<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h, branch v5.8</title>
<subtitle>Mirror of https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
</subtitle>
<id>https://git.shady.money/linux/atom?h=v5.8</id>
<link rel='self' href='https://git.shady.money/linux/atom?h=v5.8'/>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/'/>
<updated>2020-04-28T20:20:30Z</updated>
<entry>
<title>drm/amdgpu: add new unlocked flag for PTE updates</title>
<updated>2020-04-28T20:20:30Z</updated>
<author>
<name>Christian König</name>
<email>christian.koenig@amd.com</email>
</author>
<published>2020-04-07T12:54:23Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=9c466bcbda68d73c9ba26e8307b0dff11dd285d8'/>
<id>urn:sha1:9c466bcbda68d73c9ba26e8307b0dff11dd285d8</id>
<content type='text'>
For HMM support we need the ability to invalidate PTEs from
a MM callback where we can't lock the root PD.

Add a new flag to better support this instead of assuming
that all invalidation updates are unlocked.

Signed-off-by: Christian König &lt;christian.koenig@amd.com&gt;
Reviewed-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: rename direct to immediate for VM updates</title>
<updated>2020-04-28T20:20:30Z</updated>
<author>
<name>Christian König</name>
<email>christian.koenig@amd.com</email>
</author>
<published>2020-04-01T10:50:56Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=eaad0c3aa978e7ed654e574691b56cc24d2d409a'/>
<id>urn:sha1:eaad0c3aa978e7ed654e574691b56cc24d2d409a</id>
<content type='text'>
To avoid confusion with direct ring submissions rename bottom
of pipe VM table changes to immediate updates.

Signed-off-by: Christian König &lt;christian.koenig@amd.com&gt;
Reviewed-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: define the TMZ bit for the PTE</title>
<updated>2020-04-28T20:20:28Z</updated>
<author>
<name>Alex Deucher</name>
<email>alexander.deucher@amd.com</email>
</author>
<published>2019-08-08T03:31:50Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=c5efd80f48e481946ba201412c7f39b19b39a40b'/>
<id>urn:sha1:c5efd80f48e481946ba201412c7f39b19b39a40b</id>
<content type='text'>
Define the TMZ (encryption) bit in the page table entry (PTE) for
Raven and newer asics.

Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: Huang Rui &lt;ray.huang@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: fix race between pstate and remote buffer map</title>
<updated>2020-04-22T22:11:46Z</updated>
<author>
<name>Jonathan Kim</name>
<email>jonathan.kim@amd.com</email>
</author>
<published>2020-03-17T19:43:41Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=d84a430d9f7b1ce6baedf1305106d0ae706aca76'/>
<id>urn:sha1:d84a430d9f7b1ce6baedf1305106d0ae706aca76</id>
<content type='text'>
Vega20 arbitrates pstate at hive level and not device level. Last peer to
remote buffer unmap could drop P-State while another process is still
remote buffer mapped.

With this fix, P-States still needs to be disabled for now as SMU bug
was discovered on synchronous P2P transfers.  This should be fixed in the
next FW update.

Signed-off-by: Jonathan Kim &lt;Jonathan.Kim@amd.com&gt;
Reviewed-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm_amdgpu: Add job fence to resv conditionally</title>
<updated>2020-03-16T20:21:32Z</updated>
<author>
<name>xinhui pan</name>
<email>xinhui.pan@amd.com</email>
</author>
<published>2020-03-16T03:45:14Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=57210c19e4ee34f1e1986a05f6626fac55ba271f'/>
<id>urn:sha1:57210c19e4ee34f1e1986a05f6626fac55ba271f</id>
<content type='text'>
Job fence on page table should be a shared one, so add it to the root
page talbe bo resv.
last_delayed field is not needed anymore. so remove it.

