Currently, ->pd_stat() is called only when moduleparam
blkcg_debug_stats is set which prevents it from printing non-debug
policy-specific statistics. Let's move debug testing down so that
->pd_stat() can print non-debug stat too. This patch doesn't cause
any visible behavior change.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When a shared kthread needs to issue a bio for a cgroup, doing so
synchronously can lead to priority inversions as the kthread can be
trapped waiting for that cgroup. This patch implements
REQ_CGROUP_PUNT flag which makes submit_bio() punt the actual issuing
to a dedicated per-blkcg work item to avoid such priority inversions.
This will be used to fix priority inversions in btrfs compression and
should be generally useful as we grow filesystem support for
comprehensive IO control.
Cc: Chris Mason <clm@fb.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When writeback IOs are bounced through async layers, the IOs should
only be accounted against the wbc from the original bdi writeback to
avoid confusing cgroup inode ownership arbitration. Add
wbc->no_cgroup_owner to allow disabling wbc cgroup owner accounting.
This will be used make btrfs compression work well with cgroup IO
control.
v2: Renamed from no_wbc_acct to no_cgroup_owner and added comment as
per Jan.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
wbc_account_io() does a very specific job - try to see which cgroup is
actually dirtying an inode and transfer its ownership to the majority
dirtier if needed. The name is too generic and confusing. Let's
rename it to something more specific.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
btrfs is going to use css_put() and wbc helpers to improve cgroup
writeback support. Add dummy css_get() definition and export wbc
helpers to prepare for module and !CONFIG_CGROUP builds.
Reported-by: kbuild test robot <lkp@intel.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
ZAC support added sense data requesting on error for both ZAC and ATA
devices. This seems to cause erratic error handling behaviors on some
SSDs where the device reports sense data availability and then
delivers the wrong content making EH take the wrong actions. The
failure mode was sporadic on a LITE-ON ssd and couldn't be reliably
reproduced.
There is no value in requesting sense data from non-ZAC ATA devices
while there's a significant risk of introducing EH misbehaviors which
are difficult to reproduce and fix. Let's do the sense data dancing
only for ZAC devices.
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Masato Suzuki <masato.suzuki@wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
wbc_account_io() collects information on cgroup ownership of writeback
pages to determine which cgroup should own the inode. Pages can stay
associated with dead memcgs but we want to avoid attributing IOs to
dead blkcgs as much as possible as the association is likely to be
stale. However, currently, pages associated with dead memcgs
contribute to the accounting delaying and/or confusing the
arbitration.
Fix it by ignoring pages associated with dead memcgs.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When blkcg_activate_policy() is creating blkg_policy_data for existing
blkgs, it did in the wrong order - descendants first. Fix it. None
of the existing controllers seem affected by this.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blkg alloc is performed as a separate step from the rest of blkg
creation so that GFP_KERNEL allocations can be used when creating
blkgs from configuration file writes because otherwise user actions
may fail due to failures of opportunistic GFP_NOWAIT allocations.
While making blkgs use percpu_ref, 7fcf2b033b ("blkcg: change blkg
reference counting to use percpu_ref") incorrectly added unconditional
opportunistic percpu_ref_init() to blkg_create() breaking this
guarantee.
This patch moves percpu_ref_init() to blkg_alloc() so makes it use
@gfp_mask that blkg_alloc() is called with. Also, percpu_ref_exit()
is moved to blkg_free() for consistency.
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: 7fcf2b033b ("blkcg: change blkg reference counting to use percpu_ref")
Cc: Dennis Zhou <dennis@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Depending on the number of devices, blkcg stats can go over the
default seqfile buf size. seqfile normally retries with a larger
buffer but since the ->pd_stat() addition, blkcg_print_stat() doesn't
tell seqfile that overflow has happened and the output gets printed
truncated. Fix it by calling seq_commit() w/ -1 on possible
overflows.
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: 903d23f0a3 ("blk-cgroup: allow controllers to output their own stats")
Cc: stable@vger.kernel.org # v4.19+
Cc: Josef Bacik <jbacik@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If use_delay was non-zero when the latency target of a cgroup was set
to zero, it will stay stuck until io.latency is enabled on the cgroup
again. This keeps readahead disabled for the cgroup impacting
performance negatively.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Josef Bacik <jbacik@fb.com>
Fixes: d706751215 ("block: introduce blk-iolatency io controller")
Cc: stable@vger.kernel.org # v4.19+
Signed-off-by: Jens Axboe <axboe@kernel.dk>
While adding handling for dying task group leaders c03cd7738a
("cgroup: Include dying leaders with live threads in PROCS
iterations") added an inverted cset skip condition to
css_task_iter_advance_css_set(). It should skip cset if it's
completely empty but was incorrectly testing for the inverse condition
for the dying_tasks list. Fix it.
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: c03cd7738a ("cgroup: Include dying leaders with live threads in PROCS iterations")
Reported-by: syzbot+d4bba5ccd4f9a2a68681@syzkaller.appspotmail.com
b636fd38dc ("cgroup: Implement css_task_iter_skip()") introduced
css_task_iter_skip() which is used to fix task iterations skipping
dying threadgroup leaders with live threads. Skipping is implemented
as a subportion of full advancing but css_task_iter_next() forgot to
fully advance a skipped iterator before determining the next task to
visit causing it to return invalid task pointers.
Fix it by making css_task_iter_next() fully advance the iterator if it
has been skipped since the previous iteration.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: syzbot
Link: http://lkml.kernel.org/r/00000000000097025d058a7fd785@google.com
Fixes: b636fd38dc ("cgroup: Implement css_task_iter_skip()")
cgroup already uses floating point for percent[ile] numbers and there
are several controllers which want to take them as input. Add a
generic parse helper to handle inputs.
Update the interface convention documentation about the use of
percentage numbers. While at it, also clarify the default time unit.
Signed-off-by: Tejun Heo <tj@kernel.org>
CSS_TASK_ITER_PROCS currently iterates live group leaders; however,
this means that a process with dying leader and live threads will be
skipped. IOW, cgroup.procs might be empty while cgroup.threads isn't,
which is confusing to say the least.
Fix it by making cset track dying tasks and include dying leaders with
live threads in PROCS iteration.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Topi Miettinen <toiwoton@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
When a task is moved out of a cset, task iterators pointing to the
task are advanced using the normal css_task_iter_advance() call. This
is fine but we'll be tracking dying tasks on csets and thus moving
tasks from cset->tasks to (to be added) cset->dying_tasks. When we
remove a task from cset->tasks, if we advance the iterators, they may
move over to the next cset before we had the chance to add the task
back on the dying list, which can allow the task to escape iteration.
This patch separates out skipping from advancing. Skipping only moves
the affected iterators to the next pointer rather than fully advancing
it and the following advancing will recognize that the cursor has
already been moved forward and do the rest of advancing. This ensures
that when a task moves from one list to another in its cset, as long
as it moves in the right direction, it's always visible to iteration.
This doesn't cause any visible behavior changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
cgroup_release() calls cgroup_subsys->release() which is used by the
pids controller to uncharge its pid. We want to use it to manage
iteration of dying tasks which requires putting it before
__unhash_process(). Move cgroup_release() above __exit_signal().
While this makes it uncharge before the pid is freed, pid is RCU freed
anyway and the window is very narrow.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
A PF_EXITING task can stay associated with an offline css. If such
task calls task_get_css(), it can get stuck indefinitely. This can be
triggered by BSD process accounting which writes to a file with
PF_EXITING set when racing against memcg disable as in the backtrace
at the end.
After this change, task_get_css() may return a css which was already
offline when the function was called. None of the existing users are
affected by this change.
INFO: rcu_sched self-detected stall on CPU
INFO: rcu_sched detected stalls on CPUs/tasks:
...
