btrfs: reject root items with drop_progress and zero drop_level

PATCH In review ZhengYuan Huang 2 participants 6 messages Mar 12

Subsystems (2)

Stable branch Btrfs file system

lore

Share & tools

lore

ZhengYuan Huang

author

2026-03-12 00:14 UTC #0 lore raw

PATCH v2 1/1 btrfs: reject root items with drop_progress and zero drop_level

[BUG]
When recovering relocation at mount time, merge_reloc_root() and btrfs_drop_snapshot() both use BUG_ON(level == 0) to guard against an impossible state: a non-zero drop_progress combined with a zero drop_level in a root_item, which can be triggered:

------------[ cut here ]------------
kernel BUG at fs/btrfs/relocation.c:1545!
Oops: invalid opcode: 0000 [#1] SMP KASAN NOPTI
CPU: 1 UID: 0 PID: 283 ... Tainted: 6.18.0+ #16 PREEMPT(voluntary) 
Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
Hardware name: QEMU Ubuntu 24.04 PC v2, BIOS 1.16.3-debian-1.16.3-2
RIP: 0010:merge_reloc_root+0x1266/0x1650 fs/btrfs/relocation.c:1545
Code: ffff0000 00004589 d7e9acfa ffffe8a1 79bafebe 02000000
Call Trace:
 merge_reloc_roots+0x295/0x890 fs/btrfs/relocation.c:1861
 btrfs_recover_relocation+0xd6e/0x11d0 fs/btrfs/relocation.c:4195
 btrfs_start_pre_rw_mount+0xa4d/0x1810 fs/btrfs/disk-io.c:3130
 open_ctree+0x5824/0x5fe0 fs/btrfs/disk-io.c:3640
 btrfs_fill_super fs/btrfs/super.c:987 [inline]
 btrfs_get_tree_super fs/btrfs/super.c:1951 [inline]
 btrfs_get_tree_subvol fs/btrfs/super.c:2094 [inline]
 btrfs_get_tree+0x111c/0x2190 fs/btrfs/super.c:2128
 vfs_get_tree+0x9a/0x370 fs/super.c:1758
 fc_mount fs/namespace.c:1199 [inline]
 do_new_mount_fc fs/namespace.c:3642 [inline]
 do_new_mount fs/namespace.c:3718 [inline]
 path_mount+0x5b8/0x1ea0 fs/namespace.c:4028
 do_mount fs/namespace.c:4041 [inline]
 __do_sys_mount fs/namespace.c:4229 [inline]
 __se_sys_mount fs/namespace.c:4206 [inline]
 __x64_sys_mount+0x282/0x320 fs/namespace.c:4206
 ...
RIP: 0033:0x7f969c9a8fde
Code: 0f1f4000 48c7c2b0 fffffff7 d8648902 b8ffffff ffc3660f
---[ end trace 0000000000000000 ]---

[CAUSE]
A non-zero drop_progress.objectid means an interrupted btrfs_drop_snapshot() left a resume point on disk, and in that case drop_level must be greater than 0 because the checkpoint is only saved at internal node levels.

Although this invariant is enforced when the kernel writes the root item, it is not validated when the root item is read back from disk. That allows on-disk corruption to provide an invalid state with drop_progress.objectid != 0 and drop_level == 0.

When relocation recovery later processes such a root item, merge_reloc_root() reads drop_level and hits BUG_ON(level == 0). The same invalid metadata can also trigger the corresponding BUG_ON() in btrfs_drop_snapshot().

[FIX]
Fix this by validating the root_item invariant in tree-checker when reading root items from disk: if drop_progress.objectid is non-zero, drop_level must also be non-zero. Reject such malformed metadata with -EUCLEAN before it reaches merge_reloc_root() or btrfs_drop_snapshot() and triggers the BUG_ON.

The bug is reproducible on 7.0.0-rc2-next-20260310 with our dynamic metadata fuzzing tool that corrupts btrfs metadata at runtime. After the fix, the same corruption is correctly rejected by tree-checker and the BUG_ON is no longer triggered.

Fixes 259ee7754b67 ("btrfs: tree-checker: Add ROOT_ITEM check")

Cc [email protected] # 5.3+

Signed-off-by ZhengYuan Huang <[email protected]>

---
[CHANGELOG]
v2:
- Split out the error message fix from the previous patch, as requested during review.

---
 fs/btrfs/tree-checker.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

fs/btrfs/tree-checker.c +17 -0

index dd274f67ad7f..1e052c3303b3 100644
@@ -1260,6 +1260,23 @@ static int check_root_item(struct extent_buffer *leaf, struct btrfs_key *key,
 			    btrfs_root_drop_level(&ri), BTRFS_MAX_LEVEL - 1);
 		return -EUCLEAN;
 	}
+	/*
+	 * If drop_progress.objectid is non-zero, a btrfs_drop_snapshot() was
+	 * interrupted and the resume point was recorded in drop_progress and
+	 * drop_level.  In that case drop_level must be >= 1: level 0 is the
+	 * leaf level and drop_snapshot never saves a checkpoint there (it
+	 * only records checkpoints at internal node levels in DROP_REFERENCE
+	 * stage).  A zero drop_level combined with a non-zero drop_progress
+	 * objectid indicates on-disk corruption and would cause a BUG_ON in
+	 * merge_reloc_root() and btrfs_drop_snapshot() at mount time.
+	 */
+	if (unlikely(btrfs_disk_key_objectid(&ri.drop_progress) != 0 &&
+		     btrfs_root_drop_level(&ri) == 0)) {
+		generic_err(leaf, slot,
+			    "invalid root drop_level 0 with non-zero drop_progress objectid %llu",
+			    btrfs_disk_key_objectid(&ri.drop_progress));
+		return -EUCLEAN;
+	}
 
 	/* Flags check */
 	if (unlikely(btrfs_root_flags(&ri) & ~valid_root_flags)) {

Signature · 1 line

2.43.0

Qu Wenruo

Reviewed-by on v2 1/1

2026-03-12 00:23 UTC #1 lore raw

Quoted 60 lines · ZhengYuan Huang

[BUG]
When recovering relocation at mount time, merge_reloc_root() and btrfs_drop_snapshot() both use BUG_ON(level == 0) to guard against an impossible state: a non-zero drop_progress combined with a zero drop_level in a root_item, which can be triggered:
------------[ cut here ]------------
kernel BUG at fs/btrfs/relocation.c:1545!
Oops: invalid opcode: 0000 [#1] SMP KASAN NOPTI
CPU: 1 UID: 0 PID: 283 ... Tainted: 6.18.0+ #16 PREEMPT(voluntary)
Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
Hardware name: QEMU Ubuntu 24.04 PC v2, BIOS 1.16.3-debian-1.16.3-2
RIP: 0010:merge_reloc_root+0x1266/0x1650 fs/btrfs/relocation.c:1545
Code: ffff0000 00004589 d7e9acfa ffffe8a1 79bafebe 02000000
Call Trace:
  merge_reloc_roots+0x295/0x890 fs/btrfs/relocation.c:1861
  btrfs_recover_relocation+0xd6e/0x11d0 fs/btrfs/relocation.c:4195
  btrfs_start_pre_rw_mount+0xa4d/0x1810 fs/btrfs/disk-io.c:3130
  open_ctree+0x5824/0x5fe0 fs/btrfs/disk-io.c:3640
  btrfs_fill_super fs/btrfs/super.c:987 [inline]
  btrfs_get_tree_super fs/btrfs/super.c:1951 [inline]
  btrfs_get_tree_subvol fs/btrfs/super.c:2094 [inline]
  btrfs_get_tree+0x111c/0x2190 fs/btrfs/super.c:2128
  vfs_get_tree+0x9a/0x370 fs/super.c:1758
  fc_mount fs/namespace.c:1199 [inline]
  do_new_mount_fc fs/namespace.c:3642 [inline]
  do_new_mount fs/namespace.c:3718 [inline]
  path_mount+0x5b8/0x1ea0 fs/namespace.c:4028
  do_mount fs/namespace.c:4041 [inline]
  __do_sys_mount fs/namespace.c:4229 [inline]
  __se_sys_mount fs/namespace.c:4206 [inline]
  __x64_sys_mount+0x282/0x320 fs/namespace.c:4206
  ...
RIP: 0033:0x7f969c9a8fde
Code: 0f1f4000 48c7c2b0 fffffff7 d8648902 b8ffffff ffc3660f
---[ end trace 0000000000000000 ]---
[CAUSE]
A non-zero drop_progress.objectid means an interrupted btrfs_drop_snapshot() left a resume point on disk, and in that case drop_level must be greater than 0 because the checkpoint is only saved at internal node levels.
Although this invariant is enforced when the kernel writes the root item, it is not validated when the root item is read back from disk. That allows on-disk corruption to provide an invalid state with drop_progress.objectid != 0 and drop_level == 0.
When relocation recovery later processes such a root item, merge_reloc_root() reads drop_level and hits BUG_ON(level == 0). The same invalid metadata can also trigger the corresponding BUG_ON() in btrfs_drop_snapshot().
[FIX]
Fix this by validating the root_item invariant in tree-checker when reading root items from disk: if drop_progress.objectid is non-zero, drop_level must also be non-zero. Reject such malformed metadata with -EUCLEAN before it reaches merge_reloc_root() or btrfs_drop_snapshot() and triggers the BUG_ON.

