On 11/18/2021 5:52 AM, Felix Kuehling wrote:
On 2021-11-10 11:34 a.m., Felix Kuehling wrote:
Am 2021-11-10 um 11:11 a.m. schrieb Lazar, Lijo:
[Public]
(... && !amdgpu_sriov_vf(adev))
This kind of closes the door for all versions. My thought was - having
it in the same function provides a
Do ASIC reset at the moment Sx suspend aborted behind of amdgpu suspend
to keep AMDGPU in a clean reset state and that can avoid re-initialize
device improperly error.
Signed-off-by: Prike Liang
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c |
Am 2021-11-18 um 12:39 a.m. schrieb Amber Lin:
> Instead of hard coding the number of sdma engines and the number of
> sdma_xgmi engines in the device_info table, get the number of toal SDMA
> instances from amdgpu. The first two engines are sdma engines and the
> rest are sdma-xgmi engines unless
Instead of hard coding the number of sdma engines and the number of
sdma_xgmi engines in the device_info table, get the number of toal SDMA
instances from amdgpu. The first two engines are sdma engines and the
rest are sdma-xgmi engines unless the ASIC doesn't support XGMI.
v2: add kfd_ prefix to
On 11/18/2021 9:56 AM, Yang, Stanley wrote:
[AMD Official Use Only]
-邮件原件-
发件人: Lazar, Lijo
发送时间: Thursday, November 18, 2021 12:04 PM
收件人: Yang, Stanley ; amd-
g...@lists.freedesktop.org; Zhang, Hawking ;
Clements, John ; Quan, Evan
; Wang, Yang(Kevin)
主题: Re: 回复: [PATCH Review
[AMD Official Use Only]
> -邮件原件-
> 发件人: Lazar, Lijo
> 发送时间: Thursday, November 18, 2021 12:04 PM
> 收件人: Yang, Stanley ; amd-
> g...@lists.freedesktop.org; Zhang, Hawking ;
> Clements, John ; Quan, Evan
> ; Wang, Yang(Kevin)
> 主题: Re: 回复: [PATCH Review 3/4] drm/amdgpu: add message smu t
Hi Dave, Daniel,
Fixes for 5.16.
The following changes since commit fa55b7dcdc43c1aa1ba12bca9d2dd4318c2a0dbf:
Linux 5.16-rc1 (2021-11-14 13:56:52 -0800)
are available in the Git repository at:
https://gitlab.freedesktop.org/agd5f/linux.git
tags/amd-drm-fixes-5.16-2021-11-17
for you to fe
On 11/18/2021 9:29 AM, Yang, Stanley wrote:
[AMD Official Use Only]
-邮件原件-
发件人: Lazar, Lijo
发送时间: Wednesday, November 17, 2021 7:15 PM
收件人: Yang, Stanley ; amd-
g...@lists.freedesktop.org; Zhang, Hawking ;
Clements, John ; Quan, Evan
; Wang, Yang(Kevin)
主题: Re: [PATCH Review 4/4]
On 11/18/2021 9:07 AM, Yang, Stanley wrote:
[AMD Official Use Only]
-邮件原件-
发件人: Lazar, Lijo
发送时间: Wednesday, November 17, 2021 7:24 PM
收件人: Yang, Stanley ; amd-
g...@lists.freedesktop.org; Zhang, Hawking ;
Clements, John ; Quan, Evan
; Wang, Yang(Kevin)
主题: Re: [PATCH Review 3/4]
[AMD Official Use Only]
> -邮件原件-
> 发件人: Lazar, Lijo
> 发送时间: Wednesday, November 17, 2021 7:15 PM
> 收件人: Yang, Stanley ; amd-
> g...@lists.freedesktop.org; Zhang, Hawking ;
> Clements, John ; Quan, Evan
> ; Wang, Yang(Kevin)
> 主题: Re: [PATCH Review 4/4] query umc error info from ecc_tab
Use proper amdgpu_gem_prime_import function to handle all kinds of
imports. Remember the dmabuf reference to enable proper multi-GPU
attachment to multiple VMs without erroneously re-exporting the
underlying BO multiple times.
Signed-off-by: Felix Kuehling
---
.../gpu/drm/amd/amdgpu/amdgpu_amdkf
Exports a DMA buf fd of a given KFD buffer handle. This is intended for
the new upstreamable RDMA solution coming to UCX and libfabric.
The corresponding user mode change (Thunk API and kfdtest) is here:
https://github.com/RadeonOpenCompute/ROCT-Thunk-Interface/commits/fxkamd/dmabuf
Signed-off-by
[AMD Official Use Only]
> -邮件原件-
> 发件人: Lazar, Lijo
> 发送时间: Wednesday, November 17, 2021 7:24 PM
> 收件人: Yang, Stanley ; amd-
> g...@lists.freedesktop.org; Zhang, Hawking ;
> Clements, John ; Quan, Evan
> ; Wang, Yang(Kevin)
> 主题: Re: [PATCH Review 3/4] drm/amdgpu: add message smu to ge
[AMD Official Use Only]
> -邮件原件-
> 发件人: Lazar, Lijo
> 发送时间: Wednesday, November 17, 2021 7:36 PM
> 收件人: Yang, Stanley ; amd-
> g...@lists.freedesktop.org; Zhang, Hawking ;
> Clements, John ; Quan, Evan
> ; Wang, Yang(Kevin)
> 主题: Re: [PATCH Review 2/4] drm/amdgpu: add new query interfa
On 2021-11-10 11:34 a.m., Felix Kuehling wrote:
Am 2021-11-10 um 11:11 a.m. schrieb Lazar, Lijo:
[Public]
(... && !amdgpu_sriov_vf(adev))
This kind of closes the door for all versions. My thought was - having
it in the same function provides a logical grouping for how it's
handled for differ
On 2021-11-16 10:43 p.m., Philip Yang wrote:
unmap range always set svms->drain_pagefaults flag to simplify both
parent range and child range unmap. Deferred list work takes mmap write
lock to read and clear svms->drain_pagefaults, to serialize with unmap
callback.
Add atomic flag svms->draini
On 2021-11-16 10:43 p.m., Philip Yang wrote:
VMA may be removed before unmap notifier callback, restore pages take
mmap write lock to lookup VMA to avoid race,
The old code looked up the VMA after taking the mmap lock (either read
or write) and kept holding the lock afterwards. I think even wi
On 2021-11-16 10:43 p.m., Philip Yang wrote:
kfd process mmu release notifier callback drain retry fault to ensure no
retry fault comes after removing kfd process from the hash table,
otherwise svm page fault handler will fail to recover the fault and dump
GPU vm fault log.
Drain retry fault nee
On 2021-11-17 3:36 p.m., Amber Lin wrote:
Instead of hard coding the number of sdma engines and the number of
sdma_xgmi engines in the device_info table, get the number of toal SDMA
instances from amdgpu. The first two engines are sdma engines and the
rest are sdma-xgmi engines unless the ASIC do
Acked-by: Alex Deucher
On Wed, Nov 17, 2021 at 1:56 PM Luben Tuikov wrote:
>
> Print the error on command submission immediately after submitting to
> the SMU. This is rate-limited. It helps to immediately know there was an
> error on command submission, rather than leave it up to clients to rep
On Wed, Nov 17, 2021 at 1:56 PM Luben Tuikov wrote:
>
> Add prints where there are none and none are printed in the callee.
>
> Add a print in sienna_cichlid_run_btc() to help debug and to mirror other
> platforms, as no print is present in the caller, smu_smc_hw_setup().
>
> Remove the word "prev
Applied. Thanks!
On Mon, Nov 15, 2021 at 3:10 AM Yang Li wrote:
>
> Clang static analysis reports this error
>
> drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:2870:7: warning:
> Dereference of null pointer [clang-analyzer-core.NullDereference]
> if
> (top_pipe_to_program->st
Applied. Thanks!
On Mon, Nov 15, 2021 at 8:48 PM Bernard Zhao wrote:
>
> This change is to cleanup the code a bit.
>
> Signed-off-by: Bernard Zhao
> ---
> .../drm/amd/display/dc/dcn10/dcn10_resource.c | 18 ++
> 1 file changed, 6 insertions(+), 12 deletions(-)
>
> diff --git a
Applied. Thanks!