NMI backtrace for cpu 0
...
Call Trace:
<IRQ>
dump_stack+0x46/0x68
nmi_cpu_backtrace.cold.2+0x13/0x57
nmi_trigger_cpumask_backtrace+0xba/0xca
rcu_dump_cpu_stacks+0x9e/0xce
rcu_check_callbacks.cold.74+0x2af/0x433
update_process_times+0x28/0x60
tick_sched_timer+0x34/0x70
__hrtimer_run_queues+0xee/0x250
hrtimer_interrupt+0xf4/0x210
smp_apic_timer_interrupt+0x56/0x110
apic_timer_interrupt+0xf/0x20
</IRQ>
RIP: 0010:balance_dirty_pages_ratelimited+0x28f/0x3d0
...
btrfs_file_write_iter+0x31b/0x563
__vfs_write+0xfa/0x140
__kernel_write+0x4f/0x100
do_acct_process+0x495/0x580
acct_process+0xb9/0xdb
do_exit+0x748/0xa00
do_group_exit+0x3a/0xa0
get_signal+0x254/0x560
do_signal+0x23/0x5c0
exit_to_usermode_loop+0x5d/0xa0
prepare_exit_to_usermode+0x53/0x80
retint_user+0x8/0x8
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@vger.kernel.org # v4.2+
Fixes: ec438699a9 ("cgroup, block: implement task_get_css() and use it in bio_associate_current()")
cgroup_rstat_cpu_pop_updated() is used to traverse the updated cgroups
on flush. While it was only visiting updated ones in the subtree, it
was visiting @root unconditionally. We can easily check whether @root
is updated or not by looking at its ->updated_next just as with the
cgroups in the subtree.
* Remove the unnecessary cgroup_parent() test. The system root cgroup
is never updated and thus its ->updated_next is always NULL. No
need to test whether cgroup_parent() exists in addition to
->updated_next.
* Terminate traverse if ->updated_next is NULL. This can only happen
for subtree @root and there's no reason to visit it if it's not
marked updated.
This reduces cpu consumption when reading a lot of rstat backed files.
In a micro benchmark reading stat from ~1600 cgroups, the sys time was
lowered by >40%.
Signed-off-by: Tejun Heo <tj@kernel.org>
sync_inodes_sb() can race against cgwb (cgroup writeback) membership
switches and fail to writeback some inodes. For example, if an inode
switches to another wb while sync_inodes_sb() is in progress, the new
wb might not be visible to bdi_split_work_to_wbs() at all or the inode
might jump from a wb which hasn't issued writebacks yet to one which
already has.
This patch adds backing_dev_info->wb_switch_rwsem to synchronize cgwb
switch path against sync_inodes_sb() so that sync_inodes_sb() is
guaranteed to see all the target wbs and inodes can't jump wbs to
escape syncing.
v2: Fixed misplaced rwsem init. Spotted by Jiufei.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Jiufei Xue <xuejiufei@gmail.com>
Link: http://lkml.kernel.org/r/dc694ae2-f07f-61e1-7097-7c8411cee12d@gmail.com
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It can be useful to inhibit all cgroup1 hierarchies especially during
transition and for debugging. cgroup_no_v1 can block hierarchies with
controllers which leaves out the named hierarchies. Expand it to
cover the named hierarchies so that "cgroup_no_v1=all,named" disables
all cgroup1 hierarchies.
Signed-off-by: Tejun Heo <tj@kernel.org>
Suggested-by: Marcin Pawlowski <mpawlowski@fb.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
CSS_TASK_ITER_PROCS implements process-only iteration by making
css_task_iter_advance() skip tasks which aren't threadgroup leaders;
however, when an iteration is started css_task_iter_start() calls the
inner helper function css_task_iter_advance_css_set() instead of
css_task_iter_advance(). As the helper doesn't have the skip logic,
when the first task to visit is a non-leader thread, it doesn't get
skipped correctly as shown in the following example.
# ps -L 2030
PID LWP TTY STAT TIME COMMAND
2030 2030 pts/0 Sl+ 0:00 ./test-thread
2030 2031 pts/0 Sl+ 0:00 ./test-thread
# mkdir -p /sys/fs/cgroup/x/a/b
# echo threaded > /sys/fs/cgroup/x/a/cgroup.type
# echo threaded > /sys/fs/cgroup/x/a/b/cgroup.type
# echo 2030 > /sys/fs/cgroup/x/a/cgroup.procs
# cat /sys/fs/cgroup/x/a/cgroup.threads
2030
2031
# cat /sys/fs/cgroup/x/cgroup.procs
2030
# echo 2030 > /sys/fs/cgroup/x/a/b/cgroup.threads
# cat /sys/fs/cgroup/x/cgroup.procs
2031
2030
The last read of cgroup.procs is incorrectly showing non-leader 2031
in cgroup.procs output.
This can be fixed by updating css_task_iter_advance() to handle the
first advance and css_task_iters_tart() to call
css_task_iter_advance() instead of the inner helper. After the fix,
the same commands result in the following (correct) result:
# ps -L 2062
PID LWP TTY STAT TIME COMMAND
2062 2062 pts/0 Sl+ 0:00 ./test-thread
2062 2063 pts/0 Sl+ 0:00 ./test-thread
# mkdir -p /sys/fs/cgroup/x/a/b
# echo threaded > /sys/fs/cgroup/x/a/cgroup.type
# echo threaded > /sys/fs/cgroup/x/a/b/cgroup.type
# echo 2062 > /sys/fs/cgroup/x/a/cgroup.procs
# cat /sys/fs/cgroup/x/a/cgroup.threads
2062
2063
# cat /sys/fs/cgroup/x/cgroup.procs
2062
# echo 2062 > /sys/fs/cgroup/x/a/b/cgroup.threads
# cat /sys/fs/cgroup/x/cgroup.procs
2062
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: "Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>
Fixes: 8cfd8147df ("cgroup: implement cgroup v2 thread support")
Cc: stable@vger.kernel.org # v4.14+
Clearly mark the debug files and hide them by default by prefixing
".__DEBUG__.".
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Waiman Long <longman@redhat.com>
* Rename the partition file from "cpuset.sched.partition" to
"cpuset.cpus.partition".
* When writing to the partition file, drop "0" and "1" and only accept
"member" and "root".
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Waiman Long <longman@redhat.com>
A cgroup which is already a threaded domain may be converted into a
threaded cgroup if the prerequisite conditions are met. When this
happens, all threaded descendant should also have their ->dom_cgrp
updated to the new threaded domain cgroup. Unfortunately, this
propagation was missing leading to the following failure.
# cd /sys/fs/cgroup/unified
# cat cgroup.subtree_control # show that no controllers are enabled
# mkdir -p mycgrp/a/b/c
# echo threaded > mycgrp/a/b/cgroup.type
At this point, the hierarchy looks as follows:
mycgrp [d]
a [dt]
b [t]
c [inv]
Now let's make node "a" threaded (and thus "mycgrp" s made "domain threaded"):
# echo threaded > mycgrp/a/cgroup.type
By this point, we now have a hierarchy that looks as follows:
mycgrp [dt]
a [t]
b [t]
c [inv]
But, when we try to convert the node "c" from "domain invalid" to
"threaded", we get ENOTSUP on the write():
# echo threaded > mycgrp/a/b/c/cgroup.type
sh: echo: write error: Operation not supported
This patch fixes the problem by
* Moving the opencoded ->dom_cgrp save and restoration in
cgroup_enable_threaded() into cgroup_{save|restore}_control() so
that mulitple cgroups can be handled.
