As the only entry point, it's now safe and reasonable to
enforce the lock protections in amdgpu_dpm.c. And with
this, we can drop other internal used power locks.
Signed-off-by: Evan Quan
Change-Id: Iad228cad0b3d8c41927def08965a52525f3f51d3
---
drivers/gpu/drm/amd/pm/amdgpu_dpm.c| 716
Avoid cross callings which make lock protection enforcement
on amdgpu_dpm_force_performance_level() impossible.
Signed-off-by: Evan Quan
Change-Id: Ie658140f40ab906ce2ec47576a086062b61076a6
---
drivers/gpu/drm/amd/pm/amdgpu_pm.c| 29 ---
Instead of centralizing all headers in the same folder. Separate them into
different folders and place them among those source files those who really
need them.
Signed-off-by: Evan Quan
Change-Id: Id74cb4c7006327ca7ecd22daf17321e417c4aa71
---
drivers/gpu/drm/amd/pm/Makefile | 10
Those gfxoff controls added for some specific ASICs are unnecessary.
The functionalities are not affected without them. Also to align with
other ASICs, they should also be dropped.
Signed-off-by: Evan Quan
Change-Id: Ia8475ef9e97635441aca5e0a7693e2a515498523
---
This can cover the power implementation details. And as what did for
powerplay framework, we hook the smu_context to adev->powerplay.pp_handle.
Signed-off-by: Evan Quan
Change-Id: I3969c9f62a8b63dc6e4321a488d8f15022ffeb3d
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 6 --
Instead of putting them in amdgpu_dpm.c.
Signed-off-by: Evan Quan
Change-Id: Ieb7ed5fb6140401a7692b401c5a42dc53da92af8
---
drivers/gpu/drm/amd/pm/amdgpu_dpm.c| 14 --
drivers/gpu/drm/amd/pm/inc/hwmgr.h | 3 ---
Drop those unused APIs and data structures.
Signed-off-by: Evan Quan
Change-Id: I57d2a03dcda02d0b5d9c5ffbdd37bffe49945407
---
drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h | 49 -
drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h | 4 ++
2 files changed, 4 insertions(+), 49
Drop cross callings and multi-function APIs. Also avoid exposing
internal implementations details.
Signed-off-by: Evan Quan
Change-Id: I55e5ab3da6a70482f5f5d8c256eed2f754feae20
---
.../gpu/drm/amd/include/kgd_pp_interface.h| 2 +-
drivers/gpu/drm/amd/pm/Makefile | 2 +-
We should avoid having multi-function APIs. It should be up to the caller
to determine when or whether to call amdgpu_dpm_dispatch_task().
Signed-off-by: Evan Quan
Change-Id: I78ec4eb8ceb6e526a4734113d213d15a5fbaa8a4
---
drivers/gpu/drm/amd/pm/amdgpu_dpm.c | 18 ++
As it lables an internal pm state and amdgpu_pm structure is the more
proper place than amdgpu_device structure for it.
Signed-off-by: Evan Quan
Change-Id: I7890e8fe7af2ecd8591d30442340deb8773bacc3
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 -
drivers/gpu/drm/amd/pm/amdgpu_pm.c | 6
Those APIs are used only by legacy ASICs(si/kv). They cannot be
shared by other ASICs. So, we create a new holder for them.
Signed-off-by: Evan Quan
Change-Id: I555dfa37e783a267b1d3b3a7db5c87fcc3f1556f
--
v1->v2:
- move other APIs used by si/kv in amdgpu_atombios.c to the new
holder
Move them to si_dpm.c instead.
Signed-off-by: Evan Quan
Change-Id: I288205cfd7c6ba09cfb22626ff70360d61ff0c67
--
v1->v2:
- rename the API with "si_" prefix(Alex)
---
drivers/gpu/drm/amd/pm/amdgpu_dpm.c | 25 ---
drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h | 25 ---
Move it to kv_dpm.c instead.
Signed-off-by: Evan Quan
Change-Id: I554332b386491a79b7913f72786f1e2cb1f8165b
--
v1->v2:
- rename the API with "kv_" prefix(Alex)
---
drivers/gpu/drm/amd/pm/amdgpu_dpm.c | 23 -
drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h | 2 --
Move them to amdgpu_dpm.c instead.
Signed-off-by: Evan Quan
Change-Id: I59fe0efcb47c18ec7254f3624db7a2eb78d91b8c
---
drivers/gpu/drm/amd/pm/amdgpu_dpm.c | 25 +++--
drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h | 23 ---
2 files changed, 23 insertions(+),
Display is another client of our power APIs. It's not proper to spike
into power implementation details there.
