summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c194
1 files changed, 110 insertions, 84 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index a674c8a58dc2..65577eca58f1 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -760,7 +760,7 @@ static u64 gfx_v9_0_ring_get_rptr_compute(struct amdgpu_ring *ring);
static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
void *ras_error_status);
static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev,
- void *inject_if);
+ void *inject_if, uint32_t instance_mask);
static void gfx_v9_0_reset_ras_error_count(struct amdgpu_device *adev);
static void gfx_v9_0_kiq_set_resources(struct amdgpu_ring *kiq_ring,
@@ -888,7 +888,7 @@ static const struct kiq_pm4_funcs gfx_v9_0_kiq_pm4_funcs = {
static void gfx_v9_0_set_kiq_pm4_funcs(struct amdgpu_device *adev)
{
- adev->gfx.kiq.pmf = &gfx_v9_0_kiq_pm4_funcs;
+ adev->gfx.kiq[0].pmf = &gfx_v9_0_kiq_pm4_funcs;
}
static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev)
@@ -1494,7 +1494,7 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev)
mask = 1;
cu_bitmap = 0;
counter = 0;
- amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
if (cu_info->bitmap[i][j] & mask) {
@@ -1513,7 +1513,7 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev)
cu_info->ao_cu_bitmap[i][j] = cu_bitmap;
}
}
- amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
mutex_unlock(&adev->grbm_idx_mutex);
}
@@ -1535,7 +1535,7 @@ static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev)
mutex_lock(&adev->grbm_idx_mutex);
/* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/
- amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff);
/* set mmRLC_LB_PARAMS = 0x003F_1006 */
@@ -1584,7 +1584,7 @@ static void gfx_v9_4_init_lbpw(struct amdgpu_device *adev)
mutex_lock(&adev->grbm_idx_mutex);
/* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/
- amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff);
/* set mmRLC_LB_PARAMS = 0x003F_1006 */
@@ -1703,7 +1703,7 @@ static int gfx_v9_0_mec_init(struct amdgpu_device *adev)
const struct gfx_firmware_header_v1_0 *mec_hdr;
- bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+ bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
/* take ownership of the relevant compute queues */
amdgpu_gfx_compute_queue_acquire(adev);
@@ -1778,7 +1778,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd,
*(out++) = RREG32_SOC15(GC, 0, mmSQ_IND_DATA);
}
-static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
+static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
{
/* type 1 wave data */
dst[(*no_fields)++] = 1;
@@ -1799,7 +1799,7 @@ static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u
dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE);
}
-static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
uint32_t wave, uint32_t start,
uint32_t size, uint32_t *dst)
{
@@ -1808,7 +1808,7 @@ static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
start + SQIND_WAVE_SGPRS_OFFSET, size, dst);
}
-static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
uint32_t wave, uint32_t thread,
uint32_t start, uint32_t size,
uint32_t *dst)
@@ -1819,9 +1819,9 @@ static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd,
}
static void gfx_v9_0_select_me_pipe_q(struct amdgpu_device *adev,
- u32 me, u32 pipe, u32 q, u32 vm)
+ u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
{
- soc15_grbm_select(adev, me, pipe, q, vm);
+ soc15_grbm_select(adev, me, pipe, q, vm, 0);
}
static const struct amdgpu_gfx_funcs gfx_v9_0_gfx_funcs = {
@@ -1995,7 +1995,7 @@ static int gfx_v9_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
+ (ring_id * GFX9_MEC_HPD_SIZE);
- ring->vm_hub = AMDGPU_GFXHUB_0;
+ ring->vm_hub = AMDGPU_GFXHUB(0);
sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue);
irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
@@ -2095,7 +2095,7 @@ static int gfx_v9_0_sw_init(void *handle)
/* disable scheduler on the real ring */
ring->no_scheduler = true;
- ring->vm_hub = AMDGPU_GFXHUB_0;
+ ring->vm_hub = AMDGPU_GFXHUB(0);
r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq,
AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP,
AMDGPU_RING_PRIO_DEFAULT, NULL);
@@ -2113,7 +2113,7 @@ static int gfx_v9_0_sw_init(void *handle)
ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1;
ring->is_sw_ring = true;
