Date: Sat, 30 May 2009 18:14:25 +0200
From: Andi Kleen <andi@firstfloor.org>
To: Felix Blyakher <felixb@sgi.com>
Cc: Andi Kleen <andi@firstfloor.org>, Christoph Hellwig <hch@infradead.org>,
        Hedi Berriche <hedi@sgi.com>, xfs@oss.sgi.com
Subject: Re: [PATCH] xfs: prevent deadlock in xfs_qm_shake()
Message-ID: <20090530161425.GO1065@one.firstfloor.org>
References: <1243620631-10749-1-git-send-email-felixb@sgi.com> <1243620631-10749-2-git-send-email-felixb@sgi.com> <20090529192529.GA1599@infradead.org> <87d49qeuqd.fsf@basil.nowhere.org> <1A09BC73-E1E8-4EF4-AD29-FB93556C3E3B@sgi.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1A09BC73-E1E8-4EF4-AD29-FB93556C3E3B@sgi.com>
User-Agent: Mutt/1.4.2.1i

On Sat, May 30, 2009 at 09:57:20AM -0500, Felix Blyakher wrote:
> 
> (gfp_mask & __GFP_WAIT && gfp_mask & __GFP_FS)  != 0   or as Andi noted
> (gfp_mask & (__GFP_WAIT|__GFP_FS)) == (__GFP_WAIT|__GFP_FS)
> 
> I'd prefer the former, as in my original patch.
> 
> Also, I accidentally put an extra open brace in a statement. After a
> successful build I started playing with braces for more readability,
> and left it in inconsistent state.
> Seems like the preferred style in the kernel is as following:
> 
> return ((gfp_mask & __GFP_WAIT) && (gfp_mask & __GFP_FS)) != 0;

I would say it is

return (gfp_mask & __GFP_WAIT) && (gfp_mask & __GFP_FS);

The != 0 is completely superfluous.

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only.