Skip to contentSkip to conversation

btrfs: scrub: Fix use of uninitialized variable

PATCH In review 'Guanjun' 2 participants 2 messages Dec 9, 2023 Btrfs file system

lore
Share & tools
lore

'Guanjun'

author
#0 lore raw

PATCH 1/1 btrfs: scrub: Fix use of uninitialized variable

From: Guanjun <[email protected]>

'ret' will be uninitialized in case that the logical_length is 0. Even if the caller has already ensured that logical_length is greater than 0, we still need to fix this issue. Due to the compiler may complain like this:

fs/btrfs/scrub.c: In function ‘scrub_simple_mirror.constprop’:
fs/btrfs/scrub.c:2123:9: error: ‘ret’ may be used uninitialized in this function [-Werror=maybe-uninitialized]

 2123 |  return ret;
      |         ^~~
Fixes 09022b14fafc (btrfs: scrub: introduce dedicated helper to scrub simple-mirror based range)
---
 fs/btrfs/scrub.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/btrfs/scrub.c +1 -1
index a01807cbd4d4..13024131f77d 100644
@@ -2071,7 +2071,7 @@ static int scrub_simple_mirror(struct scrub_ctx *sctx,
 	struct btrfs_fs_info *fs_info = sctx->fs_info;
 	const u64 logical_end = logical_start + logical_length;
 	u64 cur_logical = logical_start;
-	int ret;
+	int ret = 0;
 
 	/* The range must be inside the bg */
 	ASSERT(logical_start >= bg->start && logical_end <= bg->start + bg->length);
Signature · 1 line
2.39.3
#1 lore raw
Quoted 7 lines · Guanjun

From: Guanjun <[email protected]>

'ret' will be uninitialized in case that the logical_length is 0. Even if the caller has already ensured that logical_length is greater than 0, we still need to fix this issue. Due to the compiler may complain like this:

Quoted 5 lines

fs/btrfs/scrub.c: In function ‘scrub_simple_mirror.constprop’:
fs/btrfs/scrub.c:2123:9: error: ‘ret’ may be used uninitialized in this function [-Werror=maybe-uninitialized]

  2123 |  return ret;
       |         ^~~

Compiler version and config please.

You know btrfs has enabled -Wmaybe-uninitialized already and all warnings would be treated as error. This means if this is really valid, tons of testers would already hit it.

Thus I'm wondering if it's some internal out-of-date toolchain on your side.

Furthermore, if you really want to fix the problem, I strongly discourage from blindly setting the @ret to 0.

But change all the break calls of the loop to return directly, so that the final return out of the loop can always return 0, so that @ret can be defined inside the loop, and be much safer.

Thanks,
Qu

Quoted 19 lines
Fixes 09022b14fafc (btrfs: scrub: introduce dedicated helper to scrub simple-mirror based range)
---
  fs/btrfs/scrub.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
fs/btrfs/scrub.c +1 -1
index a01807cbd4d4..13024131f77d 100644
@@ -2071,7 +2071,7 @@ static int scrub_simple_mirror(struct scrub_ctx *sctx,
  	struct btrfs_fs_info *fs_info = sctx->fs_info;
  	const u64 logical_end = logical_start + logical_length;
  	u64 cur_logical = logical_start;
-	int ret;
+	int ret = 0;
 
  	/* The range must be inside the bg */
  	ASSERT(logical_start >= bg->start && logical_end <= bg->start + bg->length);