Signed-off-by: Evan Quan
Change-Id: Ic897131e16473ed29d3d7586d822a55c64e6574a
---
.../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 6 +-
.../amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c
amdgpu_pm.c holds all the user sysfs/hwmon interfaces. It's another
client of our power APIs. It's not proper to spike into power
implementation details there.
Signed-off-by: Evan Quan
Change-Id: I397853ddb13eacfce841366de2a623535422df9a
---
drivers/gpu/drm/amd/pm/amdgpu_dpm.c | 458
Those implementation details(whether swsmu supported, some ppt_funcs supported,
accessing internal statistics ...)should be kept internally. It's not a good
practice and even error prone to expose implementation details.
Signed-off-by: Evan Quan
Change-Id:
There are several problems with current power implementations:
1. Too many internal details are exposed to other blocks. Thus to interact with
power, they need to know which power framework is used(powerplay vs swsmu)
or even whether some API is implemented.
2. A lot of cross callings exist
Am 2021-11-29 um 9:40 p.m. schrieb shaoyunl:
> This change revert previous commit
> 7079e7d5c6bf: drm/amd/amdgpu: fix the kfd pre_reset sequence in sriov
> cd547b93c62a: drm/amdgpu: move kfd post_reset out of reset_sriov function
It looks like this is not a straight revert. It moves the
To maintain system error state when SMU errors occurred,
which will aid in debugging SMU firmware issues,
add SMU debug option support.
It can be enabled or disabled via amdgpu_smu_debug
debugfs file. When enabled, it makes SMU errors fatal.
It is disabled by default.
== Command Guide ==
1,
This change revert previous commit
7079e7d5c6bf: drm/amd/amdgpu: fix the kfd pre_reset sequence in sriov
cd547b93c62a: drm/amdgpu: move kfd post_reset out of reset_sriov function
Some register access(GRBM_GFX_CNTL) only be allowed on full access
mode. Move kfd_pre_reset and kfd_post_reset back
Am 2021-11-29 um 4:23 p.m. schrieb Philip Yang:
> Refactor userptr and pin_bo path to make it less confusing, move
> err_pin_bo label up to remove mem from process_info kfd_bo_list.
>
> Signed-off-by: Philip Yang
The series is
Reviewed-by: Felix Kuehling
> ---
>
Am 2021-11-29 um 9:59 a.m. schrieb Graham Sider:
> Initializes kfd->device_info given either asic_type (enum) if GFX
> version is less than GFX9, or GC IP version if greater. Also takes in vf
> and the target compiler gfx version. Uses SDMA version to determine
> num_sdma_queues_per_engine.
>
>
Refactor userptr and pin_bo path to make it less confusing, move
err_pin_bo label up to remove mem from process_info kfd_bo_list.
Signed-off-by: Philip Yang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 14 ++
1 file changed, 6 insertions(+), 8 deletions(-)
diff --git
process_info->lock is used to protect kfd_bo_list, vm_list_head, n_vms
and userptr valid/inval list, svm_range_restore_work and
svm_range_set_attr don't access those, so do not need to take
process_info lock. This will avoid potential circular locking issue.
Signed-off-by: Philip Yang
---
Hi Rodrigo,
That will really be helpful!
I know drawing the line is a difficult problem (and can even make things
harder when searching), but maybe it would make sense to keep generic
acronyms not specific to amdgpu in a separate list. I bet a number of
them would be useful in the scope of
amdgpu_ucode_get_load_type() does not interpret this parameter as
documented. It is ignored for many ASIC types (which presumably
only support one load_type), and when not ignored it is only used
to force direct loading instead of PSP loading. SMU loading is
only available for ASICs for which
On Thu, Nov 25, 2021 at 10:40 AM Rodrigo Siqueira
wrote:
>
> In the DC driver, we have multiple acronyms that are not obvious most of
> the time. This commit introduces a DC glossary in order to make it
> easier to navigate through our driver.
>
> Signed-off-by: Rodrigo Siqueira
> ---
>
Am 2021-11-29 um 2:14 p.m. schrieb Philip Yang:
> drm_gem_object_put calls release_notify callback to free the mem
> structure and unreserve_mem_limit, move it down after the last access
> of mem and make it conditional call.
>
> Signed-off-by: Philip Yang
Reviewed-by: Felix Kuehling
> ---
>
drm_gem_object_put calls release_notify callback to free the mem
structure and unreserve_mem_limit, move it down after the last access
of mem and make it conditional call.
