For a channel to make use of SVM features, it requires a different GPU MMU
configuration than we would normally use, which is not desirable to switch
to unless a client is actively going to use SVM.
In order to supporting SVM without more extensive changes to the userspace
interfaces, the SVM_INIT ioctl needs to replace the previous configuration
safely.
The only way we can currently do this safely, accounting for some unlikely
failure conditions, is to allocate the new VMM without destroying the last
one, and prioritising the SVM-enabled configuration in the code that cares.
This will get cleaned up again further down the track.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The function ttm_bo_put releases a reference to a TTM buffer object. The
function's name is more aligned to the Linux kernel convention of naming
ref-counting function _get and _put.
A call to ttm_bo_unref takes the address of the TTM BO object's pointer and
clears the pointer's value to NULL. This is not necessary in most cases and
sometimes even worked around by the calling code. A call to ttm_bo_put only
releases the reference without clearing the pointer.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch unifies the naming of DRM functions for reference counting
of struct drm_gem_object. The resulting code is more aligned with the
rest of the Linux kernel interfaces.
Signed-off-by: Thomas Zimmermann <tdz@users.sourceforge.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The bo array has req->nr_buffers elements so the > should be >= so we
don't read beyond the end of the array.
Fixes: a1606a9596 ("drm/nouveau: new gem pushbuf interface, bump to 0.0.16")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
As VMAs are per-client, unlike buffers, this allows us to avoid referencing
foreign fences (those that belong to another client/driver) from the client
deferred work handler, and prevent some not-fun race conditions that can be
triggered when a fence stalls.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
We previously only did this for push buffers, but an upcoming patch will
need to attach fences to all VMAs to resolve another issue.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Just another alias for ttm_bo_unreserve.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
VMAs are about to not take references on the VMM they belong to, which
means more care is required when handling delayed unmapping.
Queuing it on the client workqueue ensures all pending VMA unmaps will
have completed before the VMM is destroyed.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
There are many code paths opencoding kvmalloc. Let's use the helper
instead. The main difference to kvmalloc is that those users are
usually not considering all the aspects of the memory allocator. E.g.
allocation requests <= 32kB (with 4kB pages) are basically never failing
and invoke OOM killer to satisfy the allocation. This sounds too
disruptive for something that has a reasonable fallback - the vmalloc.
On the other hand those requests might fallback to vmalloc even when the
memory allocator would succeed after several more reclaim/compaction
attempts previously. There is no guarantee something like that happens
though.
This patch converts many of those places to kv[mz]alloc* helpers because
they are more conservative.
Link: http://lkml.kernel.org/r/20170306103327.2766-2-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> # Xen bits
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Andreas Dilger <andreas.dilger@intel.com> # Lustre
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com> # KVM/s390
Acked-by: Dan Williams <dan.j.williams@intel.com> # nvdim
Acked-by: David Sterba <dsterba@suse.com> # btrfs
Acked-by: Ilya Dryomov <idryomov@gmail.com> # Ceph
Acked-by: Tariq Toukan <tariqt@mellanox.com> # mlx4
Acked-by: Leon Romanovsky <leonro@mellanox.com> # mlx5
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Anton Vorontsov <anton@enomsg.org>
Cc: Colin Cross <ccross@android.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Kent Overstreet <kent.overstreet@gmail.com>
Cc: Santosh Raspatur <santosh@chelsio.com>
Cc: Hariprasad S <hariprasad@chelsio.com>
Cc: Yishai Hadas <yishaih@mellanox.com>
Cc: Oleg Drokin <oleg.drokin@intel.com>
Cc: "Yan, Zheng" <zyan@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
gcc-4.9 notices that the validate_init() function returns unintialized
data when called with a zero 'nr_buffers' argument, when called with the
-Wmaybe-uninitialized flag:
drivers/gpu/drm/nouveau/nouveau_gem.c: In function ‘validate_init.isra.6’:
drivers/gpu/drm/nouveau/nouveau_gem.c:457:5: error: ‘ret’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
However, the only caller of this function always passes a nonzero
argument, and gcc-6 is clever enough to take this into account and
not warn about it any more.
