summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/nouveau/nouveau_bo.c
AgeCommit message (Collapse)AuthorFilesLines
2014-12-02drm/nouveau: teach nouveau_bo_pin() how to force a contig vram allocationBen Skeggs1-7/+32
We have the ability to move buffers around in the kernel if necessary, and should probably use it rather than failing if userspace passes us a non-contig buffer for a plane. The NOUVEAU_GEM_TILE_NONCONTIG flag from userspace will become a mere initial placement hint once all the relevant paths have been updated. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-12-02drm/nouveau: we need pin_refcnt for nouveau_bo_placement_set()Ben Skeggs1-5/+9
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-12-02drm/nouveau: fix pin refcnt leak in failure pathBen Skeggs1-12/+12
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-12-02drm/nouveau: synchronize BOs when requiredAlexandre Courbot1-0/+42
On architectures for which access to GPU memory is non-coherent, caches need to be flushed and invalidated explicitly when BO control changes between CPU and GPU. This patch adds buffer synchronization functions which invokes the correct API (PCI or DMA) to ensure synchronization is effective. Based on the TTM DMA cache helper patches by Lucas Stach. Signed-off-by: Lucas Stach <dev@lynxeye.de> Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-12-02drm/nouveau: implement explicitly coherent BOsAlexandre Courbot1-8/+72
Allow nouveau_bo_new() to recognize the TTM_PL_FLAG_UNCACHED flag, which means that we want the allocated BO to be perfectly coherent between the CPU and GPU. This is useful on non-coherent architectures for which we do not want to manually sync some rarely-accessed buffers: typically, fences and pushbuffers. A TTM BO allocated with the TTM_PL_FLAG_UNCACHED on a non-coherent architecture will be populated using the DMA API, and accesses to it performed using the coherent mapping performed by dma_alloc_coherent(). Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-12-02drm/nouveau: warn when moving a pinned objectAlexandre Courbot1-2/+9
Pinned BOs are supposed to remain in their current location until unpinned. Display a warning for the supposedly-erroneous case where we are trying to move such objects. Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-10-01drm/nouveau: export reservation_object from dmabuf to ttmMaarten Lankhorst1-2/+2
Adds an extra argument to nouveau_bo_new, which is only used in nouveau_prime.c. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
2014-09-30drm/ttm: add reservation_object as argument to ttm_bo_initMaarten Lankhorst1-1/+1
This allows importing reservation objects from dma-bufs. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
2014-09-30drm/nouveau: specify if interruptible wait is desired in nouveau_fence_syncMaarten Lankhorst1-1/+1
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
2014-09-02drm/nouveau: use shared fences for readable objectsMaarten Lankhorst1-4/+7
nouveau keeps track in userspace whether a buffer is being written to or being read, but it doesn't use that information. Change this to allow multiple readers on the same bo. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com>
2014-09-02drm/ttm: flip the switch, and convert to dma_fenceMaarten Lankhorst1-42/+6
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
2014-09-02drm/nouveau: rework to new fence interfaceMaarten Lankhorst1-2/+4
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com>
2014-09-01drm/ttm: kill fence_lockMaarten Lankhorst1-17/+8
No users are left, kill it off! :D Conversion to the reservation api is next on the list, after that the functionality can be restored with rcu. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
2014-09-01drm/nouveau: require reservations for nouveau_fence_sync and nouveau_bo_fenceMaarten Lankhorst1-0/+2
This will ensure we always hold the required lock when calling those functions. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-27drm/ttm: move fpfn and lpfn into each placement v2Christian König1-15/+37
This allows us to more fine grained specify where to place the buffer object. v2: rebased on drm-next, add bochs changes as well Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2014-08-15drm/nouveau: kill unused variable warning if !__OS_HAS_AGPBen Skeggs1-2/+1
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: use ram info from nvif_deviceBen Skeggs1-2/+1
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: remove as much direct use of core headers as possibleBen Skeggs1-1/+1
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: remove (most) hardcoded object handle usageBen Skeggs1-5/+5