* Updating all threaded descendants' ->dom_cgrp to point to the new
dom_cgrp when enabling threaded mode.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: "Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>
Reported-by: Amin Jamali <ajamali@pivotal.io>
Reported-by: Joao De Almeida Pereira <jpereira@pivotal.io>
Link: https://lore.kernel.org/r/CAKgNAkhHYCMn74TCNiMJ=ccLd7DcmXSbvw3CbZ1YREeG7iJM5g@mail.gmail.com
Fixes: 454000adaa ("cgroup: introduce cgroup->dom_cgrp and threaded css_set handling")
Cc: stable@vger.kernel.org # v4.14+
Dennis rewrote a significant portion of the percpu allocator and has
shown that he can respond in a timely and helpful manner when issues
are reported against percpu allocator.
Let's make Dennis the percpu tree maintainer.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Dennis Zhou <dennis@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
While forking, if delayacct init fails due to memory shortage, it
continues expecting all delayacct users to check task->delays pointer
against NULL before dereferencing it, which all of them used to do.
Commit c96f5471ce ("delayacct: Account blkio completion on the correct
task"), while updating delayacct_blkio_end() to take the target task
instead of always using %current, made the function test NULL on
%current->delays and then continue to operated on @p->delays. If
%current succeeded init while @p didn't, it leads to the following
crash.
BUG: unable to handle kernel NULL pointer dereference at 0000000000000004
IP: __delayacct_blkio_end+0xc/0x40
PGD 8000001fd07e1067 P4D 8000001fd07e1067 PUD 1fcffbb067 PMD 0
Oops: 0000 [#1] SMP PTI
CPU: 4 PID: 25774 Comm: QIOThread0 Not tainted 4.16.0-9_fbk1_rc2_1180_g6b593215b4d7 #9
RIP: 0010:__delayacct_blkio_end+0xc/0x40
Call Trace:
try_to_wake_up+0x2c0/0x600
autoremove_wake_function+0xe/0x30
__wake_up_common+0x74/0x120
wake_up_page_bit+0x9c/0xe0
mpage_end_io+0x27/0x70
blk_update_request+0x78/0x2c0
scsi_end_request+0x2c/0x1e0
scsi_io_completion+0x20b/0x5f0
blk_mq_complete_request+0xa2/0x100
ata_scsi_qc_complete+0x79/0x400
ata_qc_complete_multiple+0x86/0xd0
ahci_handle_port_interrupt+0xc9/0x5c0
ahci_handle_port_intr+0x54/0xb0
ahci_single_level_irq_intr+0x3b/0x60
__handle_irq_event_percpu+0x43/0x190
handle_irq_event_percpu+0x20/0x50
handle_irq_event+0x2a/0x50
handle_edge_irq+0x80/0x1c0
handle_irq+0xaf/0x120
do_IRQ+0x41/0xc0
common_interrupt+0xf/0xf
Fix it by updating delayacct_blkio_end() check @p->delays instead.
Link: http://lkml.kernel.org/r/20180724175542.GP1934745@devbig577.frc2.facebook.com
Fixes: c96f5471ce ("delayacct: Account blkio completion on the correct task")
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Dave Jones <dsj@fb.com>
Debugged-by: Dave Jones <dsj@fb.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Josh Snyder <joshs@netflix.com>
Cc: <stable@vger.kernel.org> [4.15+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add tracking of REQ_OP_DISCARD ios to the per-cgroup io.stat. Two
fields, dbytes and dios, to respectively count the total bytes and
number of discards are added.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Andy Newell <newella@fb.com>
Cc: Michael Callahan <michaelcallahan@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
c11f0c0b5b ("block/mm: make bdev_ops->rw_page() take a bool for
read/write") replaced @op with boolean @is_write, which limited the
amount of information going into ->rw_page() and more importantly
page_endio(), which removed the need to expose block internals to mm.
Unfortunately, we want to track discards separately and @is_write
isn't enough information. This patch updates bdev_ops->rw_page() to
take REQ_OP instead but leaves page_endio() to take bool @is_write.
This allows the block part of operations to have enough information
while not leaking it to mm.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Memory allocations can induce swapping via kswapd or direct reclaim. If
we are having IO done for us by kswapd and don't actually go into direct
reclaim we may never get scheduled for throttling. So instead check to
see if our cgroup is congested, and if so schedule the throttling.
Before we return to user space the throttling stuff will only throttle
if we actually required it.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
For backcharging we need to know who the page belongs to when swapping
it out. We don't worry about things that do ->rw_page (zram etc) at the
moment, we're only worried about pages that actually go to a block
device.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently an attempt to set swap.max into a value lower than the actual
swap usage fails, which causes configuration problems as there's no way
of lowering the configuration below the current usage short of turning
off swap entirely. This makes swap.max difficult to use and allows
delegatees to lock the delegator out of reducing swap allocation.
This patch updates swap_max_write() so that the limit can be lowered
below the current usage. It doesn't implement active reclaiming of swap
entries for the following reasons.
* mem_cgroup_swap_full() already tells the swap machinary to
aggressively reclaim swap entries if the usage is above 50% of
limit, so simply lowering the limit automatically triggers gradual
reclaim.
* Forcing back swapped out pages is likely to heavily impact the
workload and mess up the working set. Given that swap usually is a
lot less valuable and less scarce, letting the existing usage
dissipate over time through the above gradual reclaim and as they're
falted back in is likely the better behavior.
Link: http://lkml.kernel.org/r/20180523185041.GR1718769@devbig577.frc2.facebook.com
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Acked-by: Rik van Riel <riel@surriel.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Shaohua Li <shli@fb.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "mm, memcontrol: Implement memory.swap.events", v2.
This patchset implements memory.swap.events which contains max and fail
events so that userland can monitor and respond to swap running out.
This patch (of 2):
get_swap_page() is always followed by mem_cgroup_try_charge_swap().
This patch moves mem_cgroup_try_charge_swap() into get_swap_page() and
makes get_swap_page() call the function even after swap allocation
failure.
This simplifies the callers and consolidates memcg related logic and
will ease adding swap related memcg events.
Link: http://lkml.kernel.org/r/20180416230934.GH1911913@devbig577.frc2.facebook.com
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Rik van Riel <riel@surriel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If a connection gets aborted while congested, FUSE can leave
nr_wb_congested[] stuck until reboot causing wait_iff_congested() to
wait spuriously which can lead to severe performance degradation.
The leak is caused by gating congestion state clearing with
fc->connected test in request_end(). This was added way back in 2009
by 26c3679101 ("fuse: destroy bdi on umount"). While the commit
description doesn't explain why the test was added, it most likely was
to avoid dereferencing bdi after it got destroyed.
Since then, bdi lifetime rules have changed many times and now we're
always guaranteed to have access to the bdi while the superblock is
alive (fc->sb).
Drop fc->connected conditional to avoid leaking congestion states.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Joshua Miller <joshmiller@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: stable@vger.kernel.org # v2.6.29+
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
From 0aa2e9b921d6db71150633ff290199554f0842a8 Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj@kernel.org>
Date: Wed, 23 May 2018 10:29:00 -0700
cgwb_release() punts the actual release to cgwb_release_workfn() on
system_wq. Depending on the number of cgroups or block devices, there
can be a lot of cgwb_release_workfn() in flight at the same time.
We're periodically seeing close to 256 kworkers getting stuck with the
following stack trace and overtime the entire system gets stuck.
[<ffffffff810ee40c>] _synchronize_rcu_expedited.constprop.72+0x2fc/0x330
[<ffffffff810ee634>] synchronize_rcu_expedited+0x24/0x30
[<ffffffff811ccf23>] bdi_unregister+0x53/0x290
[<ffffffff811cd1e9>] release_bdi+0x89/0xc0
[<ffffffff811cd645>] wb_exit+0x85/0xa0
[<ffffffff811cdc84>] cgwb_release_workfn+0x54/0xb0
[<ffffffff810a68d0>] process_one_work+0x150/0x410
[<ffffffff810a71fd>] worker_thread+0x6d/0x520
[<ffffffff810ad3dc>] kthread+0x12c/0x160
[<ffffffff81969019>] ret_from_fork+0x29/0x40
[<ffffffffffffffff>] 0xffffffffffffffff
The events leading to the lockup are...
