Instead of having a confusing wrapper or call the underlying ww_mutex
function directly.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-and-Tested-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Needed for the next patch and makes the code quite a bit easier to
understand.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-and-Tested-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Fixes a use-after-free due to a race condition in
ttm_bo_cleanup_refs_and_unlock, which allows one task to reserve a BO
and destroy its ttm_resv while another task is waiting for it to signal
in reservation_object_wait_timeout_rcu.

v2:
* Always initialize bo->ttm_resv in ttm_bo_init_reserved
 (Christian König)

Fixes: 0d2bd2ae "drm/ttm: fix memory leak while individualizing BOs"
Reviewed-by: Chunming Zhou <david1.zhou@amd.com> # v1
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

after individualize we need manually call reservation_object_fini()
if all fences on resv signaled during test, otherwise kmemory leak
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

We need to free the reservation object before we take the BO
from the delayed delete list.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NMonk Liu <monk.liu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

With shared reservation objects __ttm_bo_reserve() can easily fail even on
destroyed BOs. This prevents correct handling when we need to individualize
the reservation object.

Fix this by individualizing the object before even trying to reserve it.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Acked-by: NChunming Zhou <david1.zhou@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

we observe abnormal number from:
/sys/devices/virtual/drm/amdttm/buffer_objects/bo_count

bo_count is atomic_inc which is "int" type,
shouldn't explicitly turn it to unsigned long.
Signed-off-by: NMonk Liu <monk.liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

we observe abnormal number from:
/sys/devices/virtual/drm/amdttm/buffer_objects/bo_count

bo_count is atomic_inc which is "int" type,
shouldn't explicitly turn it to unsigned long.
Signed-off-by: NMonk Liu <monk.liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Fixes a false positive from might_sleep(). The reservation object is freshly
initialized, so nobody else can hold the mutex but the function is
called from atomic context.

v2: Correctly invert the check as well.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Provide the drm printer directly instead of just the callback.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Use the BOs reservation object when it is put on the ddelete list. This way we
avoid delaying freeing up the BO because of new fences on the reservation object.

This is used by dma-buf and amdgpu's VM page tables.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

With shared reservation objects the assumption that no fence
could have been added isn't true any more.

Additional to that the BO is about to be destroyed, so removing the
fences now has no advantage whatsoever.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

We unref the man->move fence in ttm_bo_clean_mm() and then call
ttm_bo_force_list_clean() which waits on it, except the refcount is now
zero so a warning is generated (or worse):