hw_prio = amdgpu_sw_ring_priority(i);
- ring->vm_hub = AMDGPU_GFXHUB_0;
+ ring->vm_hub = AMDGPU_GFXHUB(0);
r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq,
AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP, hw_prio,
NULL);
@@ -2144,7 +2144,8 @@ static int gfx_v9_0_sw_init(void *handle)
for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
- if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, j))
+ if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i,
+ k, j))
continue;
r = gfx_v9_0_compute_ring_init(adev,
@@ -2158,19 +2159,19 @@ static int gfx_v9_0_sw_init(void *handle)
}
}
- r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE);
+ r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE, 0);
if (r) {
DRM_ERROR("Failed to init KIQ BOs!\n");
return r;
}
- kiq = &adev->gfx.kiq;
- r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq);
+ kiq = &adev->gfx.kiq[0];
+ r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq, 0);
if (r)
return r;
/* create MQD for all compute queues as wel as KIQ for SRIOV case */
- r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v9_mqd_allocation));
+ r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v9_mqd_allocation), 0);
if (r)
return r;
@@ -2205,9 +2206,9 @@ static int gfx_v9_0_sw_fini(void *handle)
for (i = 0; i < adev->gfx.num_compute_rings; i++)
amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
- amdgpu_gfx_mqd_sw_fini(adev);
- amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring);
- amdgpu_gfx_kiq_fini(adev);
+ amdgpu_gfx_mqd_sw_fini(adev, 0);
+ amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring);
+ amdgpu_gfx_kiq_fini(adev, 0);
gfx_v9_0_mec_fini(adev);
amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj,
@@ -2230,7 +2231,7 @@ static void gfx_v9_0_tiling_mode_table_init(struct amdgpu_device *adev)
}
void gfx_v9_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num,
- u32 instance)
+ u32 instance, int xcc_id)
{
u32 data;
@@ -2279,19 +2280,42 @@ static void gfx_v9_0_setup_rb(struct amdgpu_device *adev)
mutex_lock(&adev->grbm_idx_mutex);
for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
- amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
data = gfx_v9_0_get_rb_active_bitmap(adev);
active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
rb_bitmap_width_per_sh);
}
}
- amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
mutex_unlock(&adev->grbm_idx_mutex);
adev->gfx.config.backend_enable_mask = active_rbs;
adev->gfx.config.num_rbs = hweight32(active_rbs);
}
+static void gfx_v9_0_debug_trap_config_init(struct amdgpu_device *adev,
+ uint32_t first_vmid,
+ uint32_t last_vmid)
+{
+ uint32_t data;
+ uint32_t trap_config_vmid_mask = 0;
+ int i;
+
+ /* Calculate trap config vmid mask */
+ for (i = first_vmid; i < last_vmid; i++)
+ trap_config_vmid_mask |= (1 << i);
+
+ data = REG_SET_FIELD(0, SPI_GDBG_TRAP_CONFIG,
+ VMID_SEL, trap_config_vmid_mask);
+ data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG,
+ TRAP_EN, 1);
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_CONFIG), data);
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA0), 0);
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA1), 0);
+}
+
#define DEFAULT_SH_MEM_BASES (0x6000)
static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev)
{
@@ -2313,12 +2337,12 @@ static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev)
mutex_lock(&adev->srbm_mutex);
for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
- soc15_grbm_select(adev, 0, 0, 0, i);
+ soc15_grbm_select(adev, 0, 0, 0, i, 0);
/* CP and shaders */
WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
}
- soc15_grbm_select(adev, 0, 0, 0, 0);
+ soc15_grbm_select(adev, 0, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
/* Initialize all compute VMIDs to have no GDS, GWS, or OA
@@ -2356,8 +2380,8 @@ static void gfx_v9_0_init_sq_config(struct amdgpu_device *adev)
switch (adev->ip_versions[GC_HWIP][0]) {
case IP_VERSION(9, 4, 1):
tmp = RREG32_SOC15(GC, 0, mmSQ_CONFIG);
- tmp = REG_SET_FIELD(tmp, SQ_CONFIG,
- DISABLE_BARRIER_WAITCNT, 1);
+ tmp = REG_SET_FIELD(tmp, SQ_CONFIG, DISABLE_BARRIER_WAITCNT,
+ !READ_ONCE(adev->barrier_has_auto_waitcnt));
WREG32_SOC15(GC, 0, mmSQ_CONFIG, tmp);
break;
default:
@@ -2382,8 +2406,8 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
/* XXX SH_MEM regs */
/* where to put LDS, scratch, GPUVM in FSA64 space */
