<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux/drivers/accel/habanalabs/include, branch master</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=master</id>
<link rel='self' href='https://git.shady.money/linux/atom?h=master'/>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/'/>
<updated>2024-06-23T06:53:07Z</updated>
<entry>
<title>accel/habanalbs/gaudi2: reduce interrupt count to 128</title>
<updated>2024-06-23T06:53:07Z</updated>
<author>
<name>Ofir Bitton</name>
<email>obitton@habana.ai</email>
</author>
<published>2024-03-31T12:37:32Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=f8422017b2e9331876efcfa07bb7579d5bc3e671'/>
<id>urn:sha1:f8422017b2e9331876efcfa07bb7579d5bc3e671</id>
<content type='text'>
Some systems allow a maximum number of 128 MSI-X interrupts.
Hence we reduce the interrupt count to 128 instead of 512.

Reviewed-by: Tomer Tayar &lt;ttayar@habana.ai&gt;
Signed-off-by: Ofir Bitton &lt;obitton@habana.ai&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: add GAUDI2D revision support</title>
<updated>2024-06-23T06:53:04Z</updated>
<author>
<name>Farah Kassabri</name>
<email>fkassabri@habana.ai</email>
</author>
<published>2024-03-14T10:36:04Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=cbac6f4088db9e697dee4a9456fd7ff5fbeaacb9'/>
<id>urn:sha1:cbac6f4088db9e697dee4a9456fd7ff5fbeaacb9</id>
<content type='text'>
Gaudi2 with PCI revision ID with the value of '4' represents Gaudi2D
device and should be detected and initialized as Gaudi2.

Signed-off-by: Farah Kassabri &lt;fkassabri@habana.ai&gt;
Reviewed-by: Ofir Bitton &lt;obitton@habana.ai&gt;
Signed-off-by: Ofir Bitton &lt;obitton@habana.ai&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: align interrupt names to table</title>
<updated>2024-06-23T06:53:03Z</updated>
<author>
<name>Ariel Suller</name>
<email>asuller@habana.ai</email>
</author>
<published>2024-03-06T09:04:01Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=990e0d91bf7d2e6dbacfa8af9048b9c4cedaa0f5'/>
<id>urn:sha1:990e0d91bf7d2e6dbacfa8af9048b9c4cedaa0f5</id>
<content type='text'>
when reporting tpc events, the dcore and tpc in dcore should
be reported and propagated, and not the generatl tpc number

Signed-off-by: Ariel Suller &lt;asuller@habana.ai&gt;
Reviewed-by: Ofir Bitton &lt;obitton@habana.ai&gt;
Signed-off-by: Ofir Bitton &lt;obitton@habana.ai&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: align embedded specs headers</title>
<updated>2024-06-23T06:44:45Z</updated>
<author>
<name>Ofir Bitton</name>
<email>obitton@habana.ai</email>
</author>
<published>2024-02-19T11:43:55Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=467cfe945656df044c8cf9121e5cdbe5b977b497'/>
<id>urn:sha1:467cfe945656df044c8cf9121e5cdbe5b977b497</id>
<content type='text'>
Align embedded headers to latest release.

Reviewed-by: Tomer Tayar &lt;ttayar@habana.ai&gt;
Signed-off-by: Ofir Bitton &lt;obitton@habana.ai&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: update interrupts related headers</title>
<updated>2024-06-23T06:44:17Z</updated>
<author>
<name>Farah Kassabri</name>
<email>fkassabri@habana.ai</email>
</author>
<published>2024-02-18T11:04:31Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=5a558f369ef89c6fd8170ee1137274fcc08517ae'/>
<id>urn:sha1:5a558f369ef89c6fd8170ee1137274fcc08517ae</id>
<content type='text'>
Align the interrupts related headers to latest release.

