Commit Graph

93 Commits

Author SHA1 Message Date
Christian König
32b41ac21f drm/amdgpu: Revert "add mutex for ba_va->valids/invalids"
Not needed any more because we need to protect the elements on the list anyway.

This reverts commit 38bf516c75b4ef0f5c716e05fa9baab7c52d6c39.

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>
2016-03-09 13:04:02 -05:00
Christian König
2025021582 drm/amdgpu: Revert "add lock for interval tree in vm"
Not needed any more because we need to protect the elements on the list anyway.

This reverts commit fe237ed7efec8ac147a4572fdf81173a7f8ddda7.

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>
2016-03-09 13:04:02 -05:00
Christian König
e17841b975 drm/amdgpu: Revert "add spin lock to protect freed list in vm (v3)"
Not needed any more because we need to protect the elements on the list anyway.

This reverts commit dae6ecf9e6c9b677e577826c3ac665c6dd9c490b.

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>
2016-03-09 13:04:01 -05:00
Felix Kuehling
fb29b57c34 drm/amdgpu: Fix two bugs in amdgpu_vm_bo_split_mapping
Off-by-one: last is inclusive, so the maximum is start + max_size - 1
Wrong unit: addr is in bytes, max_size is in pages

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-09 13:04:00 -05:00
Christian König
2f568dbd6b drm/amdgpu: move get_user_pages out of amdgpu_ttm_tt_pin_userptr v6
That avoids lock inversion between the BO reservation lock
and the anon_vma lock.

v2:
* Changed amdgpu_bo_list_entry.user_pages to an array of pointers
* Lock mmap_sem only for get_user_pages
* Added invalidation of unbound userpointer BOs
* Fixed memory leak and page reference leak

v3 (chk):
* Revert locking mmap_sem only for_get user_pages
* Revert adding invalidation of unbound userpointer BOs
* Sanitize and fix error handling

v4 (chk):
* Init userpages pointer everywhere.
* Fix error handling when get_user_pages() fails.
* Add invalidation of unbound userpointer BOs again.

v5 (chk):
* Add maximum number of tries.

v6 (chk):
* Fix error handling when we run out of tries.

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> (v4)
Acked-by: Alex Deucher <alexander.deucher@amd.com>
2016-03-08 11:01:50 -05:00
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
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
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
Christian König
2d55e45a03 drm/amdgpu: use SDMA round robin for VM updates v3
Distribute the load on both rings.

v2: use a loop for the initialization
v3: agd: rebase on upstream

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-12 15:38:16 -05:00
Christian König
2bd9ccfa75 drm/amdgpu: use per VM entity for page table updates (v2)
Updates from different VMs can be processed independently.

v2: agd: rebase on upstream

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-12 15:35:16 -05:00
Christian König
e86f9ceee1 drm/amdgpu: move sync into job object
No need to keep that for every IB.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:17:24 -05:00
Christian König
d71518b5aa drm/amdgpu: cleanup in kernel job submission
Add a job_alloc_with_ib helper and proper job submission.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:17:22 -05:00
Christian König
b07c60c065 drm/amdgpu: move ring from IBs into job
We can't submit to multiple rings at the same time anyway.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:17:20 -05:00
Christian König
9e5d53094c drm/amdgpu: make pad_ib a ring function v3
The padding depends on the firmware version and we need that for BO moves as
well, not only for VM updates.

v2: new approach of making pad_ib a ring function
v3: fix typo in macro name

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:17:20 -05:00
Christian König
b6ea2f37a2 drm/amdgpu: fix size estimation for clear IB
We only need a few dw here.

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>
2016-02-10 14:17:13 -05:00
Christian König
31f6c1fedb drm/amdgpu: optimize amdgpu_vm_update_ptes a bit
Don't calculate the end address multiple times.

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>
2016-02-10 14:17:07 -05:00
Christian König
8843dbbbf3 drm/amdgpu: cleanup comments in VM code
Neither the global nor the local mutex exists any more and
amdgpu doesn't support cayman.

