[AMD Official Use Only - General]
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: Gao, Likun
Sent: Friday, April 21, 2023 13:38
To: amd-gfx list
Cc: Zhang, Hawking ; Xu, Feifei
Subject: [PATCH] drm/amdgpu: support psp vbflash sysfs for MP0 13_0_10
[AMD Official
[AMD Official Use Only - General]
From: Likun Gao
Add support for PSP vbflash sysfs interface with MP0 version v13.0.10.
Signed-off-by: Likun Gao
---
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
Hi Christian,
Thanks for working on this.
Series is,
Reviewed-by: Luben Tuikov
Regards,
Luben
On 2023-04-20 07:57, Christian König wrote:
> When a hw fence is signaled with an error properly forward that to the
> finished fence.
>
> Signed-off-by: Christian König
> ---
>
tree/branch:
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
branch HEAD: 44bf136283e567b2b62653be7630e7511da41da2 Add linux-next specific
files for 20230420
Error/Warning reports:
https://lore.kernel.org/oe-kbuild-all/202304172004.r3iph5ja-...@intel.com
https
[AMD Official Use Only - General]
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: amd-gfx On Behalf Of Alex Deucher
Sent: Wednesday, April 12, 2023 05:14
To: amd-gfx@lists.freedesktop.org
Cc: Deucher, Alexander
Subject: [PATCH V2] drm/amdgpu/gfx11: update
From: Aric Cyr
This DC version brings along:
- Improvements in the SubVP feature
- Keep disable aux-i delay as 0
- Add p-state debugging and improvements
- Fix in secure display context creation
- add option to use custom backlight caps
- Lowering min Z8 residency time
- Restore rptr/wptr for
From: Alvin Lee
[Description]
- Add debug option to enable SubVP for high refresh rate displays
- For now limit the enabled modes based on a table in debug options
- Currently disabled by default
Reviewed-by: Jun Lei
Acked-by: Rodrigo Siqueira
Signed-off-by: Alvin Lee
---
From: Michael Strauss
[WHY]
Current Aux-I sequence checks for local_sink which isn't populated on
MST links
[HOW]
Leave disable aux-i delay as 0 for MST cases
Cc: sta...@vger.kernel.org
Reviewed-by: George Shen
Reviewed-by: Aric Cyr
Acked-by: Rodrigo Siqueira
Signed-off-by: Michael Strauss
From: Alvin Lee
- We want to disable SubVP if Graphics Error Correction/Correcting Code
(GECC) is enabled.
- After reading feature caps from DMCUB during init, use the GECC
enable/disable info to determine if SubVP can be enabled or not.
Reviewed-by: Nicholas Kazlauskas
Reviewed-by: Jun
From: Sung Lee
[WHY]
P-State related issues are fairly common but currently
there is no way to debug these issues after the fact.
[HOW]
Add helpful registers to HW state queries
Reviewed-by: Aric Cyr
Reviewed-by: Jun Lei
Acked-by: Rodrigo Siqueira
Signed-off-by: Sung Lee
---
From: "Leo (Hanghong) Ma"
[Why]
Our QA found visual confirm color is not as expected for Auto
Color Management feature test after enable it.
[How]
Calculate scaler recout data when visual confirm enabled to update
the visual confirm bar on the display.
Reviewed-by: Aric Cyr
Acked-by: Rodrigo
From: Gabe Teeger
[Why]
Flickering and underflow was observed when testing extended
blank on dcn314.
[What]
Vstartup is contrainted by vblank_nom, so adjusting it to include
non-adjusted vtotal in its calculation during freesync video
means that Vstartup is not changed when vtotal changes.
This
From: Alvin Lee
- Previously update_clocks was overriding pstate support if
it checked that there were no planes
- However, P-State support should be determined in validation
phase instead
- This fixes an issue where a transition from FPO -> no planes
expects UCLK MAX, but update_clocks
From: Alan Liu
[Why & How]
We need to store CRTC information in secure_display_ctx, so postpone
the call to amdgpu_dm_crtc_secure_display_create_contexts() until we
initialize all CRTCs.
Cc: sta...@vger.kernel.org
Reviewed-by: Wayne Lin
Acked-by: Rodrigo Siqueira
Signed-off-by: Alan Liu
---
From: Josip Pavic
[Why & How]
Provide option for vendors to specify a custom brightness-to-backlight
conversion profile.
