While running a new fstest that races a readonly remount with scrub
running in repair mode, I observed the kernel tripping over debugging
assertions in the log quiesce code that were checking that the CIL was
empty. When the sysadmin runs scrub in repair mode, the scrub code
allocates real transactions (with reservations) to change things, but
doesn't increment the superblock writers count to block a readonly
remount attempt while it is running.
We don't require the userspace caller to have a writable file descriptor
to run repairs, so we have to call mnt_want_write_file to obtain freeze
protection and increment the writers count. It's ok to remove the call
to sb_start_write for the dry-run case because commit 8321ddb2fa
removed the behavior where scrub and fsfreeze fight over the buffer LRU.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Fix the weird split of responsibilities between xfs_can_free_eofblocks
and xfs_free_eofblocks by moving the chunk of code that looks for any
actual post-EOF space mappings from the second function into the first.
This clears the way for deferred inode inactivation to be able to decide
if an inode needs inactivation work before committing the released inode
to the inactivation code paths (vs. marking it for reclaim).
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
In xfs_inode_free_eofblocks, move the xfs_can_free_eofblocks call
further down in the function to the point where we have taken the
IOLOCK. This is preparation for the next patch, where we will need that
lock (or equivalent) so that we can check if there are any post-eof
blocks to clean out.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Default attr fork offset is based on inode size, so is a fixed
geometry parameter of the inode. Move it to the xfs_ino_geometry
structure and stop calculating it on every call to
xfs_default_attroffset().
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Tested-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Device inodes have a non-default data fork size of 8 bytes
as checked/enforced by xfs_repair. xfs_default_attroffset() doesn't
handle this, so lets do a minor refactor so it does.
Fixes: e6a688c332 ("xfs: initialise attr fork on inode create")
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Tested-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Due to confusion on when the XFS_IFEXTENT needs to be set, the
changes in e6a688c332 ("xfs: initialise attr fork on inode
create") failed to set the flag when initialising the empty
attribute fork at inode creation. Set this flag the same way
xfs_bmap_add_attrfork() does after attry fork allocation.
Fixes: e6a688c332 ("xfs: initialise attr fork on inode create")
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Tested-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
The pitfalls of regression testing on a machine without realising
that selinux was disabled. Only set the attr fork during inode
allocation if the attr feature bits are already set on the
superblock.
Fixes: e6a688c332 ("xfs: initialise attr fork on inode create")
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Tested-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
xchk_btree_check_minrecs() checks if the contents of the immediate child of a
bmbt root block can fit within the root block. This check could fail on inodes
with an attr fork since xfs_bmap_add_attrfork_btree() used to demote the
current root node of the data fork as the child of a newly allocated root node
if it found that the size of "struct xfs_btree_block" along with the space
required for records exceeded that of space available in the data fork.
xfs_bmap_add_attrfork_btree() should have used "struct xfs_bmdr_block" instead
of "struct xfs_btree_block" for the above mentioned space requirement
calculation. This commit disables the check for unoptimized (in terms of
disk space usage) data fork bmbt trees since there could be filesystems
in use that already have such a layout.
Suggested-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
The incore data fork of an inode stores the bmap btree root node as 'struct
xfs_btree_block'. However, the ondisk version of the inode stores the bmap
btree root node as a 'struct xfs_bmdr_block'.
xfs_bmap_add_attrfork_btree() checks if the btree root node fits inside the
data fork of the inode. However, it incorrectly uses 'struct xfs_btree_block'
to compute the size of the bmap btree root node. Since size of 'struct
xfs_btree_block' is larger than that of 'struct xfs_bmdr_block',
xfs_bmap_add_attrfork_btree() could end up unnecessarily demoting the current
root node as the child of newly allocated root node.
This commit optimizes space usage by modifying xfs_bmap_add_attrfork_btree()
to use 'struct xfs_bmdr_block' to check if the bmap btree root node fits
inside the data fork of the inode.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Use of the flag has had no effect since kernel commit 288699feca
("xfs: drop dmapi hooks"), which removed all dmapi related code, so
deprecate it.
