Re: [PATCH 07/12] drm/xe: Remove memirq status and source checks for engine interrupts

["Summers, Stuart" <stuart.summers@intel.com>]

On Sat, 2026-06-06 at 13:18 +0200, Michal Wajdeczko wrote:
> > 
> > 
> > On 6/6/2026 1:21 AM, Stuart Summers wrote:
> > > > For engine-specific, memory-based interrupts, hardware will not
> > > > > > fill in
> > > > the source and status offsets, assuming that software will
> > > > utilize > > the
> > > > vector ID to determine the destination for a particular
> > > > interrupt.
> > > > 
> > > > GuC and VF based interrupts are still handled the same -
> > > > explicitly
> > > > checking these offsets.
> > > > 
> > > > Bspec: 62316
> > > > 
> > 
> > no empty lines here

Is that a rule? I thought we wanted the blank line between the commit
body and the signatures?

> > > > Signed-off-by: Stuart Summers <stuart.summers@intel.com>
> > > > Assisted-by: Copilot:claude-sonnet-4.6
> > > > ---
> > > >  drivers/gpu/drm/xe/xe_memirq.c | 26 ++++++++------------------
> > > >  1 file changed, 8 insertions(+), 18 deletions(-)
> > > > 
> > > > diff --git a/drivers/gpu/drm/xe/xe_memirq.c > >
> > > > b/drivers/gpu/drm/xe/xe_memirq.c
> > > > index 9dfe965cb46e..208f44436c66 100644
> > > > --- a/drivers/gpu/drm/xe/xe_memirq.c
> > > > +++ b/drivers/gpu/drm/xe/xe_memirq.c
> > > > @@ -447,18 +447,18 @@ static void memirq_assume_received(struct
> > > > > > xe_memirq *memirq, const char *source,
> > > >         memirq_debug(memirq, "ASSUME %s %s(%u)\n", source,
> > > > status, > > offset);
> > > >  }
> > > >  
> > > > -static void memirq_dispatch_engine(struct xe_memirq *memirq, >
> > > > > struct iosys_map *status,
> > > > +static void memirq_dispatch_engine(struct xe_memirq *memirq,
> > > >                                    struct xe_hw_engine *hwe)
> > > >  {
> > > > -       memirq_debug(memirq, "STATUS %s %*ph\n", hwe->name, 16,
> > > > > > status->vaddr);
> > > > +       memirq_debug(memirq, "dispatching engine %s\n", hwe-
> > > > >name);
> > > >  
> > > >         /*
> > > > -        * The programming note says to assume that > >
> > > > GT_MI_USER_INTERRUPT is always
> > > > -        * set. Check and clear related status byte just for a
> > > > > > debug.
> > > > +        * On MSI-X platforms hardware does not fill in the
> > > > source > > and status
> > > > +        * fields for engine-based interrupts (only GuC and VF
> > > > > > interrupts have
> > > > +        * a valid source/status). The dma-fence check for the
> > > > > > fence completion
> > > > +        * is opportunistic, unconditionally pass
> > > > MI_USER_INTERRUPT > > to issue
> > > > +        * that check.
> > > >          */
> > > > -       if (IS_ENABLED(CONFIG_DRM_XE_DEBUG_MEMIRQ) &&
> > > > -           !memirq_received(memirq, status, > >
> > > > ilog2(GT_MI_USER_INTERRUPT), hwe->name))
> > > > -               memirq_assume_received(memirq, hwe->name, > >
> > > > ilog2(GT_MI_USER_INTERRUPT), "USER");
> > > >         xe_hw_engine_handle_irq(hwe, GT_MI_USER_INTERRUPT);
> > > >  }
> > > >  
> > > > @@ -499,17 +499,7 @@ static void memirq_dispatch_guc(struct > >
> > > > xe_memirq *memirq, struct iosys_map *stat
> > > >   */
> > > >  void xe_memirq_hwe_handler(struct xe_memirq *memirq, struct >
> > > > > xe_hw_engine *hwe)
> > > >  {
> > > > -       struct iosys_map source =
> > > > -               IOSYS_MAP_INIT_OFFSET(&memirq->bo->vmap,
> > > > -                                     > >
> > > > XE_MEMIRQ_SOURCE_OFFSET(hwe->irq_page));
> > > > -
> > > > -       if (memirq_received(memirq, &source, hwe->irq_offset, >
> > > > > "SRC")) {
> > > > -               struct iosys_map status =
> > > > -                       IOSYS_MAP_INIT_OFFSET(&memirq->bo-
> > > > >vmap,
> > > > -                                             > >
> > > > XE_MEMIRQ_VECTOR_OFFSET(hwe->irq_page,
> > > > -
> > > >                                                                 
> > > >   > >    hwe->irq_offset));
> > > > -               memirq_dispatch_engine(memirq, &status, hwe);
> > > > -       }
> > 
> > maybe put the above new comment here (or update the kernel-doc) and
> > > call xe_hw_engine_handle_irq() directly?
> > then the !MSI-X || VF case will still have a status debug dump
> > option > available

Yeah I can drop the memirq_dispatch_engine() and just call the other
directly. You're right it's just extra code at this point.

Thanks,
Stuart

> > 
> > > > +       memirq_dispatch_engine(memirq, hwe);
> > > >  }
> > > >  
> > > >  /**
> >