1. A lot of cgwb_release_workfn() is queued at the same time and all
system_wq kworkers are assigned to execute them.
2. They all end up calling synchronize_rcu_expedited(). One of them
wins and tries to perform the expedited synchronization.
3. However, that invovles queueing rcu_exp_work to system_wq and
waiting for it. Because #1 is holding all available kworkers on
system_wq, rcu_exp_work can't be executed. cgwb_release_workfn()
is waiting for synchronize_rcu_expedited() which in turn is waiting
for cgwb_release_workfn() to free up some of the kworkers.
We shouldn't be scheduling hundreds of cgwb_release_workfn() at the
same time. There's nothing to be gained from that. This patch
updates cgwb release path to use a dedicated percpu workqueue with
@max_active of 1.
While this resolves the problem at hand, it might be a good idea to
isolate rcu_exp_work to its own workqueue too as it can be used from
various paths and is prone to this sort of indirect A-A deadlocks.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: stable@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
cgroup_enable_task_cg_lists() incorrectly nests non-irq-safe
tasklist_lock inside irq-safe css_set_lock triggering the following
lockdep warning.
WARNING: possible irq lock inversion dependency detected
4.17.0-rc1-00027-gb37d049 #6 Not tainted
--------------------------------------------------------
systemd/1 just changed the state of lock:
00000000fe57773b (css_set_lock){..-.}, at: cgroup_free+0xf2/0x12a
but this lock took another, SOFTIRQ-unsafe lock in the past:
(tasklist_lock){.+.+}
and interrupts could create inverse lock ordering between them.
other info that might help us debug this:
Possible interrupt unsafe locking scenario:
CPU0 CPU1
---- ----
lock(tasklist_lock);
local_irq_disable();
lock(css_set_lock);
lock(tasklist_lock);
<Interrupt>
lock(css_set_lock);
*** DEADLOCK ***
The condition is highly unlikely to actually happen especially given
that the path is executed only once per boot.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Boqun Feng <boqun.feng@gmail.com>
The worker struct could already be freed when wq_worker_comm() tries
to access it for reporting. This patch protects PF_WQ_WORKER
modifications with wq_pool_attach_mutex and makes wq_worker_comm()
test the flag before dereferencing worker from kthread_data(), which
ensures that it only dereferences when the worker struct is valid.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Lai Jiangshan <jiangshanlai@gmail.com>
Fixes: 6b59808bfe ("workqueue: Show the latest workqueue name in /proc/PID/{comm,stat,status}")
There can be a lot of workqueue workers and they all show up with the
cryptic kworker/* names making it difficult to understand which is
doing what and how they came to be.
# ps -ef | grep kworker
root 4 2 0 Feb25 ? 00:00:00 [kworker/0:0H]
root 6 2 0 Feb25 ? 00:00:00 [kworker/u112:0]
root 19 2 0 Feb25 ? 00:00:00 [kworker/1:0H]
root 25 2 0 Feb25 ? 00:00:00 [kworker/2:0H]
root 31 2 0 Feb25 ? 00:00:00 [kworker/3:0H]
...
This patch makes workqueue workers report the latest workqueue it was
executing for through /proc/PID/{comm,stat,status}. The extra
information is appended to the kthread name with intervening '+' if
currently executing, otherwise '-'.
# cat /proc/25/comm
kworker/2:0-events_power_efficient
# cat /proc/25/stat
25 (kworker/2:0-events_power_efficient) I 2 0 0 0 -1 69238880 0 0...
# grep Name /proc/25/status
Name: kworker/2:0-events_power_efficient
Unfortunately, ps(1) truncates comm to 15 characters,
# ps 25
PID TTY STAT TIME COMMAND
25 ? I 0:00 [kworker/2:0-eve]
making it a lot less useful; however, this should be an easy fix from
ps(1) side.
Signed-off-by: Tejun Heo <tj@kernel.org>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Craig Small <csmall@enc.com.au>
proc shows task->comm in three places - comm, stat, status - and each
is fetching and formatting task->comm slighly differently. This patch
renames task_name() to proc_task_name(), makes it more generic, and
updates all three paths to use it.
This will enable expanding comm reporting for workqueue workers.
Signed-off-by: Tejun Heo <tj@kernel.org>
Work functions can use set_worker_desc() to improve the visibility of
what the worker task is doing. Currently, the desc field is unset at
the beginning of each execution and there is a separate field to track
the field is set during the current execution.
Instead of leaving empty till desc is set, worker->desc can be used to
remember the last workqueue the worker worked on by default and users
that use set_worker_desc() can override it to something more
informative as necessary.
This simplifies desc handling and helps tracking the last workqueue
that the worker exected on to improve visibility.
Signed-off-by: Tejun Heo <tj@kernel.org>
For historical reasons, the worker attach/detach functions don't
currently manage worker->pool and the callers are manually and
inconsistently updating it.
This patch moves worker->pool updates into the worker attach/detach
functions. This makes worker->pool consistent and clearly defines how
worker->pool updates are synchronized.
This will help later workqueue visibility improvements by allowing
safe access to workqueue information from worker->task.
Signed-off-by: Tejun Heo <tj@kernel.org>
To improve workqueue visibility, we want to be able to access
workqueue information from worker tasks. The per-pool attach mutex
makes that difficult because there's no way of stabilizing task ->
worker pool association without knowing the pool first.
Worker attach/detach is a slow path and there's no need for different
pools to be able to perform them concurrently. This patch replaces
the per-pool attach_mutex with global wq_pool_attach_mutex to prepare
for visibility improvement changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Sandisk SSDs SD7SN6S256G and SD8SN8U256G are regularly locking up
regularly under sustained moderate load with NCQ enabled. Blacklist
for now.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Dave Jones <davej@codemonkey.org.uk>
Cc: stable@vger.kernel.org
cgroup_rstat_updated() ensures that the cgroup's rstat is linked to
the parent. If there's no parent, it never gets linked and the
function ends up grabbing and releasing the cgroup_rstat_lock each
time for no reason which can be expensive.
This hasn't been a problem till now because nobody was calling the
function for the root cgroup but rstat is gonna be exposed to
controllers and use cases, so let's get ready. Make
cgroup_rstat_updated() an no-op for the root cgroup.
Signed-off-by: Tejun Heo <tj@kernel.org>
cgroup_rstat_updated() has a small race window where an updated
signaling can race with flush and could be lost till the next update.
This wasn't a problem for the existing usages, but we plan to use
rstat to track counters which need to be accurate.
This patch plugs the race window by synchronizing
cgroup_rstat_updated() and flush path with memory barriers around
cgroup_rstat_cpu->updated_next pointer.
Signed-off-by: Tejun Heo <tj@kernel.org>
This patch adds cgroup_subsys->css_rstat_flush(). If a subsystem has
this callback, its csses are linked on cgrp->css_rstat_list and rstat
will call the function whenever the associated cgroup is flushed.
Flush is also performed when such csses are released so that residual
counts aren't lost.
Combined with the rstat API previous patches factored out, this allows
controllers to plug into rstat to manage their statistics in a
scalable way.
Signed-off-by: Tejun Heo <tj@kernel.org>
Currently, rstat flush path is protected with a mutex which is fine as
all the existing users are from interface file show path. However,
rstat is being generalized for use by controllers and flushing from
atomic contexts will be necessary.