Signed-off-by: Anthony Iliopoulos <ailiop@suse.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Merge _xfs_dic2xflags into its only caller.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Move the crtime field from struct xfs_icdinode into stuct xfs_inode and
remove the now entirely unused struct xfs_icdinode.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the flags2
field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the flags
field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the
forkoff field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
The i_cowextsize field is only used for v3 inodes, and the i_flushiter
field is only used for v1/v2 inodes. Use a union to pack the inode a
littler better after adding a few missing guards around their usage.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Clean up xfs_ioctl_setattr a bit by using XFS_B_TO_FSB.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Add a local xfs_mount variable, and use the XFS_FSB_TO_B helper.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the
flushiter field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the
cowextsize field into the containing xfs_inode structure. Also
switch to use the xfs_extlen_t instead of a uint32_t.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the extsize
field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the nblocks
field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the on-disk
size field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
In preparation of removing the historic icinode struct, move the projid
field into the containing xfs_inode structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
The xfs_icdinode structure just contains a random mix of inode field,
which are all read from the on-disk inode and mostly not looked at
before reading the inode or initializing a new inode cluster. The
only exceptions are the forkoff and blocks field, which are used
in sanity checks for freshly allocated inodes.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
The legacy DMAPI fields were never set by upstream Linux XFS, and have no
way to be read using the kernel APIs. So instead of bloating the in-core
inode for them just copy them from the on-disk inode into the log when
logging the inode. The only caveat is that we need to make sure to zero
the fields for newly read or deleted inodes, which is solved using a new
flag in the inode.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
The crtime only exists for v5 inodes, so only copy it for those.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Make sure di_flags2 is always initialized. We currently get this implicitly
by clearing the dinode core on allocating the in-core inode, but that is
about to go away.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Split looking up the dinode from xfs_imap_to_bp, which can be
significantly simplified as a result.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
A directory with one directory block which in turns consists of two or more fs
blocks is incorrectly flagged as corrupt by scrub since it assumes that
"Block" format directories have a data fork single extent spanning the file
offset range of [0, Dir block size - 1].
This commit fixes the bug by removing the incorrect check.
Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
xfs/538 can cause the following call trace to be printed when executing on a
multi-block directory configuration,
WARNING: CPU: 1 PID: 2578 at fs/xfs/libxfs/xfs_bmap.c:717 xfs_bmap_extents_to_btree+0x520/0x5d0
Call Trace:
? xfs_buf_rele+0x4f/0x450
xfs_bmap_add_extent_hole_real+0x747/0x960
xfs_bmapi_allocate+0x39a/0x440
xfs_bmapi_write+0x507/0x9e0
xfs_da_grow_inode_int+0x1cd/0x330
? up+0x12/0x60
xfs_dir2_grow_inode+0x62/0x110
? xfs_trans_log_inode+0x234/0x2d0
xfs_dir2_sf_to_block+0x103/0x940
? xfs_dir2_sf_check+0x8c/0x210
? xfs_da_compname+0x19/0x30
? xfs_dir2_sf_lookup+0xd0/0x3d0
xfs_dir2_sf_addname+0x10d/0x910
xfs_dir_createname+0x1ad/0x210
xfs_create+0x404/0x620
xfs_generic_create+0x24c/0x320
path_openat+0xda6/0x1030
do_filp_open+0x88/0x130
? kmem_cache_alloc+0x50/0x210
? __cond_resched+0x16/0x40
? kmem_cache_alloc+0x50/0x210
do_sys_openat2+0x97/0x150
__x64_sys_creat+0x49/0x70
do_syscall_64+0x33/0x40
entry_SYSCALL_64_after_hwframe+0x44/0xae
This occurs because xfs_bmap_exact_minlen_extent_alloc() initializes
xfs_alloc_arg->total to xfs_bmalloca->minlen. In the context of
xfs_bmap_exact_minlen_extent_alloc(), xfs_bmalloca->minlen has a value of 1
and hence the space allocator could choose an AG which has less than
xfs_bmalloca->total number of free blocks available. As the transaction
proceeds, one of the future space allocation requests could fail due to
non-availability of free blocks in the AG that was originally chosen.
This commit fixes the bug by assigning xfs_alloc_arg->total to the value of
xfs_bmalloca->total.
Fixes: 3015196746 ("xfs: Introduce error injection to allocate only minlen size extents for files")
Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
With dax enabled filesystems, a direct write operation into an existing
unwritten extent results in xfs_iomap_write_direct() zero-ing and converting
the extent into a normal extent before the actual data is copied from the
userspace buffer.
The inode extent count can increase by 2 if the extent range being written to
maps to the middle of the existing unwritten extent range. Hence this commit
uses XFS_IEXT_WRITE_UNWRITTEN_CNT as the extent count delta when such a write
operation is being performed.