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>
2016-02-10 14:17:07 -05:00
Christian König
a1e08d3b83 drm/amdgpu: optimize VM fencing
No need to fence every page table, just the page directory is enough.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:17:06 -05:00
Christian König
a14faa6573 drm/amdgpu: split VM mappings into smaller operations (v3)
If we can't copy entries from the GTT or fill them with one command split
up the mapping operation into multiple ones.

v2: agd: rebase on upstream
v3: squash in Christian's fix

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>
2016-02-10 14:17:05 -05:00
Christian König
9ab2146289 drm/amdgpu: use BOs GART instance for mapping addresses v4
That allows the VM code to use GART BOs from other driver instances.

v2: don't use copy optimization for foreign GARTs, that won't work.
v3: some more comment cleanups
v4: agd: rebase on upstream

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>
2016-02-10 14:17:05 -05:00
Christian König
b07c9d2a73 drm/amdgpu: move more logic into amdgpu_vm_map_gart v3
No need to duplicate that code over and over again. Also stop using the
flags to determine if we need to map the addresses.

v2: constify the pages_addr
v3: rebased, fix typo in commit message

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>
2016-02-10 14:17:04 -05:00
Christian König
599f434817 drm/amdgpu: remove nonsense IB size checks
Those are just leftovers from the time we wrote the VM
updates directly to the ring.

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>
2016-02-10 14:17:04 -05:00
Christian König
a9a78b329a drm/amdgpu: use a global LRU list for VMIDs
With the scheduler enabled managing per ring LRUs don't
make much sense any more.

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>
2016-02-10 14:16:58 -05:00
Christian König
94dd0a4ae0 drm/amdgpu: merge vm_grab_id and vm_fence v2
No need for an extra function any more.

v2: comment cleanups

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>
2016-02-10 14:16:57 -05:00
Christian König
165e4e07c2 drm/amdgpu: add VM pointer to id trace
Because of the scheduler all traces come from the same thread now and
can't be distincted otherwise.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:56 -05:00
Chunming Zhou
cadf97b196 drm/amdgpu: clean up non-scheduler code path (v2)
Non-scheduler code is longer supported.

v2: agd: rebased on upstream

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Ken Wang  <Qingqing.Wang@amd.com>
Reviewed-by: Monk Liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:50 -05:00
Christian König
1ea863fd73 drm/amdgpu: keep the prefered/allowed domains in the BO
Stop copying that to the bo list entry, it doesn't change anyway.

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>
2016-02-10 14:16:47 -05:00
Michel Dänzer
9571e1d840 drm/amdgpu: Use drm_calloc_large for VM page_tables array
It can be big, depending on the VM address space size, which is tunable
via the vm_size module parameter.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93721
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-01-19 12:25:09 -05:00
Christian König
eceb8a1562 drm/amdgpu: move VM page tables to the LRU end on CS v2
This makes it less likely to run into an ENOMEM because
VM page tables are evicted last.

v2: move the BOs in the LRU tail after validation

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-01-13 12:22:17 -05:00
Dave Airlie
fd3e14ffbd Merge branch 'drm-next-4.5' of git://people.freedesktop.org/~agd5f/linux into drm-next
[airlied: fixup build problems on arm - added errno.h include]
* 'drm-next-4.5' of git://people.freedesktop.org/~agd5f/linux: (152 commits)
  amd/powerplay: fix copy paste typo in hardwaremanager.c
  amd/powerplay: disable powerplay by default initially
  amd/powerplay: don't enable ucode fan control if vbios has no fan table
  drm/amd/powerplay: show gpu load when print gpu performance for Cz. (v2)
  drm/amd/powerplay: check whether need to enable thermal control. (v2)
  drm/amd/powerplay: add point check to avoid NULL point hang.
  drm/amdgpu/powerplay: Program a calculated value as Deep Sleep clock.
  drm/amd/powerplay: Don't return an error if fan table is missing
  drm/powerplay/hwmgr: log errors in tonga_hwmgr_backend_init
  drm/powerplay: add debugging output to processpptables.c
  drm/powerplay: add debugging output to tonga_processpptables.c
  amd/powerplay: Add structures required to report configuration change
  amd/powerplay: Fix get dal power level
  amd\powerplay Implement get dal power level
  drm/amd/powerplay: display gpu load when print performance for tonga.
  drm/amdgpu/powerplay: enable sysfs and debugfs interfaces late
  drm/amd/powerplay: move shared function of vi to hwmgr. (v2)
  drm/amd/powerplay: check whether enable dpm in powerplay.
  drm/amd/powerplay: fix bug that dpm funcs in debugfs/sysfs missing.
  drm/amd/powerplay: fix boolreturn.cocci warnings
  ...
2015-12-23 14:15:26 +10:00
Felix Kuehling
005ae95e6e drm/amdgpu: Fix off-by-one errors in amdgpu_vm_bo_map
eaddr is sometimes treated as the last address inside the address
range, and sometimes as the first address outside the range. This
was resulting in errors when a test filled up the entire address
space. Make it consistent to always be the last address within the
range.

