On 7/17/2021 2:54 PM, Sierra Guiza, Alejandro (Alex) wrote:
On 7/16/2021 5:14 PM, Felix Kuehling wrote:
Am 2021-07-16 um 11:07 a.m. schrieb Theodore Y. Ts'o:
On Wed, Jun 23, 2021 at 05:49:55PM -0400, Felix Kuehling wrote:
I can think of two ways to test the changes for
MEMORY_DEVICE_GENERIC
On 7/21/21 9:20 PM, Evan Quan wrote:
The customized OD settings can be divided into two parts: those
committed ones and non-committed ones.
- For those changes which had been fed to SMU before S3/S4/Runpm
suspend kicked, they are committed changes. They should be properly
restored
[AMD Official Use Only]
Thanks Alex! I will make these changes.
Best,
Ryan
From: Alex Deucher
Sent: Friday, July 23, 2021 7:33 AM
To: Taylor, Ryan
Cc: Maling list - DRI developers ; amd-gfx
list ; kernel test robot ;
Daniel Vetter ; Siqueira, Rodrigo
;
Good one, that makes better use of flags.
Thanks,
Lijo
From: Chen, Guchun
Sent: Friday, July 23, 2021 7:14:58 PM
To: Lazar, Lijo ; Quan, Evan ;
amd-gfx@lists.freedesktop.org
Cc: Deucher, Alexander
Subject: RE: [PATCH V2 1/2] drm/amd/pm: restore user customized
[Public]
Just curious that in the patch, it adds a variable *user_od* serving the check
of applying user customized OD setting. Instead of this, does it make sense to
use the *flag* in struct smu_user_dpm_profile for this? As we have below bit in
pm/inc/amdgpu_smu.h, can we add another bit for
[AMD Official Use Only]
I haven't had a chance to look at the patches too closely, but if it could be
done in a generic may, that makes sense to me. Maybe as a follow up patch?
Alex
From: Lazar, Lijo
Sent: Friday, July 23, 2021 6:09 AM
To: Quan, Evan ;
On 7/23/2021 6:18 PM, Pratik Vishwakarma wrote:
[Why]
User might change the suspend behaviour from OS.
[How]
Check with pm for target suspend state and set s0ix
flag only for s2idle state.
v2: User might change default suspend state, use target state
Suggested-by: Lijo Lazar
Signed-off-by:
[Why]
User might change the suspend behaviour from OS.
[How]
Check with pm for target suspend state and set s0ix
flag only for s2idle state.
v2: User might change default suspend state, use target state
Suggested-by: Lijo Lazar
Signed-off-by: Pratik Vishwakarma
---
On Wed, Jul 21, 2021 at 1:07 PM Ryan Taylor wrote:
>
> Move dce_virtual into amdgpu_vkms and update all references to
> dce_virtual with amdgpu_vkms.
>
> v2: Removed more references to dce_virtual.
>
> v3: Restored display modes from previous implementation.
>
> Reported-by: kernel test robot
>
On Wed, Jul 21, 2021 at 1:07 PM Ryan Taylor wrote:
>
> Modify the VKMS driver into an api that dce_virtual can use to create
> virtual displays that obey drm's atomic modesetting api.
>
> v2: Made local functions static.
>
> Reported-by: kernel test robot
> Signed-off-by: Ryan Taylor
> ---
>
The series looks good to me, though I prefer to use a common logic to
restore od settings so that smuv12,smuv13 gets the restore feature by
default once they add the user table logic. Don't have strong argument
for it unless Alex, Kenneth or others have some comments.
Anyway, the series is
On Fri, Jul 23, 2021 at 9:08 AM Lazar, Lijo wrote:
>
>
>
> On 7/23/2021 6:18 PM, Pratik Vishwakarma wrote:
> > [Why]
> > User might change the suspend behaviour from OS.
> >
> > [How]
> > Check with pm for target suspend state and set s0ix
> > flag only for s2idle state.
> >
> > v2: User might
On Fri Jul 23, 2021 at 10:45:49AM +0200, Christian König wrote:
> Am 23.07.21 um 09:07 schrieb Jingwen Chen:
> > [SNIP]
> > Hi Christian,
> >
> > The thing is vm flush fence has no job passed to amdgpu_fence_emit, so
> > go through the jobs cannot help find the vm flush fence. And keep the
> >
[Public]
I probably get your point. Please check the update V2.