Reviewed-by: Anthony Koo
Acked-by: Rodrigo Siqueira
Signed-off-by: Josip Pavic
---
.../amd/display/modules/power/power_helpers.c | 43 +++
From: Leo Chen
[Why & How]
Per HW team request, we're lowering the minimum Z8
residency time to 2000us. This enables Z8 support for additional
modes we were previously blocking like 2k>60hz
Cc: sta...@vger.kernel.org
Reviewed-by: Nicholas Kazlauskas
Acked-by: Rodrigo Siqueira
Signed-off-by:
From: Alvin Lee
Enable SubVP on PSR panels now that we have FW support
Reviewed-by: Jun Lei
Acked-by: Rodrigo Siqueira
Signed-off-by: Alvin Lee
---
drivers/gpu/drm/amd/display/dc/dc.h | 1 +
drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.c | 1 +
From: Dmytro Laktyushkin
This feature is meant to unblock PSTATE for certain high end display
configs on dcn315. This is achieved by allocating CRB to detile buffer
based on display requirements to meet pstate latency hiding needs.
Reviewed-by: Charlene Liu
Acked-by: Rodrigo Siqueira
From: Alvin Lee
- Due to hardware related QoS issues, we need to limit certain
SKUs with less memory channels to DPM1 and above.
- At DPM0 + workload running, the urgent return latency can
exceed 15us (the expected maximum is 4us) which results in underflow
Cc: sta...@vger.kernel.org
From: "JinZe.Xu"
[Why]
States may be desync after resume.
[How]
Sync sw state with hw state.
Reviewed-by: Nicholas Kazlauskas
Acked-by: Rodrigo Siqueira
Signed-off-by: JinZe.Xu
---
drivers/gpu/drm/amd/display/dmub/dmub_srv.h | 14 ++
From: Alex Hung
[Why]
When userspace (IGT) inserts EDID with audio data (ELD), EDID
is not updated.
[How]
Implements force function (amdgpu_dm_connector_funcs_force) in
amdgpu_dm_connector_funcs to create emulated sink and to handle
EDID.
Reviewed-by: Hersen Wu
Acked-by: Rodrigo Siqueira
From: Alvin Lee
Reorder FW feature caps struct variable order to ensure backwards
compatability is maintained for older FW
Reviewed-by: Sung Lee
Acked-by: Rodrigo Siqueira
Signed-off-by: Alvin Lee
---
drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h | 3 ++-
1 file changed, 2 insertions(+),
From: Aric Cyr
Acked-by: Rodrigo Siqueira
Signed-off-by: Aric Cyr
---
drivers/gpu/drm/amd/display/dc/dc.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h
b/drivers/gpu/drm/amd/display/dc/dc.h
index 97747f5fde56..d6f0ecb708cd 100644
---
From: Anthony Koo
- Add feature caps for Subvp + PSR so driver is back compatible
Acked-by: Rodrigo Siqueira
Signed-off-by: Anthony Koo
---
drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
From: Michael Strauss
[WHY]
32ms delay was added to resolve issue with a specific sink, however this same
delay also introduces erroneous link training failures with certain sink
devices.
[HOW]
Only apply the 32ms delay for offending devices instead of globally.
Reviewed-by: Jun Lei
Acked-by:
From: Hersen Wu
[Why] implementation change of drm_edid_override_set since linux
kernel 6.1, edid from debugfs is saved into connector->edid_override
immediatey, not saved to connector->edid_blob_ptr at the same time.
[How] call new drm_edid function drm_connector_update_edid_property
to assign
This DC patchset brings improvements in multiple areas. In summary, we
highlight:
- Improvements in the SubVP feature
- Keep disabling aux-i delay as 0
- Add p-state debugging and improvements
- Fix in secure display context creation
- add an option to use custom backlight caps
- Lowering min Z8
From: Xiaogang Chen
amdgpu_ttm_tt_get_user_pages can fail(-EFAULT). If it failed mem has no
associated
hmm range or user_pages associated. Keep it at process_info->userptr_inval_list
and
mark mem->invalid until following scheduled attempts can valid it.
