From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kara Subject: Re: [PATCH] ext4: Remove extent tree purging from ext4_da_page_release_reservation() Date: Thu, 25 Jul 2013 16:05:28 +0200 Message-ID: <20130725140528.GD9388@quack.suse.cz> References: <1374099015-6829-1-git-send-email-jack@suse.cz> <20130719004439.GA21615@gmail.com> <20130724200536.GC27307@quack.suse.cz> <20130725115234.GA26044@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , Ted Tso , linux-ext4@vger.kernel.org, Zheng Liu To: Zheng Liu Return-path: Received: from cantor2.suse.de ([195.135.220.15]:47818 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755450Ab3GYOFc (ORCPT ); Thu, 25 Jul 2013 10:05:32 -0400 Content-Disposition: inline In-Reply-To: <20130725115234.GA26044@gmail.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu 25-07-13 19:52:34, Zheng Liu wrote: > On Wed, Jul 24, 2013 at 10:05:36PM +0200, Jan Kara wrote: > > Hi Zheng, > > > > On Fri 19-07-13 08:44:39, Zheng Liu wrote: > > > On Thu, Jul 18, 2013 at 12:10:15AM +0200, Jan Kara wrote: > > > > ext4_da_page_release_reservation() gets called from > > > > ext4_da_invalidatepage(). This function is used when we are truncating > > > > page cache for punch hole or truncate operations. In either case these > > > > operations take care of removing extents from the extent tree. This is > > > > more efficient and the code in ext4_da_page_release_reservation() is > > > > actually buggy anyway. So just remove it. > > > > > > I remember that I try to remove the entry from extent status tree here > > > because at the end of this function it tries to relase the reserved > > > space for delalloc. For 4k block we can simply release it because > > > ->s_cluster_ratio == 1. But when bigalloc is enabled, we need to > > > determine whether we can release the reserved space according to the > > > result of ext4_find_delalloc_cluster() as the comment described. If we > > > don't remove the entry from extent status tree here, we could lost some > > > spaces that could be reused by other files. If I remember correctly, I > > > have hitted a warning message when I run xfstests to test it. These > > > days I try to trigger it using xfstests but I failed. Have you seen a > > > prblem that is caused by this code? Maybe we need to refactor out the > > > code and release the reserved space outside this function. > > Ah, I see. No, I didn't observe any problem due to this code, I just > > didn't understand why is it there. Also when blocksize < pagesize, the code > > is wrong because delayed buffers to release need not be contiguous so > > ext4_es_remove_extent(inode, lblk, to_release) may not free all the buffers > > we want. But subsequent extent tree truncation in ext4_ext_truncate() hides > > this problem. > > > > So I think we might just change the condition: > > > > if (to_release) { > > > > to > > > > if (to_release && sbi->s_cluster_ratio > 1) { > > > > and add explanatory comment why cluster_ratio > 1 needs the truncation and > > other cases don't. It will also save some needlessly burned CPU cycles > > spent when manipulating extent tree. > > Yes, thanks for pointing it out. I attach a patch below. Could you > please review it? Thanks for writing the patch! It looks good. BTW, feel free to take authorship of it. I've just pointed out the problem. One typo correction below: > Subject: [PATCH 1/2] ext4: remove the entry from es tree when bigalloc is enabled > > From: Jan Kara > > Now in ext4_da_page_release_reservation() we remove the entry from es > tree if to_release != 0. But there are two issues. One is that it is > wrong when blocksize != pagesize, another is that we don't need to do > this if ->s_cluster_ratio == 1 because we will remove the entry in > ext4_truncate/ext4_punch_hole. Here we need to do this just because > when ->s_cluster_ratio > 1 we will determine whether we can release > the reserved space according to ext4_find_delalloc_cluster(). > > This commit tries to fix these problems. Now we remove the entry from > es tree only if ->s_cluster_ratio > 1. > > Signed-off-by: Jan Kara > Signed-off-by: Zheng Liu > --- > fs/ext4/inode.c | 10 +++++++++- > 1 file changed, 9 insertions(+), 1 deletion(-) > > diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c > index ba33c67..e0c8623 100644 > --- a/fs/ext4/inode.c > +++ b/fs/ext4/inode.c > @@ -1387,7 +1387,15 @@ static void ext4_da_page_release_reservation(struct page *page, > curr_off = next_off; > } while ((bh = bh->b_this_page) != head); > > - if (to_release) { > + /* > + * Here we need to remove the entry from es tree because when bigalloc > + * is enabled we need to determine whether we can release the reserved > + * space according to the result of ext4_find_delalloc_cluster(). > + * > + * If bigalloc is disabled, we don't need to do this here because these > + * extries in es tree will be removed in ext4_truncate/ext4_punch_hole. ^^^ entries > + */ > + if (sbi->s_cluster_ratio > 1 && to_release) { > lblk = page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits); > ext4_es_remove_extent(inode, lblk, to_release); > } Honza -- Jan Kara SUSE Labs, CR