ext4: fix dirtyclusters double decrement on fs shutdown

fstests test generic/388 occasionally reproduces a warning in ext4_put_super() associated with the dirty clusters count:

  WARNING: CPU: 7 PID: 76064 at fs/ext4/super.c:1324 ext4_put_super+0x48c/0x590 [ext4]

Tracing the failure shows that the warning fires due to an s_dirtyclusters_counter value of -1. IOW, this appears to be a spurious decrement as opposed to some sort of leak. Further tracing of the dirty cluster count deltas and an LLM scan of the resulting output identified the cause as a double decrement in the error path between ext4_mb_mark_diskspace_used() and the caller ext4_mb_new_blocks().

First, note that generic/388 is a shutdown vs. fsstress test and so produces a random set of operations and shutdown injections. In the problematic case, the shutdown triggers an error return from the ext4_handle_dirty_metadata() call(s) made from ext4_mb_mark_context(). The changed value is non-zero at this point, so ext4_mb_mark_diskspace_used() does not exit after the error bubbles up from ext4_mb_mark_context(). Instead, the former decrements both cluster counters and returns the error up to ext4_mb_new_blocks(). The latter falls into the !ar->len out path which decrements the dirty clusters counter a second time, creating the inconsistency.

To avoid this problem and simplify ownership of the cluster reservation in this codepath, lift the counter reduction to a single place in the caller. This makes it more clear that ext4_mb_new_blocks() is responsible for acquiring cluster reservation (via ext4_claim_free_clusters()) in the !delalloc case as well as releasing it, regardless of whether it ends up consumed or returned due to failure.

Fixes 0087d9fb3f29 ("ext4: Fix s_dirty_blocks_counter if block allocation failed with nodelalloc")

Signed-off-by Brian Foster <[email protected]>

---

v2:
- Condense counter update logic instead of modifying return flow.
- Added Fixes: tag.
v1: https://lore.kernel.org/linux-ext4/[email protected]/

 fs/ext4/mballoc.c | 21 +++++----------------
 1 file changed, 5 insertions(+), 16 deletions(-)

fs/ext4/mballoc.c +2 -10

index 56d50fd3310b..b31d7ddc52a9 100644
@@ -4185,8 +4185,7 @@ ext4_mb_mark_context(handle_t *handle, struct super_block *sb, bool state,
  * Returns 0 if success or error code
  */
 static noinline_for_stack int
-ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
-				handle_t *handle, unsigned int reserv_clstrs)
+ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, handle_t *handle)
 {
 	struct ext4_group_desc *gdp;
 	struct ext4_sb_info *sbi;

@@ -4241,13 +4240,6 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 	BUG_ON(changed != ac->ac_b_ex.fe_len);
 #endif
 	percpu_counter_sub(&sbi->s_freeclusters_counter, ac->ac_b_ex.fe_len);
-	/*
-	 * Now reduce the dirty block count also. Should not go negative
-	 */
-	if (!(ac->ac_flags & EXT4_MB_DELALLOC_RESERVED))
-		/* release all the reserved blocks if non delalloc */
-		percpu_counter_sub(&sbi->s_dirtyclusters_counter,
-				   reserv_clstrs);
 
 	return err;
 }

@@ -6332,7 +6324,7 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle,
 			ext4_mb_pa_put_free(ac);
 	}
 	if (likely(ac->ac_status == AC_STATUS_FOUND)) {
-		*errp = ext4_mb_mark_diskspace_used(ac, handle, reserv_clstrs);
+		*errp = ext4_mb_mark_diskspace_used(ac, handle);
 		if (*errp) {
 			ext4_discard_allocated_blocks(ac);
 			goto errout;

@@ -6363,12 +6355,9 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle,
 out:

fs/ext4/mballoc.c +1 -0

+		percpu_counter_sub(&sbi->s_dirtyclusters_counter, reserv_clstrs);
 
 	trace_ext4_allocate_blocks(ar, (unsigned long long)block);
 

Quoted 33 lines · Brian Foster

fstests test generic/388 occasionally reproduces a warning in ext4_put_super() associated with the dirty clusters count:

  WARNING: CPU: 7 PID: 76064 at fs/ext4/super.c:1324 ext4_put_super+0x48c/0x590 [ext4]

Tracing the failure shows that the warning fires due to an s_dirtyclusters_counter value of -1. IOW, this appears to be a spurious decrement as opposed to some sort of leak. Further tracing of the dirty cluster count deltas and an LLM scan of the resulting output identified the cause as a double decrement in the error path between ext4_mb_mark_diskspace_used() and the caller ext4_mb_new_blocks().

