From: Mike Snitzer Subject: Re: [PATCH 1/2] Revert "Revert "math64: New div64_u64_rem helper"" Date: Wed, 4 Sep 2013 13:55:29 -0400 Message-ID: <20130904175529.GA25526@redhat.com> References: <1376078164-23147-1-git-send-email-snitzer@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: dm-devel@redhat.com, LKML , Mikulas Patocka , Stanislaw Gruszka , Frederic Weisbecker , Ingo Molnar , linux-ext4@vger.kernel.org To: Anatol Pomozov Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Wed, Sep 04 2013 at 12:11pm -0400, Anatol Pomozov wrote: > Hi, > > On Fri, Aug 9, 2013 at 12:56 PM, Mike Snitzer wrote: > > This reverts commit f3002134158092178be81339ec5a22ff80e6c308. > > > > div64_u64_rem was removed because there were no other users. > > In the light of this discussion https://lkml.org/lkml/2013/8/30/417 we > will have more users of div64_u64_rem > > For example code from ext4_mb_find_by_goal() function uses function > do_div() with divisor sbi->s_stripe that has "unsigned long" type. > do_div() silently truncates the type to 32bits and do_div() returns > incorrect result (or crash if lower 32bits are zero). It should be > either replaced with div64_u64_rem() posted here or with just "a % b". Hi, As Ingo suggested, I went ahead and introduced a completely new div64_u64_rem, see: https://lkml.org/lkml/2013/8/21/452 I've staged this change in linux-next for inclusion in the current 3.12 merge (via the device-mapper tree), see: https://git.kernel.org/cgit/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=for-next&id=eb18cba78c2b9250663021e17e1e9cc34630e92a So once this is in ext4 can make use of div64_u64_rem(). Mike