linux-kernel - Re: [PATCH] ubifs: Fix ubifs_tnc_lookup() usage in do_kill

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <875zdibasg.fsf@vostro.fn.ogness.net>
Date:   Wed, 29 Apr 2020 16:56:31 +0200
From:   John Ogness <john.ogness@...utronix.de>
To:     Richard Weinberger <richard@....at>
Cc:     linux-mtd@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] ubifs: Fix ubifs_tnc_lookup() usage in do_kill_orphans()

Hi Richard,

Could you CC this patch to stable? It fixes a serious problem that I am
seeing on real devices (i.e. Linux not being able to mount its root
filesystem after a power cut). Thanks.

John Ogness

On 2020-01-19, Richard Weinberger <richard@....at> wrote:
> Orphans are allowed to point to deleted inodes.
> So -ENOENT is not a fatal error.
>
> Reported-by: Кочетков Максим <fido_max@...ox.ru>
> Reported-and-tested-by: "Christian Berger" <Christian.Berger@...bosch.com>
> Fixes: ee1438ce5dc4 ("ubifs: Check link count of inodes when killing orphans.")
> Signed-off-by: Richard Weinberger <richard@....at>
> ---
>  fs/ubifs/orphan.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c
> index 54d6db61106f..2645917360b9 100644
> --- a/fs/ubifs/orphan.c
> +++ b/fs/ubifs/orphan.c
> @@ -688,14 +688,14 @@ static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
>  
>  			ino_key_init(c, &key1, inum);
>  			err = ubifs_tnc_lookup(c, &key1, ino);
> -			if (err)
> +			if (err && err != -ENOENT)
>  				goto out_free;
>  
>  			/*
>  			 * Check whether an inode can really get deleted.
>  			 * linkat() with O_TMPFILE allows rebirth of an inode.
>  			 */
> -			if (ino->nlink == 0) {
> +			if (err == 0 && ino->nlink == 0) {
>  				dbg_rcvry("deleting orphaned inode %lu",
>  					  (unsigned long)inum);