Re: [PATCH] drm/xe: Implement print vfunc for page fault info logging

[Matthew Brost <matthew.brost@intel.com>]

On Fri, Sep 25, 2026 at 11:38:52PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Low] xe: log interleaving and redundant prefixes due to split print calls
> --
> 
> commit 114cd7fd731f9ee67d8c36081ae1b0f44d9be562
> Author: Brian Nguyen <brian3.nguyen@intel.com>
> 
> drm/xe: Implement print vfunc for page fault info logging
> 
> This commit adds a print vfunc to xe_pagefault_ops to let producer
> backends log specific relevant debug info. This keeps the consumer
> layer free of backend-specific formatting logic, with consumer-specific
> prints handled separately.
> 
> > diff --git a/drivers/gpu/drm/xe/xe_pagefault.c b/drivers/gpu/drm/xe/xe_pagefault.c
> > --- a/drivers/gpu/drm/xe/xe_pagefault.c
> > +++ b/drivers/gpu/drm/xe/xe_pagefault.c
> > @@ -607,35 +606,9 @@ static const char *xe_pagefault_error_to_str(enum xe_pagefault_error error)
> >  
> >  static void xe_pagefault_print(struct xe_pagefault *pf)
> >  {
> 
> [ ... ]
> 
> > -		   FIELD_GET(XE_PAGEFAULT_ENGINE_INSTANCE_MASK,
> > -			     pf->consumer.engine_class_instance),
> > -		   FIELD_GET(XE_PAGEFAULT_SRCID_MASK,
> > -			     pf->consumer.id),
> > +	xe_gt_assert(pf->gt, pf->producer.ops->print);
> > +	pf->producer.ops->print(pf);
> > +	xe_gt_info(pf->gt, "\tError: %s\n",
> >  		   xe_pagefault_error_to_str(xe_pagefault_get_error(pf)));
> >  }
> 
> [Severity: Low]
> Does this split print call cause log interleaving and redundant prefixes?
> 
> Because the error string is now logged with a separate call to xe_gt_info()
> in xe_pagefault_print(), it will be printed as an isolated log entry with
> its own prefix, rather than forming a continuous atomic block with the fault
> details from the producer print vfunc.
> 
> This disrupts the formatting of the page fault debug info, introducing
> redundant logging prefixes in the middle of what used to be a single
> continuous multiline log block. It also creates a risk of concurrent kernel
> messages interleaving with the output.
> 

This is valid too. So either make print accept the output of
xe_pagefault_error_to_str as an argument and print the entire thing in
one shot.

Matt

> -- 
> Sashiko AI review · https://sashiko.dev/#/patchset/20260925232757.3456172-2-brian3.nguyen@intel.com?part=1