On Tue 11-04-23 11:37:26, Carlos Maiolino wrote:
> On Wed, Apr 05, 2023 at 01:42:45PM +0200, Jan Kara wrote:
> > > @@ -3763,6 +3878,9 @@ static void shmem_put_super(struct super_block *sb)
> > > {
> > > struct shmem_sb_info *sbinfo = SHMEM_SB(sb);
> > >
> > > +#ifdef CONFIG_TMPFS_QUOTA
> > > + shmem_disable_quotas(sb);
> > > +#endif
> > > free_percpu(sbinfo->ino_batch);
> > > percpu_counter_destroy(&sbinfo->used_blocks);
> > > mpol_put(sbinfo->mpol);
> > > @@ -3841,6 +3959,17 @@ static int shmem_fill_super(struct super_block *sb, struct fs_context *fc)
> > > #endif
> > > uuid_gen(&sb->s_uuid);
> > >
> > > +#ifdef CONFIG_TMPFS_QUOTA
> > > + if (ctx->seen & SHMEM_SEEN_QUOTA) {
> > > + sb->dq_op = &shmem_quota_operations;
> > > + sb->s_qcop = &dquot_quotactl_sysfile_ops;
> > > + sb->s_quota_types = QTYPE_MASK_USR | QTYPE_MASK_GRP;
> >
> > s_quota_types should rather be copied from ctx, shouldn't it? Or why is
> > s_quota_types inconsistent with ctx->quota_types?
>
> I believe s_qupta_types here is a bitmask of supported quota types, while
> ctx->quota_types refers to the mount options being passed from the user.
>
> So we should enable in sb->s_quota_types which quota types the filesystem
> supports, not which were enabled by the user.
Oh, right.
Honza
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