Signed-off-by: Felix.Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
2015-12-21 16:39:14 -05:00
Christian König
ee1782c3f2 drm/amdgpu: keep the PTs validation list in the VM v2
This avoids allocating it on the fly.

v2: fix grammar in comment

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>
2015-12-18 17:29:46 -05:00
Christian König
56467ebfb2 drm/amdgpu: split VM PD and PT handling during CS
This way we avoid the extra allocation for the page directory entry.

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>
2015-12-18 17:29:45 -05:00
Christian König
3c0eea6c35 drm/amdgpu: put VM page tables directly into duplicates list
They share the reservation object with the page directory anyway.

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>
2015-12-18 17:29:45 -05:00
jimqu
81d75a30c6 drm/amdgpu: add spin lock to protect freed list in vm (v2)
there is a protection fault about freed list when OCL test.
add a spin lock to protect it.

v2: drop changes in vm_fini

Signed-off-by: JimQu <jim.qu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-12-04 13:08:13 -05:00
jimqu
9c4153b1ee drm/amdgpu: add spin lock to protect freed list in vm (v2)
there is a protection fault about freed list when OCL test.
add a spin lock to protect it.

v2: drop changes in vm_fini

Signed-off-by: JimQu <jim.qu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-12-04 12:23:38 -05:00
Christian König
82b9c55b1e drm/amdgpu: fix VM page table reference counting
We use the reservation object of the page directory for the page tables as
well, because of this the page directory should be freed last. Ensure that
by keeping a reference from the page tables to the directory.

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-11-30 14:41:33 -05:00
Chunming Zhou
e98c1b0de6 drm/amdgpu: remove vm->mutex
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-11-20 18:22:28 -05:00
Chunming Zhou
69b576a1bc drm/amdgpu: add mutex for ba_va->valids/invalids
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-11-20 18:22:23 -05:00
Chunming Zhou
49b02b180a drm/amdgpu: reserve/unreserve objects out of map/unmap operations
Change-Id: Id6514f2fb6e002437fdbe99353d5d35f4ac736c7
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-11-18 11:41:20 -05:00
Chunming Zhou
ef9f0a83d6 drm/amdgpu: move bo_reserve out of amdgpu_vm_clear_bo
Change-Id: Ifbb0c06680494bfa04d0be5e5941d31ae2e5ef28
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-11-18 11:41:02 -05:00
Chunming Zhou
c25867dfab drm/amdgpu: add lock for interval tree in vm
Change-Id: I62b892a22af37b32e6b4aefca80a25cf45426ed2
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-11-18 11:40:55 -05:00
Christian König
1c16c0a7b2 drm/amdgpu: keep the owner for VMIDs
We don't need the last VM use any more, keep the owner directly.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <davdi1.zhou@amd.com>
2015-11-18 11:40:37 -05:00
Christian König
ea89f8c9e8 drm/amdgpu: move VM manager clean into the VM code again
It's not a good idea to duplicate that code.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <davdi1.zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-11-18 11:40:27 -05:00
Christian König
16ae42feb0 drm/amdgpu: use common fence for amdgpu_vm_fence
Just cleanup the function parameters.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-11-16 11:05:49 -05:00
Christian König
b56c22853e drm/amdgpu: use fence_is_later() for vm_flush as well v2
v2: remove superfluous check

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>
2015-11-16 11:05:48 -05:00