[PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[Raag Jadav]

Currently 'none' recovery method for wedged event expects no explicit
recovery from userspace consumer but it expects driver to have reset
the device, which may not always be the case. Update documentation and
log to reflect this expectation.

v2: Rephrase documentation (Rodrigo)

Raag Jadav (3):
  drm/doc: Update documentation for 'none' recovery method
  drm: Update log for 'none' recovery method
  drm/xe: Send 'none' recovery method for
    XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET

 Documentation/gpu/drm-uapi.rst | 14 +++++++-------
 drivers/gpu/drm/drm_drv.c      |  2 +-
 drivers/gpu/drm/xe/xe_device.c |  9 ++++++++-
 3 files changed, 16 insertions(+), 9 deletions(-)

-- 
2.43.0

[PATCH v2 1/3] drm/doc: Update documentation for 'none' recovery method

[Raag Jadav]

Expand 'none' recovery method for wedged event to include debug cases
where driver wants to hint "no recovery" without resetting the device
from driver context.

Signed-off-by: Raag Jadav <raag.jadav@intel.com>
---
 Documentation/gpu/drm-uapi.rst | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/Documentation/gpu/drm-uapi.rst b/Documentation/gpu/drm-uapi.rst
index d98428a592f1..579e87cb9ff7 100644
--- a/Documentation/gpu/drm-uapi.rst
+++ b/Documentation/gpu/drm-uapi.rst
@@ -438,14 +438,14 @@ following expectations.
     unknown         consumer policy
     =============== ========================================
 
-The only exception to this is ``WEDGED=none``, which signifies that the device
-was temporarily 'wedged' at some point but was recovered from driver context
-using device specific methods like reset. No explicit recovery is expected from
-the consumer in this case, but it can still take additional steps like gathering
-telemetry information (devcoredump, syslog). This is useful because the first
-hang is usually the most critical one which can result in consequential hangs or
-complete wedging.
+No Recovery
+-----------
 
+Here ``WEDGED=none`` signifies that no recovery is expected from the consumer
+but it can still try to gather telemetry information (devcoredump, syslog) for
+debug purpose in order to root cause the hang. This is useful because the first
+hang is usually the most critical one which can result in consequential hangs
+or complete wedging.
 
 Vendor Specific Recovery
 ------------------------
-- 
2.43.0

[PATCH v2 2/3] drm: Update log for 'none' recovery method

[Raag Jadav]

Update log for 'none' recovery method for wedged event where driver wants
to hint "no recovery" without resetting the device from driver context.

Signed-off-by: Raag Jadav <raag.jadav@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/drm_drv.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 2915118436ce..72e7e09225c7 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -586,7 +586,7 @@ int drm_dev_wedged_event(struct drm_device *dev, unsigned long method,
 		snprintf(event_string, sizeof(event_string), "%s", "WEDGED=unknown");
 
 	drm_info(dev, "device wedged, %s\n", method == DRM_WEDGE_RECOVERY_NONE ?
-		 "but recovered through reset" : "needs recovery");
+		 "but no recovery needed" : "needs recovery");
 
 	if (info && (info->comm[0] != '\0') && (info->pid >= 0)) {
 		snprintf(pid_string, sizeof(pid_string), "PID=%u", info->pid);
-- 
2.43.0

[PATCH v2 3/3] drm/xe: Send 'none' recovery method for XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET

[Raag Jadav]

XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET is intended for debugging hangs,
so wedge the device with 'none' recovery method and have it available
to the user for debugging.

