<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c, branch v6.3</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=v6.3</id>
<link rel='self' href='https://git.shady.money/linux/atom?h=v6.3'/>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/'/>
<updated>2023-02-14T21:04:48Z</updated>
<entry>
<title>drm/amdgpu: Revert programming GRBM_GFX_* in RLCG interface to support GFX9</title>
<updated>2023-02-14T21:04:48Z</updated>
<author>
<name>Yifan Zha</name>
<email>Yifan.Zha@amd.com</email>
</author>
<published>2023-02-08T08:47:35Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=fa9b4155c3d4e8441ea57ad10ec98504845e54bb'/>
<id>urn:sha1:fa9b4155c3d4e8441ea57ad10ec98504845e54bb</id>
<content type='text'>
[Why]
Regression of commit 72fef4980ddf ("drm/amdgpu: Remove writing GRBM_GFX_CNTL in RLCG interface under SRIOV") on GFX9.
According to GFX9 VF using different method to access GC registers including MMIO(direct) and RLCG(indirect),
removing GRBM_GFX_* writing would make PIPE/ME/VM/QUEUE selection chaos leading to some OCL benchmark failure.

For example,
using RLCG interface to program GRBM_GFX_CNTL/INDEX for selecting MEC(actually the value is only in scratch2/3),
then using MMIO directly program a MEC register in VF driver.
The register programming are invalid due to GC switched to incorrect ME.

[How]
With checking RLCG accessing flag, keep writing GRBM_GFX_* as a legacy way.
But it is still skipped on GFX10+ to avoid violation occurrence.

Signed-off-by: Yifan Zha &lt;Yifan.Zha@amd.com&gt;
Reviewed-by: Hawking Zhang &lt;Hawking.Zhang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: Remove writing GRBM_GFX_CNTL in RLCG interface under SRIOV</title>
<updated>2023-02-03T20:41:36Z</updated>
<author>
<name>Yifan Zha</name>
<email>Yifan.Zha@amd.com</email>
</author>
<published>2023-01-31T07:31:27Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=72fef4980ddf20792fa095f88b008d896313c2b1'/>
<id>urn:sha1:72fef4980ddf20792fa095f88b008d896313c2b1</id>
<content type='text'>
[Why]
Accessing GRBM_GFX_CNTL in full access time has risk when VF is doing MMIO attacking.
Therefore, VF writing GRBM_GFX_CNTL are blocked by L1 Policy.
For RLCG interface, RLCG use SCRATCH_REG2 which is copied from GRBM_GFX_CNTL.

[How]
Remove writing GRBM_GFX_CNTL in amdgpu_virt_rlcg_reg_rw.

v2:
Remove directly writing GRBM_GFX_INDEX in amdgpu_virt_rlcg_reg_rw
as RLCG interface no need to use it.

Signed-off-by: Yifan Zha &lt;Yifan.Zha@amd.com&gt;
Reviewed-by: Hawking Zhang &lt;Hawking.Zhang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations</title>
<updated>2023-01-03T21:49:54Z</updated>
<author>
<name>Christian König</name>
<email>christian.koenig@amd.com</email>
</author>
<published>2022-01-14T15:49:44Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=58ab2c08d708ca8309768545b75741636c53a336'/>
<id>urn:sha1:58ab2c08d708ca8309768545b75741636c53a336</id>
<content type='text'>
Technically all of those can use GTT as well, no need to force things
into VRAM.

Signed-off-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: Luben Tuikov &lt;luben.tuikov@amd.com&gt;
Acked-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: Remove unnecessary domain argument</title>
<updated>2022-12-14T21:48:00Z</updated>
<author>
<name>Luben Tuikov</name>
<email>luben.tuikov@amd.com</email>
</author>
<published>2022-12-14T08:56:03Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=3273f11675ef11959d25a56df3279f712bcd41b7'/>
<id>urn:sha1:3273f11675ef11959d25a56df3279f712bcd41b7</id>
<content type='text'>
Remove the "domain" argument to amdgpu_bo_create_kernel_at() since this
function takes an "offset" argument which is the offset off of VRAM, and as
such allocation always takes place in VRAM. Thus, the "domain" argument is
unnecessary.

