Re: [PATCH v1 1/3] drm/xe/cri: Setup hardware error routing

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

On Thu, Mar 26, 2026 at 06:19:00PM +0530, Mallesh, Koujalagi wrote:
> On 17-02-2026 05:16 pm, Raag Jadav wrote:
> > Setup hardware error routing based on XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
> > which will be used for debug cases on Crescent Island.
> > 
> > Signed-off-by: Raag Jadav <raag.jadav@intel.com>
> > ---
> >   drivers/gpu/drm/xe/regs/xe_hw_error_regs.h |  5 +++++
> >   drivers/gpu/drm/xe/xe_hw_error.c           | 22 +++++++++++++++++++++-
> >   drivers/gpu/drm/xe/xe_hw_error.h           |  1 +
> >   3 files changed, 27 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/xe/regs/xe_hw_error_regs.h b/drivers/gpu/drm/xe/regs/xe_hw_error_regs.h
> > index c146b9ef44eb..88593f0a48fd 100644
> > --- a/drivers/gpu/drm/xe/regs/xe_hw_error_regs.h
> > +++ b/drivers/gpu/drm/xe/regs/xe_hw_error_regs.h
> > @@ -11,6 +11,11 @@
> >   #define HEC_UNCORR_FW_ERR_DW0(base)                    XE_REG((base) + 0x124)
> > +#define DEV_ERR_ROUTING_CTRL			XE_REG(0x100170)
> > +#define   FATAL_ERR_ROUTING			BIT(2)
> > +#define   NONFATAL_ERR_ROUTING			BIT(1)
> > +#define   CORR_ERR_ROUTING			BIT(0)
> > +
> >   #define DEV_ERR_STAT_NONFATAL			0x100178
> >   #define DEV_ERR_STAT_CORRECTABLE		0x10017c
> >   #define DEV_ERR_STAT_REG(x)			XE_REG(_PICK_EVEN((x), \
> > diff --git a/drivers/gpu/drm/xe/xe_hw_error.c b/drivers/gpu/drm/xe/xe_hw_error.c
> > index 8c65291f36fc..eb557266d083 100644
> > --- a/drivers/gpu/drm/xe/xe_hw_error.c
> > +++ b/drivers/gpu/drm/xe/xe_hw_error.c
> > @@ -146,6 +146,23 @@ void xe_hw_error_irq_handler(struct xe_tile *tile, const u32 master_ctl)
> >   			hw_error_source_handler(tile, hw_err);
> >   }
> > +/**
> > + * xe_hw_error_irq_route - irq routing of hw errors
> > + * @xe: xe device instance
> > + *
> > + * Set irq routing of hw errors.
> > + */
> > +void xe_hw_error_irq_route(struct xe_device *xe)
> > +{
> > +	u32 mask = CORR_ERR_ROUTING | NONFATAL_ERR_ROUTING | FATAL_ERR_ROUTING;
> > +	bool irq = xe->wedged.mode == XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET;
> 
> What happen when wedge mode change during execution?

This needs to be called when it happens. The only place I could find as
of now is through debugfs, will add in this in v2.

Raag

> > +	struct xe_tile *tile;
> > +	u8 id;
> > +
> > +	for_each_tile(tile, xe, id)
> > +		xe_mmio_rmw32(&tile->mmio, DEV_ERR_ROUTING_CTRL, mask, irq ? mask : 0);
> > +}
> > +
> >   /*
> >    * Process hardware errors during boot
> >    */
> > @@ -178,5 +195,8 @@ void xe_hw_error_init(struct xe_device *xe)
> >   	INIT_WORK(&tile->csc_hw_error_work, csc_hw_error_work);
> > -	process_hw_errors(xe);
> > +	if (xe->info.platform == XE_CRESCENTISLAND)
> > +		xe_hw_error_irq_route(xe);
> > +	else
> > +		process_hw_errors(xe);
> >   }
> > diff --git a/drivers/gpu/drm/xe/xe_hw_error.h b/drivers/gpu/drm/xe/xe_hw_error.h
> > index d86e28c5180c..8f155b66b5f1 100644
> > --- a/drivers/gpu/drm/xe/xe_hw_error.h
> > +++ b/drivers/gpu/drm/xe/xe_hw_error.h
> > @@ -11,5 +11,6 @@ struct xe_tile;
> >   struct xe_device;
> >   void xe_hw_error_irq_handler(struct xe_tile *tile, const u32 master_ctl);
> > +void xe_hw_error_irq_route(struct xe_device *xe);
> >   void xe_hw_error_init(struct xe_device *xe);
> >   #endif