Commit Graph

754 Commits

Author SHA1 Message Date
Christian König d564a06e1c drm/amdgpu: if a GDS switch is needed emit a pipeline sync as well
Otherwise we might change the GDS settings while they are still in use.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2016-03-08 11:01:50 -05:00
Christian König b8c7b39ec1 drm/amdgpu: split pipeline sync and vm flush
This allows us to use the pipeline sync for other tasks as well.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2016-03-08 11:01:49 -05:00
Christian König 971fe9a941 drm/amdgpu: switch the GDS only on demand v2
Switching the GDS space to often seems to be problematic.

This patch together with the following can avoid VM faults on context switch.

v2: extend commit message a bit

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Reviewed-by: Chunming Zhou <david1.zhou@amd.com> (v1)
2016-03-08 11:01:49 -05:00
Christian König cffadc83c7 drm/amdgpu: move the GDS switch into vm flush as well
After all it's an operation on the VMID.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2016-03-08 11:01:48 -05:00
Christian König 364beb2cc4 drm/amdgpu: return the common fence from amdgpu_fence_emit
Try to avoid using the hardware specific fences even more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2016-03-08 11:01:48 -05:00
Christian König 257bf15a4b drm/amdgpu: add slap cache for sync objects as well
We need them all the time.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2016-03-08 11:01:47 -05:00
Christian König 336d1f5efe drm/amdgpu: remove HW fence owner
Not used any more since we now always use the sheduler.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2016-03-08 11:01:47 -05:00
Christian König bcc634f4a8 drm/amdgpu: cleanup the sync code
No need for two functions doing the same, remove one and
add comments what those functions actually do.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
2016-03-08 11:01:46 -05:00
Christian König 637dd3b5ca drm/amdgpu: prevent get_user_pages recursion
Remember the tasks which are inside get_user_pages()
and ignore MMU callbacks from there.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:46 -05:00
Christian König 211dff5518 drm/amdgpu: group userptr in the BO list v2
We need them together with the next patch.

v2: Don't take bo reference twice

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:45 -05:00
Christian König a8bd1bec7c drm/amdgpu: sync to the active user on reusing a VMID
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:45 -05:00
Chunming Zhou 13459bd0fb drm/amdgpu: Clear HDP_MISC_CNTL.HDP_FLUSH_INVALIDATE_CACHE
So that we can invalidate and flush the HDP independently

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:44 -05:00
Chunming Zhou cc958e67aa drm/amdgpu: add hdp invalidation for sdma v3_0
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:44 -05:00
Chunming Zhou 6ad550c35d drm/amdgpu: add hdp invalidation for sdma v2_4
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:43 -05:00
Chunming Zhou 498dd97ddb drm/amdgpu: add hdp invalidation for cik sdma
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:42 -05:00
Chunming Zhou d35db5617a drm/amdgpu: add hdp invalidation for gfx8
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:42 -05:00
Chunming Zhou 0955860b87 drm/amdgpu: add hdp invalidation for gfx7
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:41 -05:00
Chunming Zhou 11afbde85e drm/amdgpu: add hdp_invalidate function
It's called after emitting ibs.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:41 -05:00
Chunming Zhou 5c55db83b7 drm/amdgpu: wait engine idle before vm flush for sdma
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:40 -05:00
Daniel Vetter b6bae58d8b drm/amdgpu: Switch to drm_vblank_on/off (v2)
Equivalent change to the radeon driver.

Note that with radeon this caught a bug in the dri3 DDX
implementation, which asked for vblank interrupts when the pipe is
off. That bug needs to be fixed before we can merge this patch (if
amdgpu is affected too). Michel discovered this one.

v2: agd: switch dce8 as well.

Acked-by: Michel Dänzer <michel.daenzer@amd.com>
Cc: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:40 -05:00
Nils Wallménius 0bf5df3b58 drm/amdgpu: delete set-but-not-read member has_uvd from amdgpu_device
Clean up leftover from radeon code.