On Tue, Nov 16, 2021 at 4:19 AM Christian König
wrote:
>
> Am 16.11.21 um 02:34 schrieb Bernard Zhao:
> > In function dc_sink_destruct, kfree will check pointer, no need
> > to check again.
> > This change is to cleanup the code a bit.
> >
> > Signed-off-by: Bernard Zhao
>
> Th
Applied. Thanks!
Alex
On Mon, Nov 15, 2021 at 10:56 AM Felix Kuehling wrote:
>
> Am 2021-11-14 um 9:58 p.m. schrieb Bernard Zhao:
> > In function amdgpu_get_xgmi_hive, when kobject_init_and_add failed
> > There is a potential memleak if not call kobject_put.
> >
> > Signed-off-by: Bernard Zhao
Applied. Thanks!
Alex
On Mon, Nov 15, 2021 at 7:09 AM Bernard Zhao wrote:
>
> This change is to cleanup the code style a bit.
>
> Signed-off-by: Bernard Zhao
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 21 +
> 1 file changed, 13 insertions(+), 8 deletions(-)
>
> dif
Applied thanks. If you want to make the numbering more sequential,
please also update the other dce files if you make that change.
Alex
On Mon, Nov 15, 2021 at 2:14 AM Bernard Zhao wrote:
>
> This change is to remove useless break after return.
>
> Signed-off-by: Bernard Zhao
> ---
> drivers/
Hi Arunpravin,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on drm/drm-next]
[also build test ERROR on drm-intel/for-linux-next v5.16-rc1]
[cannot apply to drm-tip/drm-tip next-2027]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when
Instead of hard coding the number of sdma engines and the number of
sdma_xgmi engines in the device_info table, get the number of toal SDMA
instances from amdgpu. The first two engines are sdma engines and the
rest are sdma-xgmi engines unless the ASIC doesn't support XGMI.
v2: Move get_num_*_sdma
On Wed, Nov 17, 2021 at 2:12 AM Lazar, Lijo wrote:
>
>
>
> On 11/17/2021 11:50 AM, Darren Powell wrote:
> > Use of sysfs_emit by each of the specific device implementations is
> > problematic.
> > To remove this back to a higher level, this patch adds a new function
> > "get_clock_levels"
>
On 16/11/2021 20:18, Arunpravin wrote:
On contiguous allocation, we round up the size
to the *next* power of 2, implement a function
to free the unused pages after the newly allocate block.
v2(Matthew Auld):
- replace function name 'drm_buddy_free_unused_pages' with
drm_buddy_block_trim
Print the error on command submission immediately after submitting to
the SMU. This is rate-limited. It helps to immediately know there was an
error on command submission, rather than leave it up to clients to report
the error, as sometimes they do not.
Cc: Alex Deucher
Signed-off-by: Luben Tuiko
Add prints where there are none and none are printed in the callee.
Add a print in sienna_cichlid_run_btc() to help debug and to mirror other
platforms, as no print is present in the caller, smu_smc_hw_setup().
Remove the word "previous" from comment and print to make it shorter and
avoid confusi
On 16/11/2021 20:18, Arunpravin wrote:
- Make drm_buddy_alloc a single function to handle
range allocation and non-range allocation demands
- Implemented a new function alloc_range() which allocates
the requested power-of-two block comply with range limitations
- Moved order computation a
On 16/11/2021 20:18, Arunpravin wrote:
Move the base i915 buddy allocator code into drm
- Move i915_buddy.h to include/drm
- Move i915_buddy.c to drm root folder
- Rename "i915" string with "drm" string wherever applicable
- Rename "I915" string with "DRM" string wherever applicable
- Fix header
Hi Christian,
I will make this a separate module.
Thanks,
Arun
On 17/11/21 1:33 pm, Christian König wrote:
> I've looked a bit more into this and I think we should just follow
> Thomas Zimmermann's idea to make this a separate module.
>
> Otherwise we just have the code around all the time even
Hi Roman,
On 17.11.2021 15:26, Li, Roman wrote:
> [Public]
>
> Hi Samuel,
>
> Can you please try: https://patchwork.freedesktop.org/patch/463485/ ?