Quoted 2 lines

The bug is reproducible on 7.0.0-rc2-next-20260310 with our dynamic metadata fuzzing tool that corrupts btrfs metadata at runtime.

I'll move this part into the BUG section as it's about how the bug is triggered.

Quoted 5 lines

After
the fix, the same corruption is correctly rejected by tree-checker and the BUG_ON is no longer triggered.
Fixes 259ee7754b67 ("btrfs: tree-checker: Add ROOT_ITEM check")

Again, it's not a bug fix. You're just adding a new check.

This fixes tag should only go with the error message fix.

You don't need to send a new update, I'll do all the update at merge time.

Otherwise looks good to me.

Reviewed-by Qu Wenruo <[email protected]>

Thanks,
Qu

Quoted 39 lines

Cc [email protected] # 5.3+

Signed-off-by ZhengYuan Huang <[email protected]>

---
[CHANGELOG]
v2:
- Split out the error message fix from the previous patch, as requested during review.

---
  fs/btrfs/tree-checker.c | 17 +++++++++++++++++
  1 file changed, 17 insertions(+)

fs/btrfs/tree-checker.c +17 -0

index dd274f67ad7f..1e052c3303b3 100644
@@ -1260,6 +1260,23 @@ static int check_root_item(struct extent_buffer *leaf, struct btrfs_key *key,
  			    btrfs_root_drop_level(&ri), BTRFS_MAX_LEVEL - 1);
  		return -EUCLEAN;
  	}
+	/*
+	 * If drop_progress.objectid is non-zero, a btrfs_drop_snapshot() was
+	 * interrupted and the resume point was recorded in drop_progress and
+	 * drop_level.  In that case drop_level must be >= 1: level 0 is the
+	 * leaf level and drop_snapshot never saves a checkpoint there (it
+	 * only records checkpoints at internal node levels in DROP_REFERENCE
+	 * stage).  A zero drop_level combined with a non-zero drop_progress
+	 * objectid indicates on-disk corruption and would cause a BUG_ON in
+	 * merge_reloc_root() and btrfs_drop_snapshot() at mount time.
+	 */
+	if (unlikely(btrfs_disk_key_objectid(&ri.drop_progress) != 0 &&
+		     btrfs_root_drop_level(&ri) == 0)) {
+		generic_err(leaf, slot,
+			    "invalid root drop_level 0 with non-zero drop_progress objectid %llu",
+			    btrfs_disk_key_objectid(&ri.drop_progress));
+		return -EUCLEAN;
+	}
 
  	/* Flags check */
  	if (unlikely(btrfs_root_flags(&ri) & ~valid_root_flags)) {

Qu Wenruo

2026-03-12 01:07 UTC #2 lore raw

在 2026/3/12 10:53, Qu Wenruo 写道:
[...]

Quoted 4 lines

Quoted 2 lines
Fixes 259ee7754b67 ("btrfs: tree-checker: Add ROOT_ITEM check")
Again, it's not a bug fix. You're just adding a new check.

More info about when to use fixes tag:

https://www.kernel.org/doc/html/v6.19/process/submitting-patches.html#using-reported-by-tested-by-reviewed-by-suggested-by-and-fixes

It's more common to use it for a regression, aka, before that commit everything works, but at that commit something is broken.

And in your case, the commit is not causing the BUG_ON() thus it's incorrect. Furthermore you'd better not put the commit introducing the BUG_ON() as the fixes target.

In your case, you're just enhancing the tree-checker to address a fuzzed image (which I guess you'll continue submitting such patches), thus getting the fixes tag done correctly will save everyone time.