BR
Evan
> -Original Message-
> From: Lazar, Lijo
> Sent: Friday, July 23, 2021 3:39 PM
> To: Quan, Evan ; amd-gfx@lists.freedesktop.org
> Cc: Deucher, Alexander
> Subject: RE: [PATCH 1/2] drm/amd/pm: restore user
Properly restore those committed and non-committed user customized OD
settings.
Change-Id: I25396df0b3ecdd7a0d9fc77ed220b0abf1fde020
Signed-off-by: Evan Quan
---
.../amd/pm/swsmu/smu11/sienna_cichlid_ppt.c | 37 ++-
1 file changed, 28 insertions(+), 9 deletions(-)
diff --git
The customized OD settings can be divided into two parts: those
committed ones and non-committed ones.
- For those changes which had been fed to SMU before S3/S4/Runpm
suspend kicked, they are committed changes. They should be properly
restored and fed to SMU on S3/S4/Runpm resume.
-
[AMD Official Use Only]
Hi Christian
How about this way:
#define AMDGPU_DMA_FENCE_FLAG_INSIDE_JOB DMA_FENCE_FLAG_USER_BITS
And we can use AMDGPU_DMA_FENCE_FLAG_INSIDE_JOB instead of
DMA_FENCE_FLAG_USER_BITS everywhere
Thanks
--
Monk Liu |
[AMD Official Use Only]
Hi jingwen
In the lines like below:
+ if (job != NULL) {
+ /* mark this fence has a parent job */
+ set_bit(DMA_FENCE_FLAG_USER_BITS, >flags);
+ }
The DMA_FENCE_FLAG_USER_BITS usage looks not good enough
Please try to improve it,
[Public]
Hi Evan,
In that case, using "od_edit" table for the intermediate table
During commit command, what if it's done like
1) Copy and update table if od_edit table is different than user_od
table
2) Set the flag to false if od_edit table and boot_od table are
different
[AMD Official Use Only]
Reviewed-by: Likun Gao
Regards,
Likun
-Original Message-
From: Hawking Zhang
Sent: Friday, July 23, 2021 2:11 PM
To: amd-gfx@lists.freedesktop.org; Gao, Likun ; Deucher,
Alexander
Cc: Zhang, Hawking
Subject: [PATCH] drm/amdgpu: retire sdma v5_2 golden
[AMD Official Use Only]
Hi Lijo,
Sorry, I doubled checked. The implementation of Navi1x is right.
The original design for "restores to default settings" is a two-step process.
That means for "case PP_OD_COMMIT_DPM_TABLE:" we must distinguish whether it's
an usual customized od setting update
On Fri Jul 23, 2021 at 08:33:02AM +0200, Christian König wrote:
> Am 22.07.21 um 18:47 schrieb Jingwen Chen:
> > On Thu Jul 22, 2021 at 06:24:28PM +0200, Christian König wrote:
> > > Am 22.07.21 um 16:45 schrieb Andrey Grodzovsky:
> > > > On 2021-07-22 6:45 a.m., Jingwen Chen wrote:
> > > > > On
[AMD Official Use Only]
> -Original Message-
> From: Lazar, Lijo
> Sent: Thursday, July 22, 2021 5:03 PM
> To: Quan, Evan ; amd-gfx@lists.freedesktop.org
> Cc: Deucher, Alexander
> Subject: Re: [PATCH 1/2] drm/amd/pm: restore user customized OD settings
> properly for NV1x
>
>
>
>
They are initalized by hardware during power up phase,
starting from sdma v5_2 generation
Signed-off-by: Hawking Zhang
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c | 17 -
1 file changed, 17 deletions(-)
diff --git
Am 23.07.21 um 11:25 schrieb Jingwen Chen:
On Fri Jul 23, 2021 at 10:45:49AM +0200, Christian König wrote:
Am 23.07.21 um 09:07 schrieb Jingwen Chen:
[SNIP]
Hi Christian,
The thing is vm flush fence has no job passed to amdgpu_fence_emit, so
go through the jobs cannot help find the vm flush
Am 23.07.21 um 11:00 schrieb Daniel Vetter:
On Fri, Jul 23, 2021 at 10:33:48AM +0200, Christian König wrote:
Am 23.07.21 um 10:21 schrieb Daniel Vetter:
On Wed, Jul 14, 2021 at 10:51 AM Christian König
wrote:
Am 13.07.21 um 17:28 schrieb Alex Deucher:
On Tue, Jul 13, 2021 at 2:57 AM
On Fri, Jul 23, 2021 at 10:33:48AM +0200, Christian König wrote:
>
>
> Am 23.07.21 um 10:21 schrieb Daniel Vetter:
> > On Wed, Jul 14, 2021 at 10:51 AM Christian König
> > wrote:
> > >
> > >
> > > Am 13.07.21 um 17:28 schrieb Alex Deucher:
> > > > On Tue, Jul 13, 2021 at 2:57 AM Christian
Am 23.07.21 um 09:07 schrieb Jingwen Chen:
[SNIP]
Hi Christian,
The thing is vm flush fence has no job passed to amdgpu_fence_emit, so
go through the jobs cannot help find the vm flush fence. And keep the
rest fences in the RCU array will lead to signaling them in the ib_test
right after ASIC
[Why]
After embeded hw_fence to amdgpu_job, we need to add tdr support
for this feature.