Fixes: 727e1acd29 ("xfs: Check for extent overflow when trivally adding a new extent")
Reported-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Removal of kmem_zone_init wrappers accidentally changed a slab cache
name from "xfs_trans" to "xf_trans". Fix this so that userspace
consumers of /proc/slabinfo and /sys/kernel/slab can find it again.
Fixes: b1231760e4 ("xfs: Remove slab init wrappers")
Signed-off-by: Anthony Iliopoulos <ailiop@suse.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
per-AG resv failure after fixing up freespace is hard to test in an
effective way, so directly add an error injection path to observe
such error handling path works as expected.
Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
As the first step of shrinking, this attempts to enable shrinking
unused space in the last allocation group by fixing up freespace
btree, agi, agf and adjusting super block and use a helper
xfs_ag_shrink_space() to fixup the last AG.
This can be all done in one transaction for now, so I think no
additional protection is needed.
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
This patch introduces a helper to shrink unused space in the last AG
by fixing up the freespace btree.
Also make sure that the per-AG reservation works under the new AG
size. If such per-AG reservation or extent allocation fails, roll
the transaction so the new transaction could cancel without any side
effects.
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Move out related logic for initializing new added AGs to a new helper
in preparation for shrinking. No logic changes.
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
sb_fdblocks will be updated lazily if lazysbcount is enabled,
therefore when shrinking the filesystem sb_fdblocks could be
larger than sb_dblocks and xfs_validate_sb_write() would fail.
Even for growfs case, it'd be better to update lazy sb counters
immediately to reflect the real sb counters.
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
s/strutures/structures/
Signed-off-by: Bhaskar Chowdhury <unixbhaskar@gmail.com>
Reviewed-by: Pavel Reichl <preichl@redhat.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
s/sytemcall/syscall/
Signed-off-by: Bhaskar Chowdhury <unixbhaskar@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
s/filesytem/filesystem/
s/instrumention/instrumentation/
Signed-off-by: Bhaskar Chowdhury <unixbhaskar@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
- 21.92% __xfs_trans_commit
- 21.62% xfs_log_commit_cil
- 11.69% xfs_trans_unreserve_and_mod_sb
- 11.58% __percpu_counter_compare
- 11.45% __percpu_counter_sum
- 10.29% _raw_spin_lock_irqsave
- 10.28% do_raw_spin_lock
__pv_queued_spin_lock_slowpath
We debated just getting rid of it last time this came up and
there was no real objection to removing it. Now it's the biggest
scalability limitation for debug kernels even on smallish machines,
so let's just get rid of it.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
On debug kernels, we call xfs_dir3_leaf_check_int() multiple times
on every directory modification. The robust hash ordering checks it
does on every entry in the leaf on every call results in a massive
CPU overhead which slows down debug kernels by a large amount.
We use xfs_dir3_leaf_check_int() for the verifiers as well, so we
can't just gut the function to reduce overhead. What we can do,
however, is reduce the work it does when it is called from the
debug interfaces, just leaving the high level checks in place and
leaving the robust validation to the verifiers. This means the debug
checks will catch gross errors, but subtle bugs might not be caught
until a verifier is run.
It is easy enough to restore the existing debug behaviour if the
developer needs it (just change a call parameter in the debug code),
but overwise the overhead makes testing large directory block sizes
on debug kernels very slow.
Profile at an unlink rate of ~80k file/s on a 64k block size
filesystem before the patch:
40.30% [kernel] [k] xfs_dir3_leaf_check_int
10.98% [kernel] [k] __xfs_dir3_data_check
8.10% [kernel] [k] xfs_verify_dir_ino
4.42% [kernel] [k] memcpy
2.22% [kernel] [k] xfs_dir2_data_get_ftype
1.52% [kernel] [k] do_raw_spin_lock
Profile after, at an unlink rate of ~125k files/s (+50% improvement)
has largely dropped the leaf verification debug overhead out of the
profile.
