Fixes suspend issues with UVD.
Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
Invalid messages can crash the hw otherwise
Ported from radeon commit a1b403da70
Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
This causes problems with multiple suspend/resume cycles.
Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
Embed the scheduler into the ring structure instead of allocating it.
Use the ring name directly instead of the id.
v2: rebased, whitespace cleanup
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Chunming Zhou<david1.zhou@amd.com>
Move the fence related stuff into amdgpu_fence.c
v2: rework commit message, cause this is actually not a bug
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou<david1.zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Just to be consistent with the other members.
v2: rename the ring member as well.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com> (v1)
Reviewed-by: Chunming Zhou<david1.zhou@amd.com>
Use container_of rather than casting.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: David Zhou <david1.zhou@amd.com>
Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Use consistent naming across functions.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: David Zhou <david1.zhou@amd.com>
Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reduces the locking and fencing overhead.
v2: add comment why we need the duplicates list in the GEM op.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
This allows for multiple BOs to have the same reservation object.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Adds an extra argument to amdgpu_bo_create, which is only used in amdgpu_prime.c.
Port of radeon commit 831b6966a6.
v2: fix up kfd.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
mem->start is a long, so this can overflow on 32bit systems.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Cc: stable@vger.kernel.org
Just free the resources immediately after submitting the job.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Make sure the CP waits for the write to be confirmed before
invalidating.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Insert wait for reg mem after EOP to fix potential issue with vm context switch
v2: move wait to vm_flush() use equal instead of greater than.
Signed-off-by: Anatoli Antonovitch <anatoli.antonovitch@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Provide module parameter to enable/disable them. Still
enabled by default.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
It should be gfx_v8_0_init_compute_vmid since it's
part of the gfx block.
Acked-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Need to properly handle the instances for the idle
checks and soft reset.
Acked-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
it seems the VCE ring 1 ib test not reliable, remove it for now.
Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Tested-and-Reviewed-by: Sonny Jiang <sonny.jiang@amd.com>
Some buffers (UVD/VM page tables) must be placed in VRAM,
but the byte restriction for moving buffers didn't took this
into account.
Port of radeon commit 4b09556660.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
The logic was reversed. This feature is not enabled
at the moment, but fix it now for the future.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Using the wrong watermwark value for the secondary
watermark. Copy paste typo. Noticed by Mykola.
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Using the wrong watermwark value for the secondary
watermark. Copy paste typo. Noticed by Mykola.
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Should help avoid fragmentation of vram due to CPU access
requirements.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For kernel driver BOs, be explicit about whether we need
vram access up front. This avoids unecessary migrations and
avoids using visible vram for buffers were it's not needed.
v2: line wrap fixes
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Customize the insert_nop func for SDMA rings, and use burst NOP for
ring/IB submissions in other places as well
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
The insert_nop function is added to amdgpu_ring_funcs structure as
well as the default implementation
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
This function is added to map the ring to sdma instance
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
The burst NOP is supported for SDMA when feature_version is >= 20.
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
This is added to support the burst NOP
v2: squash the typo fix
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Instead of the array which is used for ID management.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
No copy_(to|from)_user while BO is reserved.
v2: handle default path as well
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
That isn't used any more.
v2: rebase
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>
sync_resv is to handle both amdgpu_fence and sched_fence.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
AUX addresses are 20 bits long. Send out the entire address instead of
just the low 16 bits.
Port of:
drm/radeon/atom: Send out the full AUX address
to amdgpu
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Saving the current UVD state on suspend and restoring it on resume
just doesn't work reliable. Just close cleanup all sessions on suspend.
Ported from radeon commit "12e49feadff6d7b7ebbe852b36943a71524d8d34".
v2: rebased
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
This way the scheduler doesn't wait in it's work thread any more.
v2: fix race conditions
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
The context needs to finish before everything else.
Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
This aids handling buffers moves with the scheduler.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
The fence in the array may be skipped if wait_all is false,
thus the related callback is not initialized with list head.
So removing this kind callback will cause NULL pointer reference.
Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
when eviction is happening, if don't handle
dependency, then the fence could be dead off.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
Entity don't live as long as scheduler fences.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Calling schedule() is probably the worse things we can do.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
Simply not used any more. Only keep 32bit atomic for fence sequence numbering.
v2: trivial rebase
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com> (v1)
Reviewed-by: Chunming Zhou <david1.zhou@amd.com> (v1)
Rename the function and update the related code with this modified function.
Add the new parameter of bool wait_all.
If wait_all is true, it will return when all fences are signaled or timeout.
If wait_all is false, it will return when any fence is signaled or timeout.
Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Monk Liu <monk.liu@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Use pci_alloc_consistent rather than kzalloc since we
need 256 byte aligned memory for the ring buffer.
v2: fix copy paste typo in free function noticed
by Jammy.
bug:
https://bugs.freedesktop.org/show_bug.cgi?id=91749
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
This is a port of:
DRM - radeon: Don't link train DisplayPort on HPD until we get the dpcd
to amdgpu.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Stop double freeing the the BO list by pulling the content
of amdgpu_cs_parser_prepare_job() into the IOCTL function again.
v2: better commit message
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com> (v1)
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
The problem now is that we don't necessarily call amdgpu_ib_get()
in some error paths and so work with uninitialized data.
Better require that the memory is already zeroed.
v2: better commit message
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com> (v1)
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Used by mesa, etc. for profiling.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Pagetables can be moved and therefore the page directory update can be necessary
for the current cs even if none of the the bo's are moved. In that scenario
there is no fence between the sdma0 and gfx ring, so we add one.
v2 (chk): rebased
Signed-off-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Rework run queue implementation, especially remove the odd list handling.
v2: cleanup the code only, no algorithem change.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
fix the bug that there is duplicated bo_update_mapping issued
Signed-off-by: monk.liu <monk.liu@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
bo_list_clone() will take a lot of time when bo_list hold too much
elements, like above 7000
Signed-off-by: Monk.Liu <monk.liu@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Jammy Zhou <jammy.zhou@amd.com>
It's not validated yet and causes more harm than good.
Avoids spurious resets.
Reviewed-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
remaining timeout returned by amdgpu_fence_wait_any can be larger than
max int value, thus the truncated 32 bit value in r ends up being
negative while its original long value is positive.
Signed-off-by: monk.liu <monk.liu@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <jammy.zhou@amd.com>
fix fence is released when pass to **fence sometimes.
add reference for it.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Tested-and-Reviewed-by: Leo Liu <leo.liu@amd.com>
Not used any more.
v2: remove amd_sched_emit as well.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Similar to radeon, except that amdgpu doesn't even use struct_mutex to
protect anything like the shared z buffer (sane gpu architecture,
yay!). And the code already grabs the globa adev->ring_lock, so this
code can't race with itself. Which makes struct_mutex completely
redundnant. Remove it.
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
It really doesn't protect anything which doesn't have other locks
already. Also this is run from driver unload code so not much need for
locks anyway.
Same changes as for radeon really.
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The workaround simply doesn't work because VM mappings
are controlled by userspace not the kernel.
Additional to that this is just a performance problem
which happens if you have holes in your VM mapping.
v2: adjust virtual addr alignment as well.
v3: fix trivial warning
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Monk Liu <monk.liu@amd.com> (v1)
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com> (v2)
Looks like that somehow got missed while during porting the radeon changes.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
It was just a wrapper for fence_wait anyway.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
The common kernel function does the same thing.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
scheduler fence is based on kernel fence framework.
v2: squash in Christian's build fix
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
v2: rebased
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>
Avoiding a couple of casts.
v2: rename c_entity to entity as well
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
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>
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>
Cleanup the kernel context handling.
v2: rebased
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com> (v1)
Id's are for the IOCTL ABI only.
v2: remove tgid as well
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
We didn't initialized the mutex in the cloned bo list resulting in nice
warnings from lockdep. Also fixes error handling in this function.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
This way can avoid interrupt lost, and can process sched job exactly.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
This function is used to get the next queued sequence number
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
This function is to update last_emitted_v_seq and wake up the waiters.
It should be called by driver in the run_job backend function
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
It is clean to update last_queued_v_seq in the scheduler module
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Fix the code alignment, etc.
v2: rebase the code
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
every sbumission should be able to get a fence.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <jammy.zhou@amd.com>
It is theoretically possible that a swapped out BO gets the
same GTT address, but different backing pages while being swapped in.