Signed-off-by: Farah Kassabri &lt;fkassabri@habana.ai&gt;
Reviewed-by: Ofir Bitton &lt;obitton@habana.ai&gt;
Signed-off-by: Ofir Bitton &lt;obitton@habana.ai&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: move HMMU page tables to device memory</title>
<updated>2024-02-26T07:30:40Z</updated>
<author>
<name>Farah Kassabri</name>
<email>fkassabri@habana.ai</email>
</author>
<published>2023-11-02T09:53:29Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=f728c17fc97aea7a33151d9ba64106291c62bb02'/>
<id>urn:sha1:f728c17fc97aea7a33151d9ba64106291c62bb02</id>
<content type='text'>
Currently the HMMU page tables reside in the host memory,
which will cause host access from the device for every page walk.
This can affect PCIe bandwidth in certain scenarios.

To prevent that problem, HMMU page tables will be moved to the device
memory so the miss transaction will read the hops from there instead of
going to the host.

Signed-off-by: Farah Kassabri &lt;fkassabri@habana.ai&gt;
Reviewed-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
Signed-off-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: use correct registers to dump QM CQ info</title>
<updated>2023-12-19T09:09:43Z</updated>
<author>
<name>Tomer Tayar</name>
<email>ttayar@habana.ai</email>
</author>
<published>2023-11-17T10:49:19Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=5bc155cfea605cd64aa372b44a67473b49c4726c'/>
<id>urn:sha1:5bc155cfea605cd64aa372b44a67473b49c4726c</id>
<content type='text'>
The QM CQ PTR_LO/PTR_HI/TSIZE registers are for pushing a CQ entry, and
although they are updated by HW even when descriptors are fetched by PQ
and CB addresses are fed into CQ, the correct registers to use when
dumping the CQ info are the ones with the _STS suffix.

Signed-off-by: Tomer Tayar &lt;ttayar@habana.ai&gt;
Reviewed-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
Signed-off-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: get the correct QM CQ info upon an error</title>
<updated>2023-12-19T09:09:43Z</updated>
<author>
<name>Tomer Tayar</name>
<email>ttayar@habana.ai</email>
</author>
<published>2023-11-06T16:41:35Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=ae303d885d4a0fcea65330de9327d28edfebd206'/>
<id>urn:sha1:ae303d885d4a0fcea65330de9327d28edfebd206</id>
<content type='text'>
Upon a QM error, the address/size from both the CQ and the ARC_CQ are
printed, although the instruction that led to the error was received
from only one of them.

Moreover, in case of a QM undefined opcode, only one of these
address/size sets will be captured based on the value of ARC_CQ_PTR.
However, this value can be non-zero even if currently the CQ is used, in
case the CQ/ARC_CQ are alternately used.

Under the assumption of having a stop-on-error configuration, modify to
use CP_STS.CUR_CQ field to get the relevant CQ for the QM error.

Signed-off-by: Tomer Tayar &lt;ttayar@habana.ai&gt;
Reviewed-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
Signed-off-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs: add support for Gaudi2C device</title>
<updated>2023-12-19T09:09:43Z</updated>
<author>
<name>Oded Gabbay</name>
<email>ogabbay@kernel.org</email>
</author>
<published>2023-10-30T10:23:57Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=42422993cf28d456778ee9168d73758ec037cd51'/>
<id>urn:sha1:42422993cf28d456778ee9168d73758ec037cd51</id>
<content type='text'>
Gaudi2 with PCI revision ID with the value of '3' represents Gaudi2C
device and should be detected and initialized as Gaudi2.

Signed-off-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
</content>
</entry>
<entry>
<title>accel/habanalabs/gaudi2: perform hard-reset upon PCIe AXI drain event</title>
<updated>2023-10-09T09:37:24Z</updated>
<author>
<name>Tomer Tayar</name>
<email>ttayar@habana.ai</email>
</author>
<published>2023-09-19T07:32:23Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=0426e03126fba3e3e8c549504283f7213d31831f'/>
<id>urn:sha1:0426e03126fba3e3e8c549504283f7213d31831f</id>
<content type='text'>
Non-completed transactions from PCIe towards the device are handled by
the AXI drain mechanism. This handling is in the PCIe level, but the
transactions are still there in the device consuming some queues
entries, and therefore the device must be reset.
Modify to perform hard-reset upon PCIe AXI drain events.

Signed-off-by: Tomer Tayar &lt;ttayar@habana.ai&gt;
Reviewed-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
Signed-off-by: Oded Gabbay &lt;ogabbay@kernel.org&gt;
</content>
</entry>
</feed>
