Hi,
On 2018/08/21 14:40, Qu Wenruo wrote:
> Commit c6887cd11149 ("Btrfs: don't do nocow check unless we have to")
> makes nocow check less frequent to improve performance.
>
> However for quota enabled case, such optimization could lead to extra
> unnecessary data reservation, which results failure for test case like
> btrfs/153 in fstests.
>
> Fix it by reverting to old behavior for quota enabled case.
>
> Fixes: c6887cd11149 ("Btrfs: don't do nocow check unless we have to")
> Signed-off-by: Qu Wenruo <w...@suse.com>
> ---
> changelog
> v2:
> Fix regression for quota+cow case. (Previously it will skip data
> reservation if quota is enabled, causing regression for limit case.
> Pointed out by Misono)
> ---
> fs/btrfs/file.c | 18 +++++++++++++++++-
> 1 file changed, 17 insertions(+), 1 deletion(-)
>
> diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
> index 2be00e873e92..183e5fb96f42 100644
> --- a/fs/btrfs/file.c
> +++ b/fs/btrfs/file.c
> @@ -1584,6 +1584,7 @@ static noinline ssize_t btrfs_buffered_write(struct
> kiocb *iocb,
> int ret = 0;
> bool only_release_metadata = false;
> bool force_page_uptodate = false;
> + bool quota_enabled = test_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags);
>
> nrptrs = min(DIV_ROUND_UP(iov_iter_count(i), PAGE_SIZE),
> PAGE_SIZE / (sizeof(struct page *)));
> @@ -1624,13 +1625,28 @@ static noinline ssize_t btrfs_buffered_write(struct
> kiocb *iocb,
> fs_info->sectorsize);
>
> extent_changeset_release(data_reserved);
> +
> + /*
> + * If we have quota enabled, we must do the heavy lift nocow
> + * check here to avoid reserving data space, or we can hit
> + * limitation for NOCOW files.
> + */
> + if (quota_enabled) {
> + if ((BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW |
> + BTRFS_INODE_PREALLOC)) &&
> + check_can_nocow(BTRFS_I(inode), pos,
> + &write_bytes) > 0)
> + goto reserve_meta_only;
> + }
> ret = btrfs_check_data_free_space(inode, &data_reserved, pos,
> write_bytes);
> if (ret < 0) {
> if ((BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW |
> BTRFS_INODE_PREALLOC)) &&
> check_can_nocow(BTRFS_I(inode), pos,
> - &write_bytes) > 0) {
> + &write_bytes) > 0 &&
> + !quota_enabled) {
(When we check this condition, quota_enabled must be false or otherwise
we have already goto reserve_meta_only. So it seems redundant.)
> +reserve_meta_only:
> /*
> * For nodata cow case, no need to reserve
> * data space.
>
I applied this patch on today's misc-next and it seems mostly ok, but
btrfs/022 sometimes gives following warning:
[80244.152130] WARNING: CPU: 5 PID: 14575 at fs/btrfs/extent-tree.c:9742
btrfs_free_block_groups+0x2d7/0x440 [btrfs]
[80244.152132] Modules linked in: btrfs(O) xor zstd_decompress zstd_compress
xxhash raid6_pq xt_CHECKSUM ipt_MASQUERADE tun bridge stp llc xt_conntrack
ip_set nfnetlink iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4
nf_nat nf_conntrack iptable_mangle iptable_raw iptable_security sunrpc
intel_powerclamp kvm_intel kvm gpio_ich iTCO_wdt ipmi_ssif iTCO_vendor_support
ipmi_si st irqbypass ipmi_devintf crct10dif_pclmul crc32_pclmul ipmi_msghandler
