> > Sent: Monday, March 4, 2024 12:43 PM
> > >
> > > This patch set is intended to improve the performance of sync
> > > dentry, I don't think it is a good idea to change other logic in this
> patch set.
> > Yeah, as you said, this patch set should keep the original logic
> > except for the sync related parts. The reason I left a review comment
> > is because the code before this patch set allows deleted dentries to
> > follow unused dentries.
>
> Which commit changed to allow deleted dentries to follow unused dentries?
Not changed. It has been allowed from the initial commit as follows.
5f2aa075070c ("exfat: add inode operations")
>
> The following code still exists if without this patch set. It does not
> allow deleted dentries to follow unused dentries.
It may be the same part as the code you mentioned, but remember that
the first if-statement handles both an unused dentry and
a deleted dentry together.
static int exfat_search_empty_slot(...)
{
...
if (type == TYPE_UNUSED || type == TYPE_DELETED) {
...
} else {
if (hint_femp->eidx != EXFAT_HINT_NONE &&
hint_femp->count == CNT_UNUSED_HIT) {
/* unused empty group means
* an empty group which includes
* unused dentry
*/
exfat_fs_error(sb,
"found bogus dentry(%d) beyond unused empty group(%d) (start_clu : %u, cur_clu : %u)",
dentry, hint_femp->eidx,
p_dir->dir, clu.dir);
return -EIO;
}
...
}
...
}
>
> > Please let me know if I missed anything.
> >
> > > Patch [7/10] moves the check from exfat_search_empty_slot() to
> > > exfat_validate_empty_dentry_set().
> > >
> > > - if (hint_femp->eidx !=
> > EXFAT_HINT_NONE &&
> > > - hint_femp->count ==
> > CNT_UNUSED_HIT) {
> > > - /* unused empty group
> > means
> > > - * an empty group
> > which includes
> > > - * unused dentry
> > > - */
> > > - exfat_fs_error(sb,
> > > - "found bogus
> > dentry(%d) beyond
> > > unused empty group(%d) (start_clu : %u, cur_clu : %u)",
> > > - dentry,
> > hint_femp->eidx,
> > > - p_dir->dir,
> > clu.dir);
> > >
