linux-kernel - Re: [PATCH] f2fs: Remove BUG_ON in dec_valid_node

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] [day] [month] [year] [list]

Message-ID: <20130513070556.GB21431@hli22-desktop>
Date:	Mon, 13 May 2013 15:05:56 +0800
From:	Haicheng Li <haicheng.li@...ux.intel.com>
To:	Chris Fries <ccfries@...il.com>
Cc:	jaegeuk.kim@...sung.com, linux-f2fs-devel@...ts.sourceforge.net,
	linux-kernel@...r.kernel.org, rknize2@...orola.com,
	jason.hrycay@...orola.com
Subject: Re: [PATCH] f2fs: Remove BUG_ON in dec_valid_node_count


On Sun, May 12, 2013 at 10:46:27PM -0500, Chris Fries wrote:
> From: Chris Fries <C.Fries@...orola.com>
> 
> Panic loops while running LTP fsstress has been able to get
> a disk into two different panic loops from dec_valid_node_count.
> f2fs.h:714  BUG_ON(sbi->total_valid_node_count < count);

This is interesting catch.
from the code, dec_valid_node_count() is only called by truncate_node():
	dec_valid_node_count(sbi, dn->inode, 1);

So the failure in your test means that sbi->total_valid_node_count < 1,
i.e. equal to 0. This should be an unexpected status.

I think a better solution should be to avoid such over truncate_node situation.

How do you think?

> Once, it happens during recovery itself, and the disk would cause
> a panic every time it mounted.
> 
> Another time, it would happen during garbage collection, so the disk
> would cause a panic within 200 seconds of mounting.
> 
> Removing this BUG_ON hasn't shown any side effects, so let's take it
> out and monitor.
> 
> Signed-off-by: Chris Fries <C.Fries@...orola.com>
> 
> ---
>  fs/f2fs/f2fs.h | 18 +++++++++++++++---
>  1 file changed, 15 insertions(+), 3 deletions(-)
> 
> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> index e80a87c..b8e9679 100644
> --- a/fs/f2fs/f2fs.h
> +++ b/fs/f2fs/f2fs.h
> @@ -722,9 +722,21 @@ static inline void dec_valid_node_count(struct f2fs_sb_info *sbi,
>  {
>  	spin_lock(&sbi->stat_lock);
>  
> -	BUG_ON(sbi->total_valid_block_count < count);
> -	BUG_ON(sbi->total_valid_node_count < count);
> -	BUG_ON(inode->i_blocks < count);
> +	if (sbi->total_valid_block_count < count) {
> +		WARN(1, "F2FS: total_valid_block_count too small- %d vs %d\n",
> +			(unsigned int)sbi->total_valid_block_count, count);
> +		count = sbi->total_valid_block_count;
> +	}
> +	if (sbi->total_valid_node_count < count) {
> +		WARN(1, "F2FS: total_valid_node_count too small- %d vs %d\n",
> +			sbi->total_valid_node_count, count);
> +		count = sbi->total_valid_node_count;
> +	}
> +	if (inode->i_blocks < count) {
> +		WARN(1, "F2FS: inode->i_blocks too small - %d vs %d\n",
> +			(unsigned int)inode->i_blocks, count);
> +		count = sbi->total_valid_node_count;
> +	}
>  
>  	inode->i_blocks -= count;
>  	sbi->total_valid_node_count -= count;
> -- 
> 1.8.0
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/