Skip to contentSkip to conversation

ovl: make i_ino consistent with st_ino in more cases

PATCH In review Amir Goldstein 2 participants 4 messages Jun 9, 2019 – Jun 19, 2019 Overlay filesystem

lore
Share & tools
lore

Amir Goldstein

author maintainer
#0 lore raw

PATCH 1/1 ovl: make i_ino consistent with st_ino in more cases

Relax the condition that overlayfs supports nfs export, to require that i_ino is consistent with st_ino/d_ino.

It is enough to require that st_ino and d_ino are consistent.

This fixes the failure of xfstest generic/504, due to mismatch of st_ino to inode number in the output of /proc/locks.

Fixes 12574a9f4c9c ("ovl: consistent i_ino for non-samefs with xino")
Cc <[email protected]> # v4.19

---

Miklos,

TBH, I can't remember why I put the s_export_op condition in the first place (can you?), expect for the fact the original bug fix was reported on nfsd readdirplus.

Thanks,
Amir.

 fs/overlayfs/inode.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
fs/overlayfs/inode.c +6 -6
index f7eba21effa5..f0389849fd80 100644
@@ -553,15 +553,15 @@ static void ovl_fill_inode(struct inode *inode, umode_t mode, dev_t rdev,
 	int xinobits = ovl_xino_bits(inode->i_sb);
 
 	/*
-	 * When NFS export is enabled and d_ino is consistent with st_ino
-	 * (samefs or i_ino has enough bits to encode layer), set the same
-	 * value used for d_ino to i_ino, because nfsd readdirplus compares
-	 * d_ino values to i_ino values of child entries. When called from
+	 * When d_ino is consistent with st_ino (samefs or i_ino has enough
+	 * bits to encode layer), set the same value used for st_ino to i_ino,
+	 * so inode number exposed via /proc/locks and a like will be
+	 * consistent with d_ino and st_ino values. An i_ino value inconsistent
+	 * with d_ino also causes nfsd readdirplus to fail.  When called from
 	 * ovl_new_inode(), ino arg is 0, so i_ino will be updated to real
 	 * upper inode i_ino on ovl_inode_init() or ovl_inode_update().
 	 */
-	if (inode->i_sb->s_export_op &&
-	    (ovl_same_sb(inode->i_sb) || xinobits)) {
+	if (ovl_same_sb(inode->i_sb) || xinobits) {
 		inode->i_ino = ino;
 		if (xinobits && fsid && !(ino >> (64 - xinobits)))
 			inode->i_ino |= (unsigned long)fsid << (64 - xinobits);
Signature · 1 line
2.17.1
maintainer
#1 lore raw
Quoted 5 lines · Amir Goldstein

Relax the condition that overlayfs supports nfs export, to require that i_ino is consistent with st_ino/d_ino.

It is enough to require that st_ino and d_ino are consistent.

Yes.

Quoted 7 lines

This fixes the failure of xfstest generic/504, due to mismatch of st_ino to inode number in the output of /proc/locks.

Fixes 12574a9f4c9c ("ovl: consistent i_ino for non-samefs with xino")
Cc <[email protected]> # v4.19
Quoted 7 lines

---

Miklos,

TBH, I can't remember why I put the s_export_op condition in the first place (can you?), expect for the fact the original bug fix was reported on nfsd readdirplus.

I have no recollection of any issue that would require nfs export.

I've applied the patch.

Thanks,
Miklos

author maintainer
#2 lore raw
Quoted 21 lines · Miklos Szeredi
Quoted 5 lines · Amir Goldstein

Relax the condition that overlayfs supports nfs export, to require that i_ino is consistent with st_ino/d_ino.

It is enough to require that st_ino and d_ino are consistent.

Yes.

Quoted 11 lines

This fixes the failure of xfstest generic/504, due to mismatch of st_ino to inode number in the output of /proc/locks.

Fixes 12574a9f4c9c ("ovl: consistent i_ino for non-samefs with xino")
Cc <[email protected]> # v4.19

---

Miklos,

Quoted 8 lines
Quoted 3 lines

TBH, I can't remember why I put the s_export_op condition in the first place (can you?), expect for the fact the original bug fix was reported on nfsd readdirplus.

I have no recollection of any issue that would require nfs export.

I've applied the patch.

Forgot to apply to overlayfs-next or intentionally left for 5.3?

Thanks,
Amir.

maintainer
#3 lore raw
Quoted 9 lines · Amir Goldstein
Quoted 7 lines · Miklos Szeredi
Quoted 3 lines · Amir Goldstein

TBH, I can't remember why I put the s_export_op condition in the first place (can you?), expect for the fact the original bug fix was reported on nfsd readdirplus.

Quoted 6 lines
Quoted 4 lines

I have no recollection of any issue that would require nfs export.

I've applied the patch.

Forgot to apply to overlayfs-next or intentionally left for 5.3?

Safer to leave it to 5.3 since this is not a recent regression and the only reporter is xfstests.

Thanks,
Miklos