Thanks,
Qu

Quoted 57 lines

This fixes tag should only go with the error message fix.
You don't need to send a new update, I'll do all the update at merge time.
Otherwise looks good to me.
Reviewed-by Qu Wenruo <[email protected]>
Thanks,
Qu
Quoted 43 lines
Cc [email protected] # 5.3+
Signed-off-by ZhengYuan Huang <[email protected]>
---
[CHANGELOG]
v2:
- Split out the error message fix from the previous patch, as requested during review.
---
fs/btrfs/tree-checker.c | 17 +++++++++++++++++
  1 file changed, 17 insertions(+)
fs/btrfs/tree-checker.c +2 -0
index dd274f67ad7f..1e052c3303b3 100644
@@ -1260,6 +1260,23 @@ static int check_root_item(struct extent_buffer 
 *leaf, struct btrfs_key *key,
                   btrfs_root_drop_level(&ri), BTRFS_MAX_LEVEL - 1);
           return -EUCLEAN;
       }
+    /*
+     * If drop_progress.objectid is non-zero, a btrfs_drop_snapshot() 
was
+     * interrupted and the resume point was recorded in drop_progress and
+     * drop_level. In that case drop_level must be >= 1: level 0 is the +     * leaf level and drop_snapshot never saves a checkpoint there (it +     * only records checkpoints at internal node levels in DROP_REFERENCE
+     * stage). A zero drop_level combined with a non-zero drop_progress +     * objectid indicates on-disk corruption and would cause a BUG_ON in +     * merge_reloc_root() and btrfs_drop_snapshot() at mount time. +     */
+    if (unlikely(btrfs_disk_key_objectid(&ri.drop_progress) != 0 && +             btrfs_root_drop_level(&ri) == 0)) {
+        generic_err(leaf, slot,
+                "invalid root drop_level 0 with non-zero drop_progress objectid %llu",
+                btrfs_disk_key_objectid(&ri.drop_progress)); +        return -EUCLEAN;
+    }
      /* Flags check */
      if (unlikely(btrfs_root_flags(&ri) & ~valid_root_flags)) {

ZhengYuan Huang

author

2026-03-12 02:10 UTC #3 lore raw

Thanks a lot for the explanation. I think I now understand the intended use of the
Fixes: tag.

I still have one question, though. In this case, scripts/checkpatch.pl warns that a
Fixes: tag should be added, which seems inconsistent with the submission guidelines you pointed me to.

My understanding had been that patches should generally be sent only after passing
checkpatch.pl cleanly, so I wanted to ask: is this kind of warning acceptable in practice, or does it mean my local checkpatch.pl is outdated? If it is outdated, should I generally use the latest checkpatch.pl when checking patches before submission?

Thanks again,
ZhengYuan Huang

Qu Wenruo

2026-03-12 04:16 UTC #4 lore raw

Quoted 8 lines · ZhengYuan Huang

Thanks a lot for the explanation. I think I now understand the intended use of the
Fixes: tag.
I still have one question, though. In this case, scripts/checkpatch.pl warns that a
Fixes: tag should be added, which seems inconsistent with the submission guidelines you pointed me to.

If you check the checkpatch.pl itself, the logic in it is pretty simple and can give false alerts.

It just checks if the commit message has BUG: or KASAN/UBSAN lines.

So it's false alert prune.

Quoted 4 lines

My understanding had been that patches should generally be sent only after passing
checkpatch.pl cleanly,

No, that is only a script which has its limits.
Checkpatch is good for its code style checks, but not always correct on other suggestions.

Sometimes even its code style checks may conflict with the rules inside each subsystem.

Quoted 2 lines

so I wanted to ask: is this kind of warning acceptable in practice,

Yes, unless you believe a simple perl script can be as good as human common sense.

Quoted 1 line

or does it mean my local checkpatch.pl is outdated?

Since you're already using the latest rc kernel, I believe you're already using the latest checkpatch.

Thanks,
Qu

Quoted 6 lines

If it is outdated,
should I generally use the latest checkpatch.pl when checking patches before submission?
Thanks again,
ZhengYuan Huang

ZhengYuan Huang

author

2026-03-12 04:27 UTC #5 lore raw

Quoted 10 lines · Qu Wenruo

Quoted 3 lines
My understanding had been that patches should generally be sent only after passing
checkpatch.pl cleanly,
No, that is only a script which has its limits.
Checkpatch is good for its code style checks, but not always correct on other suggestions.
Sometimes even its code style checks may conflict with the rules inside each subsystem.

Thanks for the clarification.

Understood, that makes sense. I had been treating checkpatch.pl warnings too strictly. I'll follow the documented guidelines and subsystem expectations instead.

Thanks again,
ZhengYuan Huang