[Public]
Looks good to me. Reviewed-by: Guchun Chen
> -Original Message-
> From: Quan, Evan
> Sent: Monday, July 24, 2023 11:21 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Deucher, Alexander ; Chen, Guchun
> ; Quan, Evan
> Subject: [PATCH V2] drm/amd/pm: disable the SMU13 OD feature
The existing OD interface cannot support the growing demand for more
OD features. We are in the transition to a new OD mechanism. So,
disable the SMU13 OD feature support temporarily. And this should be
reverted when the new OD mechanism online.
Signed-off-by: Evan Quan
--
v1->v2:
- comment
[AMD Official Use Only - General]
Reviewed-by: Tao Zhou
> -Original Message-
> From: amd-gfx On Behalf Of
> Stanley.Yang
> Sent: Friday, July 21, 2023 9:18 PM
> To: amd-gfx@lists.freedesktop.org; Zhang, Hawking ;
> Zhou1, Tao ; Chai, Thomas ; Li,
> Candice
> Cc: Yang, Stanley
>
[Public]
It's better to use prefix "drm/amdgpu: ..." in subject. With it fixed, this
patch is: Reviewed-by: Guchun Chen
Regards,
Guchun
> -Original Message-
> From: Bob Zhou
> Sent: Monday, July 24, 2023 10:55 AM
> To: amd-gfx@lists.freedesktop.org; Kim, Jonathan
> Cc: Chen, Guchun
[Public]
Looks it's more clear to include code line " smu->od_settings =
_table->overdrive_table;" into the #if 0 section?
Regards,
Guchun
> -Original Message-
> From: amd-gfx On Behalf Of Evan
> Quan
> Sent: Monday, July 24, 2023 10:42 AM
> To: amd-gfx@lists.freedesktop.org
> Cc:
The setting of mes_add_queue_pkt is repeat, so remove it.
Signed-off-by: Bob Zhou
---
drivers/gpu/drm/amd/amdgpu/mes_v11_0.c | 4
1 file changed, 4 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
index 8e67e965f7ea..e806331fee42
[AMD Official Use Only - General]
Gentle ping on this series.
Hi Rafael and Andrew,
Can you help to check this latest series and share your thoughts if any?
BR,
Evan
> -Original Message-
> From: Quan, Evan
> Sent: Wednesday, July 19, 2023 5:00 PM
> To: raf...@kernel.org;
The existing OD interface cannot support the growing demand for more
OD features. We are in the transition to a new OD mechanism. So,
disable the SMU13 OD feature support temporarily. And this should be
reverted when the new OD mechanism online.
Signed-off-by: Evan Quan
---
From: shanzhulig
[ Upstream commit 2e54154b9f27262efd0cb4f903cc7d5ad1fe9628 ]
fence Decrements the reference count before exiting.
Avoid Race Vulnerabilities for fence use-after-free.
v2 (chk): actually fix the use after free and not just move it.
Signed-off-by: shanzhulig
Signed-off-by:
From: shanzhulig
[ Upstream commit 2e54154b9f27262efd0cb4f903cc7d5ad1fe9628 ]
fence Decrements the reference count before exiting.
Avoid Race Vulnerabilities for fence use-after-free.
v2 (chk): actually fix the use after free and not just move it.
Signed-off-by: shanzhulig
Signed-off-by:
From: shanzhulig
[ Upstream commit 2e54154b9f27262efd0cb4f903cc7d5ad1fe9628 ]
fence Decrements the reference count before exiting.
Avoid Race Vulnerabilities for fence use-after-free.
v2 (chk): actually fix the use after free and not just move it.
Signed-off-by: shanzhulig
Signed-off-by:
From: shanzhulig
[ Upstream commit 2e54154b9f27262efd0cb4f903cc7d5ad1fe9628 ]
fence Decrements the reference count before exiting.
Avoid Race Vulnerabilities for fence use-after-free.
v2 (chk): actually fix the use after free and not just move it.
Signed-off-by: shanzhulig
Signed-off-by:
From: shanzhulig
[ Upstream commit 2e54154b9f27262efd0cb4f903cc7d5ad1fe9628 ]
fence Decrements the reference count before exiting.
Avoid Race Vulnerabilities for fence use-after-free.
v2 (chk): actually fix the use after free and not just move it.
Signed-off-by: shanzhulig
Signed-off-by:
From: shanzhulig
[ Upstream commit 2e54154b9f27262efd0cb4f903cc7d5ad1fe9628 ]
fence Decrements the reference count before exiting.