[149492.279301] refcount_t: increment on 0; use-after-free.
[149492.279309] ------------[ cut here ]------------
[149492.279315] WARNING: CPU: 3 PID: 18726 at lib/refcount.c:150 refcount_inc+0x2b/0x30
[149492.279315] Modules linked in: vhost_net vhost tun x86_pkg_temp_thermal crc32_pclmul ghash_clmulni_intel efivarfs amdgpu(
-) i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm
[149492.279326] CPU: 3 PID: 18726 Comm: rmmod Not tainted 4.12.0-rc5-drm-next-4.13-ttmpatch+ #1
[149492.279326] Hardware name: Gigabyte Technology Co., Ltd. Z97X-UD3H-BK/Z97X-UD3H-BK-CF, BIOS F6 06/17/2014
[149492.279327] task: ffff8804ddfedcc0 task.stack: ffffc90008d20000
[149492.279329] RIP: 0010:refcount_inc+0x2b/0x30
[149492.279330] RSP: 0018:ffffc90008d23c30 EFLAGS: 00010286
[149492.279331] RAX: 000000000000002b RBX: 0000000000000170 RCX: 0000000000000000
[149492.279331] RDX: 0000000000000000 RSI: ffff88051ecccbe8 RDI: ffff88051ecccbe8
[149492.279332] RBP: ffffc90008d23c30 R08: 0000000000000001 R09: 00000000000003ee
[149492.279333] R10: ffffc90008d23bb0 R11: 00000000000003ee R12: ffff88043aaac960
[149492.279333] R13: ffff8805005e28a8 R14: 0000000000000002 R15: ffff88050115e178
[149492.279334] FS:  00007fc540168700(0000) GS:ffff88051ecc0000(0000) knlGS:0000000000000000
[149492.279335] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[149492.279336] CR2: 00007fc3e8654140 CR3: 000000027ba77000 CR4: 00000000001426e0
[149492.279337] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[149492.279337] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[149492.279338] Call Trace:
[149492.279345]  ttm_bo_force_list_clean+0xb9/0x110 [ttm]
[149492.279348]  ttm_bo_clean_mm+0x7a/0xe0 [ttm]
[149492.279375]  amdgpu_ttm_fini+0xc9/0x1f0 [amdgpu]
[149492.279392]  amdgpu_bo_fini+0x12/0x40 [amdgpu]
[149492.279415]  gmc_v7_0_sw_fini+0x32/0x40 [amdgpu]
[149492.279430]  amdgpu_fini+0x2c9/0x490 [amdgpu]
[149492.279445]  amdgpu_device_fini+0x58/0x1b0 [amdgpu]
[149492.279461]  amdgpu_driver_unload_kms+0x4f/0xa0 [amdgpu]
[149492.279470]  drm_dev_unregister+0x3c/0xe0 [drm]
[149492.279485]  amdgpu_pci_remove+0x19/0x30 [amdgpu]
[149492.279487]  pci_device_remove+0x39/0xc0
[149492.279490]  device_release_driver_internal+0x155/0x210
[149492.279491]  driver_detach+0x38/0x70
[149492.279493]  bus_remove_driver+0x4c/0xa0
[149492.279494]  driver_unregister+0x2c/0x40
[149492.279496]  pci_unregister_driver+0x21/0x90
[149492.279520]  amdgpu_exit+0x15/0x406 [amdgpu]
[149492.279523]  SyS_delete_module+0x1a8/0x270
[149492.279525]  ? exit_to_usermode_loop+0x92/0xa0
[149492.279528]  entry_SYSCALL_64_fastpath+0x13/0x94
[149492.279529] RIP: 0033:0x7fc53fcb68e7
[149492.279529] RSP: 002b:00007ffcfbfaabb8 EFLAGS: 00000206 ORIG_RAX: 00000000000000b0
[149492.279531] RAX: ffffffffffffffda RBX: 0000563117adb200 RCX: 00007fc53fcb68e7
[149492.279531] RDX: 000000000000000a RSI: 0000000000000800 RDI: 0000563117adb268
[149492.279532] RBP: 0000000000000003 R08: 0000000000000000 R09: 1999999999999999
[149492.279533] R10: 0000000000000883 R11: 0000000000000206 R12: 00007ffcfbfa9ba0
[149492.279533] R13: 0000000000000000 R14: 0000000000000000 R15: 0000563117adb200
[149492.279534] Code: 55 48 89 e5 e8 77 fe ff ff 84 c0 74 02 5d c3 80 3d 40 f2 a4 00 00 75 f5 48 c7 c7 20 3c ca 81 c6 05 30 f2 a4 00 01 e8 91 f0 d7 ff <0f> ff 5d c3 90 55 48 89 fe bf 01 00 00 00 48 89 e5 e8 9f fe ff
[149492.279557] ---[ end trace 2d4e0ffcb66a1016 ]---

Unref the fence *after* waiting for it.

v2: Set man->move to NULL after dropping the last ref (Christian König)

Fixes: aff98ba1 (drm/ttm: wait for eviction in ttm_bo_force_list_clean)
Signed-off-by: NJohn Brooks <john@fastquake.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

The ret must be 0 here, otherwise, the function will return after init_mem_type.
Signed-off-by: NHuang Rui <ray.huang@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This allows drivers to specify if they need a contiguous allocation or not.

v2: use space instead of tab
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NNicolai Hähnle <nicolai.haehnle@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

