| Message ID | 20220314175532.GB8165@magnolia (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] btrfs: fix fallocate to use file_modified to update permissions consistently | expand |
On Mon, Mar 14, 2022 at 10:55:32AM -0700, Darrick J. Wong wrote: > From: Darrick J. Wong <djwong@kernel.org> > > Since the initial introduction of (posix) fallocate back at the turn of > the century, it has been possible to use this syscall to change the > user-visible contents of files. This can happen by extending the file > size during a preallocation, or through any of the newer modes (punch, > zero range). Because the call can be used to change file contents, we > should treat it like we do any other modification to a file -- update > the mtime, and drop set[ug]id privileges/capabilities. > > The VFS function file_modified() does all this for us if pass it a > locked inode, so let's make fallocate drop permissions correctly. > > Signed-off-by: Darrick J. Wong <djwong@kernel.org> > --- > v2: move up file_modified to catch a case where we could modify file > contents but fail on something else before we end up calling > file_modified Reviewed-by: Filipe Manana <fdmanana@suse.com> Thanks, it looks good Darrick. It also survived a few full fstests runs. > --- > fs/btrfs/file.c | 13 +++++++++++-- > 1 file changed, 11 insertions(+), 2 deletions(-) > > diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c > index a0179cc62913..28ddd9cf2069 100644 > --- a/fs/btrfs/file.c > +++ b/fs/btrfs/file.c > @@ -2918,8 +2918,9 @@ int btrfs_replace_file_extents(struct btrfs_inode *inode, > return ret; > } > > -static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) > +static int btrfs_punch_hole(struct file *file, loff_t offset, loff_t len) > { > + struct inode *inode = file_inode(file); > struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); > struct btrfs_root *root = BTRFS_I(inode)->root; > struct extent_state *cached_state = NULL; > @@ -2951,6 +2952,10 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) > goto out_only_mutex; > } > > + ret = file_modified(file); > + if (ret) > + goto out_only_mutex; > + > lockstart = round_up(offset, btrfs_inode_sectorsize(BTRFS_I(inode))); > lockend = round_down(offset + len, > btrfs_inode_sectorsize(BTRFS_I(inode))) - 1; > @@ -3391,7 +3396,7 @@ static long btrfs_fallocate(struct file *file, int mode, > return -EOPNOTSUPP; > > if (mode & FALLOC_FL_PUNCH_HOLE) > - return btrfs_punch_hole(inode, offset, len); > + return btrfs_punch_hole(file, offset, len); > > /* > * Only trigger disk allocation, don't trigger qgroup reserve > @@ -3413,6 +3418,10 @@ static long btrfs_fallocate(struct file *file, int mode, > goto out; > } > > + ret = file_modified(file); > + if (ret) > + goto out; > + > /* > * TODO: Move these two operations after we have checked > * accurate reserved space, or fallocate can still fail but
On Mon, Mar 14, 2022 at 10:55:32AM -0700, Darrick J. Wong wrote: > From: Darrick J. Wong <djwong@kernel.org> > > Since the initial introduction of (posix) fallocate back at the turn of > the century, it has been possible to use this syscall to change the > user-visible contents of files. This can happen by extending the file > size during a preallocation, or through any of the newer modes (punch, > zero range). Because the call can be used to change file contents, we > should treat it like we do any other modification to a file -- update > the mtime, and drop set[ug]id privileges/capabilities. > > The VFS function file_modified() does all this for us if pass it a > locked inode, so let's make fallocate drop permissions correctly. > > Signed-off-by: Darrick J. Wong <djwong@kernel.org> > --- > v2: move up file_modified to catch a case where we could modify file > contents but fail on something else before we end up calling > file_modified Added to misc-next, thanks.
Greeting,
FYI, we noticed a 29.3% improvement of fio.write_iops due to commit:
commit: 7664aed02431897aff0c2c96c683176ba7ba72f6 ("[PATCH v2] btrfs: fix fallocate to use file_modified to update permissions consistently")
url: https://github.com/0day-ci/linux/commits/Darrick-J-Wong/btrfs-fix-fallocate-to-use-file_modified-to-update-permissions-consistently/20220315-015717
base: https://git.kernel.org/cgit/linux/kernel/git/kdave/linux.git for-next
patch link: https://lore.kernel.org/linux-btrfs/20220314175532.GB8165@magnolia
in testcase: fio-basic
on test machine: 96 threads 2 sockets Ice Lake with 256G memory
with following parameters:
runtime: 300s
disk: 1HDD
fs: btrfs
nr_task: 100%
test_size: 128G
rw: randwrite
bs: 4k
ioengine: ftruncate
cpufreq_governor: performance
ucode: 0xb000280
test-description: Fio is a tool that will spawn a number of threads or processes doing a particular type of I/O action as specified by the user.
