drm/amd/sched: fix deadlock caused by unsignaled fences of deleted jobs
Highly concurrent Piglit runs can trigger a race condition where a pending
SDMA job on a buffer object is never executed because the corresponding
process is killed (perhaps due to a crash). Since the job's fences were
never signaled, the buffer object was effectively leaked. Worse, the
buffer was stuck wherever it happened to be at the time, possibly in VRAM.
The symptom was user space processes stuck in interruptible waits with
kernel stacks like:
Note: The correctness of this change depends on the earlier commit
"drm/amd/sched: move adding finish callback to amd_sched_job_begin"
v2: set an error on the finished fence
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/sched: NULL out the s_fence field after run_job
amd_sched_process_job drops the fence reference, so NULL out the s_fence
field before adding it as a callback to guard against accidentally using
s_fence after it may have be freed.
v2: add a clarifying comment
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/sched: move adding finish callback to amd_sched_job_begin
The finish callback is responsible for removing the job from the ring
mirror list, among other things. It makes sense to add it as callback
in the place where the job is added to the ring mirror list.
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The function does not actually remove the job from the FIFO, so "peek"
describes it better.
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Mon, 18 Sep 2017 12:01:45 +0000 (14:01 +0200)]
drm/amdgpu: minor coding style fix
Fix two minor 80 char issues.
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>
Christian König [Thu, 6 Jul 2017 07:59:43 +0000 (09:59 +0200)]
drm/ttm: add transparent huge page support for DMA allocations v2
Try to allocate huge pages when it makes sense.
v2: fix comment and use ifdef
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>
Christian König [Tue, 4 Jul 2017 14:56:24 +0000 (16:56 +0200)]
drm/ttm: add support for different pool sizes
Correctly handle different page sizes in the memory accounting.
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>
Christian König [Mon, 18 Sep 2017 13:45:11 +0000 (15:45 +0200)]
drm/ttm: remove unsued options from ttm_mem_global_alloc_page
Nobody is actually using that, remove it.
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>
James Zhu [Fri, 29 Sep 2017 20:47:31 +0000 (16:47 -0400)]
drm/amdgpu: add uvd enc irq
Add UVD encode IRQ handle and enable the UVD encode trap
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Fri, 29 Sep 2017 20:42:27 +0000 (16:42 -0400)]
drm/amdgpu: add uvd enc ib test
Generate create/destroy messages to test UVD encode indirect buffer function.
And enable UVD encode IB test during device initialization.
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Fri, 29 Sep 2017 20:40:12 +0000 (16:40 -0400)]
drm/amdgpu: add uvd enc ring test
Add UVD encode ring test functions. And enable UVD encode ring test
during UVD encode hardware initialization.
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Fri, 29 Sep 2017 20:37:11 +0000 (16:37 -0400)]
drm/amdgpu: add uvd enc vm functions (v2)
Add UVD encode ring vm functions to handle frame ecoding.
v2: squash in warning fix (James)
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Fri, 29 Sep 2017 19:20:23 +0000 (15:20 -0400)]
drm/amdgpu: add uvd enc into run queue
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Fri, 29 Sep 2017 21:29:06 +0000 (17:29 -0400)]
drm/amdgpu: add uvd enc rings
UVD 6.3 has two UVD encode rings. Add the ring structures and initialize the hw ring buffers.
Currently only ASIC Polaris10/11/12 uses UVD6.3 encode engine on HEVC encoding.
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Fri, 29 Sep 2017 20:14:26 +0000 (16:14 -0400)]
drm/amdgpu: add new uvd enc ring methods
Add new UVD encode ring methods get/set/emit/flush/sync to support uvd6.3 HEVC encoding
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Fri, 29 Sep 2017 18:17:34 +0000 (14:17 -0400)]
drm/amdgpu: add uvd enc command in header
Add UVD encode command interface definition for uvd6.3 HEVC encoding
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Mon, 2 Oct 2017 00:00:07 +0000 (20:00 -0400)]
drm/amdgpu: add uvd enc registers in header
Add UVD encode write/read/size/base registers definition for uvd6.3 HEVC ecoding
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-and-Tested-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Tom St Denis <tom.stdenis@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Sat, 30 Sep 2017 01:13:47 +0000 (09:13 +0800)]
drm/amd/powerplay: fix typo on avfs disable
Signed-off-by: Evan Quan <evan.quan@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Tue, 26 Sep 2017 03:35:30 +0000 (11:35 +0800)]
drm/amd/powerplay: fixed wrong return value on error (v2)
v2: squash in typo fix (Tom)
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The dead circular lock senario captured is as followed.
