Skip to contentSkip to conversation

btrfs: ctree: Dump the leaf before BUG_ON()

PATCH In review Qu Wenruo 3 participants 4 messages Apr 25, 2019 Btrfs file system

lore
Share & tools
lore

Qu Wenruo

author
#0 lore raw

PATCH v3 1/1 btrfs: ctree: Dump the leaf before BUG_ON()

We have a user reporting BUG_ON() triggered in
btrfs_set_item_key_safe().

Let's dump the leaf content before triggering BUG_ON() so that we can have some clue on what's going wrong.
The output of tree locks should help us to debug such problem.

---
Changelog:
v2:
- Dump slot/old key/new key.
v3:
- Output message update, to avoid "on-disk" wording.
- Use BUG() to replace BUG_ON(1) to make clang quite.

 fs/btrfs/ctree.c | 24 ++++++++++++++++++++++--
 1 file changed, 22 insertions(+), 2 deletions(-)
fs/btrfs/ctree.c +22 -2
index 5116c2a1f0f9..5df76c17775a 100644
@@ -3185,11 +3185,31 @@ void btrfs_set_item_key_safe(struct btrfs_fs_info *fs_info,
 	slot = path->slots[0];
 	if (slot > 0) {
 		btrfs_item_key(eb, &disk_key, slot - 1);
-		BUG_ON(comp_keys(&disk_key, new_key) >= 0);
+		if (unlikely(comp_keys(&disk_key, new_key) >= 0)) {
+			btrfs_crit(fs_info,
+		"slot %u key (%llu %u %llu) new key (%llu %u %llu)",
+				   slot, btrfs_disk_key_objectid(&disk_key),
+				   btrfs_disk_key_type(&disk_key),
+				   btrfs_disk_key_offset(&disk_key),
+				   new_key->objectid, new_key->type,
+				   new_key->offset);
+			btrfs_print_leaf(eb);
+			BUG();
+		}
 	}
 	if (slot < btrfs_header_nritems(eb) - 1) {
 		btrfs_item_key(eb, &disk_key, slot + 1);
-		BUG_ON(comp_keys(&disk_key, new_key) <= 0);
+		if (unlikely(comp_keys(&disk_key, new_key) <= 0)) {
+			btrfs_crit(fs_info,
+		"slot %u key (%llu %u %llu) new key (%llu %u %llu)",
+				   slot, btrfs_disk_key_objectid(&disk_key),
+				   btrfs_disk_key_type(&disk_key),
+				   btrfs_disk_key_offset(&disk_key),
+				   new_key->objectid, new_key->type,
+				   new_key->offset);
+			btrfs_print_leaf(eb);
+			BUG();
+		}
 	}
 
 	btrfs_cpu_key_to_disk(&disk_key, new_key);
Signature · 1 line
2.21.0
#1 lore raw
Quoted 9 lines · Qu Wenruo

We have a user reporting BUG_ON() triggered in
btrfs_set_item_key_safe().

Let's dump the leaf content before triggering BUG_ON() so that we can have some clue on what's going wrong.
The output of tree locks should help us to debug such problem.

Filip -> Filipe

David can probably amend that at commit time.

Quoted 53 lines

---
Changelog:
v2:
- Dump slot/old key/new key.
v3:
- Output message update, to avoid "on-disk" wording.
- Use BUG() to replace BUG_ON(1) to make clang quite.

