Regression on linux-next (next-20240228)
Borah, Chaitanya Kumar
chaitanya.kumar.borah at intel.com
Tue Mar 5 06:49:16 UTC 2024
Hello Mathew,
> -----Original Message-----
> From: Borah, Chaitanya Kumar
> Sent: Monday, March 4, 2024 8:18 PM
> To: Matthew Wilcox <willy at infradead.org>
> Cc: intel-gfx at lists.freedesktop.org; Kurmi, Suresh Kumar
> <Suresh.Kumar.Kurmi at intel.com>; Saarinen, Jani <jani.saarinen at intel.com>
> Subject: RE: Regression on linux-next (next-20240228)
>
> Hello Mathew,
>
> > -----Original Message-----
> > From: Matthew Wilcox <willy at infradead.org>
> > Sent: Monday, March 4, 2024 6:52 PM
> > To: Borah, Chaitanya Kumar <chaitanya.kumar.borah at intel.com>
> > Cc: intel-gfx at lists.freedesktop.org; Kurmi, Suresh Kumar
> > <suresh.kumar.kurmi at intel.com>; Saarinen, Jani
> > <jani.saarinen at intel.com>
> > Subject: Re: Regression on linux-next (next-20240228)
> >
> > On Mon, Mar 04, 2024 at 10:03:13AM +0000, Borah, Chaitanya Kumar
> > wrote:
> > > > Could you try putting the two:
> > > >
> > > > - list_del(&folio->lru);
> > > >
> > > > statements back in and see if that fixes it?
> > >
> > > That seems to fix it.
> > >
> > > if (!folio_put_testzero(folio))
> > > + list_del(&folio->lru);
> > > continue;
> >
> > Ummm ... did you put { and } around this? Otherwise the indentation
> > is misleading and what you're actually done is:
> >
> > if (!folio_put_testzero(folio))
> > list_del(&folio->lru);
> > continue;
> >
> > which will simply leak memory.
> >
>
> Oops look like a miss on our side. Let us re-do this and get back to you.
>
Issue is still seen with the following changes
void put_pages_list(struct list_head *pages)
folio_batch_init(&fbatch);
list_for_each_entry(folio, pages, lru) {
- if (!folio_put_testzero(folio))
+ if (!folio_put_testzero(folio)) {
+ list_del(&folio->lru);
continue;
+ }
if (folio_test_large(folio)) {
__folio_put_large(folio);
+ list_del(&folio->lru);
continue;
}
Regards
Chaitanya
> Regards
>
> Chaitanya
>
> > > if (folio_test_large(folio)) {
> > > __folio_put_large(folio);
> > > + list_del(&folio->lru);
> > > continue;
> > > }
> > > Regards
> > >
> > > Chaitanya
More information about the Intel-gfx
mailing list