Signed-off-by: Raag Jadav <raag.jadav@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/xe_device.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index c137dc0a1840..eeb626737760 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -1329,8 +1329,15 @@ void xe_device_declare_wedged(struct xe_device *xe)
 		xe_gt_declare_wedged(gt);
 
 	if (xe_device_wedged(xe)) {
+		/*
+		 * XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET is intended for debugging
+		 * hangs, so wedge the device with 'none' recovery method and have
+		 * it available to the user for debugging.
+		 */
+		if (xe->wedged.mode == XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET)
+			xe_device_set_wedged_method(xe, DRM_WEDGE_RECOVERY_NONE);
 		/* If no wedge recovery method is set, use default */
-		if (!xe->wedged.method)
+		else if (!xe->wedged.method)
 			xe_device_set_wedged_method(xe, DRM_WEDGE_RECOVERY_REBIND |
 						    DRM_WEDGE_RECOVERY_BUS_RESET);
 
-- 
2.43.0

Re: [PATCH v2 1/3] drm/doc: Update documentation for 'none' recovery method

[Rodrigo Vivi]

On Thu, Mar 05, 2026 at 06:36:47PM +0530, Raag Jadav wrote:
> Expand 'none' recovery method for wedged event to include debug cases
> where driver wants to hint "no recovery" without resetting the device
> from driver context.
> 
> Signed-off-by: Raag Jadav <raag.jadav@intel.com>

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

> ---
>  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
> 
> diff --git a/Documentation/gpu/drm-uapi.rst b/Documentation/gpu/drm-uapi.rst
> index d98428a592f1..579e87cb9ff7 100644
> --- a/Documentation/gpu/drm-uapi.rst
> +++ b/Documentation/gpu/drm-uapi.rst
> @@ -438,14 +438,14 @@ following expectations.
>      unknown         consumer policy
>      =============== ========================================
>  
> -The only exception to this is ``WEDGED=none``, which signifies that the device
> -was temporarily 'wedged' at some point but was recovered from driver context
> -using device specific methods like reset. No explicit recovery is expected from
> -the consumer in this case, but it can still take additional steps like gathering
> -telemetry information (devcoredump, syslog). This is useful because the first
> -hang is usually the most critical one which can result in consequential hangs or
> -complete wedging.
> +No Recovery
> +-----------
>  
> +Here ``WEDGED=none`` signifies that no recovery is expected from the consumer
> +but it can still try to gather telemetry information (devcoredump, syslog) for
> +debug purpose in order to root cause the hang. This is useful because the first
> +hang is usually the most critical one which can result in consequential hangs
> +or complete wedging.
>  
>  Vendor Specific Recovery
>  ------------------------
> -- 
> 2.43.0
> 

Re: [PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[Raag Jadav]

On Thu, Mar 05, 2026 at 06:36:46PM +0530, Raag Jadav wrote:
> Currently 'none' recovery method for wedged event expects no explicit
> recovery from userspace consumer but it expects driver to have reset
> the device, which may not always be the case. Update documentation and
> log to reflect this expectation.

Anything I can do to move this forward?

Raag

> v2: Rephrase documentation (Rodrigo)
> 
> Raag Jadav (3):
>   drm/doc: Update documentation for 'none' recovery method
>   drm: Update log for 'none' recovery method
>   drm/xe: Send 'none' recovery method for
>     XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
> 
>  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
>  drivers/gpu/drm/drm_drv.c      |  2 +-
>  drivers/gpu/drm/xe/xe_device.c |  9 ++++++++-
>  3 files changed, 16 insertions(+), 9 deletions(-)
> 
> -- 
> 2.43.0
> 

Re: [PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[Raag Jadav]

On Tue, Mar 10, 2026 at 08:56:58AM +0100, Raag Jadav wrote:
> On Thu, Mar 05, 2026 at 06:36:46PM +0530, Raag Jadav wrote:
> > Currently 'none' recovery method for wedged event expects no explicit
> > recovery from userspace consumer but it expects driver to have reset
> > the device, which may not always be the case. Update documentation and
> > log to reflect this expectation.
> 
> Anything I can do to move this forward?

Ping.