Avoid Race Vulnerabilities for fence use-after-free.
v2 (chk): actually fix the use after free and not just move it.
Signed-off-by: shanzhulig
Signed-off-by:
From: shanzhulig
[ Upstream commit 2e54154b9f27262efd0cb4f903cc7d5ad1fe9628 ]
fence Decrements the reference count before exiting.
Avoid Race Vulnerabilities for fence use-after-free.
v2 (chk): actually fix the use after free and not just move it.
Signed-off-by: shanzhulig
Signed-off-by:
From: hackyzh002
[ Upstream commit f828b681d0cd566f86351c0b913e6cb6ed8c7b9c ]
The type of size is unsigned, if size is 0x4000, there will be an
integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later
Reviewed-by: Christian
From: hackyzh002
[ Upstream commit f828b681d0cd566f86351c0b913e6cb6ed8c7b9c ]
The type of size is unsigned, if size is 0x4000, there will be an
integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later
Reviewed-by: Christian
From: hackyzh002
[ Upstream commit f828b681d0cd566f86351c0b913e6cb6ed8c7b9c ]
The type of size is unsigned, if size is 0x4000, there will be an
integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later
Reviewed-by: Christian
From: Lang Yu
[ Upstream commit 187916e6ed9d0c3b3abc27429f7a5f8c936bd1f0 ]
When using cpu to update page tables, vm update fences are unused.
Install stub fence into these fence pointers instead of NULL
to avoid NULL dereference when calling dma_fence_wait() on them.
Suggested-by: Christian
From: hackyzh002
[ Upstream commit f828b681d0cd566f86351c0b913e6cb6ed8c7b9c ]
The type of size is unsigned, if size is 0x4000, there will be an
integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later
Reviewed-by: Christian
From: Lang Yu
[ Upstream commit 187916e6ed9d0c3b3abc27429f7a5f8c936bd1f0 ]
When using cpu to update page tables, vm update fences are unused.
Install stub fence into these fence pointers instead of NULL
to avoid NULL dereference when calling dma_fence_wait() on them.
Suggested-by: Christian
From: hackyzh002
[ Upstream commit f828b681d0cd566f86351c0b913e6cb6ed8c7b9c ]
The type of size is unsigned, if size is 0x4000, there will be an
integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later
Reviewed-by: Christian
From: Daniel Miess
[ Upstream commit 17fbdbda9cc87ff5a013898de506212d25323ed7 ]
[Why and How]
Add back debug bits enabling RCO for dcn314 as underflow
associated with this change has been resolved
Acked-by: Stylon Wang
Signed-off-by: Daniel Miess
Reviewed-by: Jun Lei
Tested-by: Daniel
From: Nicholas Kazlauskas
[ Upstream commit 30f90f3c1c2c63c2fa44f61233737d27b72637c2 ]
[Why]
Hardware implements root clock gating by utilizing the DPP DTO registers
with a special case of DTO enabled, phase = 0, modulo = 1. This
conflicts with our policy to always update the DPPDTO for cases
From: Michel Dänzer
[ Upstream commit 8e1b45c578b799510f9a01a9745a737e74f43cb1 ]
This reverts commit 474f01015ffdb74e01c2eb3584a2822c64e7b2be.
Caused a regression:
Samsung Odyssey Neo G9, running at 5120x1440@240/VRR, connected to Navi
21 via DisplayPort, blanks and the GPU hangs while
From: Lang Yu
[ Upstream commit 187916e6ed9d0c3b3abc27429f7a5f8c936bd1f0 ]
When using cpu to update page tables, vm update fences are unused.
Install stub fence into these fence pointers instead of NULL
to avoid NULL dereference when calling dma_fence_wait() on them.
Suggested-by: Christian
From: Alvin Lee
[ Upstream commit 7e60ab4eb3e4ba2adac46d737fdbbc5732bebd58 ]
[Description]
- Previously we wanted to apply extra 60us of prefetch for min DCFCLK
(200Mhz), but DCFCLK can be calculated to be 201Mhz which underflows
also without the extra prefetch
- Instead, apply the the
From: Jack Xiao
[ Upstream commit 31d7c3a4fc3d312a0646990767647925d5bde540 ]
The fences associated with mes queue have to be freed
up during amdgpu_ring_fini.
