回复: [RFC] a new approach to detect which ring is the real black sheep upon TDR reported

[AMD Official Use Only - Internal Distribution Only] Fix typo: But in fact this job2 is innocent, and we should insert it back after recovery , and due to it was already deleted this innocent job’s context/process is really harmed 发件人: Liu, Monk 发送时间: 2021年2月27日 11:56 收件人: Grodzovsky, Andrey ;

回复: [RFC] a new approach to detect which ring is the real black sheep upon TDR reported

[AMD Official Use Only - Internal Distribution Only] H Andrey The scenario I hit here is not the one you mentioned, let me explain it with more details by another much easier understood example: Consider ring you have a job1 on KCQ, but the timeout of KCQ is 60 seconds (just for example) You a

回复: [RFC] a new approach to detect which ring is the real black sheep upon TDR reported

[AMD Official Use Only - Internal Distribution Only] the code I pasted is to illustrate why the innocent job is already taken out in the mirror list thus my suggested proposal won’t work unless we don’t delete the job in sched_job_timeout() routine, and the problem you stated is with my underst

[PATCH 1/1] drm/ttm: Ignore signaled move fences

Move fences that have already signaled should not prevent memory allocations with no_wait_gpu. Signed-off-by: Felix Kuehling --- drivers/gpu/drm/ttm/ttm_bo.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 3a

RE: Overlay issues

[AMD Official Use Only - Approved for External Use] Thanks for the suggestion, Simon. Improved the patch, now in the process of getting it reviewed internally. It'll be posted for upstream thereafter. -Original Message- From: Simon Ser Sent: Monday, February 22, 2021 5:31 PM To: Corn

Re: [PATCH 00/11] DC Patches Feb 26th, 2021

Hi all, This week this patchset was tested on a HP Envy 360, with Ryzen 5 4500U, on the following display types (via usb-c to dp/dvi/hdmi/vga): 4k 60z, 1440p 144hz, 1680*1050 60hz, internal eDP 1080p 60hz Tested on a Sapphire Pulse RX5700XT on the following display types (via DP): 4k60 60hz, 144

Re: [PATCH 2/2] drm/amdgpu/display: buffer INTERRUPT_LOW_IRQ_CONTEXT interrupt work

On 2021-01-22 3:55 p.m., Chen, Xiaogang wrote: On 1/19/2021 4:29 PM, Grodzovsky, Andrey wrote: On 1/15/21 2:21 AM, Chen, Xiaogang wrote: On 1/14/2021 1:24 AM, Grodzovsky, Andrey wrote: On 1/14/21 12:11 AM, Chen, Xiaogang wrote: On 1/12/2021 10:54 PM, Grodzovsky, Andrey wrote: On 1/4/21 1

[PATCH 1/1] drm/amdkfd: Use a new capability bit for SRAM ECC

Existing, buggy user mode breaks when SRAM ECC is correctly reported as "enabled". To avoid breaking existing user mode, deprecate that bit and leave it as 0. Define a new bit to report the actual SRAM ECC mode that new, correct user mode can use in the future. Fixes: 7ec177bdcfc1 ("drm/amdkfd: fi

[PATCH 11/11] drm/amd/display: Fix HSplit causing increase in DSC Slice Count

From: Dillon Varone [Why?] HSplit should not affect DSC slice count. Can cause improper timings to be applied for certain modes. [How?] No longer change DSC Slice count based on HSplit. Signed-off-by: Dillon Varone Reviewed-by: Dmytro Laktyushkin Acked-by: Rodrigo Siqueira --- drivers/gpu/d

[PATCH 10/11] drm/amd/display: treat memory as a single-channel for asymmetric memory

From: Hugo Hu [Why] 1. Driver use umachannelnumber to calculate watermarks for stutter. In asymmetric memory config, the actual bandwidth is less than dual-channel. The bandwidth should be the same as single-channel. 2. We found single rank dimm need additional delay time for stutter. [How] Get

[PATCH 08/11] drm/amd/display: [FW Promotion] Release 0.0.54

From: Anthony Koo Signed-off-by: Anthony Koo Reviewed-by: Anthony Koo Acked-by: Rodrigo Siqueira --- drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/d