The idea of the fix is moving read_user_wptr outside of
acquire_queue...release_queue critical section
[ 63.477482] WARNING: possible circular locking dependency detected
[ 63.484091] 4.12.0-kfd-ozeng #3 Not tainted
[ 63.488531] ------------------------------------------------------
[ 63.495146] HelloWorldLoop/2526 is trying to acquire lock:
[ 63.501011] (&mm->mmap_sem){++++++}, at: [<ffffffff911898ce>] __might_fault+0x3e/0x90
[ 63.509472]
but task is already holding lock:
[ 63.515716] (&adev->srbm_mutex){+.+...}, at: [<ffffffffc0484feb>] lock_srbm+0x2b/0x50 [amdgpu]
[ 63.525099]
which lock already depends on the new lock.
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Colin Ian King [Thu, 28 Sep 2017 13:46:17 +0000 (14:46 +0100)]
drm/radeon: make functions alloc_pasid and free_pasid static
The functions alloc_pasid and free_pasid are local to the
source and do not need to be in global scope, so make them static.
Cleans up sparse warnings:
warning: symbol 'alloc_pasid' was not declared. Should it be static?
warning: symbol 'free_pasid' was not declared. Should it be static?
Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Fri, 29 Sep 2017 00:47:43 +0000 (10:47 +1000)]
amdgpu/nbio: use constant nbio_hdp_flush_reg structs.
This removes the init path as well, since the init path
just did some constant init of some structs.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Fri, 29 Sep 2017 00:08:01 +0000 (10:08 +1000)]
amdgpu/soc15: make the pcie index/data registers constant.
These don't seem to change at runtime, and the initialisers
are constant data. This could be improved by not selecting
the apu/non-apu path on each pcie read/write access.
Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Fri, 29 Sep 2017 00:39:30 +0000 (10:39 +1000)]
amdgpu/pp: constify soft_dummy_pp_table.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Fri, 29 Sep 2017 00:12:53 +0000 (10:12 +1000)]
amdgfx/gfx: don't use static objects for ce/de meta. (v2)
This isn't safe if we have multiple GPUs plugged in, since
there is only one copy of this struct in the bss, just allocate
on stack, it's 40/108 bytes which should be safe.
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Felix Kuehling [Thu, 21 Sep 2017 20:26:41 +0000 (16:26 -0400)]
drm/amdgpu: Handle GPUVM fault storms
When many wavefronts cause VM faults at the same time, it can
overwhelm the interrupt handler and cause IH ring overflows before
the driver can notify or kill the faulting application.
As a workaround I'm introducing limited per-VM fault credit. After
that number of VM faults have occurred, further VM faults are
filtered out at the prescreen stage of processing.
This depends on the PASID in the interrupt packet, so it currently
only works for KFD contexts.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Thu, 28 Sep 2017 06:34:44 +0000 (16:34 +1000)]
amdgpu/pp: reduce size of vega10_fuses_default
I've no idea why this is like this, why store 64-bit fields
as a string, and then parse the strings, this is just over
engineered.
Reduce the size of the amdgpu.o by 80k.
text data bss dec hex filename 1331332 17982 1008 1350322 149ab2 amdgpu.o 1244668 17982 1008 1263658 13482a amdgpu.o
Acked-by: Harry Wentland <harry.wentland@amd.com> Acked-by: Alex Deucer <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Thu, 28 Sep 2017 06:12:28 +0000 (16:12 +1000)]
drm/amdgpu: use designated initialiser for thermal_irq_src.
This fixes the 0-day build warning.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Thu, 28 Sep 2017 04:48:38 +0000 (14:48 +1000)]
amdgpu/pp: slim down the pwr virus tables.
This is what I'd call slightly overengineered, we waste 40k on
storing a value that is write or end, when we could just use the
register value to denote end.
Remove the virus command parameter, and save
text data bss dec hex filename 1412724 17982 1008 1431714 15d8a2 ../drm-next-build/drivers/gpu/drm/amd/amdgpu/amdgpu.o 1331332 17982 1008 1350322 149ab2 ../drm-next-build/drivers/gpu/drm/amd/amdgpu/amdgpu.o
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Wed, 27 Sep 2017 23:36:55 +0000 (09:36 +1000)]
amdgpu/pp: move amdgpu_fuses_default into static const.
There is no reason that this gets passed back into the function
from outside the file, just reference the table directly.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Wed, 27 Sep 2017 23:32:49 +0000 (09:32 +1000)]
amdgpu/pp: move PhwVega10_Magic to static const.
This isn't used outside this file.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Wed, 27 Sep 2017 23:01:47 +0000 (09:01 +1000)]
amdgpu/pp: remove ci_smc/smumgr split.
This split serves no purpose, and we can make a bunch of functions
static now.
There are lots of cases of this sort of split in the powerplay code,
please start cleaning them up. Ideally the function table is in the
same file as all the implementations used in it.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Wed, 27 Sep 2017 20:51:21 +0000 (06:51 +1000)]
drm/amdgpu/vega10: static constify channel_number
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Wed, 27 Sep 2017 20:50:54 +0000 (06:50 +1000)]
drm/amdgpu/pp: constify some powerplay tables
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Wed, 27 Sep 2017 20:44:36 +0000 (06:44 +1000)]
amdgpu/powerplay: constify large struct
This moves this from being global data to global rodata, I'm
sure it would be easy to move it to being local data.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Wed, 27 Sep 2017 22:37:34 +0000 (08:37 +1000)]
amdgpu: don't ask about CHASH just default it for now.
