drm/radeon: take the mode_config mutex when dealing with hpds (v2) crashes

drm/radeon: take the mode_config mutex when dealing with hpds (v2) crashes

[Joerg Platte]

Dear Alex,

on my old P4 based non-SMP router your patch (commit 
32d12fc20e3c726ca858d0e5055fb596fce2f8bc in linux stable) crashes on 
Linux 4.1.4 and above. I was only able to take a picture of the whole 
trace https://ferdi.naasa.net/url/jplatte/IMG_3116.JPG

Reverting the patch resolves the issue.

This is my old graphics hardware:
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. 
[AMD/ATI] RV100 [Radeon 7000 / Radeon VE]

Just for the reference, here is the full patch:
commit 32d12fc20e3c726ca858d0e5055fb596fce2f8bc
Author: Alex Deucher <alexander.deucher@amd.com>
Date:   Fri May 15 11:48:52 2015 -0400

     drm/radeon: take the mode_config mutex when dealing with hpds (v2)

     commit 39fa10f7e21574a70cecf1fed0f9b36535aa68a0 upstream.

     Since we are messing with state in the worker.

     v2: drop the changes in the mst worker

     Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
     Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c 
b/drivers/gpu/drm/radeon/radeon_irq_kms.c
index 7162c93..f682e53 100644
--- a/drivers/gpu/drm/radeon/radeon_irq_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c
@@ -79,10 +79,12 @@ static void radeon_hotplug_work_func(struct 
work_struct *work)
         struct drm_mode_config *mode_config = &dev->mode_config;
         struct drm_connector *connector;

+       mutex_lock(&mode_config->mutex);
         if (mode_config->num_connector) {
                 list_for_each_entry(connector, 
&mode_config->connector_list, head)
                         radeon_connector_hotplug(connector);
         }
+       mutex_unlock(&mode_config->mutex);
         /* Just fire off a uevent and let userspace tell us what to do */
         drm_helper_hpd_irq_event(dev);
  }

Is it possible that the mutex is not defined on non-SMP systems? Can you 
help to resolve this regression?

Best regards,
Joerg

RE: drm/radeon: take the mode_config mutex when dealing with hpds (v2) crashes

[Deucher, Alexander]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="utf-8", Size: 2597 bytes --]

> -----Original Message-----
> From: Joerg Platte [mailto:lists@naasa.net]
> Sent: Sunday, August 23, 2015 11:26 PM
> To: Deucher, Alexander; linux-kernel@vger.kernel.org
> Subject: drm/radeon: take the mode_config mutex when dealing with hpds
> (v2) crashes
> 
> Dear Alex,
> 
> on my old P4 based non-SMP router your patch (commit
> 32d12fc20e3c726ca858d0e5055fb596fce2f8bc in linux stable) crashes on
> Linux 4.1.4 and above. I was only able to take a picture of the whole
> trace https://ferdi.naasa.net/url/jplatte/IMG_3116.JPG
> 
> Reverting the patch resolves the issue.
> 
> This is my old graphics hardware:
> 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc.
> [AMD/ATI] RV100 [Radeon 7000 / Radeon VE]
> 
> Just for the reference, here is the full patch:
> commit 32d12fc20e3c726ca858d0e5055fb596fce2f8bc
> Author: Alex Deucher <alexander.deucher@amd.com>
> Date:   Fri May 15 11:48:52 2015 -0400
> 
>      drm/radeon: take the mode_config mutex when dealing with hpds (v2)
> 
>      commit 39fa10f7e21574a70cecf1fed0f9b36535aa68a0 upstream.
> 
>      Since we are messing with state in the worker.
> 
>      v2: drop the changes in the mst worker
> 
>      Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>      Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> 
> diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c
> b/drivers/gpu/drm/radeon/radeon_irq_kms.c
> index 7162c93..f682e53 100644
> --- a/drivers/gpu/drm/radeon/radeon_irq_kms.c
> +++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c
> @@ -79,10 +79,12 @@ static void radeon_hotplug_work_func(struct
> work_struct *work)
>          struct drm_mode_config *mode_config = &dev->mode_config;
>          struct drm_connector *connector;
> 
> +       mutex_lock(&mode_config->mutex);
>          if (mode_config->num_connector) {
>                  list_for_each_entry(connector,
> &mode_config->connector_list, head)
>                          radeon_connector_hotplug(connector);
>          }
> +       mutex_unlock(&mode_config->mutex);
>          /* Just fire off a uevent and let userspace tell us what to do */
>          drm_helper_hpd_irq_event(dev);
>   }
> 
> Is it possible that the mutex is not defined on non-SMP systems? Can you
> help to resolve this regression?

Fixed in:
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=7f98ca454ad373fc1b76be804fa7138ff68c1d27

Alex

> 
> Best regards,
> Joerg
ÿôèº{.nÇ+‰·Ÿ®‰­†+%ŠËÿ±éݶ\x17¥Šwÿº{.nÇ+‰·¥Š{±þG«éÿŠ{ayº\x1dʇڙë,j\a­¢f£¢·hšïêÿ‘êçz_è®\x03(­éšŽŠÝ¢j"ú\x1a¶^[m§ÿÿ¾\a«þG«éÿ¢¸?™¨è­Ú&£ø§~á¶iO•æ¬z·švØ^\x14\x04\x1a¶^[m§ÿÿÃ\fÿ¶ìÿ¢¸?–I¥