Signed-off-by: Nils Wallménius <nils.wallmenius@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:39 -05:00
Alex Deucher 4223cc3de2 drm/amdgpu/ci: sync up with dpm changes from radeon
Looks like radeon commit:
d3052b8ce8
was missed.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:00:59 -05:00
Alex Deucher 12424e5528 drm/amdgpu/ci: remove redundant pcie setup
Looks like this got leftover by accident.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:00:53 -05:00
Alex Deucher 225ac13928 drm/amdgpu/ci: drop some old thermal setup
This was leftover from a long time ago and is not longer needed
since the thermal controller setup code was added.  Additional
mucking with the thermal interrupts can cause spurious thermal
events which can lead to unnecessary state changes.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 10:58:59 -05:00
Flora Cui 6157bd7a10 drm/amdgpu: fix rb bitmap & cu bitmap calculation
Fix some copy paste typos.

Signed-off-by: Flora Cui <Flora.Cui@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-03 01:00:20 -05:00
Christian König 22073fe764 drm/amdgpu: trace the pd_addr in vm_grab_id as well
Makes matching it to the flushes much easier.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-29 11:33:59 -05:00
Christian König 4ff37a83f1 drm/amdgpu: fix VM faults caused by vm_grab_id() v4
The owner must be per ring as long as we don't
support sharing VMIDs per process. Also move the
assigned VMID and page directory address into the
IB structure.

v3: assign the VMID to all IBs, not just the first one.
v4: use correct pointer for owner

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-29 11:33:46 -05:00
Dan Carpenter a65ecc4070 drm/amd: cleanup get_mfd_cell_dev()
It's simpler to just use snprintf() to print this to one buffer instead
of using strcpy() and strcat().  Also using snprintf() is slightly safer
than using sprintf().

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-26 12:43:47 -05:00
Christian König 70eacc72db drm/amdgpu: fix error handling in amdgpu_bo_list_set
Don't leak BOs in case of some error.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-26 01:50:18 -05:00
Alex Deucher aac1e3caac drm/amdgpu/gfx: fix off by one in rb rework (v2)
When I reworked this code, I messed up num rb count.

v2: use hweight32

Reviewed-by: Ken Wang <Qingquing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-25 17:44:57 -05:00
Dave Airlie 5263925c09 Merge branch 'drm-next-4.6' of git://people.freedesktop.org/~agd5f/linux into drm-next
First radeon and amdgpu pull request for 4.6.  Highlights:
- ACP support for APUs with i2s audio
- CS ioctl optimizations
- GPU scheduler optimizations
- GPUVM optimizations
- Initial GPU reset support (not enabled yet)
- New powerplay sysfs interface for manually selecting clocks
- Powerplay fixes
- Virtualization fixes
- Removal of hw semaphore support
- Lots of other misc fixes and cleanups