The PFIFO<->EVO sync buffers will be fixed up later when inter-channel sync in general is improved. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: port to nvif client/device/objectsBen Skeggs1-13/+9
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: initial pass at moving to struct nvif_deviceBen Skeggs1-29/+28
This is an attempt at isolating some of the changes necessary to port to NVIF in a separate commit. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: kill nouveau_dev() + wrap register macrosBen Skeggs1-1/+1
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: fix some usages of the wrong print functionBen Skeggs1-1/+1
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: store a pointer to vm in nouveau_cliBen Skeggs1-6/+6
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau: map pages using DMA APIAlexandre Courbot1-6/+16
The DMA API is the recommended way to map pages no matter what the underlying bus is. Use the DMA functions for page mapping and remove currently existing wrappers. Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-08-09drm/nouveau/bar: add noncached ioremap propertyAlexandre Courbot1-5/+12
Some BARs (like GK20A's) do not support being ioremapped write-combined. Add a boolean property to the BAR structure and handle that case in the Nouveau BO implementation. Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
2014-07-22drm/ttm: Fix a few sparse warningsThierry Reding1-3/+3
The final parameter to ttm_bo_reserve() is a pointer, therefore callers should use NULL instead of 0. Fixes a bunch of sparse warnings of this type: warning: Using plain integer as NULL pointer Signed-off-by: Thierry Reding <treding@nvidia.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2014-03-26support for platform devicesAlexandre Courbot1-10/+12
Upcoming mobile Kepler GPUs (such as GK20A) use the platform bus instead of PCI to which Nouveau is tightly dependent. This patch allows Nouveau to handle platform devices by: - abstracting PCI-dependent functions that were typically used for resource querying and page mapping, - introducing a nv_device_is_pci() function that allows to make PCI-dependent code conditional, - providing a nouveau_drm_platform_probe() function that takes a GPU platform device to be probed. Core code as well as engine/subdev drivers are updated wherever possible to make use of these functions. Some older drivers are too dependent on PCI to be properly updated, but all newer code on which future chips may depend should at least be runnable with platform devices. Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-02-18drm/nouveau: fix TTM_PL_TT memtype on pre-nv50Ilia Mirkin1-1/+1
Commit a55409066 ("drm/nv50-: map TTM_PL_SYSTEM through a BAR for CPU access") made it possible to work with tiled memory. However mem->mm_node is not a nouveau_mem for AGP-using pre-NV50 cards, but a drm_mm_node, as created by the ttm_bo_manager_func. As such, extend the untiled check to explicitly include all pre-nv50 cards. Reported-by: Ronald <ronald645@gmail.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=74613 Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Tested-by: Ronald Uitermark <ronald645@gmail.com> Acked-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-01-23drm/nouveau/ttm: explicitly wait for bo idle before memcpy buffer moveBen Skeggs1-19/+18
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-01-23drm/nouveau/ttm: explicity sync with kernel channel before moving bufferBen Skeggs1-28/+13
The GEM code handles this currently, but that'll be removed. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-01-23drm/nouveau/ttm: tidy up creation of temporary buffer move vmasBen Skeggs1-20/+21
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-01-23drm/nv50-: map TTM_PL_SYSTEM through a BAR for CPU accessMaarten Lankhorst1-3/+17
Moves bo's to TTM_PL_TT for BAR mapping, to hide tiling from user. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-01-23drm/nouveau: fix m2mf copy to tiled gartMaarten Lankhorst1-25/+8
Commit de7b7d59d54852c introduced tiled GART, but a linear copy is still performed. This may result in errors on eviction, fix it by checking tiling from memtype. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Cc: stable@vger.kernel.org #3.10+ Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-01-23drm/nouveau/vm: reduce number of entry-points to vm_map()Ben Skeggs1-26/+7