Signed-off-by: Xiaogang Chen
---
On Thu, Apr 20, 2023 at 2:59 PM Christian König
wrote:
> Could you try drm-misc-next as well?
If as I assume I cloned right repo
$ git clone -b drm-misc-next
git://anongit.freedesktop.org/drm/drm-misc linux-drm-misc-next
for my hardware last commit on this branch is turned out completely
Ping?
On Tue, Apr 11, 2023 at 5:14 PM Alex Deucher wrote:
>
> This code was written prior to previous updates to this
> logic for other chips. The RSC registers are part of
> SMUIO which is an always on block so there is no need
> to disable gfxoff. Additionally add the carryover and
>
clang with W=1 reports
drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn314/display_rq_dlg_calc_314.c:1003:15:
error: variable 'dispclk_delay_subtotal' set but not used
[-Werror,-Wunused-but-set-variable]
unsigned int dispclk_delay_subtotal;
^
This variable is not
On 20/04/2023 12:56, gre...@linuxfoundation.org wrote:
> [...]
> That's 3000+ emails ago for me :)
/head_exploding
this is > 1000 emails per day, wow...my sympathies to you heh
> [...]
>> tl;dr: the offender is present on 6.1.y, but this fix is not, hence I'm
>> hereby requesting the merge.
Currently, on a handful of ASICs. We allow the framebuffer for a given
plane to exist in either VRAM or GTT. However, if the plane's new
framebuffer is in a different memory domain than it's previous
framebuffer, flipping between them can cause the screen to flicker. So,
to fix this, don't perform
Reduces preemption latency.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 60bb4bba1994..f52feee1d64d 100644
---
On Thu, Apr 20, 2023 at 12:36:00PM -0300, Guilherme G. Piccoli wrote:
> On 20/04/2023 12:02, gre...@linuxfoundation.org wrote:
> >> [...]
> >>> Which "one" are you referring to here?
> >>>
> >>> confused,
> >>>
> >>> greg k-h
> >>
> >> This one, sent in this email thread.
> >
> > I don't have
On 20/04/2023 12:02, gre...@linuxfoundation.org wrote:
>> [...]
>>> Which "one" are you referring to here?
>>>
>>> confused,
>>>
>>> greg k-h
>>
>> This one, sent in this email thread.
>
> I don't have "this email thread" anymore, remember, some of us get
> thousand+ emails a day...
I don't
Am 20.04.23 um 15:44 schrieb Hamza Mahfooz:
Currently, on a handful of ASICs. We allow the framebuffer for a given
plane to exist in either VRAM or GTT. However, if the plane's new
framebuffer is in a different memory domain than it's previous
framebuffer, flipping between them can cause the
On Thu, Apr 20, 2023 at 11:04 AM gre...@linuxfoundation.org
wrote:
>
> On Thu, Apr 20, 2023 at 09:59:17AM -0300, Guilherme G. Piccoli wrote:
> > On 20/04/2023 09:42, gre...@linuxfoundation.org wrote:
> > > [...]
> > >> @Greg, can you pick this one? Thanks!
> > >
> > > Which "one" are you
Am 20.04.23 um 14:56 schrieb Alex Deucher:
On Thu, Apr 20, 2023 at 5:19 AM Feifei Xu wrote:
Extend to 120s. The default timeout value should also extend if compute
shader execution time extended. Otherwise some stress test will trigger
compute ring timeout in software.
I think that's probably
On Thu, Apr 20, 2023 at 09:59:17AM -0300, Guilherme G. Piccoli wrote:
> On 20/04/2023 09:42, gre...@linuxfoundation.org wrote:
> > [...]
> >> @Greg, can you pick this one? Thanks!
> >
> > Which "one" are you referring to here?
> >
> > confused,
> >
> > greg k-h
>
> This one, sent in this email
[Public]
Reviewed-by: Roman Li
> -Original Message-
> From: amd-gfx On Behalf Of Hamza
> Mahfooz
> Sent: Thursday, April 20, 2023 9:44 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Wang, Chao-kai (Stylon) ; Tuikov, Luben
> ; dri-de...@lists.freedesktop.org; Li, Sun peng (Leo)
> ; David
This patch introduces new IOCTL for userqueue secure semaphore.