16.53% [kernel] [k] __xfs_dir3_data_check
12.53% [kernel] [k] xfs_verify_dir_ino
7.97% [kernel] [k] memcpy
3.36% [kernel] [k] xfs_dir2_data_get_ftype
2.86% [kernel] [k] __pv_queued_spin_lock_slowpath
Create shows a similar change in profile and a +25% improvement in
performance.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
We call xfs_dir_ino_validate() for every dir entry in a directory
when doing validity checking of the directory. It calls
xfs_verify_dir_ino() then emits a corruption report if bad or does
error injection if good. It is extremely costly:
43.27% [kernel] [k] xfs_dir3_leaf_check_int
10.28% [kernel] [k] __xfs_dir3_data_check
6.61% [kernel] [k] xfs_verify_dir_ino
4.16% [kernel] [k] xfs_errortag_test
4.00% [kernel] [k] memcpy
3.48% [kernel] [k] xfs_dir_ino_validate
7% of the cpu usage in this directory traversal workload is
xfs_dir_ino_validate() doing absolutely nothing.
We don't need error injection to simulate a bad inode numbers in the
directory structure because we can do that by fuzzing the structure
on disk.
And we don't need a corruption report, because the
__xfs_dir3_data_check() will emit one if the inode number is bad.
So just call xfs_verify_dir_ino() directly here, and get rid of all
this unnecessary overhead:
40.30% [kernel] [k] xfs_dir3_leaf_check_int
10.98% [kernel] [k] __xfs_dir3_data_check
8.10% [kernel] [k] xfs_verify_dir_ino
4.42% [kernel] [k] memcpy
2.22% [kernel] [k] xfs_dir2_data_get_ftype
1.52% [kernel] [k] do_raw_spin_lock
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
From a concurrent rm -rf workload:
41.04% [kernel] [k] xfs_dir3_leaf_check_int
9.85% [kernel] [k] __xfs_dir3_data_check
5.60% [kernel] [k] xfs_verify_ino
5.32% [kernel] [k] xfs_agino_range
4.21% [kernel] [k] memcpy
3.06% [kernel] [k] xfs_errortag_test
2.57% [kernel] [k] xfs_dir_ino_validate
1.66% [kernel] [k] xfs_dir2_data_get_ftype
1.17% [kernel] [k] do_raw_spin_lock
1.11% [kernel] [k] xfs_verify_dir_ino
0.84% [kernel] [k] __raw_callee_save___pv_queued_spin_unlock
0.83% [kernel] [k] xfs_buf_find
0.64% [kernel] [k] xfs_log_commit_cil
THere's an awful lot of overhead in just range checking inode
numbers in that, but each inode number check is not a lot of code.
The total is a bit over 14.5% of the CPU time is spent validating
inode numbers.
The problem is that they deeply nested global scope functions so the
overhead here is all in function call marshalling.
text data bss dec hex filename
2077 0 0 2077 81d fs/xfs/libxfs/xfs_types.o.orig
2197 0 0 2197 895 fs/xfs/libxfs/xfs_types.o
There's a small increase in binary size by inlining all the local
nested calls in the verifier functions, but the same workload now
profiles as:
40.69% [kernel] [k] xfs_dir3_leaf_check_int
10.52% [kernel] [k] __xfs_dir3_data_check
6.68% [kernel] [k] xfs_verify_dir_ino
4.22% [kernel] [k] xfs_errortag_test
4.15% [kernel] [k] memcpy
3.53% [kernel] [k] xfs_dir_ino_validate
1.87% [kernel] [k] xfs_dir2_data_get_ftype
1.37% [kernel] [k] do_raw_spin_lock
0.98% [kernel] [k] xfs_buf_find
0.94% [kernel] [k] __raw_callee_save___pv_queued_spin_unlock
0.73% [kernel] [k] xfs_log_commit_cil
Now we only spend just over 10% of the time validing inode numbers
for the same workload. Hence a few "inline" keyworks is good enough
to reduce the validation overhead by 30%...
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
We process the buf_log_item bitmap one set bit at a time with
xfs_next_bit() so we can detect if a region crosses a memcpy
discontinuity in the buffer data address. This has massive overhead
on large buffers (e.g. 64k directory blocks) because we do a lot of
unnecessary checks and xfs_buf_offset() calls.
For example, 16-way concurrent create workload on debug kernel
running CPU bound has this at the top of the profile at ~120k
create/s on 64kb directory block size:
20.66% [kernel] [k] xfs_dir3_leaf_check_int
7.10% [kernel] [k] memcpy
6.22% [kernel] [k] xfs_next_bit
3.55% [kernel] [k] xfs_buf_offset
3.53% [kernel] [k] xfs_buf_item_format
3.34% [kernel] [k] __pv_queued_spin_lock_slowpath
3.04% [kernel] [k] do_raw_spin_lock
2.84% [kernel] [k] xfs_buf_item_size_segment.isra.0
2.31% [kernel] [k] __raw_callee_save___pv_queued_spin_unlock
1.36% [kernel] [k] xfs_log_commit_cil
(debug checks hurt large blocks)
The only buffers with discontinuities in the data address are
unmapped buffers, and they are only used for inode cluster buffers
and only for logging unlinked pointers. IOWs, it is -rare- that we
even need to detect a discontinuity in the buffer item formatting
code.
