From mboxrd@z Thu Jan 1 00:00:00 1970 From: Zheng Liu Subject: [PATCH] ext4: fix overhead calculation in bigalloc filesystem (Re: ... ) Date: Thu, 21 Feb 2013 20:15:45 +0800 Message-ID: <20130221121545.GA30821@gmail.com> References: <1361433665-16880-1-git-send-email-lczerner@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-ext4@vger.kernel.org, Theodore Ts'o To: Lukas Czerner Return-path: Received: from mail-pa0-f42.google.com ([209.85.220.42]:49619 "EHLO mail-pa0-f42.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753397Ab3BUMA6 (ORCPT ); Thu, 21 Feb 2013 07:00:58 -0500 Received: by mail-pa0-f42.google.com with SMTP id kq12so4716046pab.15 for ; Thu, 21 Feb 2013 04:00:57 -0800 (PST) Content-Disposition: inline In-Reply-To: <1361433665-16880-1-git-send-email-lczerner@redhat.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu, Feb 21, 2013 at 09:01:05AM +0100, Lukas Czerner wrote: > ext4_has_free_clusters() should tell us whether there is enough free > clusters to allocate, however number of free clusters in the file system > is converted to blocks using EXT4_C2B() which is not only wrong use of > the macro (we should have used EXT4_NUM_B2C) but it's also completely > wrong concept since everything else is in cluster units. > > Moreover when calculating number of root clusters we should be using > macro EXT4_NUM_B2C() instead of EXT4_C2B() otherwise the result will > usually be off by one. > > As a result number of free clusters is much bigger than it should have > been and ext4_has_free_clusters() would return 1 even if there is really > not enough free clusters available. > > Fix this by removing the EXT4_C2B() conversion of free clusters and > using EXT4_NUM_B2C() when calculating number of root clusters. This bug > affects number of xfstests tests covering file system ENOSPC situation > handling. With this patch most of the ENOSPC problems with bigalloc file > system disappear, especially the errors caused by delayed allocation not > having enough space when the actual allocation is finally requested. > > Signed-off-by: Lukas Czerner Great! Thanks for fixing it. After applied this patch, xfstests #15 with bigalloc and delalloc won't cause a failure. You can add Reviewed-and-tested-by: Zheng Liu BTW, xfstests (204, 219, 233, 235, 273, and 274) still cause failures in my test environment, and I still get a warning message which looks like: kernel: EXT4-fs (sda2): ext4_da_update_reserve_space: ino 3658, allocated 1 with only 0 reserved metadata blocks kernel: kernel: ------------[ cut here ]------------ kernel: WARNING: at fs/ext4/inode.c:362 ext4_da_update_reserve_space+0x10f/0x21b [ext4]() kernel: Hardware name: OptiPlex 780 kernel: Modules linked in: ext4 jbd2 crc16 cpufreq_ondemand ipv6 dm_mirror dm_region_hash dm_log dm_mod parport_pc parport cspkr i2c_i801 i2c_core serio_raw sg ehci_pci ehci_hcd button e1000e ext3 jbd sd_mod ahci libahci libata scsi_mod uhci_hcd kernel: Pid: 2628, comm: 2372.fsstress.b Tainted: G W 3.8.0+ #7 kernel: Call Trace: kernel: [] warn_slowpath_common+0x85/0x9d kernel: [] warn_slowpath_null+0x1a/0x1c kernel: [] ext4_da_update_reserve_space+0x10f/0x21b [ext4] kernel: [] ext4_ext_map_blocks+0xd83/0xf66 [ext4] kernel: [] ? release_pages+0x169/0x178 kernel: [] ? pagevec_lookup_tag+0x25/0x2e kernel: [] ? write_cache_pages_da+0x107/0x3c4 [ext4] kernel: [] ext4_map_blocks+0x135/0x1ef [ext4] kernel: [] mpage_da_map_and_submit+0x111/0x3d8 [ext4] kernel: [] ext4_da_writepages+0x37e/0x526 [ext4] kernel: [] do_writepages+0x20/0x29 kernel: [] __filemap_fdatawrite_range+0x50/0x52 kernel: [] filemap_fdatawrite+0x1f/0x21 kernel: [] filemap_write_and_wait+0x1d/0x38 kernel: [] do_vfs_ioctl+0x2db/0x47f kernel: [] sys_ioctl+0x5e/0x82 kernel: [] system_call_fastpath+0x16/0x1b kernel: ---[ end trace d96610456f905628 ]--- It is easy to trigger this warning when running xfstests #127 or #225. Moreover, it seems that there still has an improvement in ext4_calculate_overhead(). I paste the patch here. Regards, - Zheng Subject: [PATCH] ext4: fix overhead calculation in bigalloc filesystem From: Zheng Liu ext4_calculate_overhead() should compute the overhead and stash it in sbi->s_overhead. But we miss use EXT4_B2C() to calculate the number of clusters before first_data_block and the number of journal blocks. This commit use EXT4_NUM_B2C() instead of EXT4_B2C() to calculate the overhead. Signed-off-by: Zheng Liu Cc: "Theodore Ts'o" --- fs/ext4/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 3d4fb81..6165558 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3219,7 +3219,7 @@ int ext4_calculate_overhead(struct super_block *sb) /* * All of the blocks before first_data_block are overhead */ - overhead = EXT4_B2C(sbi, le32_to_cpu(es->s_first_data_block)); + overhead = EXT4_NUM_B2C(sbi, le32_to_cpu(es->s_first_data_block)); /* * Add the overhead found in each block group @@ -3235,7 +3235,7 @@ int ext4_calculate_overhead(struct super_block *sb) } /* Add the journal blocks as well */ if (sbi->s_journal) - overhead += EXT4_B2C(sbi, sbi->s_journal->j_maxlen); + overhead += EXT4_NUM_B2C(sbi, sbi->s_journal->j_maxlen); sbi->s_overhead = overhead; smp_wmb(); -- 1.7.12.rc2.18.g61b472e