Raag

> > v2: Rephrase documentation (Rodrigo)
> > 
> > Raag Jadav (3):
> >   drm/doc: Update documentation for 'none' recovery method
> >   drm: Update log for 'none' recovery method
> >   drm/xe: Send 'none' recovery method for
> >     XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
> > 
> >  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
> >  drivers/gpu/drm/drm_drv.c      |  2 +-
> >  drivers/gpu/drm/xe/xe_device.c |  9 ++++++++-
> >  3 files changed, 16 insertions(+), 9 deletions(-)
> > 
> > -- 
> > 2.43.0
> > 

Re: [PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[Christian König]

On 3/23/26 15:10, Raag Jadav wrote:
> On Tue, Mar 10, 2026 at 08:56:58AM +0100, Raag Jadav wrote:
>> On Thu, Mar 05, 2026 at 06:36:46PM +0530, Raag Jadav wrote:
>>> Currently 'none' recovery method for wedged event expects no explicit
>>> recovery from userspace consumer but it expects driver to have reset
>>> the device, which may not always be the case. Update documentation and
>>> log to reflect this expectation.
>>
>> Anything I can do to move this forward?

Invent some tech to clone certain people so that we have more reviewers.

> Ping.

I will take a look now.

Christian.

> 
> Raag
> 
>>> v2: Rephrase documentation (Rodrigo)
>>>
>>> Raag Jadav (3):
>>>   drm/doc: Update documentation for 'none' recovery method
>>>   drm: Update log for 'none' recovery method
>>>   drm/xe: Send 'none' recovery method for
>>>     XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
>>>
>>>  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
>>>  drivers/gpu/drm/drm_drv.c      |  2 +-
>>>  drivers/gpu/drm/xe/xe_device.c |  9 ++++++++-
>>>  3 files changed, 16 insertions(+), 9 deletions(-)
>>>
>>> -- 
>>> 2.43.0
>>>

Re: [PATCH v2 1/3] drm/doc: Update documentation for 'none' recovery method

[Christian König]

On 3/5/26 14:06, Raag Jadav wrote:
> Expand 'none' recovery method for wedged event to include debug cases
> where driver wants to hint "no recovery" without resetting the device
> from driver context.
> 
> Signed-off-by: Raag Jadav <raag.jadav@intel.com>

Reviewed-by: Christian König <christian.koenig@amd.com>

> ---
>  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
> 
> diff --git a/Documentation/gpu/drm-uapi.rst b/Documentation/gpu/drm-uapi.rst
> index d98428a592f1..579e87cb9ff7 100644
> --- a/Documentation/gpu/drm-uapi.rst
> +++ b/Documentation/gpu/drm-uapi.rst
> @@ -438,14 +438,14 @@ following expectations.
>      unknown         consumer policy
>      =============== ========================================
>  
> -The only exception to this is ``WEDGED=none``, which signifies that the device
> -was temporarily 'wedged' at some point but was recovered from driver context
> -using device specific methods like reset. No explicit recovery is expected from
> -the consumer in this case, but it can still take additional steps like gathering
> -telemetry information (devcoredump, syslog). This is useful because the first
> -hang is usually the most critical one which can result in consequential hangs or
> -complete wedging.
> +No Recovery
> +-----------
>  
> +Here ``WEDGED=none`` signifies that no recovery is expected from the consumer
> +but it can still try to gather telemetry information (devcoredump, syslog) for
> +debug purpose in order to root cause the hang. This is useful because the first
> +hang is usually the most critical one which can result in consequential hangs
> +or complete wedging.
>  
>  Vendor Specific Recovery
>  ------------------------

Re: [PATCH v2 2/3] drm: Update log for 'none' recovery method

[Christian König]

On 3/5/26 14:06, Raag Jadav wrote:
> Update log for 'none' recovery method for wedged event where driver wants
> to hint "no recovery" without resetting the device from driver context.
> 
> Signed-off-by: Raag Jadav <raag.jadav@intel.com>
> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

Reviewed-by: Christian König <christian.koenig@amd.com>

> ---
>  drivers/gpu/drm/drm_drv.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
> index 2915118436ce..72e7e09225c7 100644
> --- a/drivers/gpu/drm/drm_drv.c
> +++ b/drivers/gpu/drm/drm_drv.c
> @@ -586,7 +586,7 @@ int drm_dev_wedged_event(struct drm_device *dev, unsigned long method,
>  		snprintf(event_string, sizeof(event_string), "%s", "WEDGED=unknown");
>  
>  	drm_info(dev, "device wedged, %s\n", method == DRM_WEDGE_RECOVERY_NONE ?
> -		 "but recovered through reset" : "needs recovery");
> +		 "but no recovery needed" : "needs recovery");
>  
>  	if (info && (info->comm[0] != '\0') && (info->pid >= 0)) {
>  		snprintf(pid_string, sizeof(pid_string), "PID=%u", info->pid);