No need to implement the same logic twice. Also check if the busy placements
are identical to the already scanned placements before checking them.

v2: improve check even more as suggested by Michel.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NNicolai Hähnle <nicolai.haehnle@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This variant of ttm_bo_init returns the validated buffer object with
the reservation lock held when resv == NULL. This is convenient for
callers that want to use the BO immediately, e.g. for initializing its
contents.
Signed-off-by: NNicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Fixes a potential race condition in amdgpu that looks as follows:

Task 1: attempt ttm_bo_init, but ttm_bo_validate fails
Task 1: add BO to global list anyway
Task 2: grabs hold of the BO, waits on its reservation lock
Task 1: releases its reference of the BO; never gives up the
        reservation lock

The patch "drm/amdgpu: fix a potential deadlock in
amdgpu_bo_create_restricted()" attempts to fix that by releasing
the reservation lock in amdgpu code; unfortunately, it introduces
a use-after-free when this race _doesn't_ happen.

This patch should fix the race properly by never adding the BO
to the global list in the first place.

Cc: zhoucm1 <david1.zhou@amd.com>
Signed-off-by: NNicolai Hähnle <nicolai.haehnle@amd.com>
Tested-by: NSamuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Use a more common logging style.

Miscellanea:

o Coalesce formats and realign arguments
o Neaten a few macros now using pr_<level>
Signed-off-by: NJoe Perches <joe@perches.com>
Acked-by: NRob Clark <robdclark@gmail.com>
Acked-by: NSinclair Yeh <syeh@vmware.com>
Acked-by: NPatrik Jakobsson <patrik.r.jakobsson@gmail.com>
Acked-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/76355db47b31668bb64d996865ceee53bd66b11f.1488285953.git.joe@perches.com

Without the custom LRU management the callback is not used any more.

agd: fix trivial warning
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-and-Tested-by: NRoger.He <Hongbo.He@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The additional housekeeping had too much CPU overhead,
let's use the BO priorities instead.

agd: also revert hibmc changes
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-and-Tested-by: NRoger.He <Hongbo.He@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This way the driver can specify a priority for a BO which has the effect that
a BO is only evicted when all other BOs with a lower priority are evicted
first.
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NRoger.He <Hongbo.He@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Not allowing errors here is completely pointless and actually dangerous
cause trying to continue on an error can cause an endless loop.
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NRoger.He <Hongbo.He@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Ensure that the driver can listen to evictions even when they don't take the
path through ttm_bo_driver::move.

This is crucial for amdgpu, which relies on an eviction counter to skip
re-binding page tables when possible.
Signed-off-by: NNicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: NChunming Zhou <david1.zhou@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The current caching state may not be tt_cached, even though the
placement contains TTM_PL_FLAG_CACHED, because placement can contain
multiple caching flags. Trying to swap out such a BO would trip up the

	BUG_ON(ttm->caching_state != tt_cached);

in ttm_tt_swapout.

Cc: stable@vger.kernel.org
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Reviewed-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Christian König <christian.koenig@amd.com>.
Reviewed-by: NSinclair Yeh <syeh@vmware.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>

By general sentiment kref_sub() is a bad interface, make it go away.
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NIngo Molnar <mingo@kernel.org>

Since we need to change the implementation, stop exposing internals.

Provide kref_read() to read the current reference count; typically
used for debug messages.

Kills two anti-patterns:

	atomic_read(&kref->refcount)
	kref->refcount.counter
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NIngo Molnar <mingo@kernel.org>

reservation_object_wait_timeout_rcu() should enable signaling even with a
zero timeout, but ttm_bo_wait() can also be called from atomic context and
then it is not a good idea to do this.
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NSumit Semwal <sumit.semwal@linaro.org>
  [sumits: fix checkpatch warnings]
Link: http://patchwork.freedesktop.org/patch/msgid/1478553376-18575-3-git-send-email-alexander.deucher@amd.com

This way the driver can decide if it is valuable to evict a BO or not.