[PATCH 09/11] drm/amd/display: 3.2.125

From: Aric Cyr DC version 3.2.125 brings improvements in multiple areas. In summary, we highlight: - DSC fixes - Enable ASSR - Firmware relase 0.0.54 - eDP interface refactor for multiple eDP Signed-off-by: Aric Cyr Acked-by: Rodrigo Siqueira --- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1

[PATCH 07/11] drm/amd/display: Implement dmub trace event

From: Yongqiang Sun [Why & How] DMUB FW send trace event via outbox0 interrupt. Driver will handle it. Signed-off-by: Yongqiang Sun Reviewed-by: Tony Cheng Acked-by: Rodrigo Siqueira --- .../amd/display/amdgpu_dm/amdgpu_dm_helpers.c | 6 +++ drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c |

[PATCH 06/11] drm/amd/display: Return invalid state if GPINT times out

From: Wyatt Wood [Why] GPINT timeout is causing PSR_STATE_0 to be returned when it shouldn't. We must guarantee that PSR is fully disabled before doing hw programming on driver-side. [How] Return invalid state if GPINT command times out. Let existing retry logic send the GPINT until successful.

[PATCH 05/11] drm/amd/display: Refactored DC interfaces to support multiple eDP

From: Jake Wang [Why & How] Some existing DC interfaces are optimized to return a single eDP link/stream. Refactored those DC interfaces to support multiple eDP. Signed-off-by: Jake Wang Reviewed-by: Nicholas Kazlauskas Acked-by: Rodrigo Siqueira --- .../gpu/drm/amd/display/dc/clk_mgr/clk_mg

[PATCH 04/11] drm/amd/display: Add max bpc debugfs

From: Eryk Brol [Why] Useful for testing when setting a max bpc value higher than the default is required [How] Allow for reading/writing of the max_requested_bpc property of the connector Signed-off-by: Eryk Brol Signed-off-by: Ilya Bakoulin Reviewed-by: Mikita Lipski Acked-by: Rodrigo Siqu

[PATCH 01/11] drm/amd/display: Support for DMUB AUX

From: Meenakshikumar Somasundaram [WHY] To process AUX transactions with DMUB using inbox1 and outbox1 mail boxes. [HOW] 1) Added inbox1 command DMUB_CMD__DP_AUX_ACCESS to issue AUX commands to DMUB in dc_process_dmub_aux_transfer_async(). DMUB processes AUX cmd with DCN and sends reply ba

[PATCH 03/11] drm/amd/display: Add flag for building infopacket

From: "Max.Tseng" [why] Add flag to build infopacket in SDP v1.3 format Signed-off-by: Max.Tseng Reviewed-by: Anthony Koo Acked-by: Rodrigo Siqueira --- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 4 ++- .../amd/display/modules/freesync/freesync.c | 28 +-- .../amd/dis

[PATCH 02/11] drm/amd/display: Update link encoder object creation

From: Jimmy Kizito [Why] Currently the creation of link encoder objects is tightly coupled to the creation of link objects. Decoupling link encoder object creation is a preliminary step in the process of allowing link encoders to be dynamically assigned to links. [How] Add "minimal" link encoder

[PATCH 00/11] DC Patches Feb 26th, 2021

DC version 3.2.125 brings improvements in multiple areas. In summary, we highlight: - DSC fixes - Firmware relase 0.0.54 - eDP interface refactor for multiple eDP Cc: Daniel Wheeler Best Regards Siqueira Anthony Koo (1): drm/amd/display: [FW Promotion] Release 0.0.54 Aric Cyr (1): drm/am

[PATCH] drm/amdgpu: Only check for S0ix if AMD_PMC is configured

The S0ix check only makes sense if the AMD PMC driver is present. We need to use the legacy S3 pathes when the PMC driver is not present. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/dr

Re: [PATCH] drm/amd/pm: bump Navi1x driver if version and related data structures V2