test-url: https://github.com/axboe/fio
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
sudo bin/lkp install job.yaml # job file is attached in this email
bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
sudo bin/lkp run generated-yaml-file
# if come across any failure that blocks the test,
# please remove ~/.lkp and /lkp dir to run from a clean state.
=========================================================================================
bs/compiler/cpufreq_governor/disk/fs/ioengine/kconfig/nr_task/rootfs/runtime/rw/tbox_group/test_size/testcase/ucode:
4k/gcc-9/performance/1HDD/btrfs/ftruncate/x86_64-rhel-8.3/100%/debian-10.4-x86_64-20200603.cgz/300s/randwrite/lkp-icl-2sp1/128G/fio-basic/0xb000280
commit:
89a925d756 ("Merge branch 'for-next-current-v5.16-20220314' into for-next-20220314")
7664aed024 ("btrfs: fix fallocate to use file_modified to update permissions consistently")
89a925d7560844e1 7664aed02431897aff0c2c96c68
---------------- ---------------------------
%stddev %change %stddev
\ | \
25.54 -15.7 9.87 fio.latency_1000us%
9.23 ± 4% +9.9 19.17 fio.latency_10us%
9.58 +3.5 13.06 ± 2% fio.latency_20us%
0.31 ± 21% +0.3 0.62 ± 11% fio.latency_250us%
31.76 -4.3 27.44 fio.latency_2ms%
0.49 ± 29% +2.1 2.62 ± 10% fio.latency_4us%
4.48 -2.0 2.50 ± 3% fio.latency_500us%
1.29 ± 5% +0.4 1.64 ± 15% fio.latency_50us%
16.58 +5.5 22.06 fio.latency_750us%
274.60 -22.6% 212.58 fio.time.elapsed_time
274.60 -22.6% 212.58 fio.time.elapsed_time.max
4151 -25.2% 3103 ± 3% fio.time.involuntary_context_switches
39815 -19.3% 32135 ± 2% fio.time.minor_page_faults
949.33 -10.6% 848.83 fio.time.percent_of_cpu_this_job_got
2577 -31.1% 1774 fio.time.system_time
1.044e+08 -31.4% 71667735 fio.time.voluntary_context_switches
478.06 +29.3% 617.97 fio.write_bw_MBps
1318912 -6.4% 1234261 fio.write_clat_90%_us
1378986 -5.9% 1297066 fio.write_clat_95%_us
778778 -23.0% 599675 fio.write_clat_mean_us
122382 +29.3% 158201 fio.write_iops
485.84 -2.2% 475.03 pmeter.Average_Active_Power
2.374e+10 -21.6% 1.86e+10 cpuidle..time
1.498e+08 -28.4% 1.072e+08 cpuidle..usage
89.67 +1.3% 90.80 iostat.cpu.idle
10.18 -11.4% 9.02 iostat.cpu.system
81910 ± 15% -29.1% 58108 ± 21% meminfo.Active
2072 ± 2% -15.4% 1752 ± 2% meminfo.Active(file)
326.84 -19.2% 263.93 uptime.boot
28026 -18.5% 22848 uptime.idle
246.33 +22.1% 300.67 ± 5% vmstat.io.bo
748973 -11.5% 662479 vmstat.system.cs
0.01 ± 2% +0.0 0.01 ± 3% mpstat.cpu.all.iowait%
9.49 -1.1 8.35 mpstat.cpu.all.sys%
0.14 ± 5% +0.0 0.18 ± 5% mpstat.cpu.all.usr%
850.50 ± 22% +48.1% 1259 ± 6% numa-meminfo.node0.Active(file)
833.17 ± 89% +1557.5% 13809 ± 3% numa-meminfo.node0.Inactive(file)
75084 ± 11% -31.1% 51711 ± 18% numa-meminfo.node1.Active
73861 ± 11% -30.6% 51230 ± 19% numa-meminfo.node1.Active(anon)
1222 ± 15% -60.7% 480.00 ± 23% numa-meminfo.node1.Active(file)
96944 ± 20% +80.8% 175278 ± 30% numa-meminfo.node1.Inactive
83244 ± 24% +110.2% 174993 ± 30% numa-meminfo.node1.Inactive(anon)
13700 ± 5% -97.9% 284.67 ± 43% numa-meminfo.node1.Inactive(file)
25892 ± 33% +84.7% 47819 ± 26% numa-meminfo.node1.Mapped
517.00 ± 2% -15.9% 434.67 ± 3% proc-vmstat.nr_active_file
72537 +2.2% 74150 proc-vmstat.nr_anon_pages
17664 -13.5% 15278 proc-vmstat.nr_dirtied
3057 +3.8% 3174 proc-vmstat.nr_page_table_pages
16195 -9.3% 14686 ± 2% proc-vmstat.nr_written
517.00 ± 2% -15.9% 434.67 ± 3% proc-vmstat.nr_zone_active_file
1222395 ± 3% -14.6% 1043907 ± 5% proc-vmstat.pgfault
69030 -5.3% 65348 ± 4% proc-vmstat.pgpgout
59434 -20.0% 47550 proc-vmstat.pgreuse
375.33 -9.4% 340.17 turbostat.Avg_MHz
11.81 -1.1 10.72 turbostat.Busy%
803952 ± 4% -20.1% 642465 ± 11% turbostat.C1
1.474e+08 -28.6% 1.052e+08 turbostat.C1E
63.17 ± 3% +11.6% 70.50 ± 3% turbostat.CoreTmp
56497141 -22.1% 44010000 turbostat.IRQ
222267 -27.2% 161739 ± 2% turbostat.POLL
0.05 -0.0 0.04 turbostat.POLL%
63.33 ± 3% +11.6% 70.67 ± 3% turbostat.PkgTmp
206.89 +2.5% 212.16 turbostat.RAMWatt
211.00 ± 22% +49.3% 315.00 ± 6% numa-vmstat.node0.nr_active_file
1852 ± 53% +352.7% 8387 ± 5% numa-vmstat.node0.nr_dirtied