Instead just use another VA state to note updated areas.
Ported from not upstream yet radeon commit with the same name.
v2: fix some bugs in the original implementation found in the radeon code.
v3: squash in VCE/UVD fix
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
thus unnecessary wake up could be avoid between rings
v2:
move wait_queue_head to fence_drv from ring
Signed-off-by: monk.liu <monk.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
origninal method will sleep/schedule at the granurarity of HZ/2 and
based on seq signal method, the new implement is based on kernel fance
interface, no unnecessary schedule at all
v2: replace logic of original amdgpu_fence_wait_any
Signed-off-by: monk.liu <monk.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
if enabling scheduler, then the queued seq is assigned
when pushing job before emitting job.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
the job must be emitted by scheduler, otherwise scheduler is abnormal.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
This option can be used to specify the max number of submissions in the
active HW queue. The default value is 2 now.
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
This option can be used to specify the max job number in the job queue,
and it is 16 by default.
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
it is possible that the callback isn't defined sometimes.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian K?nig <christian.koenig@amd.com>
fence_process may be called from kthread, user thread and interrupt context.
it is possible to called concurrently, then will wake up fence queue multiple times.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
use kernel context to submit command for vm
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Christian K?nig <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
user mode will still use pte ring as a normal ring.
if the prepare job generates another command(update pte) on its ring in scheduler,
then will kill scheduler which is going to waiting later job but pending running job.
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Christian K?nig <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
BO validation is moved to scheduler except usrptr which must be validated
in user process
Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Christian K?nig <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
v2: agd5f: fix the rb setup.
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
v2: agd5f: fix up XDMA golden settings
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
v2: agd5f: prepare for release
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: David Zhang <david1.zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
use kmalloc_array instead of kcalloc where appropriate and other
cleanups.
Signed-off-by: monk.liu <monk.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Make the definitions common for all driver components
v2: fix kfd
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Feel free to suggest more.
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
the definitions can be shared by different IP components.
v2: fix include path
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
We need to return the sequence number to userspace
even when we don't use user fences.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
This makes assigning VM IDs independent from the use of VM IDs.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
v2 (chk): split fix from original patch
Signed-off-by: monk.liu <monk.liu@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Otherwise the first 16 fences of a context will always signal immediately.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
It's not used any more.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
This is a prerequisite for the GPU scheduler to make the order
of submission independent from the order of execution.
v2: properly implement the locking
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
This also requires some error handling from the callers of that function.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
The comment is misleading and incorrect, remove it.
Printing the id is completely meaningless and this practice
can cause a race conditions on command submission.
The flags and hangs fields are completely unused.
Give all fields a common indentation.
v2: remove fpriv reference and unused flags as well, fix debug message.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
This new interface can be used by IP components to retrieve the
firmware information from the core driver.
v2: fix one typo
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Rex Zhu <Rex.Zhou@amd.com>
Signed-off-by: Young Yang <Young.Yang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
This implements the interface for atombios command
and data table access.
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
CGS (Common Graphics Services) is an AMD cross component
abstraction layer to designed to better encapsulate
specific IP block drivers so different teams can effectively
work on differnet IP block drivers independently. It provides
a common interface for things like accessing registers,
allocating GPU memory, and registering interrupt sources.
The plan is to eventually move more and more IP drivers to
this interface. The first user is the ACP IP driver.
