<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c, 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/amdkfd: New IOCTL to allocate queue GWS (v2)</title>
<updated>2020-04-28T20:20:30Z</updated>
<author>
<name>Oak Zeng</name>
<email>Oak.Zeng@amd.com</email>
</author>
<published>2019-05-07T03:11:14Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=5bb4b78be9c67b02a7f138850e9e89825181f555'/>
<id>urn:sha1:5bb4b78be9c67b02a7f138850e9e89825181f555</id>
<content type='text'>
Add a new kfd ioctl to allocate queue GWS. Queue
GWS is released on queue destroy.

v2: re-introduce this API with the following fixes squashed in:
- drm/amdkfd: fix null pointer dereference on dev
- drm/amdkfd: Return proper error code for gws alloc API
- drm/amdkfd: Remove GPU ID in GWS queue creation

Signed-off-by: Oak Zeng &lt;Oak.Zeng@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/amdkfd: Delete unnecessary unmap queue package submissions</title>
<updated>2020-02-26T19:20:33Z</updated>
<author>
<name>Yong Zhao</name>
<email>Yong.Zhao@amd.com</email>
</author>
<published>2020-02-05T21:53:37Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=c7637c95abebc811d2479a5b924859025def544a'/>
<id>urn:sha1:c7637c95abebc811d2479a5b924859025def544a</id>
<content type='text'>
The previous way of using SDMA queue count to infer whether we should unmap
SDMA engines has bugs. The reason it did not cause issues is because MEC
firmware unmaps all queues (CP + SDMA) when a unmap package for compute
engine is received. Becasue of that, only one unmap queue package
is needed, instead of one unmap queue package for CP and each SDMA engine,
which results in much simpler driver code.

Signed-off-by: Yong Zhao &lt;Yong.Zhao@amd.com&gt;
Acked-by: Alex Deucher &lt;alexander.deucher@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/amdkfd: Delete excessive printings</title>
<updated>2020-02-26T19:20:26Z</updated>
<author>
<name>Yong Zhao</name>
<email>Yong.Zhao@amd.com</email>
</author>
<published>2020-02-05T23:22:48Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=1e21647402f953a0f1098b4ecfff751734c6c9ee'/>
<id>urn:sha1:1e21647402f953a0f1098b4ecfff751734c6c9ee</id>
<content type='text'>
Those printings are duplicated or useless.

Signed-off-by: Yong Zhao &lt;Yong.Zhao@amd.com&gt;
Acked-by: Alex Deucher &lt;alexander.deucher@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/amdkfd: Fix a memory leak in queue creation error handling</title>
<updated>2020-02-26T19:20:20Z</updated>
<author>
<name>Yong Zhao</name>
<email>Yong.Zhao@amd.com</email>
</author>
<published>2020-02-05T22:13:54Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=66f28b9a169855367d6e3ef71001969a8bffb19b'/>
<id>urn:sha1:66f28b9a169855367d6e3ef71001969a8bffb19b</id>
<content type='text'>
When the queue creation failed, some resources were not freed. Fix it.

Signed-off-by: Yong Zhao &lt;Yong.Zhao@amd.com&gt;
Acked-by: Alex Deucher &lt;alexander.deucher@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/amdkfd: Avoid ambiguity by indicating it's cp queue</title>
<updated>2020-02-26T19:20:05Z</updated>
<author>
<name>Yong Zhao</name>
<email>Yong.Zhao@amd.com</email>
</author>
<published>2020-01-30T23:35:23Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=e6945304187deae0a28ebc65008ec11277f1c0f0'/>
<id>urn:sha1:e6945304187deae0a28ebc65008ec11277f1c0f0</id>
<content type='text'>
The queues represented in queue_bitmap are only CP queues.