208.67 ± 89% +1552.1% 3447 ± 3% numa-vmstat.node0.nr_inactive_file
1806 ± 54% +360.4% 8317 ± 5% numa-vmstat.node0.nr_written
211.00 ± 22% +49.3% 315.00 ± 6% numa-vmstat.node0.nr_zone_active_file
208.67 ± 89% +1552.1% 3447 ± 3% numa-vmstat.node0.nr_zone_inactive_file
304.17 ± 15% -60.3% 120.83 ± 23% numa-vmstat.node1.nr_active_file
8877 ± 11% -92.1% 700.00 ± 80% numa-vmstat.node1.nr_dirtied
20207 ± 25% +115.1% 43458 ± 29% numa-vmstat.node1.nr_inactive_anon
3424 ± 5% -97.9% 70.83 ± 44% numa-vmstat.node1.nr_inactive_file
8813 ± 11% -92.5% 663.83 ± 83% numa-vmstat.node1.nr_written
304.17 ± 15% -60.3% 120.83 ± 23% numa-vmstat.node1.nr_zone_active_file
20208 ± 25% +115.1% 43458 ± 29% numa-vmstat.node1.nr_zone_inactive_anon
3424 ± 5% -97.9% 70.83 ± 44% numa-vmstat.node1.nr_zone_inactive_file
1.698e+09 +3.7% 1.761e+09 ± 2% perf-stat.i.branch-instructions
43843771 +4.5% 45825714 perf-stat.i.cache-misses
757981 -11.3% 672443 perf-stat.i.context-switches
4.42 -13.3% 3.83 perf-stat.i.cpi
3.73e+10 -9.6% 3.373e+10 perf-stat.i.cpu-cycles
2.303e+09 +4.8% 2.413e+09 perf-stat.i.dTLB-loads
1.214e+09 +6.4% 1.292e+09 perf-stat.i.dTLB-stores
8.441e+09 +4.5% 8.819e+09 perf-stat.i.instructions
0.23 +15.6% 0.27 perf-stat.i.ipc
18.63 +29.2% 24.08 perf-stat.i.major-faults
0.39 -9.6% 0.35 perf-stat.i.metric.GHz
55.76 +4.7% 58.39 perf-stat.i.metric.M/sec
309205 ± 3% +10.9% 342860 ± 3% perf-stat.i.node-loads
6312089 +4.1% 6571951 ± 2% perf-stat.i.node-store-misses
4.42 -13.4% 3.83 perf-stat.overall.cpi
850.93 -13.5% 736.28 perf-stat.overall.cycles-between-cache-misses
0.23 +15.5% 0.26 perf-stat.overall.ipc
69450 -19.1% 56170 perf-stat.overall.path-length
1.691e+09 +3.6% 1.752e+09 ± 2% perf-stat.ps.branch-instructions
43679256 +4.4% 45604378 perf-stat.ps.cache-misses
755089 -11.4% 669172 perf-stat.ps.context-switches
3.716e+10 -9.7% 3.357e+10 perf-stat.ps.cpu-cycles
2.295e+09 +4.7% 2.402e+09 perf-stat.ps.dTLB-loads
1.21e+09 +6.3% 1.286e+09 perf-stat.ps.dTLB-stores
8.41e+09 +4.4% 8.778e+09 perf-stat.ps.instructions
18.53 +29.1% 23.92 perf-stat.ps.major-faults
308133 ± 3% +10.8% 341314 ± 3% perf-stat.ps.node-loads
6288209 +4.0% 6540203 ± 2% perf-stat.ps.node-store-misses
2.33e+12 -19.1% 1.885e+12 perf-stat.total.instructions
21.14 -7.8 13.34 perf-profile.calltrace.cycles-pp.btrfs_search_slot.btrfs_truncate_inode_items.btrfs_setsize.btrfs_setattr.notify_change
17.67 -7.6 10.11 ± 2% perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock_irq.rwsem_down_read_slowpath.__btrfs_tree_read_lock.btrfs_read_lock_root_node
24.00 -7.5 16.50 perf-profile.calltrace.cycles-pp.btrfs_truncate_inode_items.btrfs_setsize.btrfs_setattr.notify_change.do_truncate
18.11 -7.4 10.71 ± 2% perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.rwsem_down_read_slowpath.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot
5.86 -4.8 1.05 ± 2% perf-profile.calltrace.cycles-pp.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_lookup_file_extent.btrfs_get_extent
5.96 -4.8 1.18 ± 2% perf-profile.calltrace.cycles-pp.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_lookup_file_extent.btrfs_get_extent.btrfs_cont_expand
5.61 -4.6 0.97 ± 3% perf-profile.calltrace.cycles-pp.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_truncate_inode_items.btrfs_setsize
5.74 -4.6 1.11 ± 2% perf-profile.calltrace.cycles-pp.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_truncate_inode_items.btrfs_setsize.btrfs_setattr
5.41 -4.5 0.89 ± 2% perf-profile.calltrace.cycles-pp.rwsem_down_read_slowpath.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_truncate_inode_items
7.12 -4.1 3.00 perf-profile.calltrace.cycles-pp.btrfs_lookup_file_extent.btrfs_get_extent.btrfs_cont_expand.btrfs_setsize.btrfs_setattr
7.11 -4.1 2.99 perf-profile.calltrace.cycles-pp.btrfs_search_slot.btrfs_lookup_file_extent.btrfs_get_extent.btrfs_cont_expand.btrfs_setsize
7.89 ± 2% -3.9 3.97 perf-profile.calltrace.cycles-pp.btrfs_get_extent.btrfs_cont_expand.btrfs_setsize.btrfs_setattr.notify_change
14.14 ± 2% -3.5 10.61 ± 2% perf-profile.calltrace.cycles-pp.rwsem_down_write_slowpath.__btrfs_tree_lock.btrfs_lock_root_node.btrfs_search_slot.btrfs_truncate_inode_items
14.35 ± 2% -3.5 10.86 ± 2% perf-profile.calltrace.cycles-pp.__btrfs_tree_lock.btrfs_lock_root_node.btrfs_search_slot.btrfs_truncate_inode_items.btrfs_setsize
14.40 ± 2% -3.5 10.91 ± 2% perf-profile.calltrace.cycles-pp.btrfs_lock_root_node.btrfs_search_slot.btrfs_truncate_inode_items.btrfs_setsize.btrfs_setattr
10.27 -2.4 7.83 perf-profile.calltrace.cycles-pp.rwsem_down_read_slowpath.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_lookup_file_extent
77.81 -1.1 76.68 perf-profile.calltrace.cycles-pp.do_truncate.do_sys_ftruncate.do_syscall_64.entry_SYSCALL_64_after_hwframe.ftruncate64
77.91 -1.1 76.81 perf-profile.calltrace.cycles-pp.do_sys_ftruncate.do_syscall_64.entry_SYSCALL_64_after_hwframe.ftruncate64
77.98 -1.1 76.90 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.ftruncate64
78.01 -1.1 76.95 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.ftruncate64
78.31 -1.0 77.27 perf-profile.calltrace.cycles-pp.ftruncate64
0.67 -0.1 0.56 ± 3% perf-profile.calltrace.cycles-pp.rwsem_spin_on_owner.rwsem_down_write_slowpath.__btrfs_tree_lock.btrfs_lock_root_node.btrfs_search_slot
0.55 ± 3% +0.0 0.60 ± 4% perf-profile.calltrace.cycles-pp.irq_exit_rcu.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state.cpuidle_enter
0.58 ± 5% +0.1 0.67 ± 3% perf-profile.calltrace.cycles-pp.__hrtimer_run_queues.hrtimer_interrupt.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt
0.61 ± 3% +0.2 0.76 ± 4% perf-profile.calltrace.cycles-pp.try_to_wake_up.wake_up_q.rwsem_wake.btrfs_unlock_up_safe.btrfs_search_slot
0.64 ± 3% +0.2 0.80 ± 4% perf-profile.calltrace.cycles-pp.wake_up_q.rwsem_wake.btrfs_unlock_up_safe.btrfs_search_slot.btrfs_insert_empty_items
0.68 ± 5% +0.2 0.87 ± 3% perf-profile.calltrace.cycles-pp.btrfs_del_items.btrfs_truncate_inode_items.btrfs_setsize.btrfs_setattr.notify_change
0.87 ± 3% +0.3 1.20 ± 3% perf-profile.calltrace.cycles-pp.rwsem_wake.btrfs_unlock_up_safe.btrfs_search_slot.btrfs_insert_empty_items.btrfs_insert_file_extent
0.59 ± 3% +0.3 0.92 ± 30% perf-profile.calltrace.cycles-pp.start_transaction.btrfs_setsize.btrfs_setattr.notify_change.do_truncate
0.73 ± 5% +0.3 1.06 ± 4% perf-profile.calltrace.cycles-pp.memmove.memcpy_extent_buffer.setup_items_for_insert.btrfs_insert_empty_items.btrfs_insert_file_extent
0.91 ± 3% +0.3 1.26 ± 3% perf-profile.calltrace.cycles-pp.btrfs_unlock_up_safe.btrfs_search_slot.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand
0.78 ± 6% +0.4 1.14 ± 4% perf-profile.calltrace.cycles-pp.memcpy_extent_buffer.setup_items_for_insert.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand
11.54 ± 2% +0.4 11.96 perf-profile.calltrace.cycles-pp.intel_idle.cpuidle_enter_state.cpuidle_enter.do_idle.cpu_startup_entry
0.56 ± 3% +0.5 1.08 ± 14% perf-profile.calltrace.cycles-pp.btrfs_dirty_inode.btrfs_setattr.notify_change.do_truncate.do_sys_ftruncate
0.08 ±223% +0.5 0.63 ± 3% perf-profile.calltrace.cycles-pp.btrfs_get_token_32.setup_items_for_insert.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand
0.00 +0.5 0.55 ± 3% perf-profile.calltrace.cycles-pp.schedule.rwsem_down_write_slowpath.__btrfs_tree_lock.btrfs_lock_root_node.btrfs_search_slot
14.93 ± 2% +0.6 15.55 ± 2% perf-profile.calltrace.cycles-pp.cpuidle_enter_state.cpuidle_enter.do_idle.cpu_startup_entry.start_secondary
15.22 ± 2% +0.7 15.88 ± 2% perf-profile.calltrace.cycles-pp.cpuidle_enter.do_idle.cpu_startup_entry.start_secondary.secondary_startup_64_no_verify
1.63 ± 5% +0.7 2.29 ± 3% perf-profile.calltrace.cycles-pp.setup_items_for_insert.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand.btrfs_setsize
20.42 ± 2% +0.8 21.22 perf-profile.calltrace.cycles-pp.secondary_startup_64_no_verify
20.12 ± 2% +0.8 20.93 perf-profile.calltrace.cycles-pp.do_idle.cpu_startup_entry.start_secondary.secondary_startup_64_no_verify
20.19 ± 2% +0.8 20.99 perf-profile.calltrace.cycles-pp.start_secondary.secondary_startup_64_no_verify
20.18 ± 2% +0.8 20.98 perf-profile.calltrace.cycles-pp.cpu_startup_entry.start_secondary.secondary_startup_64_no_verify
3.79 ± 3% +1.2 5.04 ± 2% perf-profile.calltrace.cycles-pp.rwsem_down_read_slowpath.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_insert_empty_items
4.00 ± 2% +1.3 5.32 ± 2% perf-profile.calltrace.cycles-pp.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_insert_empty_items.btrfs_insert_file_extent
4.11 ± 3% +1.4 5.47 ± 2% perf-profile.calltrace.cycles-pp.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand
15.63 ± 2% +1.9 17.51 perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock_irq.rwsem_down_write_slowpath.__btrfs_tree_lock.btrfs_lock_root_node
16.14 +1.9 18.06 perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.rwsem_down_write_slowpath.__btrfs_tree_lock.btrfs_lock_root_node.btrfs_search_slot
4.85 +2.3 7.14 perf-profile.calltrace.cycles-pp.__btrfs_tree_read_lock.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_lookup_file_extent.btrfs_drop_extents
4.97 ± 2% +2.3 7.29 perf-profile.calltrace.cycles-pp.btrfs_read_lock_root_node.btrfs_search_slot.btrfs_lookup_file_extent.btrfs_drop_extents.btrfs_cont_expand
6.14 +2.9 9.02 perf-profile.calltrace.cycles-pp.btrfs_search_slot.btrfs_lookup_file_extent.btrfs_drop_extents.btrfs_cont_expand.btrfs_setsize
6.15 +2.9 9.04 perf-profile.calltrace.cycles-pp.btrfs_lookup_file_extent.btrfs_drop_extents.btrfs_cont_expand.btrfs_setsize.btrfs_setattr
72.34 +3.1 75.38 perf-profile.calltrace.cycles-pp.btrfs_setsize.btrfs_setattr.notify_change.do_truncate.do_sys_ftruncate
6.84 +3.1 9.96 perf-profile.calltrace.cycles-pp.btrfs_drop_extents.btrfs_cont_expand.btrfs_setsize.btrfs_setattr.notify_change
73.03 +3.6 76.62 perf-profile.calltrace.cycles-pp.notify_change.do_truncate.do_sys_ftruncate.do_syscall_64.entry_SYSCALL_64_after_hwframe
72.95 +3.6 76.55 perf-profile.calltrace.cycles-pp.btrfs_setattr.notify_change.do_truncate.do_sys_ftruncate.do_syscall_64
21.84 ± 2% +6.9 28.79 perf-profile.calltrace.cycles-pp.rwsem_down_write_slowpath.__btrfs_tree_lock.btrfs_lock_root_node.btrfs_search_slot.btrfs_insert_empty_items
22.01 ± 2% +7.0 29.02 perf-profile.calltrace.cycles-pp.__btrfs_tree_lock.btrfs_lock_root_node.btrfs_search_slot.btrfs_insert_empty_items.btrfs_insert_file_extent
22.05 ± 2% +7.0 29.08 perf-profile.calltrace.cycles-pp.btrfs_lock_root_node.btrfs_search_slot.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand
27.98 +9.1 37.05 perf-profile.calltrace.cycles-pp.btrfs_search_slot.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand.btrfs_setsize
46.10 +9.5 55.63 perf-profile.calltrace.cycles-pp.btrfs_cont_expand.btrfs_setsize.btrfs_setattr.notify_change.do_truncate
29.65 +9.8 39.40 perf-profile.calltrace.cycles-pp.btrfs_insert_empty_items.btrfs_insert_file_extent.btrfs_cont_expand.btrfs_setsize.btrfs_setattr
29.89 +9.8 39.70 perf-profile.calltrace.cycles-pp.btrfs_insert_file_extent.btrfs_cont_expand.btrfs_setsize.btrfs_setattr.notify_change
24.00 -7.5 16.51 perf-profile.children.cycles-pp.btrfs_truncate_inode_items
23.01 -7.4 15.61 perf-profile.children.cycles-pp.btrfs_read_lock_root_node
22.50 -7.4 15.13 perf-profile.children.cycles-pp.__btrfs_tree_read_lock
21.42 -7.1 14.32 perf-profile.children.cycles-pp.rwsem_down_read_slowpath
34.64 -4.8 29.83 perf-profile.children.cycles-pp._raw_spin_lock_irq
34.32 ± 2% -4.3 30.00 ± 2% perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath
67.15 -4.1 63.10 perf-profile.children.cycles-pp.btrfs_search_slot
7.90 ± 2% -3.9 3.98 perf-profile.children.cycles-pp.btrfs_get_extent
13.27 -1.2 12.04 perf-profile.children.cycles-pp.btrfs_lookup_file_extent
77.81 -1.1 76.69 perf-profile.children.cycles-pp.do_truncate
77.91 -1.1 76.82 perf-profile.children.cycles-pp.do_sys_ftruncate
78.22 -1.1 77.15 perf-profile.children.cycles-pp.do_syscall_64
78.25 -1.1 77.19 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
78.33 -1.0 77.30 perf-profile.children.cycles-pp.ftruncate64