mutex_lock(&adev->srbm_mutex);
- for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) {
- soc15_grbm_select(adev, 0, 0, 0, i);
+ for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) {
+ soc15_grbm_select(adev, 0, 0, 0, i, 0);
/* CP and shaders */
if (i == 0) {
tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE,
@@ -2405,7 +2429,7 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, tmp);
}
}
- soc15_grbm_select(adev, 0, 0, 0, 0);
+ soc15_grbm_select(adev, 0, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
@@ -2422,7 +2446,7 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
mutex_lock(&adev->grbm_idx_mutex);
for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
- amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
for (k = 0; k < adev->usec_timeout; k++) {
if (RREG32_SOC15(GC, 0, mmRLC_SERDES_CU_MASTER_BUSY) == 0)
break;
@@ -2430,7 +2454,7 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
}
if (k == adev->usec_timeout) {
amdgpu_gfx_select_se_sh(adev, 0xffffffff,
- 0xffffffff, 0xffffffff);
+ 0xffffffff, 0xffffffff, 0);
mutex_unlock(&adev->grbm_idx_mutex);
DRM_INFO("Timeout wait for RLC serdes %u,%u\n",
i, j);
@@ -2438,7 +2462,7 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
}
}
}
- amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
mutex_unlock(&adev->grbm_idx_mutex);
mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK |
@@ -3133,7 +3157,6 @@ static int gfx_v9_0_cp_gfx_resume(struct amdgpu_device *adev)
/* start the ring */
gfx_v9_0_cp_gfx_start(adev);
- ring->sched.ready = true;
return 0;
}
@@ -3145,7 +3168,7 @@ static void gfx_v9_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
} else {
WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL,
(CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK));
- adev->gfx.kiq.ring.sched.ready = false;
+ adev->gfx.kiq[0].ring.sched.ready = false;
}
udelay(50);
}
@@ -3509,7 +3532,6 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring)
{
struct amdgpu_device *adev = ring->adev;
struct v9_mqd *mqd = ring->mqd_ptr;
- int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS;
struct v9_mqd *tmp_mqd;
gfx_v9_0_kiq_setting(ring);
@@ -3519,20 +3541,20 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring)
* driver need to re-init the mqd.
* check mqd->cp_hqd_pq_control since this value should not be 0
*/
- tmp_mqd = (struct v9_mqd *)adev->gfx.mec.mqd_backup[mqd_idx];
+ tmp_mqd = (struct v9_mqd *)adev->gfx.kiq[0].mqd_backup;
if (amdgpu_in_reset(adev) && tmp_mqd->cp_hqd_pq_control){
/* for GPU_RESET case , reset MQD to a clean status */
- if (adev->gfx.mec.mqd_backup[mqd_idx])
- memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
+ if (adev->gfx.kiq[0].mqd_backup)
+ memcpy(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(struct v9_mqd_allocation));
/* reset ring buffer */
ring->wptr = 0;
amdgpu_ring_clear_ring(ring);
mutex_lock(&adev->srbm_mutex);
- soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+ soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0);
gfx_v9_0_kiq_init_register(ring);
- soc15_grbm_select(adev, 0, 0, 0, 0);
+ soc15_grbm_select(adev, 0, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
} else {
memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
@@ -3541,14 +3563,14 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring)
if (amdgpu_sriov_vf(adev) && adev->in_suspend)
amdgpu_ring_clear_ring(ring);
mutex_lock(&adev->srbm_mutex);
- soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+ soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0);
gfx_v9_0_mqd_init(ring);
gfx_v9_0_kiq_init_register(ring);
- soc15_grbm_select(adev, 0, 0, 0, 0);
+ soc15_grbm_select(adev, 0, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
- if (adev->gfx.mec.mqd_backup[mqd_idx])
- memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation));
+ if (adev->gfx.kiq[0].mqd_backup)
+ memcpy(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(struct v9_mqd_allocation));
}
return 0;
@@ -3572,24 +3594,21 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
mutex_lock(&adev->srbm_mutex);
- soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+ soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0);
gfx_v9_0_mqd_init(ring);
- soc15_grbm_select(adev, 0, 0, 0, 0);
+ soc15_grbm_select(adev, 0, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