If we bump this up a level, we can ask about it, for now,
just default to what amdgpu does.
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Andres Rodriguez [Tue, 26 Sep 2017 21:43:14 +0000 (17:43 -0400)]
drm/amdgpu: map compute rings by least recently used pipe
This patch provides a guarantee that the first n queues allocated by
an application will be on different pipes. Where n is the number of
pipes available from the hardware.
This helps avoid ring aliasing which can result in work executing in
time-sliced mode instead of truly parallel mode.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Andres Rodriguez [Tue, 26 Sep 2017 16:22:46 +0000 (12:22 -0400)]
drm/amdgpu: add option for force enable multipipe policy for compute
Useful for testing the effects of multipipe compute without recompiling.
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Andres Rodriguez [Tue, 26 Sep 2017 16:22:45 +0000 (12:22 -0400)]
drm/amdgpu: use multipipe compute policy on non PL11 asics
A performance regression for OpenCL tests on Polaris11 had this feature
disabled for all asics.
Instead, disable it selectively on the affected asics.
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Andres Rodriguez <andresx7@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Thu, 28 Sep 2017 13:47:32 +0000 (09:47 -0400)]
drm/amdgpu: fix vf error handling
The error handling for virtual functions assumed a single
vf per VM and didn't properly account for bare metal. Make
the error arrays per device and add locking.
Reviewed-by: Gavin Wan <gavin.wan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 19 Sep 2017 14:20:27 +0000 (10:20 -0400)]
drm/amdgpu: clarify license in amdgpu_trace_points.c
It was not clear. The rest of the driver is MIT/X11.
Reviewed-by: Christian König <christian.koenig@amd.com> Acked-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Samuel Li [Tue, 22 Aug 2017 19:25:33 +0000 (15:25 -0400)]
drm/amdgpu: Add gem_prime_mmap support
v2: drop hdp invalidate/flush.
v3: honor pgoff during prime mmap. Add a barrier after cpu access.
v4: drop begin/end_cpu_access() for now, revisit later.
Signed-off-by: Samuel Li <Samuel.Li@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Mon, 11 Sep 2017 15:29:26 +0000 (17:29 +0200)]
drm/amdgpu: simplify pinning into visible VRAM
Just set the CPU access required flag when we pin it.
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>
Monk Liu [Tue, 19 Sep 2017 08:09:53 +0000 (16:09 +0800)]
drm/amdgpu:fix firmware memoryleak(v2)
this fix memory leak due to request_firmware after driver
unloaded
v2:
release gmc firmware for gmc6/7/8 as well
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Fri, 15 Sep 2017 08:43:01 +0000 (16:43 +0800)]
drm/amdgpu:fix uvd ring fini routine(v2)
fix missing finish uvd enc_ring.
v2:
since the adev pointer check in already in ring_fini
so drop the check outsider
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Thu, 21 Sep 2017 07:10:06 +0000 (15:10 +0800)]
drm/amdgpu/sriov:alloc KIQ MQD in VRAM(v2)
this way after KIQ MQD released in drv unloading, CPC
can still let KIQ access this MQD thus RLCV SAVE_VF
will not fail
v2:
always use VRAM domain for KIQ MQD no matter BM or SRIOV
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Thu, 21 Sep 2017 06:59:40 +0000 (14:59 +0800)]
drm/amdgpu:unmap KCQ in gfx hw_fini(v2)
v2:
move kcq_disable out of SRIOV, make it genearal
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Tue, 4 Jul 2017 08:40:58 +0000 (16:40 +0800)]
drm/amdgpu:halt when vm fault
only with this way we can debug the VMC page fault issue
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdgpu: Add copy_pte_num_dw member in amdgpu_vm_pte_funcs
Use it to replace the hard coded value in amdgpu_vm_bo_update_mapping().
Signed-off-by: Yong Zhao <yong.zhao@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When max_bytes is not 8 bytes aligned and bo size is larger than
max_bytes, the last 8 bytes in a ttm node may be left unchanged.
For example, on pre SDMA 4.0, max_bytes = 0x1fffff, and the bo size
is 0x200000, the problem will happen.
In order to fix the problem, we separately store the max nums of
PTEs/PDEs a single operation can set in amdgpu_vm_pte_funcs
structure, rather than inferring it from bytes limit of SDMA
constant fill, i.e. fill_max_bytes.
Together with the fix, we replace the hard code value "10" in
amdgpu_vm_bo_update_mapping() with the corresponding values from
structure amdgpu_vm_pte_funcs.
Signed-off-by: Yong Zhao <yong.zhao@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>