1.43 -0.3 1.14 ± 3% perf-profile.children.cycles-pp.generic_bin_search
1.03 ± 2% -0.3 0.76 ± 3% perf-profile.children.cycles-pp.down_read
1.00 ± 2% -0.2 0.82 perf-profile.children.cycles-pp.up_read
2.49 ± 2% -0.2 2.31 ± 3% perf-profile.children.cycles-pp.btrfs_free_path
1.61 -0.2 1.45 perf-profile.children.cycles-pp.read_block_for_search
2.66 ± 2% -0.1 2.55 ± 2% perf-profile.children.cycles-pp.btrfs_release_path
0.59 ± 3% -0.1 0.51 ± 3% perf-profile.children.cycles-pp.btrfs_get_64
0.69 ± 2% -0.1 0.63 ± 3% perf-profile.children.cycles-pp.btrfs_root_node
0.74 ± 2% -0.0 0.70 perf-profile.children.cycles-pp.pick_next_task_fair
0.06 ± 7% +0.0 0.08 ± 6% perf-profile.children.cycles-pp.sysvec_call_function_single
0.08 +0.0 0.10 ± 5% perf-profile.children.cycles-pp.asm_sysvec_call_function_single
0.05 ± 7% +0.0 0.07 ± 8% perf-profile.children.cycles-pp.syscall_exit_to_user_mode
0.04 ± 44% +0.0 0.06 ± 7% perf-profile.children.cycles-pp.btrfs_inode_safe_disk_i_size_write
0.29 ± 2% +0.0 0.31 ± 3% perf-profile.children.cycles-pp.sched_clock_cpu
0.07 +0.0 0.09 ± 10% perf-profile.children.cycles-pp.__cond_resched
0.06 ± 6% +0.0 0.08 ± 8% perf-profile.children.cycles-pp.btrfs_leaf_free_space
0.08 ± 8% +0.0 0.10 ± 4% perf-profile.children.cycles-pp.leaf_space_used
0.07 ± 12% +0.0 0.10 ± 9% perf-profile.children.cycles-pp.fill_stack_inode_item
0.10 ± 6% +0.0 0.12 ± 8% perf-profile.children.cycles-pp.__lookup_extent_mapping
0.17 ± 5% +0.0 0.20 ± 2% perf-profile.children.cycles-pp.native_irq_return_iret
0.08 ± 12% +0.0 0.10 ± 10% perf-profile.children.cycles-pp.copy_pages
0.09 ± 7% +0.0 0.12 ± 9% perf-profile.children.cycles-pp.alloc_extent_state
0.07 ± 7% +0.0 0.09 ± 11% perf-profile.children.cycles-pp.free_extent_map
0.07 ± 5% +0.0 0.10 ± 10% perf-profile.children.cycles-pp.btrfs_block_rsv_migrate
0.09 ± 4% +0.0 0.12 ± 3% perf-profile.children.cycles-pp.memset_erms
0.11 ± 10% +0.0 0.14 ± 3% perf-profile.children.cycles-pp.btrfs_get_32
0.18 ± 2% +0.0 0.21 ± 6% perf-profile.children.cycles-pp.__might_resched
0.10 ± 5% +0.0 0.13 ± 4% perf-profile.children.cycles-pp.clear_extent_bit
0.07 ± 9% +0.0 0.11 ± 8% perf-profile.children.cycles-pp.get_io_u
0.10 ± 6% +0.0 0.14 ± 2% perf-profile.children.cycles-pp.up_write
0.14 ± 7% +0.0 0.17 ± 2% perf-profile.children.cycles-pp.btrfs_mark_buffer_dirty
0.09 ± 5% +0.0 0.12 ± 4% perf-profile.children.cycles-pp.clear_state_bit
0.15 ± 6% +0.0 0.19 ± 8% perf-profile.children.cycles-pp.kmem_cache_free
0.54 ± 3% +0.0 0.59 ± 5% perf-profile.children.cycles-pp.__softirqentry_text_start
0.11 ± 6% +0.0 0.15 ± 3% perf-profile.children.cycles-pp.btrfs_inode_clear_file_extent_range
0.08 ± 6% +0.0 0.12 ± 5% perf-profile.children.cycles-pp.fio_gettime
0.36 ± 3% +0.0 0.41 ± 2% perf-profile.children.cycles-pp.read_extent_buffer
0.61 ± 3% +0.0 0.66 ± 4% perf-profile.children.cycles-pp.irq_exit_rcu
0.02 ±142% +0.1 0.08 ± 6% perf-profile.children.cycles-pp.ktime_get_real_ts64
0.07 ± 12% +0.1 0.13 ± 4% perf-profile.children.cycles-pp.btrfs_put_transaction
0.16 ± 5% +0.1 0.22 ± 5% perf-profile.children.cycles-pp.alloc_extent_map
0.17 ± 6% +0.1 0.24 ± 7% perf-profile.children.cycles-pp.__reserve_bytes
0.44 ± 6% +0.1 0.51 ± 6% perf-profile.children.cycles-pp.tick_sched_timer
0.24 ± 4% +0.1 0.31 ± 5% perf-profile.children.cycles-pp.check_setget_bounds
0.32 ± 7% +0.1 0.40 ± 2% perf-profile.children.cycles-pp.memmove_extent_buffer
0.18 ± 3% +0.1 0.26 ± 4% perf-profile.children.cycles-pp.btrfs_block_rsv_add_bytes
0.25 ± 5% +0.1 0.34 ± 4% perf-profile.children.cycles-pp.wake_q_add
0.21 ± 5% +0.1 0.30 ± 4% perf-profile.children.cycles-pp.__clear_extent_bit
0.21 ± 4% +0.1 0.30 ± 4% perf-profile.children.cycles-pp.btrfs_reserve_metadata_bytes
0.69 ± 4% +0.1 0.78 ± 3% perf-profile.children.cycles-pp.__hrtimer_run_queues
0.70 ± 2% +0.1 0.80 ± 3% perf-profile.children.cycles-pp.rwsem_mark_wake
0.41 ± 2% +0.1 0.50 ± 3% perf-profile.children.cycles-pp.kmem_cache_alloc
0.18 ± 5% +0.1 0.28 ± 3% perf-profile.children.cycles-pp.mutex_lock