The signal IOCTL called from userspace application creates a drm
syncobj and array of bo GEM handles and passed in as parameter to
the driver to install the fence into it.
The wait IOCTL gets an array of drm syncobjs, finds the
- Added support to handle the userqueue protected fence signal
hardware interrupt.
- Create a hash table which maps va address to the fence driver.
Signed-off-by: Arunpravin Paneer Selvam
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
- Add UAPI header support for userqueue Secure semaphore
v2: Worked on review comments from Christian for the
following modifications
- Add bo handles, bo flags and padding fields.
- Include value/va in a combined array.
v3: Worked on review comments from Christian
Developed a userqueue fence driver for the userqueue process shared
BO synchronization.
Create a dma fence having write pointer as the seqno and allocate a
seq64 memory for each user queue process and feed this memory address
into the firmware/hardware, thus the firmware writes the read pointer
- Add a field in struct v11_gfx_mqd for userqueue
fence address.
- Assign fence gpu VA address to the userqueue mqd
fence address fields.
v2: Remove the mask and replace with lower_32_bits (Christian)
Signed-off-by: Arunpravin Paneer Selvam
Reviewed-by: Christian König
---
Developed a new driver which allocates a 64bit memory on
each request in sequence order. At the moment, user queue
fence memory is the main consumer of this seq64 driver.
v2: Worked on review comments from Christian for the following
modifications
- Move driver name from "semaphore" to
The idea about usermode queue fencing synchronization is described
in the below link
https://patchwork.freedesktop.org/series/114385/
The core usermode queue and doorbell design patches in review are seen below
which are prerequisites for this work.
gcc with W=1 reports
drivers/gpu/drm/amd/amdgpu/../display/dc/dc_dmub_srv.c:
In function ‘dc_dmub_srv_optimized_init_done’:
drivers/gpu/drm/amd/amdgpu/../display/dc/dc_dmub_srv.c:184:26:
error: variable ‘dmub’ set but not used [-Werror=unused-but-set-variable]
184 | struct dmub_srv
gcc with W=1 reports
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c: In function
‘amdgpu_gfx_mqd_sw_fini’:
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c:451:16: error: variable
‘j’ set but not used [-Werror=unused-but-set-variable]
451 | int i, j;
|^
On 4/20/23 04:47, Arnd Bergmann wrote:
From: Arnd Bergmann
The dmub_abm_set_ambient_level() function has no caller and can
just be removed, the other ones have a declaration in the
header file and just need to see the prototype:
On 4/20/23 04:47, Arnd Bergmann wrote:
From: Arnd Bergmann
The only references to these variables were removed, so they now cause
warnings and have to be removed as well:
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn21/dcn21_hwseq.c:226:20: error:
unused variable 'cmd'
On 4/19/23 22:04, Tom Rix wrote:
gcc reports
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn21/dcn21_hwseq.c:
In function ‘dcn21_set_backlight_level’:
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn21/dcn21_hwseq.c:229:18:
error: unused variable ‘otg_inst’ [-Werror=unused-variable]
229 |
Currently, on a handful of ASICs. We allow the framebuffer for a given
plane to exist in either VRAM or GTT. However, if the plane's new
framebuffer is in a different memory domain than it's previous
framebuffer, flipping between them can cause the screen to flicker. So,
to fix this, don't perform
On 20/04/2023 09:42, gre...@linuxfoundation.org wrote:
> [...]
>> @Greg, can you pick this one? Thanks!
>
> Which "one" are you referring to here?
>
> confused,
>
> greg k-h
This one, sent in this email thread.
The title of the patch is "drm/amdgpu/vcn: Disable indirect SRAM on
Vangogh broken
gcc with W=1 reports
drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dmub_abm.c:
In function ‘dmub_abm_set_event_ex’:
drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dmub_abm.c:138:22: error: variable
‘feature_support’ set but not used [-Werror=unused-but-set-variable]
138 | unsigned int
On Thu, Apr 20, 2023 at 5:19 AM Feifei Xu wrote:
>
> Extend to 120s. The default timeout value should also extend if compute
> shader execution time extended. Otherwise some stress test will trigger
> compute ring timeout in software.
I think that's probably too long. 2 minutes is a long time
On Thu, Apr 20, 2023 at 09:36:28AM -0300, Guilherme G. Piccoli wrote:
> On 19/04/2023 17:04, Deucher, Alexander wrote:
> > [...]
