On Mon, 26 May 2014, Theodore Ts'o wrote: > Date: Mon, 26 May 2014 12:50:10 -0400 > From: Theodore Ts'o <tytso@xxxxxxx> > To: Lukáš Czerner <lczerner@xxxxxxxxxx> > Cc: Maurizio Lombardi <mlombard@xxxxxxxxxx>, adilger.kernel@xxxxxxxxx, > linux-ext4@xxxxxxxxxxxxxxx, linux-fsdevel@xxxxxxxxxxxxxxx > Subject: Re: [PATCH 2/2] ext4: fix bug in ext4_mb_normalize_request() > > On Thu, Mar 06, 2014 at 06:54:05PM +0100, Lukáš Czerner wrote: > > Yes it tries to align down the start_off of the bigger requests to the 512, > > 1024, 2048, or 4096 respectively. What the reason for it is really I have > > no idea. The fact is however that this will only affect file systems > > with bs smaller than 4k since the start_off will be always aligned to > > block size afterwards (obviously). > > > > That said this alignment is only done when the request is "big > > enough". With your change we also do it when the block group is > > "small enough" which is the change in behaviour which I think was > > not really intended. > > > > Honestly I do not think this really matters a lot but this alignment > > you've added is not necessary. > > > > All that said, I was getting to rewrite this mess a long time ago, > > it's just a reminder that it's something that needs to be done. > > Especially since the bigger requests are getting split unnecessarily > > which hurts especially in fallocate case. > > Hey Lukas, where are we with respect to your plans to fix up this > code? > > I'm trying to figure out what we should do with Maurizio's bug fix. > Should we apply it even though it's making some changes to the > existing behavior. > > As far as to why the existing code is trying to align the starting > offset to a power of two, I believe the idea is to avoid > fragmentation, since the normalize_request function will tend to round > up allocaiton requests to the same power of two. Hi Ted, I think that leaving the alignment of the start offset for the small files/allocation is not good idea. We might end up with suboptimal file layout for smaller files. While this is not a big deal for bigger files, with smaller ones it might cause some troubles. Maurizio, can you resend the patch without the alignment ? Also I started looking into normalize_request and hopefully I'll have a patch soon. Ted, do you have any suggestion for a test to make sure that I do not make things worse ? You mentioned something simple on LSF, but I do not remember what it was. Thanks! -Lukas > > - Ted > -- > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html >