Adding an explicit initialization to -EINVAL here is correct even if
the caller changed, and it avoids the warning on gcc-4.9 as well.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-By: Karol Herbst <karolherbst@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Since fence_wait_timeout_reservation_object_wait_timeout_rcu() with a
timeout of 0 becomes reservation_object_test_signaled_rcu(), we do not
need to handle such conversion in the caller. The only challenge are
those callers that wish to differentiate the error code between the
nonblocking busy check and potentially blocking wait.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Sumit Semwal <sumit.semwal@linaro.org>
Link: http://patchwork.freedesktop.org/patch/msgid/20160829070834.22296-4-chris@chris-wilson.co.uk
Nothing too exciting here, there's a larger chunk of work that still
needs more testing but not likely to get that done today - so - here's
the rest of it. Assuming nothing else goes horribly wrong, I should be
able to send the rest Monday if it isn't too late....
Changes:
- Improvements to power sensor support
- Initial attempt at GM108 support
- Minor fixes to GR init + ucode
- Make use of topology information (provided by the GPU) in various
places, should at least fix some fault recovery issues and
engine/runlist mapping confusion on newer GPUs.
* 'linux-4.7' of git://github.com/skeggsb/linux: (51 commits)
drm/nouveau/gr/gf100-: fix race condition in fecs/gpccs ucode
drm/nouveau/core: recognise GM108 chipsets
drm/nouveau/gr/gm107-: fix touching non-existent ppcs in attrib cb setup
drm/nouveau/gr/gk104-: share implementation of ppc exception init
drm/nouveau/gr/gk104-: move rop_active_fbps init to nonctx
drm/nouveau/bios/pll: check BIT table version before trying to parse it
drm/nouveau/bios/pll: prevent oops when limits table can't be parsed
drm/nouveau/volt/gk104: round up in gk104_volt_set
drm/nouveau/fb/gm200: setup mmu debug buffer registers at init()
drm/nouveau/fb/gk20a,gm20b: setup mmu debug buffer registers at init()
drm/nouveau/fb/gf100-: allocate mmu debug buffers
drm/nouveau/fb: allow chipset-specific actions for oneinit()
drm/nouveau/gr/gm200-: fix bad hardcoding of a max-tpcs-per-gpc value
drm/nouveau/gr/gm200-: rop count == ltc count
drm/nouveau/gr/gm200: modify the mask when copying mmu settings from fb
drm/nouveau/gr/gm200: move some code into init_gpc_mmu() hook
drm/nouveau/gr/gm200: make generate_main() static
drm/nouveau/gr/gf100-: abstract fetching rop count
drm/nouveau/gr/gf100-: rename magic_not_rop_nr to screen_tile_row_offset
drm/nouveau/gr/gf100-: remove hardcoded idle_timeout values
...
After drm_gem_object_lookup() was changed along with all its callers,
we have several drivers that have unused variables:
drm/armada/armada_crtc.c: In function 'armada_drm_crtc_cursor_set':
drm/armada/armada_crtc.c:900:21: error: unused variable 'dev' [-Werror=unused-variable]
drm/nouveau/nouveau_gem.c: In function 'validate_init':
drm/nouveau/nouveau_gem.c:371:21: error: unused variable 'dev' [-Werror=unused-variable]
drm/nouveau/nv50_display.c: In function 'nv50_crtc_cursor_set':
drm/nouveau/nv50_display.c:1308:21: error: unused variable 'dev' [-Werror=unused-variable]
drm/radeon/radeon_cs.c: In function 'radeon_cs_parser_relocs':
drm/radeon/radeon_cs.c:77:21: error: unused variable 'ddev' [-Werror=unused-variable]
This fixes all the instances I found with ARM randconfig builds so far.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: a8ad0bd84f ("drm: Remove unused drm_device from drm_gem_object_lookup()")
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1463587653-3035181-6-git-send-email-arnd@arndb.de
drm_gem_object_lookup() has never required the drm_device for its file
local translation of the user handle to the GEM object. Let's remove the
unused parameter and save some space.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: dri-devel@lists.freedesktop.org
Cc: Dave Airlie <airlied@redhat.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
[danvet: Fixup kerneldoc too.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Not used any more.
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Not used any more.
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If pm_runtime_get_sync() we were going to "out" but we missed freeing
vma.
Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
No longer required in a lot of cases, as objects are identified over NVIF
via an alternate mechanism since the rework.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
On nv50+, we restrict the valid domains to just the one where the buffer
was originally created. However after the buffer is evicted to system
memory, we might move it back to a different domain that was not
originally valid. When sharing the buffer and retrieving its GEM_INFO
data, we still want the domain that will be valid for this buffer in a
pushbuf, not the one where it currently happens to be.