Cc: Christian König &lt;christian.koenig@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Cc: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Suggested-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: xinhui pan &lt;xinhui.pan@amd.com&gt;
Reviewed-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: rework synchronization of VM updates v4</title>
<updated>2020-02-05T04:30:39Z</updated>
<author>
<name>Christian König</name>
<email>christian.koenig@amd.com</email>
</author>
<published>2020-01-23T13:49:45Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=9f3cc18d19b7b791b8f59bcae51f88e759f73146'/>
<id>urn:sha1:9f3cc18d19b7b791b8f59bcae51f88e759f73146</id>
<content type='text'>
If provided we only sync to the BOs reservation
object and no longer to the root PD.

v2: update comment, cleanup amdgpu_bo_sync_wait_resv
v3: use correct reservation object while clearing
v4: fix typo in amdgpu_bo_sync_wait_resv

Signed-off-by: Christian König &lt;christian.koenig@amd.com&gt;
Tested-by: Tom St Denis &lt;tom.stdenis@amd.com&gt;
Reviewed-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: Avoid reclaim fs while eviction lock</title>
<updated>2020-01-16T18:33:16Z</updated>
<author>
<name>Alex Sierra</name>
<email>alex.sierra@amd.com</email>
</author>
<published>2019-12-17T22:57:30Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=a269e44989f3e77fb2ebfb995b8d501924ed07e2'/>
<id>urn:sha1:a269e44989f3e77fb2ebfb995b8d501924ed07e2</id>
<content type='text'>
[Why]
Avoid reclaim filesystem while eviction lock is held called from
MMU notifier.

[How]
Setting PF_MEMALLOC_NOFS flags while eviction mutex is locked.
Using memalloc_nofs_save / memalloc_nofs_restore API.

Signed-off-by: Alex Sierra &lt;alex.sierra@amd.com&gt;
Reviewed-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Reviewed-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: replace vm_pte's run-queue list with drm gpu scheds list</title>
<updated>2019-12-18T21:09:12Z</updated>
<author>
<name>Nirmoy Das</name>
<email>nirmoy.das@amd.com</email>
</author>
<published>2019-12-06T15:55:49Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=0c88b43032131ff458818addc9b65b8bd915837d'/>
<id>urn:sha1:0c88b43032131ff458818addc9b65b8bd915837d</id>
<content type='text'>
drm_sched_entity_init() takes drm gpu scheduler list instead of
drm_sched_rq list. This makes conversion of drm_sched_rq list
to drm gpu scheduler list unnecessary

Signed-off-by: Nirmoy Das &lt;nirmoy.das@amd.com&gt;
Reviewed-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: add VM eviction lock v3</title>
<updated>2019-12-09T22:02:26Z</updated>
<author>
<name>Christian König</name>
<email>christian.koenig@amd.com</email>
</author>
<published>2019-12-04T12:29:49Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=b4ff0f8a85f3c523942e57b716e8722e7f6799cc'/>
<id>urn:sha1:b4ff0f8a85f3c523942e57b716e8722e7f6799cc</id>
<content type='text'>
This allows to invalidate VM entries without taking the reservation lock.

v3: use -EBUSY

Signed-off-by: Christian König &lt;christian.koenig@amd.com&gt;
Reviewed-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: stop adding VM updates fences to the resv obj</title>
<updated>2019-12-09T22:02:15Z</updated>
<author>
<name>Christian König</name>
<email>christian.koenig@amd.com</email>
</author>
<published>2019-11-29T11:44:07Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=90b69cdc5f159be09d799257661e119a26ea4233'/>
<id>urn:sha1:90b69cdc5f159be09d799257661e119a26ea4233</id>
<content type='text'>
Don't add the VM update fences to the resv object and remove
the handling to stop implicitely syncing to them.

Ongoing updates prevent page tables from being evicted and we manually
block for all updates to complete before releasing PDs and PTS.

This way we can do updates even without the resv obj locked.

Signed-off-by: Christian König &lt;christian.koenig@amd.com&gt;
Reviewed-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
</feed>