> >> So, let me check if I understood properly: there are 2 trees (-fixes and
> >> -next),
> >> and the problem is that their merge onto mainline happens apart and
On 19/04/2023 17:04, Deucher, Alexander wrote:
> [...]
>> So, let me check if I understood properly: there are 2 trees (-fixes and
>> -next),
>> and the problem is that their merge onto mainline happens apart and there
>> are kind of duplicate commits, that were first merged on -fixes, then "re-
[AMD Official Use Only - General]
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: Feifei Xu
Sent: Thursday, April 20, 2023 17:20
To: amd-gfx@lists.freedesktop.org
Cc: Xu, Feifei ; Zhang, Hawking
Subject: [PATCH] drm/amdgpu: extend the default timeout for kernel
Instead of using the VRAM lost counter add a 64bit token which indicates
if a context or job is still valid to use.
Should the VRAM be lost or the page tables need re-creation the token will
change indicating that userspace needs to act and re-create the contexts
and re-submit the work.
When some problem with the updates of page tables is detected reset the
state machine of the VM and re-create all page tables from scratch.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 81 +-
1 file changed, 65 insertions(+), 16
This allows us to insert some error codes into the bottom of the pipeline
on an engine.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 24 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c | 15 ++
Set the fence error code before trying to soft-recover it.
It gets overwritten when a hard recovery is required.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
Forward errors from previous submissions to this one.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 13 -
1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
When we force complete fences we should mark them as canceled.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index
Switch to using RCU handling for the last scheduled job and add a
function to return the error code of it.
Signed-off-by: Christian König
---
drivers/gpu/drm/scheduler/sched_entity.c | 39 +++-
include/drm/gpu_scheduler.h | 3 +-
2 files changed, 33
When a hw fence is signaled with an error properly forward that to the
finished fence.
Signed-off-by: Christian König
---
drivers/gpu/drm/scheduler/sched_entity.c | 4 +---
drivers/gpu/drm/scheduler/sched_fence.c | 4 +++-
drivers/gpu/drm/scheduler/sched_main.c | 18 --
On Thu, Apr 20, 2023 at 2:59 PM Christian König
wrote:
>
> Could you try drm-misc-next as well?
>
> Going to give drm-fixes another round of testing.
>
> Thanks,
> Christian.
Important don't give up.
https://youtu.be/25zhHBGIHJ8 [40 min]
https://youtu.be/utnDR26eYBY [50 min]
Am 19.04.23 um 21:17 schrieb Mikhail Gavrilov:
On Wed, Apr 19, 2023 at 6:15 PM Christian König
wrote:
Well exactly that's the problem, I can't reproduce it.
Have you applied any local change or config which could explain that?
I did not apply any local changes.
I just pulled all changes
Extend to 120s. The default timeout value should also extend if compute
shader execution time extended. Otherwise some stress test will trigger
compute ring timeout in software.
Signed-off-by: Feifei Xu
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 ++--
1 file changed, 2 insertions(+), 2
From: Arnd Bergmann
The only references to these variables were removed, so they now cause
warnings and have to be removed as well:
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn21/dcn21_hwseq.c:226:20: error:
unused variable 'cmd' [-Werror,-Wunused-variable]
From: Arnd Bergmann
The dmub_abm_set_ambient_level() function has no caller and can
just be removed, the other ones have a declaration in the
header file and just need to see the prototype:
drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dmub_abm_lcd.c:122:14: error: no
previous prototype for
From: Arnd Bergmann
This was left global by accident, the corresponding functions for other
hardware types are already static:
drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c:1072:6: error: no previous prototype
for function 'gfx_v9_4_3_disable_gpa_mode' [-Werror,-Wmissing-prototypes]
Fixes:
ay-dc-link-link_validation.c:warning:variable-link-set-but-not-used
|-- alpha-randconfig-r001-20230420
| |--
drivers-gpu-drm-amd-amdgpu-..-display-dc-link-link_validation.c:warning:variable-bw_needed-set-but-not-used
| |--
drivers-gpu-drm-amd-amdgpu-..-display-dc-link-link_validation.c:warni
75 matches
Mail list logo