Optimise all this by using xfs_contig_bits() to find the size of
the contiguous regions, then test for a discontiunity inside it. If
we find one, do the slow "bit at a time" method we do now. If we
don't, then just copy the entire contiguous range in one go.
Profile now looks like:
25.26% [kernel] [k] xfs_dir3_leaf_check_int
9.25% [kernel] [k] memcpy
5.01% [kernel] [k] __pv_queued_spin_lock_slowpath
2.84% [kernel] [k] do_raw_spin_lock
2.22% [kernel] [k] __raw_callee_save___pv_queued_spin_unlock
1.88% [kernel] [k] xfs_buf_find
1.53% [kernel] [k] memmove
1.47% [kernel] [k] xfs_log_commit_cil
....
0.34% [kernel] [k] xfs_buf_item_format
....
0.21% [kernel] [k] xfs_buf_offset
....
0.16% [kernel] [k] xfs_contig_bits
....
0.13% [kernel] [k] xfs_buf_item_size_segment.isra.0
So the bit scanning over for the dirty region tracking for the
buffer log items is basically gone. Debug overhead hurts even more
now...
Perf comparison
dir block creates unlink
size (kb) time rate time
Original 4 4m08s 220k 5m13s
Original 64 7m21s 115k 13m25s
Patched 4 3m59s 230k 5m03s
Patched 64 6m23s 143k 12m33s
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Otherwise it doesn't correctly calculate the number of vectors
in a logged buffer that has a contiguous map that gets split into
multiple regions because the range spans discontigous memory.
Probably never been hit in practice - we don't log contiguous ranges
on unmapped buffers (inode clusters).
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
When we modify btrees repeatedly, we regularly increase the size of
the logged region by a single chunk at a time (per transaction
commit). This results in the CIL formatting code having to
reallocate the log vector buffer every time the buffer dirty region
grows. Hence over a typical 4kB btree buffer, we might grow the log
vector 4096/128 = 32x over a short period where we repeatedly add
or remove records to/from the buffer over a series of running
transaction. This means we are doing 32 memory allocations and frees
over this time during a performance critical path in the journal.
The amount of space tracked in the CIL for the object is calculated
during the ->iop_format() call for the buffer log item, but the
buffer memory allocated for it is calculated by the ->iop_size()
call. The size callout determines the size of the buffer, the format
call determines the space used in the buffer.
Hence we can oversize the buffer space required in the size
calculation without impacting the amount of space used and accounted
to the CIL for the changes being logged. This allows us to reduce
the number of allocations by rounding up the buffer size to allow
for future growth. This can safe a substantial amount of CPU time in
this path:
- 46.52% 2.02% [kernel] [k] xfs_log_commit_cil
- 44.49% xfs_log_commit_cil
- 30.78% _raw_spin_lock
- 30.75% do_raw_spin_lock
30.27% __pv_queued_spin_lock_slowpath
(oh, ouch!)
....
- 1.05% kmem_alloc_large
- 1.02% kmem_alloc
0.94% __kmalloc
This overhead here us what this patch is aimed at. After:
- 0.76% kmem_alloc_large
- 0.75% kmem_alloc
0.70% __kmalloc
The size of 512 bytes is based on the bitmap chunk size being 128
bytes and that random directory entry updates almost never require
more than 3-4 128 byte regions to be logged in the directory block.
The other observation is for per-ag btrees. When we are inserting
into a new btree block, we'll pack it from the front. Hence the
first few records land in the first 128 bytes so we log only 128
bytes, the next 8-16 records land in the second region so now we log
256 bytes. And so on. If we are doing random updates, it will only
allocate every 4 random 128 byte regions that are dirtied instead of
every single one.
Any larger than 512 bytes and I noticed an increase in memory
footprint in my scalability workloads. Any less than this and I
didn't really see any significant benefit to CPU usage.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Gao Xiang <hsiangkao@redhat.com>
When we allocate a new inode, we often need to add an attribute to
the inode as part of the create. This can happen as a result of
needing to add default ACLs or security labels before the inode is
made visible to userspace.