if (adev->gfx.mec.mqd_backup[mqd_idx])
memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation));
- } else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
- /* reset MQD to a clean status */
+ } else {
+ /* restore MQD to a clean status */
if (adev->gfx.mec.mqd_backup[mqd_idx])
memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
-
/* reset ring buffer */
ring->wptr = 0;
atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0);
amdgpu_ring_clear_ring(ring);
- } else {
- amdgpu_ring_clear_ring(ring);
}
return 0;
@@ -3600,7 +3619,7 @@ static int gfx_v9_0_kiq_resume(struct amdgpu_device *adev)
struct amdgpu_ring *ring;
int r;
- ring = &adev->gfx.kiq.ring;
+ ring = &adev->gfx.kiq[0].ring;
r = amdgpu_bo_reserve(ring->mqd_obj, false);
if (unlikely(r != 0))
@@ -3616,7 +3635,6 @@ static int gfx_v9_0_kiq_resume(struct amdgpu_device *adev)
amdgpu_bo_kunmap(ring->mqd_obj);
ring->mqd_ptr = NULL;
amdgpu_bo_unreserve(ring->mqd_obj);
- ring->sched.ready = true;
return 0;
}
@@ -3644,7 +3662,7 @@ static int gfx_v9_0_kcq_resume(struct amdgpu_device *adev)
goto done;
}
- r = amdgpu_gfx_enable_kcq(adev);
+ r = amdgpu_gfx_enable_kcq(adev, 0);
done:
return r;
}
@@ -3764,7 +3782,7 @@ static int gfx_v9_0_hw_fini(void *handle)
/* DF freeze and kcq disable will fail */
if (!amdgpu_ras_intr_triggered())
/* disable KCQ to avoid CPC touch memory not valid anymore */
- amdgpu_gfx_disable_kcq(adev);
+ amdgpu_gfx_disable_kcq(adev, 0);
if (amdgpu_sriov_vf(adev)) {
gfx_v9_0_cp_gfx_enable(adev, false);
@@ -3782,11 +3800,11 @@ static int gfx_v9_0_hw_fini(void *handle)
*/
if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
mutex_lock(&adev->srbm_mutex);
- soc15_grbm_select(adev, adev->gfx.kiq.ring.me,
- adev->gfx.kiq.ring.pipe,
- adev->gfx.kiq.ring.queue, 0);
- gfx_v9_0_kiq_fini_register(&adev->gfx.kiq.ring);
- soc15_grbm_select(adev, 0, 0, 0, 0);
+ soc15_grbm_select(adev, adev->gfx.kiq[0].ring.me,
+ adev->gfx.kiq[0].ring.pipe,
+ adev->gfx.kiq[0].ring.queue, 0, 0);
+ gfx_v9_0_kiq_fini_register(&adev->gfx.kiq[0].ring);
+ soc15_grbm_select(adev, 0, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
}
@@ -3906,7 +3924,7 @@ static uint64_t gfx_v9_0_kiq_read_clock(struct amdgpu_device *adev)
unsigned long flags;
uint32_t seq, reg_val_offs = 0;
uint64_t value = 0;
- struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+ struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
struct amdgpu_ring *ring = &kiq->ring;
BUG_ON(!ring->funcs->emit_rreg);
@@ -4506,6 +4524,7 @@ static int gfx_v9_0_early_init(void *handle)
adev->gfx.num_gfx_rings = 0;
else
adev->gfx.num_gfx_rings = GFX9_NUM_GFX_RINGS;
+ adev->gfx.xcc_mask = 1;
adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
AMDGPU_MAX_COMPUTE_RINGS);
gfx_v9_0_set_kiq_pm4_funcs(adev);
@@ -4571,6 +4590,13 @@ static int gfx_v9_0_late_init(void *handle)
if (r)
return r;
+ if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
+ gfx_v9_4_2_debug_trap_config_init(adev,
+ adev->vm_manager.first_kfd_vmid, AMDGPU_NUM_VMID);
+ else
+ gfx_v9_0_debug_trap_config_init(adev,
+ adev->vm_manager.first_kfd_vmid, AMDGPU_NUM_VMID);
+
return 0;
}
@@ -4586,7 +4612,7 @@ static bool gfx_v9_0_is_rlc_enabled(struct amdgpu_device *adev)
return true;
}
-static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev)
+static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id)
{
uint32_t data;
unsigned i;
@@ -4603,7 +4629,7 @@ static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev)
}
}
-static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev)
+static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id)
{
uint32_t data;
@@ -4614,7 +4640,7 @@ static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev)
static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev,
bool enable)
{
- amdgpu_gfx_rlc_enter_safe_mode(adev);
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) && enable) {
gfx_v9_0_enable_gfx_cg_power_gating(adev, true);
@@ -4626,7 +4652,7 @@ static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev,
gfx_v9_0_enable_gfx_pipeline_powergating(adev, false);
}
- amdgpu_gfx_rlc_exit_safe_mode(adev);
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static void gfx_v9_0_update_gfx_mg_power_gating(struct amdgpu_device *adev,
@@ -4653,7 +4679,7 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
{
uint32_t data, def;