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In function 'amdgpu_uvd_cs_pass2':
warning: 'min_ctx_size' may be used uninitialized in this function
buf_sizes[0x4] = min_ctx_size;
^
note: 'min_ctx_size' was declared here
unsigned image_size, tmp, min_dpb_size, num_dpb_buffer, min_ctx_size;
^
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
Two small bug fixes for the code you pulled for 4.3:
- Used a SHIFT define instead of a MASK define to check if a bit is turned on
when destroying hqd. Luckily, this is in gfx7 interface file with amdgpu,
which was used only for bring-up purposes of amdgpu, so no real effect on
a running system
- Used a logical AND instead of a bitwise AND operator, when initializing
sdma virtual memory when using SDMA queues
* tag 'drm-amdkfd-next-fixes-2015-08-05' of git://people.freedesktop.org/~gabbayo/linux:
drm/amdkfd: fix bug when initializing sdma vm
drm/amdgpu: fix bug when amdkfd destroys hqd
The fw_version and feature_verion should be set correctly when the
firmwares are loaded by SMU on Tonga/Carrzio/Iceland
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Expose feature version to user space for RLC/MEC/MEC2 ucode as well
v2: fix coding style
Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
gfx_v7_0_print_status contains a for loop on variable queue which does
not update this variable between each iteration. This is bug is
reported by clang while building allmodconfig LLVMLinux on x86_64:
drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c:5126:19: error: variable
'queue' used in loop condition not modified in loop body
[-Werror,-Wloop-analysis]
for (queue = 0; queue < 8; i++) {
^~~~~
Fix this by incrementing variable queue instead of i in this loop.
Signed-off-by: Nicolas Iooss <nicolas.iooss_linux@m4x.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Always set num_rbs to 2 for CZ. The 1 RB parts are often harvest
configs. The will get sorted out in mesa when we program
PA_SC_RASTER_CONFIG[_1].
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The wrong define was used to check if the hqd is still active
v2: Don't use SHIFT as the MASK is already shifted
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
hdp flush is not needed for IBs that dispatched from kernel inside
because there is no video memory host access
Signed-off-by: monk.liu <monk.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
compute ring didn't use const engine byfar, so ignore CE things in
compute routine
Signed-off-by: monk.liu <monk.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
We recently changed the drm_amdgpu_info_device struct so now there is
a 4 byte hole at the end. We need to initialize it so we don't disclose
secret information from the stack.
Fixes: fa92754e9c ('drm/amdgpu: add VCE harvesting instance query')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If we fail during device init, record what state each
block is in so that we can tear down clearly.
Fixes various problems on device init failure.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
VCE, UVD DPM work similarly to SCLK DPM. Report the current
clock levels for UVD and VCE via debugfs.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
CZ uses a different set of registers compared to previous asics
and supports separate NB and GFX planes.
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For boards with bad VCE blocks, only configure the working
block.
v2: use the harvest info for pipe setup
v3: fix mask check as noted by Leo
v4: add dGPU support
Reviewed-by: Christian König <christian.koenig@amd.com> (v2)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This doesn't seem strictly necessary with Tonga right now, but that might
change with future power management enhancements.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Something (ATOM BIOS?) seems to be clobbering the LB_INTERRUPT_MASK
register while the CRTC is off, which caused e.g. glxgears or
gnome-shell to hang after a modeset.
Reviewed-and-Tested-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Sonny Jiang <sonny.jiang@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
This patch adds the gfx8 interface file between amdgpu and amdkfd. This
interface file is currently in use when running on a Carrizo-based
system.
The interface itself is represented by a pointer to struct
kfd_dev. The pointer is located inside amdgpu_device structure.
All the register accesses that amdkfd need are done using this
interface. This allows us to avoid direct register accesses in
amdkfd proper, while also allows us to avoid locking between
amdkfd and amdgpu.
The single exception is the doorbells that are used in both of
the drivers. However, because they are located in separate pci
bar pages, the danger of sharing registers between the drivers
is minimal.
Having said that, we are planning to move the doorbells as well
to amdgpu.
Signed-off-by: Ben Goz <ben.goz@amd.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
This patch adds the gfx7 interface file between amdgpu and amdkfd. This
interface file mirrors (some) of the functions in radeon_kfd.c
(the interface file between radeon and amdkfd).
The gfx7 interface is used when it is run on a Kaveri-based system.
This interface file was used for bring-up of amdkfd on amdgpu and for
debugging purposes. For users who would like to run HSA on Kaveri, please
use the radeon graphic driver.
Note: CONFIG_DRM_AMDGPU_CIK must be selected for amdgpu to handle Kaveri.
v2: removed MTYPE_NONCACHED enum definition as it is defined in another
patch
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
This patch adds an interface file between amdgpu and amdkfd. This
interface file is H/W agnostic, thus containing functions that
operate the same for any AMD APU/GPU H/W generation.