Signed-off-by: Yong Zhao &lt;Yong.Zhao@amd.com&gt;
Acked-by: Alex Deucher &lt;alexander.deucher@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/amdkfd: Rename queue_count to active_queue_count</title>
<updated>2020-02-26T19:19:38Z</updated>
<author>
<name>Yong Zhao</name>
<email>Yong.Zhao@amd.com</email>
</author>
<published>2020-01-30T23:25:50Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=81b820b304a07fa4de9d46f095475555881af8fe'/>
<id>urn:sha1:81b820b304a07fa4de9d46f095475555881af8fe</id>
<content type='text'>
The name is easier to understand the code.

Signed-off-by: Yong Zhao &lt;Yong.Zhao@amd.com&gt;
Acked-by: Alex Deucher &lt;alexander.deucher@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/amdkfd: Add queue information to sysfs</title>
<updated>2020-02-06T20:04:38Z</updated>
<author>
<name>Amber Lin</name>
<email>Amber.Lin@amd.com</email>
</author>
<published>2020-01-30T05:08:05Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=6d220a7e7971ecabdf713554df44aca5640efb33'/>
<id>urn:sha1:6d220a7e7971ecabdf713554df44aca5640efb33</id>
<content type='text'>
Provide compute queues information in sysfs under /sys/class/kfd/kfd/proc.
The format is /sys/class/kfd/kfd/proc/&lt;pid&gt;/queues/&lt;queue id&gt;/XX where
XX are size, type, and gpuid three files to represent queue size, queue
type, and the GPU this queue uses. &lt;queue id&gt; folder and files underneath
are generated when a queue is created. They are removed when the queue is
destroyed.

Signed-off-by: Amber Lin &lt;Amber.Lin@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/amdkfd: Avoid hanging hardware in stop_cpsch</title>
<updated>2020-01-07T16:55:04Z</updated>
<author>
<name>Felix Kuehling</name>
<email>Felix.Kuehling@amd.com</email>
</author>
<published>2019-12-20T07:46:55Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=c2a77fde10ec2cac33a23c8ac53d181bc2fe0cee'/>
<id>urn:sha1:c2a77fde10ec2cac33a23c8ac53d181bc2fe0cee</id>
<content type='text'>
Don't use the HWS if it's known to be hanging. In a reset also
don't try to destroy the HIQ because that may hang on SRIOV if the
KIQ is unresponsive.

Signed-off-by: Felix Kuehling &lt;Felix.Kuehling@amd.com&gt;
Tested-by: Emily Deng &lt;Emily.Deng@amd.com&gt;
Reviewed-by: shaoyunl  &lt;shaoyun.liu@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdkfd: Rename create_cp_queue() to init_user_queue()</title>
<updated>2019-11-13T20:29:45Z</updated>
<author>
<name>Yong Zhao</name>
<email>Yong.Zhao@amd.com</email>
</author>
<published>2019-01-16T00:23:16Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=2d030d3e97a6da66d9c6fa558e01571ec14a92b0'/>
<id>urn:sha1:2d030d3e97a6da66d9c6fa558e01571ec14a92b0</id>
<content type='text'>
create_cp_queue() could also work with SDMA queues, so we should rename
it. It only initialize the data values rather than creating queues.

Signed-off-by: Yong Zhao &lt;Yong.Zhao@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/amdkfd: Avoid using doorbell_off as offset in process doorbell pages</title>
<updated>2019-11-13T20:29:45Z</updated>
<author>
<name>Yong Zhao</name>
<email>Yong.Zhao@amd.com</email>
</author>
<published>2019-01-15T18:58:57Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=e47a8b5223033e4dea5e68a228783e3c5ae0b623'/>
<id>urn:sha1:e47a8b5223033e4dea5e68a228783e3c5ae0b623</id>
<content type='text'>
dorbell_off in the queue properties is mainly used for the doorbell dw
offset in pci bar. We should not set it to the doorbell byte offset in
process doorbell pages. This makes the code much easier to read.

Signed-off-by: Yong Zhao &lt;Yong.Zhao@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>