 fs/btrfs/ctree.c | 24 ++++++++++++++++++++++--
 1 file changed, 22 insertions(+), 2 deletions(-)
fs/btrfs/ctree.c +22 -3
index 5116c2a1f0f9..5df76c17775a 100644
@@ -3185,11 +3185,31 @@ void btrfs_set_item_key_safe(struct btrfs_fs_info *fs_info,
        slot = path->slots[0];
        if (slot > 0) {
                btrfs_item_key(eb, &disk_key, slot - 1);
-               BUG_ON(comp_keys(&disk_key, new_key) >= 0);
+               if (unlikely(comp_keys(&disk_key, new_key) >= 0)) {
+                       btrfs_crit(fs_info,
+               "slot %u key (%llu %u %llu) new key (%llu %u %llu)",
+                                  slot, btrfs_disk_key_objectid(&disk_key),
+                                  btrfs_disk_key_type(&disk_key),
+                                  btrfs_disk_key_offset(&disk_key),
+                                  new_key->objectid, new_key->type,
+                                  new_key->offset);
+                       btrfs_print_leaf(eb);
+                       BUG();
+               }
        }
        if (slot < btrfs_header_nritems(eb) - 1) {
                btrfs_item_key(eb, &disk_key, slot + 1);
-               BUG_ON(comp_keys(&disk_key, new_key) <= 0);
+               if (unlikely(comp_keys(&disk_key, new_key) <= 0)) {
+                       btrfs_crit(fs_info,
+               "slot %u key (%llu %u %llu) new key (%llu %u %llu)",
+                                  slot, btrfs_disk_key_objectid(&disk_key),
+                                  btrfs_disk_key_type(&disk_key),
+                                  btrfs_disk_key_offset(&disk_key),
+                                  new_key->objectid, new_key->type,
+                                  new_key->offset);
+                       btrfs_print_leaf(eb);
+                       BUG();
+               }
        }
 
        btrfs_cpu_key_to_disk(&disk_key, new_key);
--

2.21.0

Signature · 3 lines
Filipe David Manana,

“Whether you think you can, or you think you can't — you're right.”
#2 lore raw
In reply to Qu Wenruo
Quoted 29 lines · Qu Wenruo

We have a user reporting BUG_ON() triggered in
btrfs_set_item_key_safe().

Let's dump the leaf content before triggering BUG_ON() so that we can have some clue on what's going wrong.
The output of tree locks should help us to debug such problem.

---
Changelog:
v2:
- Dump slot/old key/new key.
v3:
- Output message update, to avoid "on-disk" wording.
- Use BUG() to replace BUG_ON(1) to make clang quite.

 fs/btrfs/ctree.c | 24 ++++++++++++++++++++++--
 1 file changed, 22 insertions(+), 2 deletions(-)
fs/btrfs/ctree.c +0 -1
index 5116c2a1f0f9..5df76c17775a 100644
@@ -3185,11 +3185,31 @@ void btrfs_set_item_key_safe(struct btrfs_fs_info *fs_info,
 	slot = path->slots[0];
 	if (slot > 0) {
 		btrfs_item_key(eb, &disk_key, slot - 1);
-		BUG_ON(comp_keys(&disk_key, new_key) >= 0);

fs/btrfs/ctree.c void btrfs_set_item_key_safe(struct btrfs_fs_info *fs_info, · line 3197

Quoted 10 lines
fs/btrfs/ctree.c +10 -0
+		if (unlikely(comp_keys(&disk_key, new_key) >= 0)) {
+			btrfs_crit(fs_info,
+		"slot %u key (%llu %u %llu) new key (%llu %u %llu)",
+				   slot, btrfs_disk_key_objectid(&disk_key),
+				   btrfs_disk_key_type(&disk_key),
+				   btrfs_disk_key_offset(&disk_key),
+				   new_key->objectid, new_key->type,
+				   new_key->offset);
+			btrfs_print_leaf(eb);
+			BUG();

This is going to produce a lot of lines, with the stacktrace at the end. We'll understand what's wrong just by seeing the stack, though the contents of the leaf will be only in the logs and not on screen. This hopefully will be available for analysis. I've seen this bug too and will be able to provide the logs eventually.

#3 lore raw
In reply to Qu Wenruo
Quoted 9 lines · Qu Wenruo

We have a user reporting BUG_ON() triggered in
btrfs_set_item_key_safe().

Let's dump the leaf content before triggering BUG_ON() so that we can have some clue on what's going wrong.
The output of tree locks should help us to debug such problem.

Quoted 7 lines

---
Changelog:
v2:
- Dump slot/old key/new key.
v3:
- Output message update, to avoid "on-disk" wording.
- Use BUG() to replace BUG_ON(1) to make clang quite.

Name fixed and I've added a sample stacktrace that I had in the logs. Patch added to misc-next, thanks.