This patch replaces cgroup_rstat_mutex with a spinlock and adds a
irq-safe flush function - cgroup_rstat_flush_irqsafe(). Explicit
yield handling is added to the flush path so that other flush
functions can yield to other threads and flushers.
Signed-off-by: Tejun Heo <tj@kernel.org>
cgroup_rstat is being generalized so that controllers can use it too.
This patch factors out and exposes the following interface functions.
* cgroup_rstat_updated(): Renamed from cgroup_rstat_cpu_updated() for
consistency.
* cgroup_rstat_flush_hold/release(): Factored out from base stat
implementation.
* cgroup_rstat_flush(): Verbatim expose.
While at it, drop assert on cgroup_rstat_mutex in
cgroup_base_stat_flush() as it crosses layers and make a minor comment
update.
v2: Added EXPORT_SYMBOL_GPL(cgroup_rstat_updated) to fix a build bug.
Signed-off-by: Tejun Heo <tj@kernel.org>
Currently, rstat.c has rstat and base stat implementations intermixed.
Collect base stat implementation at the end of the file. Also,
reorder the prototypes.
This patch doesn't make any functional changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Base resource stat accounts universial (not specific to any
controller) resource consumptions on top of rstat. Currently, its
implementation is intermixed with rstat implementation making the code
confusing to follow.
This patch clarifies the distintion by doing the followings.
* Encapsulate base resource stat counters, currently only cputime, in
struct cgroup_base_stat.
* Move prev_cputime into struct cgroup and initialize it with cgroup.
* Rename the related functions so that they start with cgroup_base_stat.
* Prefix the related variables and field names with b.
This patch doesn't make any functional changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
stat is too generic a name and ends up causing subtle confusions.
It'll be made generic so that controllers can plug into it, which will
make the problem worse. Let's rename it to something more specific -
cgroup_rstat for cgroup recursive stat.
This patch does the following renames. No other changes.
* cpu_stat -> rstat_cpu
* stat -> rstat
* ?cstat -> ?rstatc
Note that the renames are selective. The unrenamed are the ones which
implement basic resource statistics on top of rstat. This will be
further cleaned up in the following patches.
Signed-off-by: Tejun Heo <tj@kernel.org>
stat is too generic a name and ends up causing subtle confusions.
It'll be made generic so that controllers can plug into it, which will
make the problem worse. Let's rename it to something more specific -
cgroup_rstat for cgroup recursive stat.
First, rename kernel/cgroup/stat.c to kernel/cgroup/rstat.c. No
content changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
".events" files generate file modified event to notify userland of
possible new events. Some of the events can be quite bursty
(e.g. memory high event) and generating notification each time is
costly and pointless.
This patch implements a event rate limit mechanism. If a new
notification is requested before 10ms has passed since the previous
notification, the new notification is delayed till then.
As this only delays from the second notification on in a given close
cluster of notifications, userland reactions to notifications
shouldn't be delayed at all in most cases while avoiding notification
storms.
Signed-off-by: Tejun Heo <tj@kernel.org>
The "cgroup." core interface files bypass the usual interface removal
path and get removed recursively along with the cgroup itself. While
this works now, the subtle discrepancy gets in the way of implementing
common mechanisms.
This patch updates cgroup core interface file handling so that it's
consistent with controller interface files. When added, the css is
marked CSS_VISIBLE and they're explicitly removed before the cgroup is
destroyed.
This doesn't cause user-visible behavior changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
There are cases where RCU callback needs to be bounced to a sleepable
context. This is currently done by the RCU callback queueing a work
item, which can be cumbersome to write and confusing to read.
This patch introduces rcu_work, a workqueue work variant which gets
executed after a RCU grace period, and converts the open coded
bouncing in fs/aio and kernel/cgroup.
v3: Dropped queue_rcu_work_on(). Documented rcu grace period behavior
after queue_rcu_work().
v2: Use rcu_barrier() instead of synchronize_rcu() to wait for
completion of previously queued rcu callback as per Paul.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
percpu_ref internally uses sched-RCU to implement the percpu -> atomic
mode switching and the documentation suggested that this could be
depended upon. This doesn't seem like a good idea.
* percpu_ref uses sched-RCU which has different grace periods regular
RCU. Users may combine percpu_ref with regular RCU usage and
incorrectly believe that regular RCU grace periods are performed by
percpu_ref. This can lead to, for example, use-after-free due to
premature freeing.
* percpu_ref has a grace period when switching from percpu to atomic
mode. It doesn't have one between the last put and release. This
distinction is subtle and can lead to surprising bugs.
* percpu_ref allows starting in and switching to atomic mode manually
for debugging and other purposes. This means that there may not be
any grace periods from kill to release.
This patch makes it clear that the grace periods are percpu_ref's
internal implementation detail and can't be depended upon by the
users.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Kent Overstreet <kent.overstreet@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
microblaze build broke due to missing declaration of the
cond_resched() invocation added recently. Let's include linux/sched.h
explicitly.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: kbuild test robot <fengguang.wu@intel.com>
rvt_mregion uses percpu_ref for reference counting and RCU to protect
accesses from lkey_table. When a rvt_mregion needs to be freed, it
first gets unregistered from lkey_table and then rvt_check_refs() is
called to wait for in-flight usages before the rvt_mregion is freed.
rvt_check_refs() seems to have a couple issues.
* It has a fast exit path which tests percpu_ref_is_zero(). However,
a percpu_ref reading zero doesn't mean that the object can be
released. In fact, the ->release() callback might not even have
started executing yet. Proceeding with freeing can lead to
use-after-free.
* lkey_table is RCU protected but there is no RCU grace period in the
free path. percpu_ref uses RCU internally but it's sched-RCU whose
grace periods are different from regular RCU. Also, it generally
isn't a good idea to depend on internal behaviors like this.
To address the above issues, this patch removes the fast exit and adds
an explicit synchronize_rcu().
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Cc: Mike Marciniszyn <mike.marciniszyn@intel.com>
Cc: linux-rdma@vger.kernel.org
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
rvt_mregion uses percpu_ref for reference counting and RCU to protect
accesses from lkey_table. When a rvt_mregion needs to be freed, it
first gets unregistered from lkey_table and then rvt_check_refs() is
called to wait for in-flight usages before the rvt_mregion is freed.
rvt_check_refs() seems to have a couple issues.
* It has a fast exit path which tests percpu_ref_is_zero(). However,
a percpu_ref reading zero doesn't mean that the object can be
released. In fact, the ->release() callback might not even have
started executing yet. Proceeding with freeing can lead to
use-after-free.
* lkey_table is RCU protected but there is no RCU grace period in the
free path. percpu_ref uses RCU internally but it's sched-RCU whose
grace periods are different from regular RCU. Also, it generally
isn't a good idea to depend on internal behaviors like this.
To address the above issues, this patch removes the fast exit and adds
an explicit synchronize_rcu().
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Cc: Mike Marciniszyn <mike.marciniszyn@intel.com>
Cc: linux-rdma@vger.kernel.org
Cc: Linus Torvalds <torvalds@linux-foundation.org>
While converting ioctx index from a list to a table, db446a08c2
("aio: convert the ioctx list to table lookup v3") missed tagging
kioctx_table->table[] as an array of RCU pointers and using the
appropriate RCU accessors. This introduces a small window in the
lookup path where init and access may race.
Mark kioctx_table->table[] with __rcu and use the approriate RCU
accessors when using the field.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Jann Horn <jannh@google.com>
Fixes: db446a08c2 ("aio: convert the ioctx list to table lookup v3")
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: stable@vger.kernel.org # v3.12+
While fixing refcounting, e34ecee2ae ("aio: Fix a trinity splat")
incorrectly removed explicit RCU grace period before freeing kioctx.
