From: Thavatchai Makphaibulchoke Subject: Re: [PATCH 1/2] fs/ext4: adding and initalizing new members of ext4_inode_info and ext4_sb_info Date: Thu, 03 Oct 2013 17:14:30 -0600 Message-ID: <524DFA56.6000708@hp.com> References: <1380728219-60784-1-git-send-email-tmac@hp.com> <94F53E8F-6D99-4747-97A9-4FB60DD2BED2@dilger.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: T Makphaibulchoke , Theodore Ts'o , adilger.kernel@dilger.ca, "linux-ext4@vger.kernel.org List" , Linux Kernel Mailing List , aswin@hp.com To: Andreas Dilger Return-path: In-Reply-To: <94F53E8F-6D99-4747-97A9-4FB60DD2BED2@dilger.ca> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On 10/03/2013 06:37 PM, Andreas Dilger wrote: > On 2013-10-02, at 9:36 AM, T Makphaibulchoke wrote: > > What do these additional fields do to the size of struct ext4_inode_info? > I recall that Ted did a bunch of work to shrink this enough to fit nicely > into a slab, and it would be a shame to increase the inode size to overflow > the current packing and increase per-inode memory usage by 25-33%, for an > improvement that is only noticeable on a 90-core machine. > > Is there another lock that could be shared for this that is unlikely to > cause much contention? Thanks for the suggestion. I was also thinking about this earlier, not sure if it's a good practice. Looks like it is way better than increasing the inode size. Will look into this in my rework. > > Also, it isn't clear to me why this patch is separate from 2/2, because > all it does is add fields that are not used for anything. I don't think > the 8 lines of code here are so complex that they can't be part of the > same patch that is actually using them. > > Cheers, Andreas > I was debating whether to combine them into 1 or make them 2 patches. I'll combine them into one patch in my next submittal. Thanks, Mak.