Pretty much everywhere had to make the decision which to use, so it makes a lot more sense to just have one entrypoint decide the path to take instead. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-11-14drm/nouveau: do not map evicted vram buffers in nouveau_bo_vma_addMaarten Lankhorst1-1/+2
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-11-14drm/nouveau: allow nouveau_fence_ref() to be a noopBen Skeggs1-10/+3
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-11-08drm/nv10: introduce a new NV_11 card typeIlia Mirkin1-1/+2
NV11/17/1F/18 come after NV10/15/16/1A. In order to facilitate using numerical comparisons, split up the two sets into different card types. This change should be a no-op except that the relevant cards will see NV11 printed instead of NV10 for the family. Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-11-08drm/nouveau: remove pointless assignmentDave Jones1-1/+1
self-assignment of a variable doesn't make a lot of sense. Signed-off-by: Dave Jones <davej@fedoraproject.org> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-10-09drm/nouveau: embed gem object in nouveau_boDavid Herrmann1-2/+2
There is no reason to keep the gem object separately allocated. nouveau is the last user of gem_obj->driver_private, so if we embed it, we can get rid of 8bytes per gem-object. The implementation follows the radeon driver. bo->gem is only valid, iff the bo was created via the gem helpers _and_ iff the user holds a valid gem reference. That is, as the gem object holds a reference to the nouveau_bo. If you use nouveau_ref() to gain a bo reference, you are not guaranteed to also hold a gem reference. The gem object might get destroyed after the last user drops the gem-ref via drm_gem_object_unreference(). Use drm_gem_object_reference() to gain a gem-reference. For debugging, we can use bo->gem.filp != NULL to test whether a gem-bo is valid. However, this shouldn't be used for real functionality to avoid gem-internal dependencies. Note that the implementation follows the previous style. However, we no longer can check for bo->gem != NULL to test for a valid gem object. This wasn't done before, so we should be safe now. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-09-02Merge branch 'drm-next-3.12' of git://people.freedesktop.org/~agd5f/linux ↵Dave Airlie1-1/+6
into drm-next Alex writes: This is the radeon drm-next request. Big changes include: - support for dpm on CIK parts - support for ASPM on CIK parts - support for berlin GPUs - major ring handling cleanup - remove the old 3D blit code for bo moves in favor of CP DMA or sDMA - lots of bug fixes [airlied: fix up a bunch of conflicts from drm_order removal] * 'drm-next-3.12' of git://people.freedesktop.org/~agd5f/linux: (898 commits) drm/radeon/dpm: make sure dc performance level limits are valid (CI) drm/radeon/dpm: make sure dc performance level limits are valid (BTC-SI) (v2) drm/radeon: gcc fixes for extended dpm tables drm/radeon: gcc fixes for kb/kv dpm drm/radeon: gcc fixes for ci dpm drm/radeon: gcc fixes for si dpm drm/radeon: gcc fixes for ni dpm drm/radeon: gcc fixes for trinity dpm drm/radeon: gcc fixes for sumo dpm drm/radeonn: gcc fixes for rv7xx/eg/btc dpm drm/radeon: gcc fixes for rv6xx dpm drm/radeon: gcc fixes for radeon_atombios.c drm/radeon: enable UVD interrupts on CIK drm/radeon: fix init ordering for r600+ drm/radeon/dpm: only need to reprogram uvd if uvd pg is enabled drm/radeon: check the return value of uvd_v1_0_start in uvd_v1_0_init drm/radeon: split out radeon_uvd_resume from uvd_v4_2_resume radeon kms: fix uninitialised hotplug work usage in r100_irq_process() drm/radeon/audio: set up the sads on DCE3.2 asics drm/radeon: fix handling of variable sized arrays for router objects ... Conflicts: drivers/gpu/drm/i915/i915_dma.c drivers/gpu/drm/i915/i915_gem_dmabuf.c drivers/gpu/drm/i915/intel_pm.c drivers/gpu/drm/radeon/cik.c drivers/gpu/drm/radeon/ni.c drivers/gpu/drm/radeon/r600.c
2013-08-27drm: verify vma access in TTM+GEM driversDavid Herrmann1-1/+3
GEM does already a good job in tracking access to gem buffers via handles and drm_vma access management. However, TTM drivers currently do not verify this during mmap(). TTM provides the verify_access() callback to test this. So fix all drivers to actually call into gem+vma to verify access instead of always returning 0. All drivers assume that user-space can only get access to TTM buffers via GEM handles. So whenever the verify_access() callback is called from ttm_bo_mmap(), the buffer must have a valid embedded gem object. This is true for all TTM+GEM drivers. But that's why this patch doesn't touch pure TTM drivers (ie, vmwgfx). v2: Switch to drm_vma_node_verify_access() to correctly return -EACCES if access was denied. Cc: Dave Airlie <airlied@redhat.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Ben Skeggs <bskeggs@redhat.com> Cc: Maarten Lankhorst <maarten.lankhorst@canonical.com> Cc: Jerome Glisse <jglisse@redhat.com> Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-07-30drm/nouveau: fix size check for cards without vmMaarten Lankhorst1-1/+6