The intention seems to be depending on the internal RCU grace periods
of percpu_ref; however, percpu_ref uses a different flavor of RCU,
sched-RCU. This can lead to kioctx being freed while RCU read
protected dereferences are still in progress.
Fix it by updating free_ioctx() to go through call_rcu() explicitly.
v2: Comment added to explain double bouncing.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Jann Horn <jannh@google.com>
Fixes: e34ecee2ae ("aio: Fix a trinity splat")
Cc: Kent Overstreet <kent.overstreet@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: stable@vger.kernel.org # v3.13+
A tty is hung up by __tty_hangup() setting file->f_op to
hung_up_tty_fops, which is skipped on ttys whose write operation isn't
tty_write(). This means that, for example, /dev/console whose write
op is redirected_tty_write() is never actually marked hung up.
Because n_tty_read() uses the hung up status to decide whether to
abort the waiting readers, the lack of hung-up marking can lead to the
following scenario.
1. A session contains two processes. The leader and its child. The
child ignores SIGHUP.
2. The leader exits and starts disassociating from the controlling
terminal (/dev/console).
3. __tty_hangup() skips setting f_op to hung_up_tty_fops.
4. SIGHUP is delivered and ignored.
5. tty_ldisc_hangup() is invoked. It wakes up the waits which should
clear the read lockers of tty->ldisc_sem.
6. The reader wakes up but because tty_hung_up_p() is false, it
doesn't abort and goes back to sleep while read-holding
tty->ldisc_sem.
7. The leader progresses to tty_ldisc_lock() in tty_ldisc_hangup()
and is now stuck in D sleep indefinitely waiting for
tty->ldisc_sem.
The following is Alan's explanation on why some ttys aren't hung up.
http://lkml.kernel.org/r/20171101170908.6ad08580@alans-desktop
1. It broke the serial consoles because they would hang up and close
down the hardware. With tty_port that *should* be fixable properly
for any cases remaining.
2. The console layer was (and still is) completely broken and doens't
refcount properly. So if you turn on console hangups it breaks (as
indeed does freeing consoles and half a dozen other things).
As neither can be fixed quickly, this patch works around the problem
by introducing a new flag, TTY_HUPPING, which is used solely to tell
n_tty_read() that hang-up is in progress for the console and the
readers should be aborted regardless of the hung-up status of the
device.
The following is a sample hung task warning caused by this issue.
INFO: task agetty:2662 blocked for more than 120 seconds.
Not tainted 4.11.3-dbg-tty-lockup-02478-gfd6c7ee-dirty #28
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
0 2662 1 0x00000086
Call Trace:
__schedule+0x267/0x890
schedule+0x36/0x80
schedule_timeout+0x23c/0x2e0
ldsem_down_write+0xce/0x1f6
tty_ldisc_lock+0x16/0x30
tty_ldisc_hangup+0xb3/0x1b0
__tty_hangup+0x300/0x410
disassociate_ctty+0x6c/0x290
do_exit+0x7ef/0xb00
do_group_exit+0x3f/0xa0
get_signal+0x1b3/0x5d0
do_signal+0x28/0x660
exit_to_usermode_loop+0x46/0x86
do_syscall_64+0x9c/0xb0
entry_SYSCALL64_slow_path+0x25/0x25
The following is the repro. Run "$PROG /dev/console". The parent
process hangs in D state.
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/wait.h>
#include <sys/ioctl.h>
#include <fcntl.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#include <errno.h>
#include <signal.h>
#include <time.h>
#include <termios.h>
int main(int argc, char **argv)
{
struct sigaction sact = { .sa_handler = SIG_IGN };
struct timespec ts1s = { .tv_sec = 1 };
pid_t pid;
int fd;
if (argc < 2) {
fprintf(stderr, "test-hung-tty /dev/$TTY\n");
return 1;
}
/* fork a child to ensure that it isn't already the session leader */
pid = fork();
if (pid < 0) {
perror("fork");
return 1;
}
if (pid > 0) {
/* top parent, wait for everyone */
while (waitpid(-1, NULL, 0) >= 0)
;
if (errno != ECHILD)
perror("waitpid");
return 0;
}
/* new session, start a new session and set the controlling tty */
if (setsid() < 0) {
perror("setsid");
return 1;
}
fd = open(argv[1], O_RDWR);
if (fd < 0) {
perror("open");
return 1;
}
if (ioctl(fd, TIOCSCTTY, 1) < 0) {
perror("ioctl");
return 1;
}
/* fork a child, sleep a bit and exit */
pid = fork();
if (pid < 0) {
perror("fork");
return 1;
}
if (pid > 0) {
nanosleep(&ts1s, NULL);
printf("Session leader exiting\n");
exit(0);
}
/*
* The child ignores SIGHUP and keeps reading from the controlling
* tty. Because SIGHUP is ignored, the child doesn't get killed on
* parent exit and the bug in n_tty makes the read(2) block the
* parent's control terminal hangup attempt. The parent ends up in
* D sleep until the child is explicitly killed.
*/
sigaction(SIGHUP, &sact, NULL);
printf("Child reading tty\n");
while (1) {
char buf[1024];
if (read(fd, buf, sizeof(buf)) < 0) {
perror("read");
return 1;
}
}
return 0;
}
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Alan Cox <alan@llwyncelyn.cymru>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A domain cgroup isn't allowed to be turned threaded if its subtree is
populated or domain controllers are enabled. cgroup_enable_threaded()
depended on cgroup_can_be_thread_root() test to enforce this rule. A
parent which has populated domain descendants or have domain
controllers enabled can't become a thread root, so the above rules are
enforced automatically.
However, for the root cgroup which can host mixed domain and threaded
children, cgroup_can_be_thread_root() doesn't check any of those
conditions and thus first level cgroups ends up escaping those rules.
This patch fixes the bug by adding explicit checks for those rules in
cgroup_enable_threaded().
Reported-by: Michael Kerrisk (man-pages) <mtk.manpages@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: 8cfd8147df ("cgroup: implement cgroup v2 thread support")
Cc: stable@vger.kernel.org # v4.14+
Dennis rewrote the percpu area allocator some months ago, understands
most of the code base and has been responsive with the bug reports and
questions. Let's add him as a co-maintainer.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Christopher Lameter <cl@linux.com>
While adding cgroup2 interface for the cpu controller, 0d5936344f
("sched: Implement interface for cgroup unified hierarchy") forgot to
update input validation and left it to reject cpu.max config if any
descendant has set a higher value.
cgroup2 officially supports delegation and a descendant must not be
able to restrict what its ancestors can configure. For absolute
limits such as cpu.max and memory.max, this means that the config at
each level should only act as the upper limit at that level and
shouldn't interfere with what other cgroups can configure.
This patch updates config validation on cgroup2 so that the cpu
controller follows the same convention.
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: 0d5936344f ("sched: Implement interface for cgroup unified hierarchy")
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org # v4.15+
e7fd37ba12 ("cgroup: avoid copying strings longer than the buffers")
converted possibly unsafe strncpy() usages in cgroup to strscpy().
However, although the callsites are completely fine with truncated
copied, because strscpy() is marked __must_check, it led to the
following warnings.
kernel/cgroup/cgroup.c: In function ‘cgroup_file_name’:
kernel/cgroup/cgroup.c:1400:10: warning: ignoring return value of ‘strscpy’, declared with attribute warn_unused_result [-Wunused-result]
strscpy(buf, cft->name, CGROUP_FILE_NAME_MAX);
^
To avoid the warnings, 50034ed496 ("cgroup: use strlcpy() instead of
strscpy() to avoid spurious warning") switched them to strlcpy().
strlcpy() is worse than strlcpy() because it unconditionally runs
strlen() on the source string, and the only reason we switched to
strlcpy() here was because it was lacking __must_check, which doesn't
reflect any material differences between the two function. It's just
that someone added __must_check to strscpy() and not to strlcpy().