Re: [PATCH v2 3/3] drm/xe: Send 'none' recovery method for XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET

[Christian König]

On 3/5/26 14:06, Raag Jadav wrote:
> XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET is intended for debugging hangs,
> so wedge the device with 'none' recovery method and have it available
> to the user for debugging.

Looks sane to me, but I certainly can't review nor ack that patch.

It's basically up to the XE maintainers to decide if this fulfills the requirements or not.

Regards,
Christian.

> 
> Signed-off-by: Raag Jadav <raag.jadav@intel.com>
> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_device.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
> index c137dc0a1840..eeb626737760 100644
> --- a/drivers/gpu/drm/xe/xe_device.c
> +++ b/drivers/gpu/drm/xe/xe_device.c
> @@ -1329,8 +1329,15 @@ void xe_device_declare_wedged(struct xe_device *xe)
>  		xe_gt_declare_wedged(gt);
>  
>  	if (xe_device_wedged(xe)) {
> +		/*
> +		 * XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET is intended for debugging
> +		 * hangs, so wedge the device with 'none' recovery method and have
> +		 * it available to the user for debugging.
> +		 */
> +		if (xe->wedged.mode == XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET)
> +			xe_device_set_wedged_method(xe, DRM_WEDGE_RECOVERY_NONE);
>  		/* If no wedge recovery method is set, use default */
> -		if (!xe->wedged.method)
> +		else if (!xe->wedged.method)
>  			xe_device_set_wedged_method(xe, DRM_WEDGE_RECOVERY_REBIND |
>  						    DRM_WEDGE_RECOVERY_BUS_RESET);
>  

Re: [PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[Raag Jadav]

On Mon, Mar 23, 2026 at 03:11:58PM +0100, Christian König wrote:
> On 3/23/26 15:10, Raag Jadav wrote:
> > On Tue, Mar 10, 2026 at 08:56:58AM +0100, Raag Jadav wrote:
> >> On Thu, Mar 05, 2026 at 06:36:46PM +0530, Raag Jadav wrote:
> >>> Currently 'none' recovery method for wedged event expects no explicit
> >>> recovery from userspace consumer but it expects driver to have reset
> >>> the device, which may not always be the case. Update documentation and
> >>> log to reflect this expectation.
> >>
> >> Anything I can do to move this forward?
> 
> Invent some tech to clone certain people so that we have more reviewers.

If this allows my clones to also review my patches, I fully support this :)

> > Ping.
> 
> I will take a look now.

Thank you, I think it's safe to land through drm-misc (if anyone has access).

Raag

> >>> v2: Rephrase documentation (Rodrigo)
> >>>
> >>> Raag Jadav (3):
> >>>   drm/doc: Update documentation for 'none' recovery method
> >>>   drm: Update log for 'none' recovery method
> >>>   drm/xe: Send 'none' recovery method for
> >>>     XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
> >>>
> >>>  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
> >>>  drivers/gpu/drm/drm_drv.c      |  2 +-
> >>>  drivers/gpu/drm/xe/xe_device.c |  9 ++++++++-
> >>>  3 files changed, 16 insertions(+), 9 deletions(-)
> >>>
> >>> -- 
> >>> 2.43.0
> >>>
> 

Re: [PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[Rodrigo Vivi]