The functions in this interface mirror (some) of the functions in
radeon_kfd.c (the radeon<-->amdkfd interface file). The main functions
are:
- amdgpu_amdkfd_init - initialize the amdkfd module
- amdgpu_amdkfd_load_interface - load the H/W interface according to the
currently probed device
- amdgpu_amdkfd_device_probe - probe the device in amdkfd
- amdgpu_amdkfd_device_init - initialize the device in amdkfd
- amdgpu_amdkfd_interrupt - call the ISR of amdkfd
- amdgpu_amdkfd_suspend - suspend callback from amdgpu
- amdgpu_amdkfd_resume - resume callback from amdgpu
This patch also modifies the relevant amdgpu files, to use this new
interface.
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Something (ATOM BIOS?) seems to be clobbering the LB_INTERRUPT_MASK
register while the CRTC is off, which caused e.g. glxgears or
gnome-shell to hang after a modeset.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
So the selected forced level shows up properly in sysfs.
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This is necessary to properly reset the min/max limits before
clamping them otherwise we may get improper clamping depending
on what what was the last forced level.
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Always respect the harvest configuration as is.
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If we run into a part with a harvest configuration where the entire
IP block is unusable, just disable the IP block.
v2: fix logic as noted by Christian
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Just to make sure userspace don't send nonsense to the kernel.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
This is a translation of the patch ...
"drm/radeon: Handle irqs only based on irq ring, not irq status regs."
... for the vblank irq handling, to fix the same problem described
in that patch on the new driver.
Only compile tested due to lack of suitable hw.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
CC: Michel Dänzer <michel.daenzer@amd.com>
CC: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
v2: remove unrelated whitespace change, fix C comment
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
And use common fence infrastructure for the wait.
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>
If the CONFIG_DEBUG_FS is not selected, compilation of the
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c provides two warnings that
amdgpu_debugfs_regs_init and amdgpu_debugfs_regs_cleanup are used but
never defined. And as result:
ERROR: "amdgpu_debugfs_regs_cleanup" [drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined!
ERROR: "amdgpu_debugfs_regs_init" [drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined!
^
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Remove duplication across asic families and make it symmetric
with the freeing of the code in amdgpu_device.c
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
kfree(NULL) is safe and this check is probably not required
Signed-off-by: Maninder Singh <maninder1.s@samsung.com>
Reviewed-by: Vaneet Narang <v.narang@samsung.com>
Reviewed-by: Christian Konig <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Use kzalloc rather than kcalloc(1.. for allocating
one thing.
Signed-off-by: Maninder Singh <maninder1.s@samsung.com>
Reviewed-by: Vaneet Narang <v.narang@samsung.com>
Reviewed-by: Christian Konig <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to resolve compute hang at resume time.
v2: (agd5f) squash in second fix
Signed-off-by: Sonny Jiang <sonny.jiang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
We were doing it in the common code and in the IP specific code.
Remove the IP specific code. The common code handles the
ordering properly.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Enable VCE dpm and powergating. VCE dpm dynamically scales the VCE clocks on
demand.
Signed-off-by: Sonny Jiang <sonny.jiang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
smu_init allocates buffers and initializes them. It does not
touch the hw. There is no need to do it again on resume. It
should really be part of sw_init (and smu_fini should be part
of sw_fini), but we need the firmware sizes from the other IPs
for firmware loading so we have to wait until sw init is done
for all other IPs.
Reviewed-by: Sonny Jiang <Sonny.Jiang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
smu_init allocates buffers and initializes them. It does not
touch the hw. There is no need to do it again on resume. It
should really be part of sw_init (and smu_fini should be part
of sw_fini), but we need the firmware sizes from the other IPs
for firmware loading so we have to wait until sw init is done
for all other IPs.
Reviewed-by: Sonny Jiang <Sonny.Jiang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
smu_init allocates buffers and initializes them. It does not
touch the hw. There is no need to do it again on resume. It
should really be part of sw_init (and smu_fini should be part
of sw_fini), but we need the firmware sizes from the other IPs
for firmware loading so we have to wait until sw init is done
for all other IPs.
Reviewed-by: Sonny Jiang <Sonny.Jiang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
