Hello Carlos! On Tue 04-04-23 15:48:36, Carlos Maiolino wrote: > > > + if (!dquot->dq_dqb.dqb_bhardlimit && > > > + !dquot->dq_dqb.dqb_bsoftlimit && > > > + !dquot->dq_dqb.dqb_ihardlimit && > > > + !dquot->dq_dqb.dqb_isoftlimit) > > > + set_bit(DQ_FAKE_B, &dquot->dq_flags); > > > + spin_unlock(&dquot->dq_dqb_lock); > > > + > > > + /* Make sure flags update is visible after dquot has been filled */ > > > + smp_mb__before_atomic(); > > > + set_bit(DQ_ACTIVE_B, &dquot->dq_flags); > > > > I'm slightly wondering whether we shouldn't have a dquot_mark_active() > > helper for this to hide the barrier details... > > This sounds good to me, would be ok for you if I simply add this to my todo > list, and do it once this series is merged? I'd rather avoid to add more patches > to the series now adding more review overhead. > I can send a new patch later creating a new helper and replacing all > set_bit(DQ_ACTIVE_B, ...) calls with the new helper. Yes, sounds fine to me. Honza -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR
