2008-10-18 20:27:21 -07:00
|
|
|
/*
|
|
|
|
* cgroup_freezer.c - control group freezer subsystem
|
|
|
|
*
|
|
|
|
* Copyright IBM Corporation, 2007
|
|
|
|
*
|
|
|
|
* Author : Cedric Le Goater <clg@fr.ibm.com>
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify it
|
|
|
|
* under the terms of version 2.1 of the GNU Lesser General Public License
|
|
|
|
* as published by the Free Software Foundation.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it would be useful, but
|
|
|
|
* WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
|
|
|
*/
|
|
|
|
|
2011-05-23 11:51:41 -07:00
|
|
|
#include <linux/export.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 01:04:11 -07:00
|
|
|
#include <linux/slab.h>
|
2008-10-18 20:27:21 -07:00
|
|
|
#include <linux/cgroup.h>
|
|
|
|
#include <linux/fs.h>
|
|
|
|
#include <linux/uaccess.h>
|
|
|
|
#include <linux/freezer.h>
|
|
|
|
#include <linux/seq_file.h>
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
#include <linux/mutex.h>
|
2023-04-05 06:15:32 -07:00
|
|
|
#include <linux/cpu.h>
|
2008-10-18 20:27:21 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
/*
|
|
|
|
* A cgroup is freezing if any FREEZING flags are set. FREEZING_SELF is
|
|
|
|
* set if "FROZEN" is written to freezer.state cgroupfs file, and cleared
|
|
|
|
* for "THAWED". FREEZING_PARENT is set if the parent freezer is FREEZING
|
|
|
|
* for whatever reason. IOW, a cgroup has FREEZING_PARENT set if one of
|
|
|
|
* its ancestors has FREEZING_SELF set.
|
|
|
|
*/
|
2012-11-09 10:12:30 -07:00
|
|
|
enum freezer_state_flags {
|
2012-11-09 10:12:30 -07:00
|
|
|
CGROUP_FREEZER_ONLINE = (1 << 0), /* freezer is fully online */
|
2012-11-09 10:12:30 -07:00
|
|
|
CGROUP_FREEZING_SELF = (1 << 1), /* this freezer is freezing */
|
|
|
|
CGROUP_FREEZING_PARENT = (1 << 2), /* the parent freezer is freezing */
|
2012-11-09 10:12:30 -07:00
|
|
|
CGROUP_FROZEN = (1 << 3), /* this and its descendants frozen */
|
2012-11-09 10:12:30 -07:00
|
|
|
|
|
|
|
/* mask for all FREEZING flags */
|
|
|
|
CGROUP_FREEZING = CGROUP_FREEZING_SELF | CGROUP_FREEZING_PARENT,
|
2008-10-18 20:27:21 -07:00
|
|
|
};
|
|
|
|
|
|
|
|
struct freezer {
|
2012-11-09 10:12:29 -07:00
|
|
|
struct cgroup_subsys_state css;
|
2012-11-09 10:12:30 -07:00
|
|
|
unsigned int state;
|
2008-10-18 20:27:21 -07:00
|
|
|
};
|
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
static DEFINE_MUTEX(freezer_mutex);
|
|
|
|
|
2013-08-08 17:11:23 -07:00
|
|
|
static inline struct freezer *css_freezer(struct cgroup_subsys_state *css)
|
|
|
|
{
|
|
|
|
return css ? container_of(css, struct freezer, css) : NULL;
|
|
|
|
}
|
|
|
|
|
2008-10-18 20:27:21 -07:00
|
|
|
static inline struct freezer *task_freezer(struct task_struct *task)
|
|
|
|
{
|
2014-02-08 08:36:58 -07:00
|
|
|
return css_freezer(task_css(task, freezer_cgrp_id));
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
static struct freezer *parent_freezer(struct freezer *freezer)
|
|
|
|
{
|
2014-05-16 10:22:48 -07:00
|
|
|
return css_freezer(freezer->css.parent);
|
2012-11-09 10:12:30 -07:00
|
|
|
}
|
|
|
|
|
2011-11-21 13:32:25 -07:00
|
|
|
bool cgroup_freezing(struct task_struct *task)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2011-11-21 13:32:25 -07:00
|
|
|
bool ret;
|
2023-11-15 09:20:43 -07:00
|
|
|
unsigned int state;
|
2008-10-18 20:27:21 -07:00
|
|
|
|
2011-11-21 13:32:25 -07:00
|
|
|
rcu_read_lock();
|
2023-11-15 09:20:43 -07:00
|
|
|
/* Check if the cgroup is still FREEZING, but not FROZEN. The extra
|
|
|
|
* !FROZEN check is required, because the FREEZING bit is not cleared
|
|
|
|
* when the state FROZEN is reached.