Op 24-07-13 17:55, Dan Carpenter schreef: > Hello Maarten Lankhorst, > > This is a semi-automatic email about new static checker warnings. > > The patch 0108bc808107: "drm/nouveau: do not allow negative sizes for > now" from Jul 7, 2013, leads to the following Smatch complaint: > > drivers/gpu/drm/nouveau/nouveau_bo.c:222 nouveau_bo_new() > warn: variable dereferenced before check 'drm->client.base.vm' (see line 201) > > drivers/gpu/drm/nouveau/nouveau_bo.c > 200 int type = ttm_bo_type_device; > 201 int max_size = INT_MAX & ~((1 << drm->client.base.vm->vmm->lpg_shift) - 1); > ^^^^^^^^^^^^^^^^^^^ > New dereference. > > 202 > 203 if (size <= 0 || size > max_size) { > 204 nv_warn(drm, "skipped size %x\n", (u32)size); > 205 return -EINVAL; > 206 } > 207 > 208 if (sg) > 209 type = ttm_bo_type_sg; > 210 > 211 nvbo = kzalloc(sizeof(struct nouveau_bo), GFP_KERNEL); > 212 if (!nvbo) > 213 return -ENOMEM; > 214 INIT_LIST_HEAD(&nvbo->head); > 215 INIT_LIST_HEAD(&nvbo->entry); > 216 INIT_LIST_HEAD(&nvbo->vma_list); > 217 nvbo->tile_mode = tile_mode; > 218 nvbo->tile_flags = tile_flags; > 219 nvbo->bo.bdev = &drm->ttm.bdev; > 220 > 221 nvbo->page_shift = 12; > 222 if (drm->client.base.vm) { > ^^^^^^^^^^^^^^^^^^^ > Old check. > > 223 if (!(flags & TTM_PL_FLAG_TT) && size > 256 * 1024) > 224 nvbo->page_shift = drm->client.base.vm->vmm->lpg_shift; > > regards, > dan carpenter 8<----- Commit 0108bc808107: "drm/nouveau: do not allow negative sizes for now" broke older nvidia gpu's that lack a vm. Add an explicit check to handle this. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Reported-by: konrad wilk <konrad.wilk@oracle.com> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-10drm/nouveau: do not allow negative sizes for nowMaarten Lankhorst1-0/+6
The API allows up to 64-bits allocations, but size is handled as int inside nouveau almost everywhere. Until this is fixed it's better to prevent negative sizes. The 256 kB before INT_MAX is paranoia, because of the large page aligning below that could flip it above INT_MAX. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-10drm/nouveau: use dedicated channel for async moves on GT/GF chipsets.Ben Skeggs1-3/+3
The moves themselves were generally async to graphics previously, with the exception that if the "main" channel is used to synchronise a page flip at the same time, it can end up blocked for a noticable amount of time for large buffer moves. Not really critical, and there's better ways of handling this, but they are all rather invasive, so this is fine for now. Based on a patch by Maarten Lankhorst addressing the same issue. Reported-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Acked-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
2013-07-10drm/nouveau: do not unpin in nouveau_gem_object_delMaarten Lankhorst1-2/+5
This should no longer be required, and is harmful for framebuffer pinning. Also add a warning if unpin causes the pin count to drop below 0. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-10drm/nouveau: fix locking issues in page flipping pathsBen Skeggs1-1/+1
b580c9e2b7ba5030a795aa2fb73b796523d65a78 introduced additional problems while trying to solve issues that became apparent while porting to the new reservation stuff. The major problem was that the the previously mentioned patch took the client mutex earlier than previously, but the pinning of new_bo can can potentially cause a buffer move, which would result in attempting to acquire the same mutex again. This commit attempts to fix that "fix". Thanks to Maarten for the tips on keeping lockdep happy and cooking :) Reported-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Acked-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
2013-07-08drm/nouveau: fix minor thinko causing bo moves to not be async on keplerBen Skeggs1-1/+2
Reported-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau: delay busy bo vma removal until fence signalsBen Skeggs1-6/+1
As opposed to an explicit wait. Allows userspace to not stall waiting on buffer deletion. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/fb: initialise vram controller as pfb sub-objectBen Skeggs1-1/+1
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
generated by cgit v1.2.3 (git 2.39.0) at 2024-09-19 11:38:44 +0300