[PATCH] drm/amd/powerplay: a quick fix for the deadlock issue below

Alex Deucher alexdeucher at gmail.com
Wed Jan 15 19:48:35 UTC 2020


On Wed, Jan 15, 2020 at 4:21 AM Evan Quan <evan.quan at amd.com> wrote:
>
> NFO: task ocltst:2028 blocked for more than 120 seconds.
>      Tainted: G           OE     5.0.0-37-generic #40~18.04.1-Ubuntu
> echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> cltst          D    0  2028   2026 0x00000000
> all Trace:
> __schedule+0x2c0/0x870
> schedule+0x2c/0x70
> schedule_preempt_disabled+0xe/0x10
> __mutex_lock.isra.9+0x26d/0x4e0
> __mutex_lock_slowpath+0x13/0x20
> ? __mutex_lock_slowpath+0x13/0x20
> mutex_lock+0x2f/0x40
> amdgpu_dpm_set_powergating_by_smu+0x64/0xe0 [amdgpu]
> gfx_v8_0_enable_gfx_static_mg_power_gating+0x3c/0x70 [amdgpu]
> gfx_v8_0_set_powergating_state+0x66/0x260 [amdgpu]
> amdgpu_device_ip_set_powergating_state+0x62/0xb0 [amdgpu]
> pp_dpm_force_performance_level+0xe7/0x100 [amdgpu]
> amdgpu_set_dpm_forced_performance_level+0x129/0x330 [amdgpu]
>
> This was introduced by "drm/amd/powerplay: cleanup the interfaces for
> powergate setting through SMU".
>
> Change-Id: I6ae2ed89d40a9475ae37b58e22108f4d6da974bb
> Signed-off-by: Evan Quan <evan.quan at amd.com>
> Reported-by: Rui Teng <Rui.Teng at amd.com>

Please add the description of the deadlock from the comment in the
code to the patch description as well.  With that fixed:
Acked-by: Alex Deucher <alexander.deucher at amd.com>

> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c | 40 +++++++++++++++----------
>  1 file changed, 25 insertions(+), 15 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c
> index 6c7dca1da992..19a042cf673c 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c
> @@ -946,23 +946,36 @@ int amdgpu_dpm_set_powergating_by_smu(struct amdgpu_device *adev, uint32_t block
>         bool swsmu = is_support_sw_smu(adev);
>
>         switch (block_type) {
> -       case AMD_IP_BLOCK_TYPE_GFX:
>         case AMD_IP_BLOCK_TYPE_UVD:
> -       case AMD_IP_BLOCK_TYPE_VCN:
>         case AMD_IP_BLOCK_TYPE_VCE:
> -       case AMD_IP_BLOCK_TYPE_SDMA:
>                 if (swsmu) {
>                         ret = smu_dpm_set_power_gate(&adev->smu, block_type, gate);
> -               } else {
> -                       if (adev->powerplay.pp_funcs &&
> -                           adev->powerplay.pp_funcs->set_powergating_by_smu) {
> -                               mutex_lock(&adev->pm.mutex);
> -                               ret = ((adev)->powerplay.pp_funcs->set_powergating_by_smu(
> -                                       (adev)->powerplay.pp_handle, block_type, gate));
> -                               mutex_unlock(&adev->pm.mutex);
> -                       }
> +               } else if (adev->powerplay.pp_funcs &&
> +                          adev->powerplay.pp_funcs->set_powergating_by_smu) {
> +                       /*
> +                        * TODO: need a better lock mechanism
> +                        *
> +                        * Here adev->pm.mutex lock protection is enforced on
> +                        * UVD and VCE cases only. Since for other cases, there
> +                        * may be already lock protection in amdgpu_pm.c.
> +                        * This is a quick fix for the deadlock issue.
> +                        */
> +                       mutex_lock(&adev->pm.mutex);
> +                       ret = ((adev)->powerplay.pp_funcs->set_powergating_by_smu(
> +                               (adev)->powerplay.pp_handle, block_type, gate));
> +                       mutex_unlock(&adev->pm.mutex);
>                 }
>                 break;
> +       case AMD_IP_BLOCK_TYPE_GFX:
> +       case AMD_IP_BLOCK_TYPE_VCN:
> +       case AMD_IP_BLOCK_TYPE_SDMA:
> +               if (swsmu)
> +                       ret = smu_dpm_set_power_gate(&adev->smu, block_type, gate);
> +               else if (adev->powerplay.pp_funcs &&
> +                        adev->powerplay.pp_funcs->set_powergating_by_smu)
> +                       ret = ((adev)->powerplay.pp_funcs->set_powergating_by_smu(
> +                               (adev)->powerplay.pp_handle, block_type, gate));
> +               break;
>         case AMD_IP_BLOCK_TYPE_JPEG:
>                 if (swsmu)
>                         ret = smu_dpm_set_power_gate(&adev->smu, block_type, gate);
> @@ -970,12 +983,9 @@ int amdgpu_dpm_set_powergating_by_smu(struct amdgpu_device *adev, uint32_t block
>         case AMD_IP_BLOCK_TYPE_GMC:
>         case AMD_IP_BLOCK_TYPE_ACP:
>                 if (adev->powerplay.pp_funcs &&
> -                   adev->powerplay.pp_funcs->set_powergating_by_smu) {
> -                       mutex_lock(&adev->pm.mutex);
> +                   adev->powerplay.pp_funcs->set_powergating_by_smu)
>                         ret = ((adev)->powerplay.pp_funcs->set_powergating_by_smu(
>                                 (adev)->powerplay.pp_handle, block_type, gate));
> -                       mutex_unlock(&adev->pm.mutex);
> -               }
>                 break;
>         default:
>                 break;
> --
> 2.24.1
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx


More information about the amd-gfx mailing list