Re: [PATCH 1/4] drm/xe: Add build on bug to assert page fault queue works

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

On Tue, Jan 09, 2024 at 01:52:06PM -0800, Welty, Brian wrote:
> 
> 
> On 1/9/2024 9:28 AM, Matthew Brost wrote:
> > If NUM_PF_QUEUE % PF_MSG_LEN_DW != 0 then the page fault queue logic
> > does not work when wrapping occurs. Add a build bug on to assert
> > NUM_PF_QUEUE % PF_MSG_LEN_DW == 0 to enforce this restriction and
> > document the code.
> 
> I'm a bit confused.   Number of queues is independent.
> 
> Don't you mean to test against size of queue in DWs?
> So this for the check?
>    BUILD_BUG_ON(PF_QUEUE_NUM_DW % PF_MSG_LEN_DW);
> 

Yes. This the check I wanted. Thanks, will fix.

Matt

> 
> > 
> > Cc: Lucas De Marchi <lucas.demarchi@intel.com>
> > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > ---
> >   drivers/gpu/drm/xe/xe_gt_pagefault.c | 5 +++++
> >   1 file changed, 5 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c b/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > index 4489aadc7a52..16cea47d9e15 100644
> > --- a/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > +++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > @@ -328,6 +328,11 @@ int xe_guc_pagefault_handler(struct xe_guc *guc, u32 *msg, u32 len)
> >   	u32 asid;
> >   	bool full;
> > +	/*
> > +	 * The below logic doesn't work unless NUM_PF_QUEUE % PF_MSG_LEN_DW == 0
> > +	 */
> > +	BUILD_BUG_ON(NUM_PF_QUEUE % PF_MSG_LEN_DW);
> > +
> >   	if (unlikely(len != PF_MSG_LEN_DW))
> >   		return -EPROTO;