linux-kernel - Re: [PATCH] drm/amdgpu: skip disabling audio when device is unplugged

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <9a632900-4ebb-40af-8bf8-bf55f8e25c7b@amd.com>
Date: Thu, 7 Aug 2025 12:03:21 +0200
From: Christian König <christian.koenig@....com>
To: oushixiong1025@....com, Alex Deucher <alexander.deucher@....com>
Cc: David Airlie <airlied@...il.com>, Simona Vetter <simona@...ll.ch>,
 Sunil Khatri <sunil.khatri@....com>,
 Alexandre Demers <alexandre.f.demers@...il.com>,
 Boyuan Zhang <boyuan.zhang@....com>, amd-gfx@...ts.freedesktop.org,
 dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
 Shixiong Ou <oushixiong@...inos.cn>
Subject: Re: [PATCH] drm/amdgpu: skip disabling audio when device is unplugged

On 07.08.25 11:47, oushixiong1025@....com wrote:
> From: Shixiong Ou <oushixiong@...inos.cn>
> 
> When Stopping lightdm and removing amdgpu driver are executed, the following
> error is triggered probably:
> 
> Unable to handle kernel paging request at virtual address 0000000000005e00
> .....
> [ 2] [T10084] Call trace:
> [ 2] [T10084]  amdgpu_device_wreg.part.0+0x58/0x110 [amdgpu]
> [ 2] [T10084]  amdgpu_device_wreg+0x20/0x38 [amdgpu]
> [ 2] [T10084]  dce_v6_0_audio_endpt_wreg+0x64/0xd8 [amdgpu]
> [ 2] [T10084]  dce_v6_0_sw_fini+0xa0/0x118 [amdgpu]
> [ 2] [T10084]  amdgpu_device_ip_fini.isra.0+0xdc/0x1e8 [amdgpu]
> [ 2] [T10084]  amdgpu_device_fini_sw+0x2c/0x220 [amdgpu]
> [ 2] [T10084]  amdgpu_driver_release_kms+0x20/0x40 [amdgpu]
> [ 2] [T10084]  devm_drm_dev_init_release+0x8c/0xc0 [drm]
> [ 2] [T10084]  devm_action_release+0x18/0x28
> [ 2] [T10084]  release_nodes+0x5c/0xc8
> [ 2] [T10084]  devres_release_all+0xa0/0x130
> [ 2] [T10084]  device_unbind_cleanup+0x1c/0x70
> [ 2] [T10084]  device_release_driver_internal+0x1e4/0x228
> [ 2] [T10084]  driver_detach+0x90/0x100
> [ 2] [T10084]  bus_remove_driver+0x74/0x100
> [ 2] [T10084]  driver_unregister+0x34/0x68
> [ 2] [T10084]  pci_unregister_driver+0x24/0x108
> [ 2] [T10084]  amdgpu_exit+0x1c/0x3270 [amdgpu]
> [ 2] [T10084]  __do_sys_delete_module.constprop.0+0x1d0/0x330
> [ 2] [T10084]  __arm64_sys_delete_module+0x18/0x28
> [ 2] [T10084]  invoke_syscall+0x4c/0x120
> [ 2] [T10084]  el0_svc_common.constprop.0+0xc4/0xf0
> [ 2] [T10084]  do_el0_svc+0x24/0x38
> [ 2] [T10084]  el0_svc+0x24/0x88
> [ 2] [T10084]  el0t_64_sync_handler+0x134/0x150
> [ 2] [T10084]  el0t_64_sync+0x14c/0x150
> [ 2] [T10084] Code: f9401bf7 f9453e60 8b150000 d50332bf (b9000016)
> [ 2] [T10084] ---[ end trace 0000000000000000 ]---
> 
> The adev->rmmio has been unmmaped in amdgpu_device_fini_hw().
> 
> So skip disabling audio when device is unplugged.
> 
> Signed-off-by: Shixiong Ou <oushixiong@...inos.cn>
> ---
>  drivers/gpu/drm/amd/amdgpu/dce_v6_0.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
> index 276c025c4c03..48b29990da7f 100644
> --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
> @@ -23,6 +23,7 @@
>  
>  #include <linux/pci.h>
>  
> +#include <drm/drm_drv.h>
>  #include <drm/drm_edid.h>
>  #include <drm/drm_fourcc.h>
>  #include <drm/drm_modeset_helper.h>
> @@ -1459,8 +1460,10 @@ static void dce_v6_0_audio_fini(struct amdgpu_device *adev)
>  	if (!adev->mode_info.audio.enabled)
>  		return;
>  
> -	for (i = 0; i < adev->mode_info.audio.num_pins; i++)
> -		dce_v6_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
> +	if (!drm_dev_is_unplugged(adev_to_drm(adev))) {

Good catch, but that looks like a workaround for something done in the wrong place.

A *_sw_fini() function should not enable/disable HW.

Regards,
Christian.

> +		for (i = 0; i < adev->mode_info.audio.num_pins; i++)
> +			dce_v6_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
> +	}
>  
>  	adev->mode_info.audio.enabled = false;
>  }