First, note that generic/388 is a shutdown vs. fsstress test and so produces a random set of operations and shutdown injections. In the problematic case, the shutdown triggers an error return from the ext4_handle_dirty_metadata() call(s) made from ext4_mb_mark_context(). The changed value is non-zero at this point, so ext4_mb_mark_diskspace_used() does not exit after the error bubbles up from ext4_mb_mark_context(). Instead, the former decrements both cluster counters and returns the error up to ext4_mb_new_blocks(). The latter falls into the !ar->len out path which decrements the dirty clusters counter a second time, creating the inconsistency.

To avoid this problem and simplify ownership of the cluster reservation in this codepath, lift the counter reduction to a single place in the caller. This makes it more clear that ext4_mb_new_blocks() is responsible for acquiring cluster reservation (via ext4_claim_free_clusters()) in the !delalloc case as well as releasing it, regardless of whether it ends up consumed or returned due to failure.

Quoted 2 lines

Fixes 0087d9fb3f29 ("ext4: Fix s_dirty_blocks_counter if block allocation failed with nodelalloc")

Signed-off-by Brian Foster <[email protected]>

Thanks for the patch.

However, the call site in test_mark_diskspace_used_range() missed the argument update, which triggered a Kernel Test Robot warning. Also, I have one nit below.

Quoted 55 lines

---

v2:
- Condense counter update logic instead of modifying return flow.
- Added Fixes: tag.
v1: https://lore.kernel.org/linux-ext4/[email protected]/

 fs/ext4/mballoc.c | 21 +++++----------------
 1 file changed, 5 insertions(+), 16 deletions(-)

fs/ext4/mballoc.c +2 -13

index 56d50fd3310b..b31d7ddc52a9 100644
@@ -4185,8 +4185,7 @@ ext4_mb_mark_context(handle_t *handle, struct super_block *sb, bool state,
  * Returns 0 if success or error code
  */
 static noinline_for_stack int
-ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
-				handle_t *handle, unsigned int reserv_clstrs)
+ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, handle_t *handle)
 {
 	struct ext4_group_desc *gdp;
 	struct ext4_sb_info *sbi;

@@ -4241,13 +4240,6 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 	BUG_ON(changed != ac->ac_b_ex.fe_len);
 #endif
 	percpu_counter_sub(&sbi->s_freeclusters_counter, ac->ac_b_ex.fe_len);
-	/*
-	 * Now reduce the dirty block count also. Should not go negative
-	 */
-	if (!(ac->ac_flags & EXT4_MB_DELALLOC_RESERVED))
-		/* release all the reserved blocks if non delalloc */
-		percpu_counter_sub(&sbi->s_dirtyclusters_counter,
-				   reserv_clstrs);
 
 	return err;
 }

@@ -6332,7 +6324,7 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle,
 			ext4_mb_pa_put_free(ac);
 	}
 	if (likely(ac->ac_status == AC_STATUS_FOUND)) {
-		*errp = ext4_mb_mark_diskspace_used(ac, handle, reserv_clstrs);
+		*errp = ext4_mb_mark_diskspace_used(ac, handle);
 		if (*errp) {
 			ext4_discard_allocated_blocks(ac);
 			goto errout;

@@ -6363,12 +6355,9 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle,
 out:
 	if (inquota && ar->len < inquota)
 		dquot_free_block(ar->inode, EXT4_C2B(sbi, inquota - ar->len));
-	if (!ar->len) {
-		if ((ar->flags & EXT4_MB_DELALLOC_RESERVED) == 0)
-			/* release all the reserved blocks if non delalloc */

fs/ext4/mballoc.c ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, · line 6359

Quoted 5 lines

fs/ext4/mballoc.c +2 -3

-			percpu_counter_sub(&sbi->s_dirtyclusters_counter,
-						reserv_clstrs);
-	}
+	/* release all the reserved blocks if non delalloc */
+	if ((ar->flags & EXT4_MB_DELALLOC_RESERVED) == 0)

Nit: It might be better to check if (reserv_clstrs) directly. It’s more straightforward and stays robust even if the flag logic changes later.

Quoted 5 lines

fs/ext4/mballoc.c +1 -0

+		percpu_counter_sub(&sbi->s_dirtyclusters_counter, reserv_clstrs);
 
 	trace_ext4_allocate_blocks(ar, (unsigned long long)block);