linux-ext4 - Re: Broken e2fsck i

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]

Date:	Wed, 1 Sep 2010 16:49:59 -0700
From:	Andreas Dilger <adilger.kernel@...ger.ca>
To:	Justin Maggard <jmaggard10@...il.com>
Cc:	ext4 development <linux-ext4@...r.kernel.org>
Subject: Re: Broken e2fsck i_blocks repair

On 2010-09-01, at 16:01, Justin Maggard wrote:
> I've mentioned this before on here, but I never verified whether or
> not it was actually broken.  I have done so today.  I manually
> modified the block count of a 3TB file using debugfs, then ran e2fsck
> on it.  e2fsck claimed to repair it, but would throw an error on every
> subsequent e2fsck run for the same issue.  The reason is
> inode->osd2.linux2.l_i_blocks_hi is always set to 0 if e2fsck is told
> to fix it.  This is an issue in both stable 1.41.12, and in master.
> This patch fixes it for me, but is there anything else that needs to
> get checked here?

It probably makes sense to audit all uses of "i_blocks" to see if they behave similarly badly w.r.t. l_i_blocks_hi.  A casual glance indicates that there are a number of places that are similarly broken.

I see there is a helper function e2fsck/blknum.c:ext2fs_inode_i_blocks() that should be used for most i_blocks accesses (excluding those places in e2fsck that check for l_i_blocks_hi being non-zero without the appropriate feature flag being set).  It might make sense to rename this function ext2fs_inode_i_blocks_get() and add similar ext2fs_inode_i_blocks_set() routine to set it correctly.

> diff -urp e2fsprogs-1.41.12/e2fsck/pass1.c e2fsprogs-1.41.12-jm/e2fsck/pass1.c
> --- e2fsprogs-1.41.12/e2fsck/pass1.c	2010-05-14 14:51:21.000000000 -0700
> +++ e2fsprogs-1.41.12-jm/e2fsck/pass1.c	2010-09-01 15:54:42.000000000 -0700
> @@ -2044,7 +2044,7 @@ static void check_blocks(e2fsck_t ctx, s
> 		pctx->num = pb.num_blocks;
> 		if (fix_problem(ctx, PR_1_BAD_I_BLOCKS, pctx)) {
> 			inode->i_blocks = pb.num_blocks;
> -			inode->osd2.linux2.l_i_blocks_hi = 0;
> +			inode->osd2.linux2.l_i_blocks_hi = (pb.num_blocks >> 32);
> 			dirty_inode++;
> 		}
> 		pctx->num = 0;
> 
> -Justin
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html


Cheers, Andreas





--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html