Cc: Alex Deucher &lt;Alexander.Deucher@amd.com&gt;
Cc: Christian König &lt;christian.koenig@amd.com&gt;
Cc: AMD Graphics &lt;amd-gfx@lists.freedesktop.org&gt;
Signed-off-by: Luben Tuikov &lt;luben.tuikov@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 drv_vram_usage_va for virt data exchange</title>
<updated>2022-11-29T16:03:35Z</updated>
<author>
<name>Tong Liu01</name>
<email>Tong.Liu01@amd.com</email>
</author>
<published>2022-11-17T10:18:58Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=6d96ced7600e02ac1efb03a21af529fd9a95e3c6'/>
<id>urn:sha1:6d96ced7600e02ac1efb03a21af529fd9a95e3c6</id>
<content type='text'>
For vram_usagebyfirmware_v2_2, fw_vram_reserve is not used. So
fw_vram_usage_va is NULL, and cannot do virt data exchange
anymore. Should add drv_vram_usage_va to do virt data exchange
in vram_usagebyfirmware_v2_2 case. And refine some code style
checks in pre add vram reservation logic patch

Signed-off-by: Tong Liu01 &lt;Tong.Liu01@amd.com&gt;
Acked-by: Luben Tuikov &lt;luben.tuikov@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: Disable MCBP from soc21 for SRIOV</title>
<updated>2022-11-04T20:05:53Z</updated>
<author>
<name>Yiqing Yao</name>
<email>yiqing.yao@amd.com</email>
</author>
<published>2022-10-28T09:30:17Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=8a1fbb4a5ec29d9ae5f792a1d3c049c5c709851c'/>
<id>urn:sha1:8a1fbb4a5ec29d9ae5f792a1d3c049c5c709851c</id>
<content type='text'>
[why]
Start from soc21, CP does not support MCBP, so disable it.

[how]
Used amgpu_mcbp flag alone instead of checking if is in SRIOV to
enable/disable MCBP.
Only set flag to enable on asic_type prior to soc21 in SRIOV.

Signed-off-by: Yiqing Yao &lt;yiqing.yao@amd.com&gt;
Reviewed-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amd: Add IMU fw version to fw version queries</title>
<updated>2022-10-24T18:34:47Z</updated>
<author>
<name>David Francis</name>
<email>David.Francis@amd.com</email>
</author>
<published>2022-09-16T15:15:01Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=b72362962a66693cd095389cbe41dd005bfcfb44'/>
<id>urn:sha1:b72362962a66693cd095389cbe41dd005bfcfb44</id>
<content type='text'>
IMU is a new firmware for GFX11.

There are four means by which firmware version can be queried
from the driver: device attributes, vf2pf, debugfs,
and the AMDGPU_INFO_FW_VERSION option in the amdgpu info ioctl.

Add IMU as an option for those four methods.

V2: Added debugfs

Reviewed-by: Likun Gao &lt;Likun.Gao@amd.com&gt;
Reviewed-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Signed-off-by: David Francis &lt;David.Francis@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: set vm_update_mode=0 as default for Sienna Cichlid in SRIOV case</title>
<updated>2022-10-17T21:41:20Z</updated>
<author>
<name>Danijel Slivka</name>
<email>danijel.slivka@amd.com</email>
</author>
<published>2022-10-04T13:39:44Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=a7310d8de3ba60a6ec4294392daf747b8333b3b2'/>
<id>urn:sha1:a7310d8de3ba60a6ec4294392daf747b8333b3b2</id>
<content type='text'>
For asic with VF MMIO access protection avoid using CPU for VM table updates.
CPU pagetable updates have issues with HDP flush as VF MMIO access protection
blocks write to mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL register
during sriov runtime.

v3: introduce virtualization capability flag AMDGPU_VF_MMIO_ACCESS_PROTECT
which indicates that VF MMIO write access is not allowed in sriov runtime

Signed-off-by: Danijel Slivka &lt;danijel.slivka@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/vcn: Add vcn/vcn1 in white list to load its firmware under sriov</title>
<updated>2022-09-01T19:12:04Z</updated>
<author>
<name>Jane Jian</name>
<email>Jane.Jian@amd.com</email>
</author>
<published>2022-08-22T07:19:11Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=60e9c7ee3fec7e4f40b1bb8e4e58be03a12c9eb1'/>
<id>urn:sha1:60e9c7ee3fec7e4f40b1bb8e4e58be03a12c9eb1</id>
<content type='text'>
Previously since vcn0/vcn1 are not enabled, loading firmware
is skipped. Now add firmware loading back since vcn0/vcn1
has already been enabled on sriov

Acked-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: Jane Jian &lt;Jane.Jian@amd.com&gt;
Reviewed-by: Hawking Zhang &lt;Hawking.Zhang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
<entry>
<title>drm/amdgpu: Support PSP 13.0.10 on SR-IOV</title>
<updated>2022-09-01T19:11:26Z</updated>
<author>
<name>Horace Chen</name>
<email>horace.chen@amd.com</email>
</author>
<published>2022-07-29T05:44:43Z</published>
<link rel='alternate' type='text/html' href='https://git.shady.money/linux/commit/?id=f8bd73213a13b695594fac76cae67105bcfc7706'/>
<id>urn:sha1:f8bd73213a13b695594fac76cae67105bcfc7706</id>
<content type='text'>
Add support for PSP 13.0.10 for SR-IOV VF

Acked-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: Horace Chen &lt;horace.chen@amd.com&gt;
Reviewed-by: Hawking Zhang &lt;Hawking.Zhang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
</content>
</entry>
</feed>