|
|
|
|
*/
|
|
|
|
state = task_freezer(task)->state;
|
|
|
|
ret = (state & CGROUP_FREEZING) && !(state & CGROUP_FROZEN);
|
2011-11-21 13:32:25 -07:00
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
return ret;
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
static const char *freezer_state_strs(unsigned int state)
|
|
|
|
{
|
|
|
|
if (state & CGROUP_FROZEN)
|
|
|
|
return "FROZEN";
|
|
|
|
if (state & CGROUP_FREEZING)
|
|
|
|
return "FREEZING";
|
|
|
|
return "THAWED";
|
2008-10-18 20:27:21 -07:00
|
|
|
};
|
|
|
|
|
2013-08-08 17:11:23 -07:00
|
|
|
static struct cgroup_subsys_state *
|
|
|
|
freezer_css_alloc(struct cgroup_subsys_state *parent_css)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
|
|
|
struct freezer *freezer;
|
|
|
|
|
|
|
|
freezer = kzalloc(sizeof(struct freezer), GFP_KERNEL);
|
|
|
|
if (!freezer)
|
|
|
|
return ERR_PTR(-ENOMEM);
|
|
|
|
|
|
|
|
return &freezer->css;
|
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
/**
|
2013-08-08 17:11:23 -07:00
|
|
|
* freezer_css_online - commit creation of a freezer css
|
|
|
|
* @css: css being created
|
2012-11-09 10:12:30 -07:00
|
|
|
*
|
2013-08-08 17:11:23 -07:00
|
|
|
* We're committing to creation of @css. Mark it online and inherit
|
2024-04-16 20:50:28 -07:00
|
|
|
* parent's freezing state while holding cpus read lock and freezer_mutex.
|
2012-11-09 10:12:30 -07:00
|
|
|
*/
|
2013-08-08 17:11:23 -07:00
|
|
|
static int freezer_css_online(struct cgroup_subsys_state *css)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2013-08-08 17:11:23 -07:00
|
|
|
struct freezer *freezer = css_freezer(css);
|
2012-11-09 10:12:30 -07:00
|
|
|
struct freezer *parent = parent_freezer(freezer);
|
|
|
|
|
2023-06-11 06:48:12 -07:00
|
|
|
cpus_read_lock();
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_lock(&freezer_mutex);
|
2011-11-21 13:32:25 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
freezer->state |= CGROUP_FREEZER_ONLINE;
|
2012-11-09 10:12:30 -07:00
|
|
|
|
|
|
|
if (parent && (parent->state & CGROUP_FREEZING)) {
|
|
|
|
freezer->state |= CGROUP_FREEZING_PARENT | CGROUP_FROZEN;
|
2023-06-11 06:48:12 -07:00
|
|
|
static_branch_inc_cpuslocked(&freezer_active);
|
2012-11-09 10:12:30 -07:00
|
|
|
}
|
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_unlock(&freezer_mutex);
|
2023-06-11 06:48:12 -07:00
|
|
|
cpus_read_unlock();
|
2012-11-19 09:13:38 -07:00
|
|
|
return 0;
|
2012-11-09 10:12:30 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2013-08-08 17:11:23 -07:00
|
|
|
* freezer_css_offline - initiate destruction of a freezer css
|
|
|
|
* @css: css being destroyed
|
2012-11-09 10:12:30 -07:00
|
|
|
*
|
2024-04-18 05:43:49 -07:00
|
|
|
* @css is going away. Mark it dead and decrement freezer_active if
|
2013-08-08 17:11:23 -07:00
|
|
|
* it was holding one.
