linux-kernel - Re: [PATCH 2/4] ext4: fix BUG in ext4_mb_new_inode

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: <878rbao0zs.fsf@doe.com>
Date:   Fri, 21 Jul 2023 00:37:35 +0530
From:   Ritesh Harjani (IBM) <ritesh.list@...il.com>
To:     Baokun Li <libaokun1@...wei.com>, linux-ext4@...r.kernel.org
Cc:     tytso@....edu, adilger.kernel@...ger.ca, jack@...e.cz,
        ojaswin@...ux.ibm.com, linux-kernel@...r.kernel.org,
        yi.zhang@...wei.com, yangerkun@...wei.com, yukuai3@...wei.com,
        libaokun1@...wei.com
Subject: Re: [PATCH 2/4] ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow

Baokun Li <libaokun1@...wei.com> writes:

> When we calculate the end position of ext4_free_extent, this position may
> be exactly where ext4_lblk_t (i.e. uint) overflows. For example, if
> ac_g_ex.fe_logical is 4294965248 and ac_orig_goal_len is 2048, then the
> computed end is 0x100000000, which is 0. If ac->ac_o_ex.fe_logical is not
> the first case of adjusting the best extent, that is, new_bex_end > 0, the
> following BUG_ON will be triggered:
>
> =========================================================
> kernel BUG at fs/ext4/mballoc.c:5116!
> invalid opcode: 0000 [#1] PREEMPT SMP PTI
> CPU: 3 PID: 673 Comm: xfs_io Tainted: G E 6.5.0-rc1+ #279
> RIP: 0010:ext4_mb_new_inode_pa+0xc5/0x430
> Call Trace:
>  <TASK>
>  ext4_mb_use_best_found+0x203/0x2f0
>  ext4_mb_try_best_found+0x163/0x240
>  ext4_mb_regular_allocator+0x158/0x1550
>  ext4_mb_new_blocks+0x86a/0xe10
>  ext4_ext_map_blocks+0xb0c/0x13a0
>  ext4_map_blocks+0x2cd/0x8f0
>  ext4_iomap_begin+0x27b/0x400
>  iomap_iter+0x222/0x3d0
>  __iomap_dio_rw+0x243/0xcb0
>  iomap_dio_rw+0x16/0x80
> =========================================================
>
> A simple reproducer demonstrating the problem:
>
> 	mkfs.ext4 -F /dev/sda -b 4096 100M
> 	mount /dev/sda /tmp/test
> 	fallocate -l1M /tmp/test/tmp
> 	fallocate -l10M /tmp/test/file
> 	fallocate -i -o 1M -l16777203M /tmp/test/file
> 	fsstress -d /tmp/test -l 0 -n 100000 -p 8 &
> 	sleep 10 && killall -9 fsstress
> 	rm -f /tmp/test/tmp
> 	xfs_io -c "open -ad /tmp/test/file" -c "pwrite -S 0xff 0 8192"
>
> We declare new_bex_start and new_bex_end as correct types and use fex_end()
> to avoid the problems caused by the ext4_lblk_t overflow above.
>
> Fixes: 93cdf49f6eca ("ext4: Fix best extent lstart adjustment logic in ext4_mb_new_inode_pa()")
> Signed-off-by: Baokun Li <libaokun1@...wei.com>
> ---
>  fs/ext4/mballoc.c | 11 +++++------
>  1 file changed, 5 insertions(+), 6 deletions(-)
>
> diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
> index eb7f5d35ef96..2090e5e7ba58 100644
> --- a/fs/ext4/mballoc.c
> +++ b/fs/ext4/mballoc.c
> @@ -5076,8 +5076,8 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
>  	pa = ac->ac_pa;
>  
>  	if (ac->ac_b_ex.fe_len < ac->ac_orig_goal_len) {
> -		int new_bex_start;
> -		int new_bex_end;
> +		ext4_lblk_t new_bex_start;
> +		loff_t new_bex_end;
>  
>  		/* we can't allocate as much as normalizer wants.
>  		 * so, found space must get proper lstart
> @@ -5096,8 +5096,7 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
>  		 *    still cover original start
>  		 * 3. Else, keep the best ex at start of original request.
>  		 */
> -		new_bex_end = ac->ac_g_ex.fe_logical +
> -			EXT4_C2B(sbi, ac->ac_orig_goal_len);
> +		new_bex_end = fex_end(sbi, &ac->ac_g_ex, &ac->ac_orig_goal_len);
>  		new_bex_start = new_bex_end - EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
>  		if (ac->ac_o_ex.fe_logical >= new_bex_start)
>  			goto adjust_bex;

		new_bex_end = ac->ac_g_ex.fe_logical +
			EXT4_C2B(sbi, ac->ac_orig_goal_len);
		new_bex_start = new_bex_end - EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
		if (ac->ac_o_ex.fe_logical >= new_bex_start)
			goto adjust_bex;

		new_bex_start = ac->ac_g_ex.fe_logical;
		new_bex_end =
			new_bex_start + EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
		if (ac->ac_o_ex.fe_logical < new_bex_end)
			goto adjust_bex;

		new_bex_start = ac->ac_o_ex.fe_logical;
		new_bex_end =
			new_bex_start + EXT4_C2B(sbi, ac->ac_b_ex.fe_len);

I think it will be safer if we typecast all above new_bex_end
calculations. I understand that it might never happen here that
fe_logical + best found extent length will make it overflow.
Simply because we only enter here when ac_b_ex.fe_len <
ac_orig_goal_len. But I think if we change any logic tomorrow or
refactor any part of code, it will be much safer if we simply keep the
typecast in place so that we avoid any future tripping caused by
arithmatic overflow here.

Also I am not finding fex_end() or pa_end() helpers very intuitive way
of doing it. I feel typecasting in place is much simpler and reads better.

-ritesh

> @@ -5117,8 +5116,8 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
>  
>  		BUG_ON(ac->ac_o_ex.fe_logical < ac->ac_b_ex.fe_logical);
>  		BUG_ON(ac->ac_o_ex.fe_len > ac->ac_b_ex.fe_len);
> -		BUG_ON(new_bex_end > (ac->ac_g_ex.fe_logical +
> -				      EXT4_C2B(sbi, ac->ac_orig_goal_len)));
> +		BUG_ON(new_bex_end >
> +			fex_end(sbi, &ac->ac_g_ex, &ac->ac_orig_goal_len));
>  	}
>  
>  	pa->pa_lstart = ac->ac_b_ex.fe_logical;
> -- 
> 2.31.1