This resolves fdo#92504 and several others. These are due to suspend
evicting all buffers, making it more likely that they temporarily end up
in the wrong place.
Cc: stable@vger.kernel.org
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92504
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
A variety of tweaks to the NVIF library interfaces, mostly ripping out
things that turned out to be not so useful.
- Removed refcounting from nvif_object, callers are expected to not be
stupid instead.
- nvif_client is directly reachable from anything derived from nvif_object,
removing the need for heuristics to locate it
- _new() versions of interfaces, that allocate memory for the object
they construct, have been removed. The vast majority of callers used
the embedded _init() interfaces.
- No longer storing constructor arguments (and the data returned from
nvkm) inside nvif_object, it's more or less unused and just wastes
memory.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Use kvfree() instead of open-coding it.
Signed-off-by: Pekka Enberg <penberg@kernel.org>
Cc: David Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
User-space use mappable BOs notably for fences, and expects that a
value update by the GPU will be immediatly visible through the
user-space mapping.
ARM has a property that may prevent this from happening though: memory
can be mapped multiple times only if the different mappings share the
same caching properties. However all the lowmem memory is already
identity-mapped into the kernel with cache enabled, so when user-space
requests an uncached mapping, we actually get an "undefined caching
policy" one and this has strange side-effects described on Freedesktop
bug 86690.
To prevent this from happening, allow user-space to explicitly specify
which objects should be coherent, and create such objects with the
TTM_PL_FLAG_UNCACHED flag. This will make TTM allocate memory using the
DMA API, which will fix the identify mapping and allow us to safely map
the objects to user-space uncached.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Reviewed-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The namespace of NVKM is being changed to nvkm_ instead of nouveau_,
which will be used for the DRM part of the driver. This is being
done in order to make it very clear as to what part of the driver a
given symbol belongs to, and as a minor step towards splitting the
DRM driver out to be able to stand on its own (for virt).
Because there's already a large amount of churn here anyway, this is
as good a time as any to also switch to NVIDIA's device and chipset
naming to ease collaboration with them.
A comparison of objdump disassemblies proves no code changes.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
NVKM is having it's namespace switched to nvkm_, which will conflict
with these functions (which are workarounds for the fact that as of
yet, we still aren't able to split DRM and NVKM completely).
A comparison of objdump disassemblies proves no code changes.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Remove the function domain_to_ttm() that is not used anywhere.
This was partially found by using a static code analysis program
called cppcheck.
Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
- Fix BUG() on !SMP builds
- Fix for OOPS on pre-NV50 that snuck into -next
- MCP7[789A] hang fix where firmware hasn't already setup NISO pollers
- NV4x IGP MSI disable, it doesn't appear to work correctly
- Add GK208B to recognised boards (no code change aside from adding
chipset recognition)
* 'linux-3.19' of git://anongit.freedesktop.org/git/nouveau/linux-2.6:
drm/nouveau/nouveau: Do not BUG_ON(!spin_is_locked()) on UP
drm/nv4c/mc: disable msi
drm/nouveau/fb/ram/mcp77: enable NISO poller
drm/nouveau/fb/ram/mcp77: use carveout reg to determine size
drm/nouveau/fb/ram/mcp77: subclass nouveau_ram
drm/nouveau: wake up the card if necessary during gem callbacks
drm/nouveau/device: Add support for GK208B, resolves bug 86935
drm/nouveau: fix missing return statement in nouveau_ttm_tt_unpopulate
drm/nouveau/bios: fix oops on pre-nv50 chipsets
This reverts commit 355a701838.
This had some bad side effects under normal operation, and should
have been dropped earlier.
Signed-off-by: Dave Airlie <airlied@redhat.com>
The failure paths if we fail to wake the card are less than desirable,
but there's not really a graceful way to handle this case currently.
I'll keep this situation in mind when I get to fixing other vm-related
issues.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
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>
It happens on occasion that developers of generic user-space applications
abuse the dumb buffer API to get hold of drm buffers that they can both
mmap() and use for GPU acceleration, using the assumptions that dumb buffers
and buffers available for GPU are
a) The same type and can be aribtrarily type-casted.
b) fully coherent.
This patch makes the most widely used drivers warn nicely when that happens,
the next step will be to fail.
v2: Move drmP.h changes to drm_gem.h. Fix Radeon dumb mmap breakage.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