On Tue, Mar 24, 2026 at 03:36:24AM +0100, Raag Jadav wrote:
> On Mon, Mar 23, 2026 at 03:11:58PM +0100, Christian König wrote:
> > On 3/23/26 15:10, Raag Jadav wrote:
> > > On Tue, Mar 10, 2026 at 08:56:58AM +0100, Raag Jadav wrote:
> > >> On Thu, Mar 05, 2026 at 06:36:46PM +0530, Raag Jadav wrote:
> > >>> Currently 'none' recovery method for wedged event expects no explicit
> > >>> recovery from userspace consumer but it expects driver to have reset
> > >>> the device, which may not always be the case. Update documentation and
> > >>> log to reflect this expectation.
> > >>
> > >> Anything I can do to move this forward?
> > 
> > Invent some tech to clone certain people so that we have more reviewers.
> 
> If this allows my clones to also review my patches, I fully support this :)
> 
> > > Ping.
> > 
> > I will take a look now.
> 
> Thank you, I think it's safe to land through drm-misc (if anyone has access).

Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

But it is not applying cleanly on drm-misc-next.

drm-misc maintainers, could you please do a backmerge of drm-next?

Or ack these patches to go through drm-xe-next

> 
> Raag
> 
> > >>> v2: Rephrase documentation (Rodrigo)
> > >>>
> > >>> Raag Jadav (3):
> > >>>   drm/doc: Update documentation for 'none' recovery method
> > >>>   drm: Update log for 'none' recovery method
> > >>>   drm/xe: Send 'none' recovery method for
> > >>>     XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
> > >>>
> > >>>  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
> > >>>  drivers/gpu/drm/drm_drv.c      |  2 +-
> > >>>  drivers/gpu/drm/xe/xe_device.c |  9 ++++++++-
> > >>>  3 files changed, 16 insertions(+), 9 deletions(-)
> > >>>
> > >>> -- 
> > >>> 2.43.0
> > >>>
> > 

Re: [PATCH v2 0/3] Update 'none' recovery method for DRM wedged event

[Raag Jadav]

On Tue, Mar 24, 2026 at 09:52:07AM -0400, Rodrigo Vivi wrote:
> On Tue, Mar 24, 2026 at 03:36:24AM +0100, Raag Jadav wrote:
> > On Mon, Mar 23, 2026 at 03:11:58PM +0100, Christian König wrote:
> > > On 3/23/26 15:10, Raag Jadav wrote:
> > > > On Tue, Mar 10, 2026 at 08:56:58AM +0100, Raag Jadav wrote:
> > > >> On Thu, Mar 05, 2026 at 06:36:46PM +0530, Raag Jadav wrote:
> > > >>> Currently 'none' recovery method for wedged event expects no explicit
> > > >>> recovery from userspace consumer but it expects driver to have reset
> > > >>> the device, which may not always be the case. Update documentation and
> > > >>> log to reflect this expectation.
> > > >>
> > > >> Anything I can do to move this forward?
> > > 
> > > Invent some tech to clone certain people so that we have more reviewers.
> > 
> > If this allows my clones to also review my patches, I fully support this :)
> > 
> > > > Ping.
> > > 
> > > I will take a look now.
> > 
> > Thank you, I think it's safe to land through drm-misc (if anyone has access).
> 
> Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
> 
> But it is not applying cleanly on drm-misc-next.
> 
> drm-misc maintainers, could you please do a backmerge of drm-next?
> 
> Or ack these patches to go through drm-xe-next

It's already on misc thanks to Brost. Thank you all for the support.

Raag

> > > >>> v2: Rephrase documentation (Rodrigo)
> > > >>>
> > > >>> Raag Jadav (3):
> > > >>>   drm/doc: Update documentation for 'none' recovery method
> > > >>>   drm: Update log for 'none' recovery method
> > > >>>   drm/xe: Send 'none' recovery method for
> > > >>>     XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
> > > >>>
> > > >>>  Documentation/gpu/drm-uapi.rst | 14 +++++++-------
> > > >>>  drivers/gpu/drm/drm_drv.c      |  2 +-
> > > >>>  drivers/gpu/drm/xe/xe_device.c |  9 ++++++++-
> > > >>>  3 files changed, 16 insertions(+), 9 deletions(-)
> > > >>>
> > > >>> -- 
> > > >>> 2.43.0
> > > >>>
> > >