[How]
1. Add a resubmit_flag for resubmit jobs.
2. Clear job fence from RCU and force complete vm flush fences in
pre_asic_reset
3. skip dma_fence_get for resubmit jobs and add a dma_fence_put
for
From: Jack Zhang
Why: Previously hw fence is alloced separately with job.
It caused historical lifetime issues and corner cases.
The ideal situation is to take fence to manage both job
and fence's lifetime, and simplify the design of gpu-scheduler.
How:
We propose to embed hw_fence into
[Why]
After embeded hw_fence to amdgpu_job, we need to add tdr support
for this feature.
[How]
1. Add a resubmit_flag for resubmit jobs.
2. Clear job fence from RCU and force complete vm flush fences in
pre_asic_reset
3. skip dma_fence_get for resubmit jobs and add a dma_fence_put
for
From: Jack Zhang
Why: Previously hw fence is alloced separately with job.
It caused historical lifetime issues and corner cases.
The ideal situation is to take fence to manage both job
and fence's lifetime, and simplify the design of gpu-scheduler.
How:
We propose to embed hw_fence into
On Fri Jul 23, 2021 at 12:06:32AM -0400, Andrey Grodzovsky wrote:
>
> On 2021-07-22 8:20 p.m., Jingwen Chen wrote:
> > On Thu Jul 22, 2021 at 01:50:09PM -0400, Andrey Grodzovsky wrote:
> > > On 2021-07-22 1:27 p.m., Jingwen Chen wrote:
> > > > On Thu Jul 22, 2021 at 01:17:13PM -0400, Andrey
Am 23.07.21 um 10:21 schrieb Daniel Vetter:
On Wed, Jul 14, 2021 at 10:51 AM Christian König
wrote:
Am 13.07.21 um 17:28 schrieb Alex Deucher:
On Tue, Jul 13, 2021 at 2:57 AM Christian König
wrote:
Am 13.07.21 um 00:06 schrieb Felix Kuehling:
KFD Thunk maps invisible VRAM BOs with
On Wed, Jul 14, 2021 at 10:51 AM Christian König
wrote:
>
>
>
> Am 13.07.21 um 17:28 schrieb Alex Deucher:
> > On Tue, Jul 13, 2021 at 2:57 AM Christian König
> > wrote:
> >>
> >>
> >> Am 13.07.21 um 00:06 schrieb Felix Kuehling:
> >>> KFD Thunk maps invisible VRAM BOs with PROT_NONE,
On 7/22/2021 10:57 AM, Pratik Vishwakarma wrote:
[Why]
User might set mem_sleep as deep and it will result
in amdgpu resume errors.
[How]
Check with pm for default suspend state
Signed-off-by: Pratik Vishwakarma
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
1 file changed, 2
On 2021-07-22 8:20 p.m., Jingwen Chen wrote:
On Thu Jul 22, 2021 at 01:50:09PM -0400, Andrey Grodzovsky wrote:
On 2021-07-22 1:27 p.m., Jingwen Chen wrote:
On Thu Jul 22, 2021 at 01:17:13PM -0400, Andrey Grodzovsky wrote:
On 2021-07-22 12:47 p.m., Jingwen Chen wrote:
On Thu Jul 22, 2021 at
Am 22.07.21 um 18:47 schrieb Jingwen Chen:
On Thu Jul 22, 2021 at 06:24:28PM +0200, Christian König wrote:
Am 22.07.21 um 16:45 schrieb Andrey Grodzovsky:
On 2021-07-22 6:45 a.m., Jingwen Chen wrote:
On Wed Jul 21, 2021 at 12:53:51PM -0400, Andrey Grodzovsky wrote:
On 2021-07-20 11:13 p.m.,
38 matches
Mail list logo