* 'drm-next-4.6' of git://people.freedesktop.org/~agd5f/linux: (118 commits)
  drm/amdgpu: Don't call interval_tree_remove in amdgpu_mn_destroy
  drm/amdgpu: Fix race condition in amdgpu_mn_unregister
  drm/amdgpu: cleanup gem init/finit
  drm/amdgpu: rework GEM info printing
  drm/amdgpu: print the GPU offset as well in gem_info
  drm/amdgpu: optionally print the pin count in gem_info as well
  drm/amdgpu: print the BO size only once in amdgpu_gem_info
  drm/amdgpu: print pid as integer
  drm/amdgpu: remove page flip work queue v3
  drm/amdgpu: stop blocking for page filp fences
  drm/amdgpu: stop calling amdgpu_gpu_reset from the flip code
  drm/amdgpu: remove fence reset detection leftovers
  drm/amdgpu: Fix race condition in MMU notifier release
  drm/radeon: Fix WARN_ON if DRM_DP_AUX_CHARDEV is enabled
  drm/amdgpu/vi: move uvd tiling config setup into uvd code
  drm/amdgpu/vi: move sdma tiling config setup into sdma code
  drm/amdgpu/cik: move uvd tiling config setup into uvd code
  drm/amdgpu/cik: move sdma tiling config setup into sdma code
  drm/amdgpu/gfx7: rework gpu_init()
  drm/amdgpu/gfx: clean up harvest configuration (v2)
  ...
2016-02-19 11:13:01 +10:00
Felix Kuehling 390be2824f drm/amdgpu: Don't call interval_tree_remove in amdgpu_mn_destroy
rbtree_postorder_for_each_entry_safe can skip over some entries if
the tree is rebalanced in interval_tree_remove. interval_tree_remove
is also redundant when the tree is just about to be freed.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2016-02-17 16:10:56 -05:00
Felix Kuehling b8ea378346 drm/amdgpu: Fix race condition in amdgpu_mn_unregister
Exchange locking order of adev->mn_lock and mm_sem, so that
rmn->mm->mmap_sem can be taken safely, protected by adev->mn_lock,
when amdgpu_mn_destroy runs concurrently.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2016-02-17 16:10:42 -05:00
Christian König 418aa0c296 drm/amdgpu: cleanup gem init/finit
Remove the double housekeeping and use something sane to
forcefuly delete BOs on unload.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:46:24 -05:00
Christian König 7ea2356533 drm/amdgpu: rework GEM info printing
Print BOs grouped per client.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:45:32 -05:00
Christian König ebb36d19a0 drm/amdgpu: print the GPU offset as well in gem_info
To easily find which memory is used.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:44:43 -05:00
Christian König b22e3ce8a6 drm/amdgpu: optionally print the pin count in gem_info as well
Usefull when debugging page flipping.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:43:54 -05:00
Christian König de0549008a drm/amdgpu: print the BO size only once in amdgpu_gem_info
Splitting it into KB/MB is just confusing.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:43:16 -05:00
Christian König 3e4ae1dcbb drm/amdgpu: print pid as integer
Not sure why somebody thought that this is a long.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:31:15 -05:00
Christian König 87d58c11ec drm/amdgpu: remove page flip work queue v3
Just use the system queue now that we don't block any more.

v2: handle DAL as well.
v3: agd: split DAL changes out

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Mykola Lysenko <mykola.lysenko@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com> (v1)
2016-02-16 17:25:38 -05:00
Christian König c3874b758a drm/amdgpu: stop blocking for page filp fences
Just register an callback and reschedule the work item if necessary.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:24:07 -05:00
Christian König 6800e2ea2e drm/amdgpu: stop calling amdgpu_gpu_reset from the flip code
We don't return -EDEADLK any more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:23:31 -05:00
Christian König 9b3896686f drm/amdgpu: remove fence reset detection leftovers
wait_event() never returns before the fence was signaled.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-16 17:22:16 -05:00
Felix Kuehling fa5b500030 drm/amdgpu: Fix race condition in MMU notifier release
The release notifier can get called a second time from
mmu_notifier_unregister depending on a race between
__mmu_notifier_release and amdgpu_mn_destroy. Use
mmu_notifier_unregister_no_release to avoid this.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2016-02-16 17:17:14 -05:00
Carlos Palminha b6e740837c drm/amdgpu: removed optional dummy encoder mode_fixup function.
mode_fixup function for encoder drivers became optional with patch
http://patchwork.freedesktop.org/patch/msgid/1455106522-32307-1-git-send-email-palminha@synopsys.com

This patch set nukes all the dummy mode_fixup implementations.

(made on top of Daniel topic/drm-misc branch)

Signed-off-by: Carlos Palminha <palminha@synopsys.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/8c4710ee930d2ab3a505b8d3cd724f4f4c79b3df.1455540137.git.palminha@synopsys.com
2016-02-16 15:31:38 +01:00
Alex Deucher 549300ceae drm/amdgpu/vi: move uvd tiling config setup into uvd code
Split uvd and gfx programming.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-12 15:53:14 -05:00
Alex Deucher c458fe9425 drm/amdgpu/vi: move sdma tiling config setup into sdma code
Split sdma and gfx programming.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-12 15:53:04 -05:00
Alex Deucher 76ed6cb017 drm/amdgpu/cik: move uvd tiling config setup into uvd code
Split uvd and gfx programming.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-12 15:52:53 -05:00
Alex Deucher 2b3a765d83 drm/amdgpu/cik: move sdma tiling config setup into sdma code
Split sdma and gfx programming.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-12 15:52:43 -05:00
Alex Deucher d93f3ca706 drm/amdgpu/gfx7: rework gpu_init()
Split the sw and hw parts into separate functions.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-12 15:52:03 -05:00