Re: [PATCH] jfs: avoid undefined behavior from left-shifting by 32 bits

Re: [PATCH] jfs: avoid undefined behavior from left-shifting by 32 bits

[Dave Kleikamp]

On 01/05/2013 01:19 PM, Nickolai Zeldovich wrote:
> Shifting a 32-bit int by 32 bits is undefined behavior in C, and
> results in different behavior on different architectures (e.g., x86
> and PowerPC).  diAlloc() in fs/jfs/jfs_imap.c computes a mask using
> 0xffffffffu<<(32-bitno), which can left-shift by 32 bits.  To avoid
> unexpected behavior, explicitly check for bitno==0 and use a 0 mask.

Thanks. Pushed to
git://github.com/kleikamp/linux-shaggy.git jfs-next

Shaggy

> 
> Signed-off-by: Nickolai Zeldovich <nickolai@csail.mit.edu>
> ---
>  fs/jfs/jfs_imap.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c
> index 6ba4006..f7e042b 100644
> --- a/fs/jfs/jfs_imap.c
> +++ b/fs/jfs/jfs_imap.c
> @@ -1493,7 +1493,7 @@ int diAlloc(struct inode *pip, bool dir, struct inode *ip)
>  		/* mask any prior bits for the starting words of the
>  		 * summary map.
>  		 */
> -		mask = ONES << (EXTSPERSUM - bitno);
> +		mask = (bitno == 0) ? 0 : (ONES << (EXTSPERSUM - bitno));
>  		inosmap = le32_to_cpu(iagp->inosmap[sword]) | mask;
>  		extsmap = le32_to_cpu(iagp->extsmap[sword]) | mask;
>  
>