ghash_clmulni_intel pcspkr acpi_power_meter i2c_i801 pcc_cpufreq i7core_edac
lpc_ich acpi_cpufreq xfs libcrc32c mgag200 drm_kms_helper syscopyarea
sysfillrect sysimgblt fb_sys_fops ttm drm igb sr_mod hwmon uas ptp crc32c_intel
cdrom usb_storage pps_core ata_generic megaraid_sas dca pata_acpi i2c_algo_bit
ipv6 [last unloaded: xor]
[80244.152185] CPU: 5 PID: 14575 Comm: umount Tainted: G W IO
4.18.0-rc8+ #98
[80244.152187] Hardware name: FUJITSU-SV PRIMERGY RX300
S6 /D2619, BIOS 6.00 Rev. 1.09.2619.N1 12/13/2010
[80244.152205] RIP: 0010:btrfs_free_block_groups+0x2d7/0x440 [btrfs]
[80244.152206] Code: 85 20 cb 00 00 48 39 c6 0f 84 b9 00 00 00 49 bf 00 01 00
00 00 00 ad de 48 8b 9d 20 cb 00 00 48 83 7b a0 00 0f 84 0d 01 00 00 <0f> 0b 48
8d 73 88 31 c9 31 d2 48 89 ef e8 27 7a ff ff 48 89 df e8
[80244.152235] RSP: 0018:ffff8ea10393fdb0 EFLAGS: 00010286
[80244.152237] RAX: ffff8c1025819e78 RBX: ffff8c1025819e78 RCX: 0000000000000000
[80244.152238] RDX: 0000000000000001 RSI: ffff8c115329cb20 RDI: ffff8c1025818e00
[80244.152239] RBP: ffff8c1153290000 R08: 0000000000000000 R09: 0000000000000000
[80244.152240] R10: ffff8c1025818e98 R11: 0000000000000002 R12: ffff8c11532900a0
[80244.152241] R13: 0000000000000000 R14: dead000000000200 R15: dead000000000100
[80244.152243] FS: 00007ff0cc3d8fc0(0000) GS:ffff8c1137d40000(0000)
knlGS:0000000000000000
[80244.152244] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[80244.152245] CR2: 00007ffe75e53c88 CR3: 0000000235d16001 CR4: 00000000000206e0
[80244.152246] Call Trace:
[80244.152270] close_ctree+0x146/0x320 [btrfs]
[80244.152276] ? kthread_stop+0x42/0xf0
[80244.152280] generic_shutdown_super+0x6c/0x110
[80244.152283] kill_anon_super+0xe/0x20
[80244.152298] btrfs_kill_super+0x13/0x100 [btrfs]
[80244.152301] deactivate_locked_super+0x3f/0x70
[80244.152303] cleanup_mnt+0x3b/0x70
[80244.152305] task_work_run+0x84/0xa0
[80244.152308] do_syscall_64+0x143/0x4cd
[80244.152311] ? do_page_fault+0x31/0x130
[80244.152314] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[80244.152316] RIP: 0033:0x7ff0cb43c1a7
[80244.152317] Code: ad 2b 00 f7 d8 64 89 01 48 83 c8 ff c3 66 0f 1f 44 00 00
31 f6 e9 09 00 00 00 66 0f 1f 84 00 00 00 00 00 b8 a6 00 00 00 0f 05 <48> 3d 01
f0 ff ff 73 01 c3 48 8b 0d c9 ac 2b 00 f7 d8 64 89 01 48
[80244.152346] RSP: 002b:00007ffe75e554b8 EFLAGS: 00000246 ORIG_RAX:
00000000000000a6
[80244.152348] RAX: 0000000000000000 RBX: 0000563ccced82d0 RCX: 00007ff0cb43c1a7
[80244.152349] RDX: 0000000000000001 RSI: 0000000000000000 RDI: 0000563ccced84b0
[80244.152350] RBP: 0000563ccced84b0 R08: 0000000000000005 R09: 0000563ccced84d0
[80244.152351] R10: 00007ff0cb4ba320 R11: 0000000000000246 R12: 00007ff0cc1d0184
[80244.152352] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
[80244.152354] ---[ end trace b5975ec96174c60c ]---
[80244.152358] BTRFS info (device sdh2): space_info 1 has 1022132224 free, is
not full
[80244.152360] BTRFS info (device sdh2): space_info total=1082130432,
used=60039168, pinned=0, reserved=0, may_use=18446744073709510656, readonly=0
Obviously, may_use value is underflowed.
Thanks,
Misono