On Fri, Feb 26, 2021 at 1:09 AM Evan Quan wrote: > > New changes were involved for the SmuMetrics structure. > > Change-Id: Ib45443db03977ccd18618bcfdfd3574ac13d50d1 > Signed-off-by: Evan Quan Reviewed-by: Alex Deucher > --- > .../drm/amd/pm/inc/smu11_driver_if_navi10.h | 98 ++- > drivers

Re: [PATCH] drm/amd/display/dc/dce/dmub_psr: Remove unnecessary conversion to bool

On Fri, Feb 26, 2021 at 4:42 AM Jiapeng Chong wrote: > > Fix the following coccicheck warnings: > > ./drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c:273:16-21: WARNING: > conversion to bool not needed here. > > Reported-by: Abaci Robot > Signed-off-by: Jiapeng Chong Applied. Thanks! Alex > --

Re: [PATCH] drm/amd/pm/swsmu: Remove unnecessary conversion to bool

On Fri, Feb 26, 2021 at 4:17 AM Jiapeng Chong wrote: > > Fix the following coccicheck warnings: > > ./drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c:924:47-52: > WARNING: conversion to bool not needed here. > > Reported-by: Abaci Robot > Signed-off-by: Jiapeng Chong Applied. Thanks!

Re: [bug] tab crash on media playback

On 2/26/21 11:15 AM, Nirmoy wrote: Please try this[1] debug patch and let us know if you see anything interesting in dmesg. [1] https://github.com/nirmoy/linux/commit/72c7a7e16284b30905753241b1e90112188822c9 Regards, Nirmoy Dmesg output when triggering the crash: [   75.851792] amdgp

Re: [RFC] a new approach to detect which ring is the real black sheep upon TDR reported

On 2021-02-26 6:54 a.m., Liu, Monk wrote: [AMD Official Use Only - Internal Distribution Only] See in line Thanks -- Monk Liu | Cloud-GPU Core team -- *From:* Koenig, Christian *Sent:* Friday, February 26, 20

Re: [PATCH 2/3] drm/amdgpu: Make noretry the default on Aldebaran

Never mind. Those previous patches are part of the HMM work, which hasn't been merged yet. This-patch is good on its own for amd-staging-drm-next. Thanks,   Felix Am 2021-02-26 um 11:24 a.m. schrieb Felix Kuehling: > This depends on two previous patches: > > 2e464a7fb883 drm/amdgpu: Enable retry

Re: [PATCH 2/3] drm/amdgpu: Make noretry the default on Aldebaran

This depends on two previous patches: 2e464a7fb883 drm/amdgpu: Enable retry faults unconditionally on Aldebaran 628c8d3a1602 drm/amdgpu: Retry fault handling based on per-process XNACK mode Regards,   Felix Am 2021-02-25 um 1:56 p.m. schrieb Alex Deucher: > From: Felix Kuehling > > This is need

Re: [bug] tab crash on media playback

Hi Cory, On 2/26/21 3:04 PM, Cory Bolar wrote: On 2/26/21 4:59 AM, Christian König wrote: Can I send you a debugging patch which prints a few lines into dmesg when problems happen? I would be happy to apply and test a debug patch. Please try this[1] debug patch and let us know if you see a

Re: [PATCH v3] drm/scheduler: Fix hang when sched_entity released

On 2021-02-26 3:04 a.m., Christian König wrote: Am 25.02.21 um 22:37 schrieb Andrey Grodzovsky: Problem: If scheduler is already stopped by the time sched_entity is released and entity's job_queue not empty I encountred a hang in drm_sched_entity_flush. This is because drm_sched_entity_is_id

Re: [bug] tab crash on media playback

On 2/26/21 4:59 AM, Christian König wrote: Can I send you a debugging patch which prints a few lines into dmesg when problems happen? I would be happy to apply and test a debug patch. Thanks, Cory ___ amd-gfx mailing list amd-gfx@lists.freedesktop

[PATCH] drm/amd/display/dc/dce/dmub_psr: Remove unnecessary conversion to bool

Fix the following coccicheck warnings: ./drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c:273:16-21: WARNING: conversion to bool not needed here. Reported-by: Abaci Robot Signed-off-by: Jiapeng Chong --- drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c | 3 +-- 1 file changed, 1 insertion(+), 2 dele