This is highly inefficient right now. We do the create transaction
to allocate the inode, then we do an "add attr fork" transaction to
modify the just created empty inode to set the inode fork offset to
allow attributes to be stored, then we go and do the attribute
creation.
This means 3 transactions instead of 1 to allocate an inode, and
this greatly increases the load on the CIL commit code, resulting in
excessive contention on the CIL spin locks and performance
degradation:
18.99% [kernel] [k] __pv_queued_spin_lock_slowpath
3.57% [kernel] [k] do_raw_spin_lock
2.51% [kernel] [k] __raw_callee_save___pv_queued_spin_unlock
2.48% [kernel] [k] memcpy
2.34% [kernel] [k] xfs_log_commit_cil
The typical profile resulting from running fsmark on a selinux enabled
filesytem is adds this overhead to the create path:
- 15.30% xfs_init_security
- 15.23% security_inode_init_security
- 13.05% xfs_initxattrs
- 12.94% xfs_attr_set
- 6.75% xfs_bmap_add_attrfork
- 5.51% xfs_trans_commit
- 5.48% __xfs_trans_commit
- 5.35% xfs_log_commit_cil
- 3.86% _raw_spin_lock
- do_raw_spin_lock
__pv_queued_spin_lock_slowpath
- 0.70% xfs_trans_alloc
0.52% xfs_trans_reserve
- 5.41% xfs_attr_set_args
- 5.39% xfs_attr_set_shortform.constprop.0
- 4.46% xfs_trans_commit
- 4.46% __xfs_trans_commit
- 4.33% xfs_log_commit_cil
- 2.74% _raw_spin_lock
- do_raw_spin_lock
__pv_queued_spin_lock_slowpath
0.60% xfs_inode_item_format
0.90% xfs_attr_try_sf_addname
- 1.99% selinux_inode_init_security
- 1.02% security_sid_to_context_force
- 1.00% security_sid_to_context_core
- 0.92% sidtab_entry_to_string
- 0.90% sidtab_sid2str_get
0.59% sidtab_sid2str_put.part.0
- 0.82% selinux_determine_inode_label
- 0.77% security_transition_sid
0.70% security_compute_sid.part.0
And fsmark creation rate performance drops by ~25%. The key point to
note here is that half the additional overhead comes from adding the
attribute fork to the newly created inode. That's crazy, considering
we can do this same thing at inode create time with a couple of
lines of code and no extra overhead.
So, if we know we are going to add an attribute immediately after
creating the inode, let's just initialise the attribute fork inside
the create transaction and chop that whole chunk of code out of
the create fast path. This completely removes the performance
drop caused by enabling SELinux, and the profile looks like:
- 8.99% xfs_init_security
- 9.00% security_inode_init_security
- 6.43% xfs_initxattrs
- 6.37% xfs_attr_set
- 5.45% xfs_attr_set_args
- 5.42% xfs_attr_set_shortform.constprop.0
- 4.51% xfs_trans_commit
- 4.54% __xfs_trans_commit
- 4.59% xfs_log_commit_cil
- 2.67% _raw_spin_lock
- 3.28% do_raw_spin_lock
3.08% __pv_queued_spin_lock_slowpath
0.66% xfs_inode_item_format
- 0.90% xfs_attr_try_sf_addname
- 0.60% xfs_trans_alloc
- 2.35% selinux_inode_init_security
- 1.25% security_sid_to_context_force
- 1.21% security_sid_to_context_core
- 1.19% sidtab_entry_to_string
- 1.20% sidtab_sid2str_get
- 0.86% sidtab_sid2str_put.part.0
- 0.62% _raw_spin_lock_irqsave
- 0.77% do_raw_spin_lock
__pv_queued_spin_lock_slowpath
- 0.84% selinux_determine_inode_label
- 0.83% security_transition_sid
0.86% security_compute_sid.part.0
Which indicates the XFS overhead of creating the selinux xattr has
been halved. This doesn't fix the CIL lock contention problem, just
means it's not a limiting factor for this workload. Lock contention
in the security subsystems is going to be an issue soon, though...
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
[djwong: fix compilation error when CONFIG_SECURITY=n]
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Gao Xiang <hsiangkao@redhat.com>
Darrick J. Wong