These basic string copy operations are used in variety of ways, and
one of not-so-uncommon use cases is safely handling truncated copies,
where the caller naturally doesn't care about the return value. The
__must_check doesn't match the actual use cases and forces users to
opt for inferior variants which lack __must_check by happenstance or
spread ugly (void) casts.
Remove __must_check from strscpy() and restore strscpy() usages in
cgroup.
Signed-off-by: Tejun Heo <tj@kernel.org>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Ma Shimiao <mashimiao.fnst@cn.fujitsu.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
The RCU protection has been expanded to cover both queueing and
completion paths making ->queue_rq_srcu a misnomer. Rename it to
->srcu as suggested by Bart.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Bart Van Assche <Bart.VanAssche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After the recent updates to use generation number and state based
synchronization, we can easily replace REQ_ATOM_STARTED usages by
adding an extra state to distinguish completed but not yet freed
state.
Add MQ_RQ_COMPLETE and replace REQ_ATOM_STARTED usages with
blk_mq_rq_state() tests. REQ_ATOM_STARTED no longer has any users
left and is removed.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After the recent updates to use generation number and state based
synchronization, blk-mq no longer depends on REQ_ATOM_COMPLETE except
to avoid firing the same timeout multiple times.
Remove all REQ_ATOM_COMPLETE usages and use a new rq_flags flag
RQF_MQ_TIMEOUT_EXPIRED to avoid firing the same timeout multiple
times. This removes atomic bitops from hot paths too.
v2: Removed blk_clear_rq_complete() from blk_mq_rq_timed_out().
v3: Added RQF_MQ_TIMEOUT_EXPIRED flag.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: "jianchao.wang" <jianchao.w.wang@oracle.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
With issue/complete and timeout paths now using the generation number
and state based synchronization, blk_abort_request() is the only one
which depends on REQ_ATOM_COMPLETE for arbitrating completion.
There's no reason for blk_abort_request() to be a completely separate
path. This patch makes blk_abort_request() piggyback on the timeout
path instead of trying to terminate the request directly.
This removes the last dependency on REQ_ATOM_COMPLETE in blk-mq.
Note that this makes blk_abort_request() asynchronous - it initiates
abortion but the actual termination will happen after a short while,
even when the caller owns the request. AFAICS, SCSI and ATA should be
fine with that and I think mtip32xx and dasd should be safe but not
completely sure. It'd be great if people who know the drivers take a
look.
v2: - Add comment explaining the lack of synchronization around
->deadline update as requested by Bart.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Asai Thambi SP <asamymuthupa@micron.com>
Cc: Stefan Haberland <sth@linux.vnet.ibm.com>
Cc: Jan Hoeppner <hoeppner@linux.vnet.ibm.com>
Cc: Bart Van Assche <Bart.VanAssche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_check_inflight() and blk_mq_poll_hybrid_sleep() test
REQ_ATOM_COMPLETE to determine the request state. Both uses are
speculative and we can test REQ_ATOM_STARTED and blk_mq_rq_state() for
equivalent results. Replace the tests. This will allow removing
REQ_ATOM_COMPLETE usages from blk-mq.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently, blk-mq timeout path synchronizes against the usual
issue/completion path using a complex scheme involving atomic
bitflags, REQ_ATOM_*, memory barriers and subtle memory coherence
rules. Unfortunately, it contains quite a few holes.
There's a complex dancing around REQ_ATOM_STARTED and
REQ_ATOM_COMPLETE between issue/completion and timeout paths; however,
they don't have a synchronization point across request recycle
instances and it isn't clear what the barriers add.
blk_mq_check_expired() can easily read STARTED from N-2'th iteration,
deadline from N-1'th, blk_mark_rq_complete() against Nth instance.
In fact, it's pretty easy to make blk_mq_check_expired() terminate a
later instance of a request. If we induce 5 sec delay before
time_after_eq() test in blk_mq_check_expired(), shorten the timeout to
2s, and issue back-to-back large IOs, blk-mq starts timing out
requests spuriously pretty quickly. Nothing actually timed out. It
just made the call on a recycle instance of a request and then
terminated a later instance long after the original instance finished.
The scenario isn't theoretical either.
This patch replaces the broken synchronization mechanism with a RCU
and generation number based one.
1. Each request has a u64 generation + state value, which can be
updated only by the request owner. Whenever a request becomes
in-flight, the generation number gets bumped up too. This provides
the basis for the timeout path to distinguish different recycle
instances of the request.
Also, marking a request in-flight and setting its deadline are
protected with a seqcount so that the timeout path can fetch both
values coherently.
2. The timeout path fetches the generation, state and deadline. If
the verdict is timeout, it records the generation into a dedicated
request abortion field and does RCU wait.
3. The completion path is also protected by RCU (from the previous
patch) and checks whether the current generation number and state
match the abortion field. If so, it skips completion.
4. The timeout path, after RCU wait, scans requests again and
terminates the ones whose generation and state still match the ones
requested for abortion.
By now, the timeout path knows that either the generation number
and state changed if it lost the race or the completion will yield
to it and can safely timeout the request.
While it's more lines of code, it's conceptually simpler, doesn't
depend on direct use of subtle memory ordering or coherence, and
hopefully doesn't terminate the wrong instance.
While this change makes REQ_ATOM_COMPLETE synchronization unnecessary
between issue/complete and timeout paths, REQ_ATOM_COMPLETE isn't
removed yet as it's still used in other places. Future patches will
move all state tracking to the new mechanism and remove all bitops in
the hot paths.
Note that this patch adds a comment explaining a race condition in
BLK_EH_RESET_TIMER path. The race has always been there and this
patch doesn't change it. It's just documenting the existing race.
v2: - Fixed BLK_EH_RESET_TIMER handling as pointed out by Jianchao.
- s/request->gstate_seqc/request->gstate_seq/ as suggested by Peter.
- READ_ONCE() added in blk_mq_rq_update_state() as suggested by Peter.
v3: - Fixed possible extended seqcount / u64_stats_sync read looping
spotted by Peter.
- MQ_RQ_IDLE was incorrectly being set in complete_request instead
of free_request. Fixed.
v4: - Rebased on top of hctx_lock() refactoring patch.
- Added comment explaining the use of hctx_lock() in completion path.
v5: - Added comments requested by Bart.
- Note the addition of BLK_EH_RESET_TIMER race condition in the
commit message.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: "jianchao.wang" <jianchao.w.wang@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Bart Van Assche <Bart.VanAssche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently, blk-mq protects only the issue path with RCU. This patch
puts the completion path under the same RCU protection. This will be
used to synchronize issue/completion against timeout by later patches,
which will also add the comments.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Workqueues can be created early during boot before workqueue subsystem
in fully online - work items are queued waiting for later full
initialization. However, early init wasn't supported for
WQ_MEM_RECLAIM workqueues causing unnecessary annoyances for a subset
of users. Expand early init support to include WQ_MEM_RECLAIM
workqueues.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
Separate out init_rescuer() from __alloc_workqueue_key() to prepare
for early init support for WQ_MEM_RECLAIM. This patch doesn't
introduce any functional changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
We haven't yet figured out what to do with RT threads on cgroup2.
Document the limitation.
v2: Included the warning about system management software behavior as
suggested by Michael.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: "Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>
This reverts commit aa24163b2e.
This and the following commit led to another circular locking scenario
and the scenario which is fixed by this commit no longer exists after
e8b3f8db7a ("workqueue/hotplug: simplify workqueue_offline_cpu()")
which removes work item flushing from hotplug path.
Revert it for now.
Signed-off-by: Tejun Heo <tj@kernel.org>
This reverts commit 1599a185f0.
