_CTRL0_GL1_SEL); - /* Release reset for all VOU modules */ zx_writel(vou->vouctl + VOU_SOFT_RST, ~0); - /* Select main clock for GL0 and aux clock for GL1 module */ - zx_writel_mask(vou->vouctl + VOU_CLK_SEL, VOU_CLK_GL0_SEL, 0); - zx_writel_mask(vou->vouctl + VOU_CLK_SEL, VOU_CLK_GL1_SEL, - VOU_CLK_GL1_SEL); - /* Enable clock auto-gating for all VOU modules */ zx_writel(vou->vouctl + VOU_CLK_REQEN, ~0); @@ -600,6 +856,8 @@ static int zx_crtc_bind(struct device *dev, struct device *master, void *data) goto disable_ppu_clk; } + zx_overlay_init(drm, vou); + return 0; disable_ppu_clk: diff --git a/drivers/gpu/drm/zte/zx_vou.h b/drivers/gpu/drm/zte/zx_vou.h index 349e06cd86f4..57e3c31ee6a5 100644 --- a/drivers/gpu/drm/zte/zx_vou.h +++ b/drivers/gpu/drm/zte/zx_vou.h @@ -23,24 +23,48 @@ enum vou_inf_id { VOU_VGA = 5, }; -enum vou_inf_data_sel { - VOU_YUV444 = 0, - VOU_RGB_101010 = 1, - VOU_RGB_888 = 2, - VOU_RGB_666 = 3, +enum vou_inf_hdmi_audio { + VOU_HDMI_AUD_SPDIF = BIT(0), + VOU_HDMI_AUD_I2S = BIT(1), + VOU_HDMI_AUD_DSD = BIT(2), + VOU_HDMI_AUD_HBR = BIT(3), + VOU_HDMI_AUD_PARALLEL = BIT(4), }; -struct vou_inf { - enum vou_inf_id id; - enum vou_inf_data_sel data_sel; - u32 clocks_en_bits; - u32 clocks_sel_bits; +void vou_inf_hdmi_audio_sel(struct drm_crtc *crtc, + enum vou_inf_hdmi_audio aud); +void vou_inf_enable(enum vou_inf_id id, struct drm_crtc *crtc); +void vou_inf_disable(enum vou_inf_id id, struct drm_crtc *crtc); + +enum vou_div_id { + VOU_DIV_VGA, + VOU_DIV_PIC, + VOU_DIV_TVENC, + VOU_DIV_HDMI_PNX, + VOU_DIV_HDMI, + VOU_DIV_INF, + VOU_DIV_LAYER, +}; + +enum vou_div_val { + VOU_DIV_1 = 0, + VOU_DIV_2 = 1, + VOU_DIV_4 = 3, + VOU_DIV_8 = 7, }; -void vou_inf_enable(const struct vou_inf *inf, struct drm_crtc *crtc); -void vou_inf_disable(const struct vou_inf *inf, struct drm_crtc *crtc); +struct vou_div_config { + enum vou_div_id id; + enum vou_div_val val; +}; + +void zx_vou_config_dividers(struct drm_crtc *crtc, + struct vou_div_config *configs, int num); int zx_vou_enable_vblank(struct drm_device *drm, unsigned int pipe); void zx_vou_disable_vblank(struct drm_device *drm, unsigned int pipe); +void zx_vou_layer_enable(struct drm_plane *plane); +void zx_vou_layer_disable(struct drm_plane *plane); + #endif /* __ZX_VOU_H__ */ diff --git a/drivers/gpu/drm/zte/zx_vou_regs.h b/drivers/gpu/drm/zte/zx_vou_regs.h index f44e7a4ae441..c066ef123434 100644 --- a/drivers/gpu/drm/zte/zx_vou_regs.h +++ b/drivers/gpu/drm/zte/zx_vou_regs.h @@ -22,6 +22,15 @@ #define AUX_HBSC_OFFSET 0x860 #define AUX_RSZ_OFFSET 0x800 +#define OSD_VL0_OFFSET 0x040 +#define