- amdgpu_gfx_rlc_enter_safe_mode(adev);
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
/* It is disabled by HW by default */
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
@@ -4720,7 +4746,7 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
}
}
- amdgpu_gfx_rlc_exit_safe_mode(adev);
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
@@ -4731,7 +4757,7 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
if (!adev->gfx.num_gfx_rings)
return;
- amdgpu_gfx_rlc_enter_safe_mode(adev);
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
/* Enable 3D CGCG/CGLS */
if (enable) {
@@ -4775,7 +4801,7 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D, data);
}
- amdgpu_gfx_rlc_exit_safe_mode(adev);
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
@@ -4783,7 +4809,7 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
{
uint32_t def, data;
- amdgpu_gfx_rlc_enter_safe_mode(adev);
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE);
@@ -4827,7 +4853,7 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL, data);
}
- amdgpu_gfx_rlc_exit_safe_mode(adev);
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static int gfx_v9_0_update_gfx_clock_gating(struct amdgpu_device *adev,
@@ -5425,7 +5451,7 @@ static int gfx_v9_0_ring_preempt_ib(struct amdgpu_ring *ring)
{
int i, r = 0;
struct amdgpu_device *adev = ring->adev;
- struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+ struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
struct amdgpu_ring *kiq_ring = &kiq->ring;
unsigned long flags;
@@ -6384,7 +6410,7 @@ static const struct soc15_ras_field_entry gfx_v9_0_ras_fields[] = {
};
static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev,
- void *inject_if)
+ void *inject_if, uint32_t instance_mask)
{
struct ras_inject_if *info = (struct ras_inject_if *)inject_if;
int ret;
@@ -6423,7 +6449,7 @@ static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev,
block_info.value = info->value;
mutex_lock(&adev->grbm_idx_mutex);
- ret = psp_ras_trigger_error(&adev->psp, &block_info);
+ ret = psp_ras_trigger_error(&adev->psp, &block_info, instance_mask);
mutex_unlock(&adev->grbm_idx_mutex);
return ret;
@@ -6651,7 +6677,7 @@ static void gfx_v9_0_reset_ras_error_count(struct amdgpu_device *adev)
for (i = 0; i < ARRAY_SIZE(gfx_v9_0_edc_counter_regs); i++) {
for (j = 0; j < gfx_v9_0_edc_counter_regs[i].se_num; j++) {
for (k = 0; k < gfx_v9_0_edc_counter_regs[i].instance; k++) {
- amdgpu_gfx_select_se_sh(adev, j, 0x0, k);
+ amdgpu_gfx_select_se_sh(adev, j, 0x0, k, 0);
RREG32(SOC15_REG_ENTRY_OFFSET(gfx_v9_0_edc_counter_regs[i]));
}
}
@@ -6713,7 +6739,7 @@ static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
for (i = 0; i < ARRAY_SIZE(gfx_v9_0_edc_counter_regs); i++) {
for (j = 0; j < gfx_v9_0_edc_counter_regs[i].se_num; j++) {
for (k = 0; k < gfx_v9_0_edc_counter_regs[i].instance; k++) {
- amdgpu_gfx_select_se_sh(adev, j, 0, k);
+ amdgpu_gfx_select_se_sh(adev, j, 0, k, 0);
reg_value =
RREG32(SOC15_REG_ENTRY_OFFSET(gfx_v9_0_edc_counter_regs[i]));
if (reg_value)
@@ -6728,7 +6754,7 @@ static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
err_data->ce_count += sec_count;
err_data->ue_count += ded_count;
- amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
mutex_unlock(&adev->grbm_idx_mutex);
gfx_v9_0_query_utc_edc_status(adev, err_data);
@@ -7010,7 +7036,7 @@ static void gfx_v9_0_set_ring_funcs(struct amdgpu_device *adev)
{
int i;
- adev->gfx.kiq.ring.funcs = &gfx_v9_0_ring_funcs_kiq;
+ adev->gfx.kiq[0].ring.funcs = &gfx_v9_0_ring_funcs_kiq;
for (i = 0; i < adev->gfx.num_gfx_rings; i++)
adev->gfx.gfx_ring[i].funcs = &gfx_v9_0_ring_funcs_gfx;
@@ -7191,7 +7217,7 @@ static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev,
mask = 1;
ao_bitmap = 0;
counter = 0;
- amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
gfx_v9_0_set_user_cu_inactive_bitmap(
adev, disable_masks[i * adev->gfx.config.max_sh_per_se + j]);
bitmap = gfx_v9_0_get_cu_active_bitmap(adev);
@@ -7224,7 +7250,7 @@ static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev,
cu_info->ao_cu_bitmap[i % 4][j + i / 4] = ao_bitmap;
}
}
- amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+ amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
mutex_unlock(&adev->grbm_idx_mutex);
cu_info->number = active_cu_number;
generated by cgit v1.2.3 (git 2.39.0) at 2024-10-11 04:52:59 +0300