Signed-off-by: Jack Xiao
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
Signed-off-by: Sasha Levin
---
From: hackyzh002
[ Upstream commit 87c2213e85bd81e4a9a4d0880c256568794ae388 ]
The type of size is unsigned int, if size is 0x4000, there will
be an integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later.
Reviewed-by:
From: hackyzh002
[ Upstream commit f828b681d0cd566f86351c0b913e6cb6ed8c7b9c ]
The type of size is unsigned, if size is 0x4000, there will be an
integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later
Reviewed-by: Christian
From: Longlong Yao
[ Upstream commit 01382501509871d0799bab6bd412c228486af5bf ]
The following call trace is observed when removing the amdgpu driver, which
is caused by that BOs allocated for psp are not freed until removing.
[61811.450562] RIP: 0010:amddrm_buddy_fini.cold+0x29/0x47
From: Wesley Chalmers
[ Upstream commit e101bf95ea87ccc03ac2f48dfc0757c6364ff3c7 ]
[WHY]
Writing to DRR registers such as OTG_V_TOTAL_MIN on the same frame as a
pipe commit can cause underflow.
[HOW]
Move DMUB p-state delegate into optimze_bandwidth; enabling FAMS sets
optimized_required.
From: Daniel Miess
[ Upstream commit 17fbdbda9cc87ff5a013898de506212d25323ed7 ]
[Why and How]
Add back debug bits enabling RCO for dcn314 as underflow
associated with this change has been resolved
Acked-by: Stylon Wang
Signed-off-by: Daniel Miess
Reviewed-by: Jun Lei
Tested-by: Daniel
From: Nicholas Kazlauskas
[ Upstream commit 30f90f3c1c2c63c2fa44f61233737d27b72637c2 ]
[Why]
Hardware implements root clock gating by utilizing the DPP DTO registers
with a special case of DTO enabled, phase = 0, modulo = 1. This
conflicts with our policy to always update the DPPDTO for cases
From: Lang Yu
[ Upstream commit 5daff15cd013422bc6d1efcfe82b586800025384 ]
Root PD BO should be reserved before unmap and remove
a bo_va from VM otherwise lockdep will complain.
v2: check fpriv->csa_va is not NULL instead of amdgpu_mcbp (christian)
[14616.936827] WARNING: CPU: 6 PID: 1711 at
From: Lijo Lazar
[ Upstream commit 1718e973e3d23b653cd77994073a9deda3875689 ]
Populate metrics data table for SMU v13.0.6. Add PCIe link speed/width
information also.
Signed-off-by: Lijo Lazar
Reviewed-by: Le Ma
Signed-off-by: Alex Deucher
Signed-off-by: Sasha Levin
---
From: Michel Dänzer
[ Upstream commit 360930985ec9f394c82ba0b235403b4a366d1560 ]
This reverts commit e101bf95ea87ccc03ac2f48dfc0757c6364ff3c7.
Caused a regression:
Samsung Odyssey Neo G9, running at 5120x1440@240/VRR, connected to Navi
21 via DisplayPort, blanks and the GPU hangs while
From: Alvin Lee
[ Upstream commit 7e60ab4eb3e4ba2adac46d737fdbbc5732bebd58 ]
[Description]
- Previously we wanted to apply extra 60us of prefetch for min DCFCLK
(200Mhz), but DCFCLK can be calculated to be 201Mhz which underflows
also without the extra prefetch
- Instead, apply the the
From: Saaem Rizvi
[ Upstream commit 3e8d74cb128fb1a4d56270ffbecea6056c55739a ]
[WHY]
Currently, there is an intermittent issue where a screen can either go
blank or be corrupted.
[HOW]
To resolve the issue we trigger the ramping logic for DIO FIFO so that
it goes back up to the correct speed.
From: Daniel Miess
[ Upstream commit 3a31e8b89b7240d9a17ace8a1ed050bdcb560f9e ]
[Why]
Calls to dcn20_adjust_freesync_v_startup are no longer
needed as of dcn3+ and can cause underflow in some cases
[How]
Move calls to dcn20_adjust_freesync_v_startup up into
validate_bandwidth for dcn2.x
From: Lang Yu
[ Upstream commit 187916e6ed9d0c3b3abc27429f7a5f8c936bd1f0 ]
When using cpu to update page tables, vm update fences are unused.
Install stub fence into these fence pointers instead of NULL
to avoid NULL dereference when calling dma_fence_wait() on them.