The current implementation is added as default to all existing drivers.

v2: fix some typos found during internal testing
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

I plan to usurp the short name of struct fence for a core kernel struct,
and so I need to rename the specialised fence/timeline for DMA
operations to make room.

A consensus was reached in
https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html
that making clear this fence applies to DMA operations was a good thing.
Since then the patch has grown a bit as usage increases, so hopefully it
remains a good thing!

(v2...: rebase, rerun spatch)
v3: Compile on msm, spotted a manual fixup that I broke.
v4: Try again for msm, sorry Daniel

coccinelle script:
@@

@@
- struct fence
+ struct dma_fence
@@

@@
- struct fence_ops
+ struct dma_fence_ops
@@

@@
- struct fence_cb
+ struct dma_fence_cb
@@

@@
- struct fence_array
+ struct dma_fence_array
@@

@@
- enum fence_flag_bits
+ enum dma_fence_flag_bits
@@

@@
(
- fence_init
+ dma_fence_init
|
- fence_release
+ dma_fence_release
|
- fence_free
+ dma_fence_free
|
- fence_get
+ dma_fence_get
|
- fence_get_rcu
+ dma_fence_get_rcu
|
- fence_put
+ dma_fence_put
|
- fence_signal
+ dma_fence_signal
|
- fence_signal_locked
+ dma_fence_signal_locked
|
- fence_default_wait
+ dma_fence_default_wait
|
- fence_add_callback
+ dma_fence_add_callback
|
- fence_remove_callback
+ dma_fence_remove_callback
|
- fence_enable_sw_signaling
+ dma_fence_enable_sw_signaling
|
- fence_is_signaled_locked
+ dma_fence_is_signaled_locked
|
- fence_is_signaled
+ dma_fence_is_signaled
|
- fence_is_later
+ dma_fence_is_later
|
- fence_later
+ dma_fence_later
|
- fence_wait_timeout
+ dma_fence_wait_timeout
|
- fence_wait_any_timeout
+ dma_fence_wait_any_timeout
|
- fence_wait
+ dma_fence_wait
|
- fence_context_alloc
+ dma_fence_context_alloc
|
- fence_array_create
+ dma_fence_array_create
|
- to_fence_array
+ to_dma_fence_array
|
- fence_is_array
+ dma_fence_is_array
|
- trace_fence_emit
+ trace_dma_fence_emit
|
- FENCE_TRACE
+ DMA_FENCE_TRACE
|
- FENCE_WARN
+ DMA_FENCE_WARN
|
- FENCE_ERR
+ DMA_FENCE_ERR
)
 (
 ...
 )
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Acked-by: NSumit Semwal <sumit.semwal@linaro.org>
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk

A bit pointless to search for the first bit set manually.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NChunming Zhou <david1.zhou@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Fixes hangs under memory pressure, e.g. running the piglit test
tex3d-maxsize concurrently with other tests.

Fixes: 17d33bc9 ("drm/ttm: drop waiting for idle in ttm_bo_evict.")
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Fixes hangs under memory pressure, e.g. running the piglit test
tex3d-maxsize concurrently with other tests.

Fixes: 17d33bc9 ("drm/ttm: drop waiting for idle in ttm_bo_evict.")
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The ttm_tt_destroy() function tests whether its argument is NULL
and then returns immediately. Thus the test around the call is not needed.

This issue was detected by using the Coccinelle software.
Signed-off-by: NMarkus Elfring <elfring@users.sourceforge.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/54338f58-830c-a8b4-4554-5d4459bcd321@users.sourceforge.net

Now that we can pipeline evictions we need to wait for
them to finish when we cleanup a memory domain.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Free up the memory immediately, remember the last eviction for each domain and
make new allocations depend on the last eviction to be completed.
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

As far as I can see no need for a custom implementation any more.
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Instead of using the flag just remember the fence of the last move operation.

This avoids waiting for command submissions pipelined after the move, but
before accessing the BO with the CPU again.
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>