This and the previous commit led to another circular locking scenario
and the scenario which is fixed by this commit no longer exists after
e8b3f8db7a ("workqueue/hotplug: simplify workqueue_offline_cpu()")
which removes work item flushing from hotplug path.
Revert it for now.
Signed-off-by: Tejun Heo <tj@kernel.org>
This reverts commit b5149873a0.
It conflicts with the following isolcpus change from the sched branch.
edb9382175 ("sched/isolation: Move isolcpus= handling to the housekeeping code")
Let's revert for now.
Signed-off-by: Tejun Heo <tj@kernel.org>
The local variable @cgrp isn't used if !CONFIG_CGROUP_SCHED. Mark the
variable with __maybe_unused to avoid a compile warning.
Reported-by: "kbuild-all@01.org" <kbuild-all@01.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
The basic cpu stat is currently shown with "cpu." prefix in
cgroup.stat, and the same information is duplicated in cpu.stat when
cpu controller is enabled. This is ugly and not very scalable as we
want to expand the coverage of stat information which is always
available.
This patch makes cgroup core always create "cpu.stat" file and show
the basic cpu stat there and calls the cpu controller to show the
extra stats when enabled. This ensures that the same information
isn't presented in multiple places and makes future expansion of basic
stats easier.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Josef reported a HARDIRQ-safe -> HARDIRQ-unsafe lock order detected by
lockdep:
[ 1270.472259] WARNING: HARDIRQ-safe -> HARDIRQ-unsafe lock order detected
[ 1270.472783] 4.14.0-rc1-xfstests-12888-g76833e8 #110 Not tainted
[ 1270.473240] -----------------------------------------------------
[ 1270.473710] kworker/u5:2/5157 [HC0[0]:SC0[0]:HE0:SE1] is trying to acquire:
[ 1270.474239] (&(&lock->wait_lock)->rlock){+.+.}, at: [<ffffffff8da253d2>] __mutex_unlock_slowpath+0xa2/0x280
[ 1270.474994]
[ 1270.474994] and this task is already holding:
[ 1270.475440] (&pool->lock/1){-.-.}, at: [<ffffffff8d2992f6>] worker_thread+0x366/0x3c0
[ 1270.476046] which would create a new lock dependency:
[ 1270.476436] (&pool->lock/1){-.-.} -> (&(&lock->wait_lock)->rlock){+.+.}
[ 1270.476949]
[ 1270.476949] but this new dependency connects a HARDIRQ-irq-safe lock:
[ 1270.477553] (&pool->lock/1){-.-.}
...
[ 1270.488900] to a HARDIRQ-irq-unsafe lock:
[ 1270.489327] (&(&lock->wait_lock)->rlock){+.+.}
...
[ 1270.494735] Possible interrupt unsafe locking scenario:
[ 1270.494735]
[ 1270.495250] CPU0 CPU1
[ 1270.495600] ---- ----
[ 1270.495947] lock(&(&lock->wait_lock)->rlock);
[ 1270.496295] local_irq_disable();
[ 1270.496753] lock(&pool->lock/1);
[ 1270.497205] lock(&(&lock->wait_lock)->rlock);
[ 1270.497744] <Interrupt>
[ 1270.497948] lock(&pool->lock/1);
, which will cause a irq inversion deadlock if the above lock scenario
happens.
The root cause of this safe -> unsafe lock order is the
mutex_unlock(pool->manager_arb) in manage_workers() with pool->lock
held.
Unlocking mutex while holding an irq spinlock was never safe and this
problem has been around forever but it never got noticed because the
only time the mutex is usually trylocked while holding irqlock making
actual failures very unlikely and lockdep annotation missed the
condition until the recent b9c16a0e1f ("locking/mutex: Fix
lockdep_assert_held() fail").
Using mutex for pool->manager_arb has always been a bit of stretch.
It primarily is an mechanism to arbitrate managership between workers
which can easily be done with a pool flag. The only reason it became
a mutex is that pool destruction path wants to exclude parallel
managing operations.
This patch replaces the mutex with a new pool flag POOL_MANAGER_ACTIVE
and make the destruction path wait for the current manager on a wait
queue.
v2: Drop unnecessary flag clearing before pool destruction as
suggested by Boqun.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: stable@vger.kernel.org
There are a couple interface issues which can be addressed in cgroup2
interface.
* Stats from cpuacct being reported separately from the cpu stats.
* Use of different time units. Writable control knobs use
microseconds, some stat fields use nanoseconds while other cpuacct
stat fields use centiseconds.
* Control knobs which can't be used in the root cgroup still show up
in the root.
* Control knob names and semantics aren't consistent with other
controllers.
This patchset implements cpu controller's interface on cgroup2 which
adheres to the controller file conventions described in
Documentation/cgroups/cgroup-v2.txt. Overall, the following changes
are made.
* cpuacct is implictly enabled and disabled by cpu and its information
is reported through "cpu.stat" which now uses microseconds for all
time durations. All time duration fields now have "_usec" appended
to them for clarity.
Note that cpuacct.usage_percpu is currently not included in
"cpu.stat". If this information is actually called for, it will be
added later.
* "cpu.shares" is replaced with "cpu.weight" and operates on the
standard scale defined by CGROUP_WEIGHT_MIN/DFL/MAX (1, 100, 10000).
The weight is scaled to scheduler weight so that 100 maps to 1024
and the ratio relationship is preserved - if weight is W and its
scaled value is S, W / 100 == S / 1024. While the mapped range is a
bit smaller than the orignal scheduler weight range, the dead zones
on both sides are relatively small and covers wider range than the
nice value mappings. This file doesn't make sense in the root
cgroup and isn't created on root.
* "cpu.weight.nice" is added. When read, it reads back the nice value
which is closest to the current "cpu.weight". When written, it sets
"cpu.weight" to the weight value which matches the nice value. This
makes it easy to configure cgroups when they're competing against
threads in threaded subtrees.
* "cpu.cfs_quota_us" and "cpu.cfs_period_us" are replaced by "cpu.max"
which contains both quota and period.
v4: - Use cgroup2 basic usage stat as the information source instead
of cpuacct.
v3: - Added "cpu.weight.nice" to allow using nice values when
configuring the weight. The feature is requested by PeterZ.
- Merge the patch to enable threaded support on cpu and cpuacct.
- Dropped the bits about getting rid of cpuacct from patch
description as there is a pretty strong case for making cpuacct
an implicit controller so that basic cpu usage stats are always
available.
- Documentation updated accordingly. "cpu.rt.max" section is
dropped for now.
v2: - cpu_stats_show() was incorrectly using CONFIG_FAIR_GROUP_SCHED
for CFS bandwidth stats and also using raw division for u64.
Use CONFIG_CFS_BANDWITH and do_div() instead. "cpu.rt.max" is
not included yet.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Make the following changes in preparation for the cpu controller
interface implementation for cgroup2. This patch doesn't cause any
functional differences.
* s/cpu_stats_show()/cpu_cfs_stat_show()/
* s/cpu_files/cpu_legacy_files/
v2: Dropped cpuacct changes as it won't be used by cpu controller
interface anymore.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
cfb766da54 ("sched/cputime: Expose cputime_adjust()") made
cputime_adjust() public for cgroup basic cpu stat support; however,
the commit forgot to add a dummy implementaiton for
CONFIG_VIRT_CPU_ACCOUNTING_NATIVE leading to compiler errors on some
s390 configurations.
Fix it by adding the missing dummy implementation.
Reported-by: “kbuild-all@01.org” <kbuild-all@01.org>
Fixes: cfb766da54 ("sched/cputime: Expose cputime_adjust()")
Signed-off-by: Tejun Heo <tj@kernel.org>
Tejun Heo