Yup, that did the trick. Works as before. Thank you very much.
Samuel
>
> Thanks,
> Roman
>
> > -Original Message-
> > From: Samuel Čav
On 17.11.2021 10:22, roman...@amd.com wrote:
> From: Roman Li
>
> [Why]
> After commit ("drm/amd/display: add support for multiple backlights")
> number of eDPs is defined while registering backlight device.
> However the panel's extended caps get updated once before register call.
> That leads t
On Wed, Nov 17, 2021 at 10:22 AM wrote:
>
> From: Roman Li
>
> [Why]
> After commit ("drm/amd/display: add support for multiple backlights")
> number of eDPs is defined while registering backlight device.
> However the panel's extended caps get updated once before register call.
> That leads to r
[Public]
Hi Samuel,
Can you please try: https://patchwork.freedesktop.org/patch/463485/ ?
Thanks,
Roman
> -Original Message-
> From: Samuel Čavoj
> Sent: Tuesday, November 16, 2021 8:33 AM
> To: Alex Deucher
> Cc: Deucher, Alexander ; Li, Sun peng (Leo)
> ; Li, Roman ; Maling list - D
From: Roman Li
[Why]
After commit ("drm/amd/display: add support for multiple backlights")
number of eDPs is defined while registering backlight device.
However the panel's extended caps get updated once before register call.
That leads to regression with extended caps like oled brightness contro
On Mon, Nov 15, 2021 at 01:30:18PM -0600, Alex Sierra wrote:
> @@ -5695,8 +5695,8 @@ static int mem_cgroup_move_account(struct page *page,
> * 2(MC_TARGET_SWAP): if the swap entry corresponding to this pte is a
> * target for charge migration. if @target is not NULL, the entry is
> store
On 11/17/2021 3:41 PM, Stanley.Yang wrote:
add message smu to query error information
Signed-off-by: Stanley.Yang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 16 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h | 4 +
drivers/gpu/drm/amd/amdgpu/umc_v6_7.c | 161
On 11/17/2021 3:41 PM, Stanley.Yang wrote:
support ECC TABLE message, this table include unc ras error count
and error address
Signed-off-by: Stanley.Yang
---
drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h | 7
.../drm/amd/pm/swsmu/smu13/aldebaran_ppt.c| 38 +++
On 11/17/2021 3:41 PM, Stanley.Yang wrote:
if smu support ECCTABLE, driver can message smu to get ecc_table
then query umc error info from ECCTABLE
apply pmfw version check to ensure backward compatibility
Signed-off-by: Stanley.Yang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 42 +
if smu support ECCTABLE, driver can message smu to get ecc_table
then query umc error info from ECCTABLE
apply pmfw version check to ensure backward compatibility
Signed-off-by: Stanley.Yang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 42 ---
drivers/gpu/drm/amd/amdgpu/amdgpu_ras
add message smu to query error information
Signed-off-by: Stanley.Yang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 16 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h | 4 +
drivers/gpu/drm/amd/amdgpu/umc_v6_7.c | 161
3 files changed, 181 insertions(+)
diff --git a/
support ECC TABLE message, this table include unc ras error count
and error address
Signed-off-by: Stanley.Yang
---
drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h | 7
.../drm/amd/pm/swsmu/smu13/aldebaran_ppt.c| 38 +++
.../gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c| 2
update smu driver if version to 0x08 to avoid mismatch log
A version mismatch can still happen with an older FW
Change-Id: I97f2bc4ed9a9cba313b744e2ff6812c90b244935
Signed-off-by: Stanley.Yang
---
.../drm/amd/pm/inc/smu13_driver_if_aldebaran.h | 18 +-
drivers/gpu/drm/amd/pm/inc/
I've looked a bit more into this and I think we should just follow
Thomas Zimmermann's idea to make this a separate module.
Otherwise we just have the code around all the time even if it is unused
and implementing this should be trivial.
See how DRM_GEM_CMA_HELPER or DRM_VRAM_HELPER are done
50 matches
Mail list logo