0.99 +0.1 1.09 ± 2% perf-profile.children.cycles-pp.rwsem_spin_on_owner
0.22 ± 6% +0.1 0.32 ± 3% perf-profile.children.cycles-pp.btrfs_set_token_32
0.46 ± 4% +0.1 0.56 perf-profile.children.cycles-pp.down_write
0.20 ± 6% +0.1 0.31 ± 6% perf-profile.children.cycles-pp.btrfs_update_root_times
0.41 ± 2% +0.1 0.54 ± 4% perf-profile.children.cycles-pp.btrfs_drop_extent_cache
0.37 ± 2% +0.2 0.52 ± 4% perf-profile.children.cycles-pp.btrfs_block_rsv_add
2.61 +0.2 2.79 ± 2% perf-profile.children.cycles-pp.wake_up_q
0.34 ± 8% +0.2 0.52 ± 7% perf-profile.children.cycles-pp.mutex_unlock
0.69 ± 5% +0.2 0.89 ± 2% perf-profile.children.cycles-pp.btrfs_del_items
0.18 ± 4% +0.2 0.38 ± 78% perf-profile.children.cycles-pp.wait_current_trans
0.67 ± 5% +0.2 0.88 ± 2% perf-profile.children.cycles-pp.btrfs_get_token_32
0.22 ± 5% +0.2 0.46 ± 4% perf-profile.children.cycles-pp.btrfs_trans_release_metadata
0.28 ± 3% +0.2 0.53 ± 26% perf-profile.children.cycles-pp.join_transaction
0.59 ± 2% +0.3 0.84 ± 3% perf-profile.children.cycles-pp._raw_spin_lock_irqsave
0.24 ± 5% +0.3 0.49 ± 4% perf-profile.children.cycles-pp.btrfs_block_rsv_release
0.92 ± 2% +0.3 1.26 ± 2% perf-profile.children.cycles-pp.btrfs_unlock_up_safe
3.17 ± 2% +0.4 3.52 ± 2% perf-profile.children.cycles-pp.rwsem_wake
0.88 ± 6% +0.4 1.28 ± 4% perf-profile.children.cycles-pp.memcpy_extent_buffer
11.72 ± 2% +0.4 12.15 perf-profile.children.cycles-pp.intel_idle
1.10 ± 5% +0.4 1.54 ± 3% perf-profile.children.cycles-pp.memmove
0.48 ± 4% +0.4 0.93 ± 3% perf-profile.children.cycles-pp.__btrfs_end_transaction
0.56 ± 4% +0.5 1.09 ± 14% perf-profile.children.cycles-pp.btrfs_dirty_inode
15.41 ± 2% +0.7 16.06 ± 2% perf-profile.children.cycles-pp.cpuidle_enter
15.38 ± 2% +0.7 16.03 ± 2% perf-profile.children.cycles-pp.cpuidle_enter_state
1.66 ± 5% +0.7 2.33 ± 3% perf-profile.children.cycles-pp.setup_items_for_insert
1.12 +0.7 1.83 ± 24% perf-profile.children.cycles-pp.start_transaction
20.41 ± 2% +0.8 21.20 perf-profile.children.cycles-pp.do_idle
20.42 ± 2% +0.8 21.22 perf-profile.children.cycles-pp.secondary_startup_64_no_verify
20.42 ± 2% +0.8 21.22 perf-profile.children.cycles-pp.cpu_startup_entry
20.19 ± 2% +0.8 20.99 perf-profile.children.cycles-pp.start_secondary
72.34 +3.1 75.40 perf-profile.children.cycles-pp.btrfs_setsize
6.85 +3.1 9.96 perf-profile.children.cycles-pp.btrfs_drop_extents
36.09 ± 2% +3.5 39.62 perf-profile.children.cycles-pp.rwsem_down_write_slowpath
36.46 ± 2% +3.5 40.00 perf-profile.children.cycles-pp.btrfs_lock_root_node
72.96 +3.6 76.55 perf-profile.children.cycles-pp.btrfs_setattr
73.03 +3.6 76.63 perf-profile.children.cycles-pp.notify_change
36.52 ± 2% +3.6 40.16 perf-profile.children.cycles-pp.__btrfs_tree_lock
46.11 +9.5 55.64 perf-profile.children.cycles-pp.btrfs_cont_expand
29.65 +9.7 39.40 perf-profile.children.cycles-pp.btrfs_insert_empty_items
29.89 +9.8 39.70 perf-profile.children.cycles-pp.btrfs_insert_file_extent
34.31 ± 2% -4.3 29.99 ± 2% perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath
1.24 -0.3 0.96 ± 3% perf-profile.self.cycles-pp.generic_bin_search
0.98 ± 2% -0.3 0.72 ± 3% perf-profile.self.cycles-pp.down_read
1.27 ± 2% -0.2 1.03 ± 2% perf-profile.self.cycles-pp.rwsem_down_read_slowpath
0.99 ± 2% -0.2 0.81 perf-profile.self.cycles-pp.up_read
1.74 -0.1 1.61 perf-profile.self.cycles-pp.btrfs_search_slot
0.48 ± 2% -0.1 0.40 ± 3% perf-profile.self.cycles-pp.btrfs_get_64
1.01 ± 2% -0.1 0.95 ± 3% perf-profile.self.cycles-pp._raw_spin_lock_irq
0.68 ± 3% -0.1 0.63 ± 3% perf-profile.self.cycles-pp.btrfs_root_node
0.05 +0.0 0.07 ± 7% perf-profile.self.cycles-pp.btrfs_drop_extent_cache
0.08 ± 7% +0.0 0.10 ± 7% perf-profile.self.cycles-pp.schedule
0.21 ± 4% +0.0 0.23 ± 3% perf-profile.self.cycles-pp.read_tsc
0.07 ± 11% +0.0 0.09 ± 8% perf-profile.self.cycles-pp.btrfs_setsize
0.17 ± 5% +0.0 0.20 ± 2% perf-profile.self.cycles-pp.try_to_wake_up
0.17 ± 5% +0.0 0.20 ± 2% perf-profile.self.cycles-pp.native_irq_return_iret
0.07 ± 11% +0.0 0.10 ± 9% perf-profile.self.cycles-pp.setup_items_for_insert
