Re: [PATCH v3 4/4] drm/xe: Handle PUNIT errors by requesting cold-reset recovery

[Raag Jadav <raag.jadav@intel.com>]

On Mon, Apr 06, 2026 at 07:53:30PM +0530, Mallesh Koujalagi wrote:
> When PUNIT (power management unit) errors are detected that persist across
> warm resets, mark the device as wedged with DRM_WEDGE_RECOVERY_COLD_RESET
> and notify userspace that a complete device cold reset is required to
> restore normal operation.
> 
> v3:
> - Use PUNIT instead of PMU. (Riana)
> - Use consistent wordingi.
> - Remove log. (Raag)
> 
> Signed-off-by: Mallesh Koujalagi <mallesh.koujalagi@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_ras.c | 21 ++++++++++++++++++++-
>  drivers/gpu/drm/xe/xe_ras.h |  1 +
>  2 files changed, 21 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_ras.c b/drivers/gpu/drm/xe/xe_ras.c
> index 437811845c01..e2e1ab3fb4ce 100644
> --- a/drivers/gpu/drm/xe/xe_ras.c
> +++ b/drivers/gpu/drm/xe/xe_ras.c
> @@ -5,6 +5,7 @@
>  
>  #include "xe_assert.h"
>  #include "xe_device_types.h"
> +#include "xe_device.h"
>  #include "xe_printk.h"
>  #include "xe_ras.h"
>  #include "xe_ras_types.h"
> @@ -93,6 +94,24 @@ static enum xe_ras_recovery_action handle_compute_errors(struct xe_device *xe,
>  	return XE_RAS_RECOVERY_ACTION_RECOVERED;
>  }
>  
> +/**
> + * xe_punit_error_handler - Handler for Punit errors requiring cold reset
> + * @xe: device instance
> + *
> + * Handles Punit errors that affect the device and cannot be recovered
> + * through driver reload, PCIe reset, etc.
> + *
> + * Marks the device as wedged with DRM_WEDGE_RECOVERY_COLD_RESET method
> + * and notifies userspace that a device cold reset is required.
> + */
> +void xe_punit_error_handler(struct xe_device *xe)

Should this be static?

> +{
> +	xe_err(xe, "Recovery: Device cold reset required\n");

Rather print it inside the event helper so we don't have to go around
logging each case. Let me see what can be done here.

Raag

> +	xe_device_set_wedged_method(xe, DRM_WEDGE_RECOVERY_COLD_RESET);
> +	xe_device_declare_wedged(xe);
> +}
> +
>  static enum xe_ras_recovery_action handle_soc_internal_errors(struct xe_device *xe,
>  							      struct xe_ras_error_array *arr)
>  {
> @@ -132,7 +151,7 @@ static enum xe_ras_recovery_action handle_soc_internal_errors(struct xe_device *
>  			xe_err(xe, "[RAS]: PUNIT %s error detected: 0x%x\n",
>  			       severity_to_str(xe, common.severity),
>  			       ieh_error->global_error_status);
> -			/** TODO: Add PUNIT error handling */
> +			xe_punit_error_handler(xe);
>  			return XE_RAS_RECOVERY_ACTION_DISCONNECT;
>  		}
>  	}
> diff --git a/drivers/gpu/drm/xe/xe_ras.h b/drivers/gpu/drm/xe/xe_ras.h
> index e191ab80080c..ab1fde200625 100644
> --- a/drivers/gpu/drm/xe/xe_ras.h
> +++ b/drivers/gpu/drm/xe/xe_ras.h
> @@ -11,6 +11,7 @@
>  struct xe_device;
>  
>  void xe_ras_init(struct xe_device *xe);
> +void xe_punit_error_handler(struct xe_device *xe);
>  enum xe_ras_recovery_action  xe_ras_process_errors(struct xe_device *xe);
>  
>  #endif
> -- 
> 2.34.1
>