Re: [PATCH v2 08/11] drm/xe/xe_ras: Add support for Uncorrectable Core-Compute errors

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

On Tue, Mar 31, 2026 at 09:54:32PM +0530, Tauro, Riana wrote:
> On 3/4/2026 10:22 PM, Raag Jadav wrote:
> > On Mon, Mar 02, 2026 at 03:52:03PM +0530, Riana Tauro wrote:
> > > Uncorrectable Core-Compute errors are classified into Global and Local
> > > errors.

...

> > > +	} while (response.additional_errors);
> > I know we're not NASA but I'd try to have some timeout instead of blindly
> > trusting the hardware.
> 
> additional_errors is a indication. So will be 0/1.
> Timeout is already present in sysctrl_send_command. So unnecessary here
> again.
> 
> Also before every sysctrl command we are setting response to 0.

If the firmware returns the same values repeatedly, we'll be stuck here
indefinitely. I know it's extreme corner case and probably a firmware
bug, but that's not an excuse for the driver to not handle it.

I've handled it[1] as flooding, not ideal but convenient for now.

[1] https://lore.kernel.org/intel-xe/20260331102346.1034100-3-raag.jadav@intel.com/

Raag