Suggested-by: Christian
From: Aurabindo Pillai
[ Upstream commit f38129bb081758176dd78304faaee95007fb8838 ]
This reverts commit 80c6d6804f31451848a3956a70c2bcb1f07cfcb0.
The orignal commit was intended as a workaround to prevent underflow and
flickering when using one normal monitor and the other high refresh rate
From: Jack Xiao
[ Upstream commit 31d7c3a4fc3d312a0646990767647925d5bde540 ]
The fences associated with mes queue have to be freed
up during amdgpu_ring_fini.
Signed-off-by: Jack Xiao
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
Signed-off-by: Sasha Levin
---
From: hackyzh002
[ Upstream commit 87c2213e85bd81e4a9a4d0880c256568794ae388 ]
The type of size is unsigned int, if size is 0x4000, there will
be an integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later.
Reviewed-by:
From: Longlong Yao
[ Upstream commit 01382501509871d0799bab6bd412c228486af5bf ]
The following call trace is observed when removing the amdgpu driver, which
is caused by that BOs allocated for psp are not freed until removing.
[61811.450562] RIP: 0010:amddrm_buddy_fini.cold+0x29/0x47
From: hackyzh002
[ Upstream commit f828b681d0cd566f86351c0b913e6cb6ed8c7b9c ]
The type of size is unsigned, if size is 0x4000, there will be an
integer overflow, size will be zero after size *= sizeof(uint32_t),
will cause uninitialized memory to be referenced later
Reviewed-by: Christian
From: Alvin Lee
[ Upstream commit 128c1ca0303fe764a4cde5f761e72810d9e40b6e ]
[Why]
- Implement interface to program DTBCLK DTO’s
according to reference DTBCLK returned by PMFW
- This is required because DTO programming
requires exact DTBCLK reference freq or it could
result in underflow
From: Wesley Chalmers
[ Upstream commit e101bf95ea87ccc03ac2f48dfc0757c6364ff3c7 ]
[WHY]
Writing to DRR registers such as OTG_V_TOTAL_MIN on the same frame as a
pipe commit can cause underflow.
[HOW]
Move DMUB p-state delegate into optimze_bandwidth; enabling FAMS sets
optimized_required.
Return -ENOMEM, when there is no sufficient dynamically allocated memory
Cc: Christian König
Cc: Alex Deucher
Signed-off-by: Srinivasan Shanmugam
---
v2:
- Added -ENOMEM - when it failed to create MQD backup,
while creating MQD for each KCQ
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 17
Return -ENOMEM, when there is no sufficient dynamically allocated memory
Cc: Christian König
Cc: Alex Deucher
Signed-off-by: Srinivasan Shanmugam
---
drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c | 1 +
drivers/gpu/drm/amd/amdgpu/mes_v10_1.c | 4 +++-
drivers/gpu/drm/amd/amdgpu/mes_v11_0.c | 4
Expressions under 'else' branch in function 'dm_vblank_get_counter' &
'amdgpu_dm_backlight_get_level' are executed whenever the expression in
'if' is False. Otherwise, return from case occurs. Therefore, there is
no need in 'else', and it has been removed.
Fixes the following:
WARNING: else is
Fixes the following:
WARNING: min() should probably be min_t(size_t, size, sizeof(ip))
+ ret = copy_to_user(out, , min((size_t)size, sizeof(ip)));
And other style fixes:
WARNING: Prefer 'unsigned int' to bare use of 'unsigned'
WARNING: Missing a blank line after declarations
Cc:
For a constant format without additional arguments, use seq_puts()
instead of seq_printf(). Also, it fixes the following warning.
WARNING: Prefer seq_puts to seq_printf
And other style fixes:
WARNING: Prefer 'unsigned int' to bare use of 'unsigned'
WARNING: Block comments should align the * on
Expressions under 'else' branch under case 'CHIP_SIENNA_CICHLID' in
function 'is_fru_eeprom_supported' are executed whenever the expression
in 'if' is False. Otherwise, return from case occurs. Therefore, there
is no need in 'else', and it has been removed.
Fixes the following:
WARNING: else is
ENOSYS should be used for nonexistent syscalls only, replace ENOSYS with
EINVAL & other style fixes
WARNING: ENOSYS means 'invalid syscall nr' and nothing else
+ if (r == -ENOSYS)
WARNING: ENOSYS means 'invalid syscall nr' and nothing else
+ if (r == -ENOSYS)
Cc: Christian König
55 matches
Mail list logo