diff mbox series

tune2fs: fix a NULL dereference on failed journal replay

Message ID 20220923132735.1701587-1-lkundrak@v3.sk
State Superseded
Headers show
Series tune2fs: fix a NULL dereference on failed journal replay | expand

Commit Message

Lubomir Rintel Sept. 23, 2022, 1:27 p.m. UTC
Got a crash:

  Starting program: e2fsprogs-1.46.5/misc/tune2fs -O ^has_journal -ff /dev/sdh2
  tune2fs 1.46.5 (30-Dec-2021)
  Recovering journal.
  tune2fs: Unknown code ____ 251 while recovering journal.
  Please run e2fsck -fy /dev/sdh2.

  Program received signal SIGSEGV, Segmentation fault.
  0x00007ffff7f8565a in ext2fs_mmp_stop (fs=0x0) at ../mmp.c:405
  405		if (!ext2fs_has_feature_mmp(fs->super) ||
  (gdb) bt
  #0  0x00007ffff7f8565a in ext2fs_mmp_stop (fs=0x0) at ../mmp.c:405
  #1  0x000055555555acd8 in main (argc=<optimized out>, argv=<optimized out>) at /home/lkundrak/fedora/e2fsprogs/e2fsprogs-1.46.5/misc/tune2fs.c:3441
  (gdb)

Turns out, ext2fs_run_ext3_journal() can close fs and then fail. If that
happened, we shall not jump to closefs:, quit right away instead.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
---
 misc/tune2fs.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

Comments

Zhiqiang Liu Sept. 24, 2022, 2:50 a.m. UTC | #1
Nice catch
But Li Jinlin has also found this problem, and sent a patch:
"[PATCH] tune2fs: exit directly when fs freed in ext2fs_run_ext3_journal"

On 2022/9/23 21:27, Lubomir Rintel wrote:
> Got a crash:
> 
>   Starting program: e2fsprogs-1.46.5/misc/tune2fs -O ^has_journal -ff /dev/sdh2
>   tune2fs 1.46.5 (30-Dec-2021)
>   Recovering journal.
>   tune2fs: Unknown code ____ 251 while recovering journal.
>   Please run e2fsck -fy /dev/sdh2.
> 
>   Program received signal SIGSEGV, Segmentation fault.
>   0x00007ffff7f8565a in ext2fs_mmp_stop (fs=0x0) at ../mmp.c:405
>   405		if (!ext2fs_has_feature_mmp(fs->super) ||
>   (gdb) bt
>   #0  0x00007ffff7f8565a in ext2fs_mmp_stop (fs=0x0) at ../mmp.c:405
>   #1  0x000055555555acd8 in main (argc=<optimized out>, argv=<optimized out>) at /home/lkundrak/fedora/e2fsprogs/e2fsprogs-1.46.5/misc/tune2fs.c:3441
>   (gdb)
> 
> Turns out, ext2fs_run_ext3_journal() can close fs and then fail. If that
> happened, we shall not jump to closefs:, quit right away instead.
> 
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> ---
>  misc/tune2fs.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/misc/tune2fs.c b/misc/tune2fs.c
> index 088f87e5..96cfd001 100644
> --- a/misc/tune2fs.c
> +++ b/misc/tune2fs.c
> @@ -3345,7 +3345,10 @@ _("Warning: The journal is dirty. You may wish to replay the journal like:\n\n"
>  				"while recovering journal.\n");
>  			printf(_("Please run e2fsck -fy %s.\n"), argv[1]);
>  			rc = 1;
> -			goto closefs;
> +			if (fs)
> +				goto closefs;
> +			else
> +				return 1;
>  		}
>  		sb = fs->super;
>  	}
>
diff mbox series

Patch

diff --git a/misc/tune2fs.c b/misc/tune2fs.c
index 088f87e5..96cfd001 100644
--- a/misc/tune2fs.c
+++ b/misc/tune2fs.c
@@ -3345,7 +3345,10 @@  _("Warning: The journal is dirty. You may wish to replay the journal like:\n\n"
 				"while recovering journal.\n");
 			printf(_("Please run e2fsck -fy %s.\n"), argv[1]);
 			rc = 1;
-			goto closefs;
+			if (fs)
+				goto closefs;
+			else
+				return 1;
 		}
 		sb = fs->super;
 	}