Re: [PATCH] drm/amdgpu: enable one vf mode on sienna cichlid vf

[AMD Official Use Only - Internal Distribution Only] Acked-by: Alex Deucher From: Horace Chen Sent: Friday, February 26, 2021 1:31 AM To: amd-gfx@lists.freedesktop.org Cc: Grodzovsky, Andrey ; Quan, Evan ; Chen, Horace ; Tuikov, Luben ; Koenig, Christian ; Deu

Re: [RFC] a new approach to detect which ring is the real black sheep upon TDR reported

Yeah that is exactly the stuff which doesn't works at all. We got feedback for multiple people that this whole approach of tying the job to the tdr was not a good idea at all. What we should do instead is to have a pointer in the scheduler fence to which job it belongs. Freeing up the job when

RE: [RFC] a new approach to detect which ring is the real black sheep upon TDR reported

[AMD Official Use Only - Internal Distribution Only] static void drm_sched_job_timedout(struct work_struct *work) 279 { 280 struct drm_gpu_scheduler *sched; 281 struct drm_sched_job *job; 282 283 sched = container_of(work, struct drm_gpu_scheduler, work_tdr.work); 284 285 /* Protec

RE: [RFC] a new approach to detect which ring is the real black sheep upon TDR reported

[AMD Official Use Only - Internal Distribution Only] See in line Thanks -- Monk Liu | Cloud-GPU Core team -- From: Koenig, Christian Sent: Friday, February 26, 2021 3:58 PM To: Liu, Monk ; amd-gfx@lists.freedesktop

Re: [bug] tab crash on media playback

Hi Cory, I'm the author of the patch you bisected, but that looks like a bit strange bisect result. The patch in question shouldn't have an functional change, it just restructures the code. Can I send you a debugging patch which prints a few lines into dmesg when problems happen? Thanks

[PATCH] drm/amd/pm/swsmu: Remove unnecessary conversion to bool

Fix the following coccicheck warnings: ./drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c:924:47-52: WARNING: conversion to bool not needed here. Reported-by: Abaci Robot Signed-off-by: Jiapeng Chong --- drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c | 2 +- 1 file changed, 1 i

[PATCH] radeon: ERROR: space prohibited before that ','

drm_property_create_range(rdev->ddev, 0 , "coherent", 0, 1); Signed-off-by: wangjingyu --- drivers/gpu/drm/radeon/radeon_display.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 3a6fedad

[bug] tab crash on media playback

Hello, When upgrading the kernel from 5.10 to 5.11 my Chromium browser's tab crashes when playing video files, for example, any youtube video.  Other media related programs and browsers appear unaffected and I have tried multiple versions of Chromium but this has not resolved the issue. I bi

Re: [PATCH] drm/amdgpu: remove unused variable in amdgpu_dma_buf_unmap()

Am 26.02.21 um 05:24 schrieb Kevin Wang: clean up unsued variable in amdgpu_dma_buf_unmap(). Fixes: drm/amdgpu: Remove amdgpu_device arg from free_sgt api Signed-off-by: Kevin Wang Reviewed-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c | 4 1 file changed, 4

Re: [PATCH v3] drm/scheduler: Fix hang when sched_entity released

Am 25.02.21 um 22:37 schrieb Andrey Grodzovsky: Problem: If scheduler is already stopped by the time sched_entity is released and entity's job_queue not empty I encountred a hang in drm_sched_entity_flush. This is because drm_sched_entity_is_idle never becomes false. Fix: In drm_sched_fini det

Re: [PATCH v2] drm/scheduler: Fix hang when sched_entity released

Am 25.02.21 um 22:27 schrieb Andrey Grodzovsky: On 2021-02-25 1:42 p.m., Christian König wrote: Am 25.02.21 um 17:03 schrieb Andrey Grodzovsky: On 2021-02-25 2:53 a.m., Christian König wrote: Am 24.02.21 um 16:13 schrieb Andrey Grodzovsky: Ping Sorry, I've been on vacation this week.