[PATCH v2 0/2] e2fsprogs: fix memory leaks detected by ASAN

[PATCH v2 0/2] e2fsprogs: fix memory leaks detected by ASAN

[Wu Guanghao]

v1 -> v2:
 - Check the return value of ext2fs_close_free()

Wu Guanghao (2):
  fsck: fix memory leak of inst->type
  resize: fix memory leak when exiting normally

 misc/fsck.c   | 1 +
 resize/main.c | 2 ++
 2 files changed, 3 insertions(+)

-- 
2.27.0

[PATCH v2 1/2] fsck: fix memory leak of inst->type

[Wu Guanghao]

The function free_instance() does not release i->type, resulting in a
memory leak.

Signed-off-by: Wu Guanghao <wuguanghao3@huawei.com>
---
 misc/fsck.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/misc/fsck.c b/misc/fsck.c
index 64d0e7c0..a06f2668 100644
--- a/misc/fsck.c
+++ b/misc/fsck.c
@@ -235,6 +235,7 @@ static void parse_escape(char *word)
 static void free_instance(struct fsck_instance *i)
 {
 	free(i->prog);
+	free(i->type);
 	free(i->device);
 	free(i->base_device);
 	free(i);
-- 
2.27.0

[PATCH v2 2/2] resize: fix memory leak when exiting normally

[Wu Guanghao]

The main() function only releases fs when it exits through the errout or
success_exit labels. When completes normally, it does not release fs.

Signed-off-by: Wu Guanghao <wuguanghao3@huawei.com>
---
 resize/main.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/resize/main.c b/resize/main.c
index 08a4bbaf..e7940307 100644
--- a/resize/main.c
+++ b/resize/main.c
@@ -702,6 +702,14 @@ int main (int argc, char ** argv)
 	}
 	if (fd > 0)
 		close(fd);
+
+	retval = ext2fs_close_free(&fs);
+	if (retval) {
+		com_err(program_name, retval,
+			_("ext2fs_close"));
+		exit(1);
+	}
+
 	remove_error_table(&et_ext2_error_table);
 	return 0;
 errout:
-- 
2.27.0

Re: [PATCH v2 1/2] fsck: fix memory leak of inst->type

[Darrick J. Wong]

On Fri, Nov 21, 2025 at 11:36:11AM +0800, Wu Guanghao wrote:
> The function free_instance() does not release i->type, resulting in a
> memory leak.
> 
> Signed-off-by: Wu Guanghao <wuguanghao3@huawei.com>

Looks good!
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> ---
>  misc/fsck.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/misc/fsck.c b/misc/fsck.c
> index 64d0e7c0..a06f2668 100644
> --- a/misc/fsck.c
> +++ b/misc/fsck.c
> @@ -235,6 +235,7 @@ static void parse_escape(char *word)
>  static void free_instance(struct fsck_instance *i)
>  {
>  	free(i->prog);
> +	free(i->type);
>  	free(i->device);
>  	free(i->base_device);
>  	free(i);
> -- 
> 2.27.0
> 
> 

Re: [PATCH v2 2/2] resize: fix memory leak when exiting normally

[Theodore Tso]

On Fri, Nov 21, 2025 at 11:36:12AM +0800, Wu Guanghao wrote:
> The main() function only releases fs when it exits through the errout or
> success_exit labels. When completes normally, it does not release fs.
> 
> Signed-off-by: Wu Guanghao <wuguanghao3@huawei.com>

I'm guessing you only tested the online resize code path?

	if (mount_flags & EXT2_MF_MOUNTED) {
		retval = online_resize_fs(fs, mtpt, &new_size, flags);
	} else {
	       ...
		retval = resize_fs(fs, &new_size, flags,
				   ((flags & RESIZE_PERCENT_COMPLETE) ?
				    resize_progress_func : 0));
	}

The reason why I ask this is that resize_fs() frees fs on the success path:

	rfs->old_fs = fs;
	...
	ext2fs_free(rfs->old_fs);
	
... although if we return when an error, we do *not* free ext2fs_free(rfs->old_fs).

So if you were to test with this applied when resizing a non-mounted
file system, I believe you'd get a double free failure.

Cheers,

							- Ted