linux-kernel - Re: [PATCH] fs: fix a bug when new_insert

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: <20160801160510.4a48a02d68aa5d89a0435b52@linux-foundation.org>
Date:	Mon, 1 Aug 2016 16:05:10 -0700
From:	Andrew Morton <akpm@...ux-foundation.org>
To:	zhongjiang <zhongjiang@...wei.com>
Cc:	<linux-mm@...ck.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] fs: fix a bug when new_insert_key is not initialization

On Sat, 30 Jul 2016 11:51:09 +0800 zhongjiang <zhongjiang@...wei.com> wrote:

> From: zhong jiang <zhongjiang@...wei.com>
> 
> when compile the kenrel code, I happens to the following warn.
> fs/reiserfs/ibalance.c:1156:2: warning: ___new_insert_key___ may be used
> uninitialized in this function.
> memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE);
> 
> The patch fix it by check the new_insert_ptr. if new_insert_ptr is not
> NULL, we ensure that new_insert_key is assigned. therefore, memcpy will
> saftly exec the operatetion.
> 
> --- a/fs/reiserfs/ibalance.c
> +++ b/fs/reiserfs/ibalance.c
> @@ -1153,8 +1153,10 @@ int balance_internal(struct tree_balance *tb,
>  				       insert_ptr);
>  	}
>  
> -	memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE);
> -	insert_ptr[0] = new_insert_ptr;
> +	if (new_insert_ptr) {
> +		memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE);
> +		insert_ptr[0] = new_insert_ptr;
> +	}
>  
>  	return order;

Jeff has aleady fixed this with an equivalent patch.  It's in -mm at
present.

From: Jeff Mahoney <jeffm@...e.com>
Subject: reiserfs: fix "new_insert_key may be used uninitialized ..."

new_insert_key only makes any sense when it's associated with a
new_insert_ptr, which is initialized to NULL and changed to a buffer_head
when we also initialize new_insert_key.  We can key off of that to avoid
the uninitialized warning.

Link: http://lkml.kernel.org/r/5eca5ffb-2155-8df2-b4a2-f162f105efed@suse.com
Signed-off-by: Jeff Mahoney <jeffm@...e.com>
Cc: Arnd Bergmann <arnd@...db.de>
Cc: Jan Kara <jack@...e.cz>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>
Signed-off-by: Andrew Morton <akpm@...ux-foundation.org>
---

 fs/reiserfs/ibalance.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff -puN fs/reiserfs/ibalance.c~reiserfs-fix-new_insert_key-may-be-used-uninitialized fs/reiserfs/ibalance.c
--- a/fs/reiserfs/ibalance.c~reiserfs-fix-new_insert_key-may-be-used-uninitialized
+++ a/fs/reiserfs/ibalance.c
@@ -1153,8 +1153,9 @@ int balance_internal(struct tree_balance
 				       insert_ptr);
 	}
 
-	memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE);
 	insert_ptr[0] = new_insert_ptr;
+	if (new_insert_ptr)
+		memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE);
 
 	return order;
 }
_