OSD_VL_OFFSET(i) (OSD_VL0_OFFSET + 0x050 * (i)) + +#define HBSC_VL0_OFFSET 0x760 +#define HBSC_VL_OFFSET(i) (HBSC_VL0_OFFSET + 0x040 * (i)) + +#define RSZ_VL1_U0 0xa00 +#define RSZ_VL_OFFSET(i) (RSZ_VL1_U0 + 0x200 * (i)) + /* OSD (GPC_GLOBAL) registers */ #define OSD_INT_STA 0x04 #define OSD_INT_CLRSTA 0x08 @@ -42,6 +51,12 @@ ) #define OSD_INT_ENABLE (OSD_INT_ERROR | OSD_INT_AUX_UPT | OSD_INT_MAIN_UPT) #define OSD_CTRL0 0x10 +#define OSD_CTRL0_VL0_EN BIT(13) +#define OSD_CTRL0_VL0_SEL BIT(12) +#define OSD_CTRL0_VL1_EN BIT(11) +#define OSD_CTRL0_VL1_SEL BIT(10) +#define OSD_CTRL0_VL2_EN BIT(9) +#define OSD_CTRL0_VL2_SEL BIT(8) #define OSD_CTRL0_GL0_EN BIT(7) #define OSD_CTRL0_GL0_SEL BIT(6) #define OSD_CTRL0_GL1_EN BIT(5) @@ -60,6 +75,8 @@ #define CHN_SCREEN_H_SHIFT 5 #define CHN_SCREEN_H_MASK (0x1fff << CHN_SCREEN_H_SHIFT) #define CHN_UPDATE 0x08 +#define CHN_INTERLACE_BUF_CTRL 0x24 +#define CHN_INTERLACE_EN BIT(2) /* TIMING_CTRL registers */ #define TIMING_TC_ENABLE 0x04 @@ -102,6 +119,19 @@ #define TIMING_MAIN_SHIFT 0x2c #define TIMING_AUX_SHIFT 0x30 #define H_SHIFT_VAL 0x0048 +#define V_SHIFT_VAL 0x0001 +#define SCAN_CTRL 0x34 +#define AUX_PI_EN BIT(19) +#define MAIN_PI_EN BIT(18) +#define AUX_INTERLACE_SEL BIT(1) +#define MAIN_INTERLACE_SEL BIT(0) +#define SEC_V_ACTIVE 0x38 +#define SEC_VACT_MAIN_SHIFT 0 +#define SEC_VACT_MAIN_MASK (0xffff << SEC_VACT_MAIN_SHIFT) +#define SEC_VACT_AUX_SHIFT 16 +#define SEC_VACT_AUX_MASK (0xffff << SEC_VACT_AUX_SHIFT) +#define SEC_MAIN_V_TIMING 0x3c +#define SEC_AUX_V_TIMING 0x40 #define TIMING_MAIN_PI_SHIFT 0x68 #define TIMING_AUX_PI_SHIFT 0x6c #define H_PI_SHIFT_VAL 0x000f @@ -146,10 +176,31 @@ #define VOU_INF_DATA_SEL 0x08 #define VOU_SOFT_RST 0x14 #define VOU_CLK_SEL 0x18 +#define VGA_AUX_DIV_SHIFT 29 +#define VGA_MAIN_DIV_SHIFT 26 +#define PIC_MAIN_DIV_SHIFT 23 +#define PIC_AUX_DIV_SHIFT 20 +#define VOU_CLK_VL2_SEL BIT(8) +#define VOU_CLK_VL1_SEL BIT(7) +#define VOU_CLK_VL0_SEL BIT(6) #define VOU_CLK_GL1_SEL BIT(5) #define VOU_CLK_GL0_SEL BIT(4) +#define VOU_DIV_PARA 0x1c +#define DIV_PARA_UPDATE BIT(31) +#define TVENC_AUX_DIV_SHIFT 28 +#define HDMI_AUX_PNX_DIV_SHIFT 25 +#define HDMI_MAIN_PNX_DIV_SHIFT 22 +#define HDMI_AUX_DIV_SHIFT 19 +#define HDMI_MAIN_DIV_SHIFT 16 +#define TVENC_MAIN_DIV_SHIFT 13 +#define INF_AUX_DIV_SHIFT 9 +#define INF_MAIN_DIV_SHIFT 6 +#define LAYER_AUX_DIV_SHIFT 3 +#define LAYER_MAIN_DIV_SHIFT 0 #define VOU_CLK_REQEN 0x20 #define VOU_CLK_EN 0x24 +#define VOU_INF_HDMI_CTRL 0x30 +#define VOU_HDMI_AUD_MASK 0x1f /* OTFPPU_CTRL registers */ #define OTFPPU_RSZ_DATA_SOURCE 0x04 |