Signed-off-by: Philip Yang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 8 +---
1 file changed, 5
Am 2021-11-29 um 9:05 a.m. schrieb Philip Yang:
> To silence the following Smatch static checker warning:
>
> drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_svm.c:2615
> svm_range_restore_pages()
> warn: missing error code here? 'get_task_mm()' failed. 'r' = '0'
>
> Signed-off-by: Philip Yang
>
Am 2021-11-26 um 6:58 p.m. schrieb Philip Yang:
> drm_gem_object_put calls release_notify callback to free the mem
> structure and unreserve_mem_limit, move it down after the last access
> of mem and make it conditional call.
>
> Signed-off-by: Philip Yang
> ---
>
On Sun, Nov 28, 2021 at 11:31 AM Yann Dirson wrote:
>
> amdgpu_ucode_get_load_type() does not interpret this parameter as
> documented. It is ignored for many ASIC types (which presumably
> only support one load_type), and when not ignored it is only used
> to force direct loading instead of PSP
[Public]
Yes, that makes more sense.
Alex
From: Chen, Guchun
Sent: Wednesday, November 24, 2021 9:21 PM
To: Chen, Guchun ; Alex Deucher ;
Jian, Jane
Cc: Deucher, Alexander ; Chen, JingWen
; amd-gfx list
Subject: RE: [PATCH] drm/amdgpu/sriov/vcn: skip ip
Initializes kfd->device_info given either asic_type (enum) if GFX
version is less than GFX9, or GC IP version if greater. Also takes in vf
and the target compiler gfx version. Uses SDMA version to determine
num_sdma_queues_per_engine.
Convert device_info to a non-pointer member of kfd, change
With device_info initialization being handled in kfd_device_info_init,
these structs may be removed. Also add comments to help matching IP
versions to asic names.
Signed-off-by: Graham Sider
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 469 +---
device_info->asic_name and amdgpu_asic_name[adev->asic_type] both
provide asic name strings, with the only difference being casing.
Remove asic_name from device_info and replace sysfs entry with lowercase
amdgpu_asic_name[]. Ensures string is null-terminated so that this
doesn't break if
Hi all,
This week this patchset was tested on the following systems:
Lenovo Thinkpad T14s Gen2 with AMD Ryzen 5 5650U, with the following display
types: eDP 1080p 60hz, 4k 60hz (via USB-C to DP/HDMI), 1440p 144hz (via USB-C
to DP/HDMI), 1680*1050 60hz (via USB-C to DP and then DP to
To silence the following Smatch static checker warning:
drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_svm.c:2615
svm_range_restore_pages()
warn: missing error code here? 'get_task_mm()' failed. 'r' = '0'
Signed-off-by: Philip Yang
Suggested-by: Dan Carpenter
---
On Fri, 26 Nov 2021, Daniel Vetter wrote:
> On Thu, Nov 25, 2021 at 10:38:25AM -0500, Rodrigo Siqueira wrote:
>> Display core documentation is not well organized, and it is hard to find
>> information due to the lack of sections. This commit reorganizes the
>> documentation layout, and it is
Hi Bokun,
please remove the change-id in your commit message when submitting this patch.
Acked-by: Jingwen Chen
On 2021/11/27 上午8:57, Bokun Zhang wrote:
> From: Bokun Zhang
>
> In the patch about advanced TDR mode, we force to always set
> amdgpu_gpu_recovery=2 under SRIOV. This is not
[AMD Official Use Only]
Good catch.
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: Lazar, Lijo
Sent: Monday, November 29, 2021 16:12
To: amd-gfx@lists.freedesktop.org
Cc: Zhang, Hawking ; Deucher, Alexander
; Yang, Stanley ; Clements,
John ; Zhou1, Tao
On aldebaran, RLC also controls GFXCLK. Skip halting RLC during GFX IP suspend
and keep it running till PMFW disables all DPMs.
[ 578.019986] amdgpu :23:00.0: amdgpu: GPU reset begin!
[ 583.245566] amdgpu :23:00.0: amdgpu: Failed to disable smu features.
[ 583.245621]
Hi all,
On 25/11/2021 11:09, Thorsten Leemhuis wrote:
Hi, this is your Linux kernel regression tracker speaking.
On 24.11.21 20:14, Mark Boddington wrote:
Hi all,
TL;DR - git bisection points to
'kfd->gtt_sa_bitmap' is a bitmap. So use 'bitmap_zalloc()' to simplify
code, improve the semantic and avoid some open-coded arithmetic in
allocator arguments.
Also change the corresponding 'kfree()' into 'bitmap_free()' to keep
consistency.
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 12
All uses of the 'kfd->gtt_sa_bitmap' bitmap are protected with the
'kfd->gtt_sa_lock' mutex.
So:
- prefer the non-atomic '__set_bit()' function
- use the non-atomic 'bitmap_[set|clear]()' functions instead of
equivalent 'for' loops. These functions can work on several bits at a
46 matches
Mail list logo
