Re: [PATCH v4] selftests/mm: skip hugetlb_dio tests when DIO alignment is incompatible

[Li Wang <liwang@redhat.com>]

On Wed, Apr 01, 2026 at 10:34:40AM +0200, David Hildenbrand (Arm) wrote:
> On 4/1/26 02:39, Li Wang wrote:
> >>> +static void run_test(int fd, unsigned int start_off,
> >>> +		     unsigned int end_off, unsigned int align)
> >>
> >> Nit: prefer to-tab alignment in MM land.
> > 
> > Yes, I noticed Andrew (thanks!) corrected it and added to -mm mm-new branch.
> > 
> >>> +{
> >>> +	if (!check_dio_alignment(start_off, end_off, align))
> >>> +		return;
> >>
> >> Is there a reason we just perform that at the beginning of
> >> run_dio_using_hugetlb(), avoiding run_test() entirely?
> > 
> > That will make the run_dio_using_hugetlb() have too many args and
> > uncomfortable to read, yes, it's more likely personal preference.
> 
> Nah, just squash it. There is no need for this separate function, really.

Okay, let me do it in patch v5.

-- 
Regards,
Li Wang