0.08 ± 12% +0.0 0.10 ± 10% perf-profile.self.cycles-pp.copy_pages
0.06 ± 13% +0.0 0.09 ± 8% perf-profile.self.cycles-pp.btrfs_cont_expand
0.06 ± 7% +0.0 0.09 ± 11% perf-profile.self.cycles-pp.free_extent_map
0.09 +0.0 0.12 ± 3% perf-profile.self.cycles-pp.memset_erms
0.08 ± 13% +0.0 0.11 ± 5% perf-profile.self.cycles-pp.__btrfs_release_delayed_node
0.10 ± 10% +0.0 0.13 ± 6% perf-profile.self.cycles-pp.btrfs_get_32
0.16 ± 2% +0.0 0.20 ± 6% perf-profile.self.cycles-pp.__might_resched
0.10 ± 6% +0.0 0.14 ± 3% perf-profile.self.cycles-pp.up_write
0.07 ± 11% +0.0 0.10 ± 9% perf-profile.self.cycles-pp.get_io_u
0.07 ± 6% +0.0 0.11 ± 6% perf-profile.self.cycles-pp.fio_gettime
0.02 ±141% +0.0 0.06 ± 8% perf-profile.self.cycles-pp._raw_write_lock
0.19 ± 3% +0.0 0.23 ± 5% perf-profile.self.cycles-pp.kmem_cache_alloc
0.14 ± 8% +0.0 0.18 ± 7% perf-profile.self.cycles-pp.kmem_cache_free
0.19 ± 7% +0.0 0.24 ± 5% perf-profile.self.cycles-pp.check_setget_bounds
0.13 ± 6% +0.0 0.18 ± 2% perf-profile.self.cycles-pp.wake_up_q
0.36 ± 2% +0.0 0.40 ± 3% perf-profile.self.cycles-pp.read_extent_buffer
0.01 ±223% +0.0 0.06 ± 9% perf-profile.self.cycles-pp.btrfs_truncate_inode_items
0.06 ± 8% +0.0 0.10 ± 9% perf-profile.self.cycles-pp.btrfs_block_rsv_release
0.48 ± 4% +0.1 0.53 ± 2% perf-profile.self.cycles-pp.cpuidle_enter_state
0.01 ±223% +0.1 0.06 ± 7% perf-profile.self.cycles-pp.btrfs_reserve_metadata_bytes
0.08 ± 5% +0.1 0.15 ± 4% perf-profile.self.cycles-pp.btrfs_update_inode
0.07 ± 7% +0.1 0.13 ± 4% perf-profile.self.cycles-pp.btrfs_put_transaction
0.17 ± 6% +0.1 0.25 ± 4% perf-profile.self.cycles-pp.btrfs_set_token_32
0.17 ± 4% +0.1 0.25 ± 6% perf-profile.self.cycles-pp.start_transaction
0.14 ± 6% +0.1 0.22 ± 5% perf-profile.self.cycles-pp.mutex_lock
0.25 ± 5% +0.1 0.34 ± 4% perf-profile.self.cycles-pp.wake_q_add
0.38 ± 4% +0.1 0.48 ± 2% perf-profile.self.cycles-pp.down_write
0.32 ± 4% +0.1 0.41 ± 5% perf-profile.self.cycles-pp._raw_spin_lock_irqsave
0.97 ± 2% +0.1 1.06 ± 2% perf-profile.self.cycles-pp.rwsem_spin_on_owner
0.18 ± 6% +0.1 0.28 ± 3% perf-profile.self.cycles-pp.join_transaction
0.85 ± 3% +0.1 0.96 ± 2% perf-profile.self.cycles-pp.rwsem_down_write_slowpath
0.16 ± 2% +0.1 0.29 ± 2% perf-profile.self.cycles-pp.__btrfs_end_transaction
0.61 ± 5% +0.2 0.79 ± 3% perf-profile.self.cycles-pp.btrfs_get_token_32
0.33 ± 8% +0.2 0.52 ± 7% perf-profile.self.cycles-pp.mutex_unlock
11.69 ± 2% +0.4 12.12 perf-profile.self.cycles-pp.intel_idle
1.10 ± 5% +0.4 1.53 ± 3% perf-profile.self.cycles-pp.memmove
1.63 ± 2% +0.5 2.11 ± 2% perf-profile.self.cycles-pp._raw_spin_lock
Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index a0179cc62913..28ddd9cf2069 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -2918,8 +2918,9 @@ int btrfs_replace_file_extents(struct btrfs_inode *inode, return ret; } -static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) +static int btrfs_punch_hole(struct file *file, loff_t offset, loff_t len) { + struct inode *inode = file_inode(file); struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); struct btrfs_root *root = BTRFS_I(inode)->root; struct extent_state *cached_state = NULL; @@ -2951,6 +2952,10 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) goto out_only_mutex; } + ret = file_modified(file); + if (ret) + goto out_only_mutex; + lockstart = round_up(offset, btrfs_inode_sectorsize(BTRFS_I(inode))); lockend = round_down(offset + len, btrfs_inode_sectorsize(BTRFS_I(inode))) - 1; @@ -3391,7 +3396,7 @@ static long btrfs_fallocate(struct file *file, int mode, return -EOPNOTSUPP; if (mode & FALLOC_FL_PUNCH_HOLE) - return btrfs_punch_hole(inode, offset, len); + return btrfs_punch_hole(file, offset, len); /* * Only trigger disk allocation, don't trigger qgroup reserve @@ -3413,6 +3418,10 @@ static long btrfs_fallocate(struct file *file, int mode, goto out; } + ret = file_modified(file); + if (ret) + goto out; + /* * TODO: Move these two operations after we have checked * accurate reserved space, or fallocate can still fail but