|
2012-11-09 10:12:30 -07:00
|
|
|
*/
|
2013-08-08 17:11:23 -07:00
|
|
|
static void freezer_css_offline(struct cgroup_subsys_state *css)
|
2012-11-09 10:12:30 -07:00
|
|
|
{
|
2013-08-08 17:11:23 -07:00
|
|
|
struct freezer *freezer = css_freezer(css);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
2023-06-11 06:48:12 -07:00
|
|
|
cpus_read_lock();
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_lock(&freezer_mutex);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
if (freezer->state & CGROUP_FREEZING)
|
2023-06-11 06:48:12 -07:00
|
|
|
static_branch_dec_cpuslocked(&freezer_active);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
|
|
|
freezer->state = 0;
|
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_unlock(&freezer_mutex);
|
2023-06-11 06:48:12 -07:00
|
|
|
cpus_read_unlock();
|
2012-11-09 10:12:30 -07:00
|
|
|
}
|
|
|
|
|
2013-08-08 17:11:23 -07:00
|
|
|
static void freezer_css_free(struct cgroup_subsys_state *css)
|
2012-11-09 10:12:30 -07:00
|
|
|
{
|
2013-08-08 17:11:23 -07:00
|
|
|
kfree(css_freezer(css));
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2008-10-18 20:27:22 -07:00
|
|
|
/*
|
2012-10-16 15:03:15 -07:00
|
|
|
* Tasks can be migrated into a different freezer anytime regardless of its
|
|
|
|
* current state. freezer_attach() is responsible for making new tasks
|
|
|
|
* conform to the current state.
|
|
|
|
*
|
|
|
|
* Freezer state changes and task migration are synchronized via
|
|
|
|
* @freezer->lock. freezer_attach() makes the new tasks conform to the
|
|
|
|
* current state and all following state changes can see the new tasks.
|
2008-10-18 20:27:22 -07:00
|
|
|
*/
|
cgroup: fix handling of multi-destination migration from subtree_control enabling
Consider the following v2 hierarchy.
P0 (+memory) --- P1 (-memory) --- A
\- B
P0 has memory enabled in its subtree_control while P1 doesn't. If
both A and B contain processes, they would belong to the memory css of
P1. Now if memory is enabled on P1's subtree_control, memory csses
should be created on both A and B and A's processes should be moved to
the former and B's processes the latter. IOW, enabling controllers
can cause atomic migrations into different csses.
The core cgroup migration logic has been updated accordingly but the
controller migration methods haven't and still assume that all tasks
migrate to a single target css; furthermore, the methods were fed the
css in which subtree_control was updated which is the parent of the
target csses. pids controller depends on the migration methods to
move charges and this made the controller attribute charges to the
wrong csses often triggering the following warning by driving a
counter negative.
WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40()
Modules linked in:
CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29
...
ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000
ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00
ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8
Call Trace:
[<ffffffff81551ffc>] dump_stack+0x4e/0x82
[<ffffffff810de202>] warn_slowpath_common+0x82/0xc0
[<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20
[<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40
[<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0
[<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330
[<ffffffff81188e05>] cgroup_migrate+0xf5/0x190
[<ffffffff81189016>] cgroup_attach_task+0x176/0x200
[<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460
[<ffffffff81189684>] cgroup_procs_write+0x14/0x20
[<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0
[<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190
[<ffffffff81265f88>] __vfs_write+0x28/0xe0
[<ffffffff812666fc>] vfs_write+0xac/0x1a0
[<ffffffff81267019>] SyS_write+0x49/0xb0
[<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76
This patch fixes the bug by removing @css parameter from the three
migration methods, ->can_attach, ->cancel_attach() and ->attach() and
updating cgroup_taskset iteration helpers also return the destination
css in addition to the task being migrated. All controllers are
updated accordingly.
* Controllers which don't care whether there are one or multiple
target csses can be converted trivially. cpu, io, freezer, perf,
netclassid and netprio fall in this category.
* cpuset's current implementation assumes that there's single source
and destination and thus doesn't support v2 hierarchy already. The
only change made by this patchset is how that single destination css
is obtained.
* memory migration path already doesn't do anything on v2. How the
single destination css is obtained is updated and the prep stage of
mem_cgroup_can_attach() is reordered to accomodate the change.
* pids is the only controller which was affected by this bug. It now
correctly handles multi-destination migrations and no longer causes
counter underflow from incorrect accounting.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 08:18:21 -07:00
|
|
|
static void freezer_attach(struct cgroup_taskset *tset)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2011-12-12 19:12:21 -07:00
|
|
|
struct task_struct *task;
|
cgroup: fix handling of multi-destination migration from subtree_control enabling
Consider the following v2 hierarchy.
P0 (+memory) --- P1 (-memory) --- A
\- B
P0 has memory enabled in its subtree_control while P1 doesn't. If
both A and B contain processes, they would belong to the memory css of
P1. Now if memory is enabled on P1's subtree_control, memory csses
should be created on both A and B and A's processes should be moved to
the former and B's processes the latter. IOW, enabling controllers
can cause atomic migrations into different csses.
The core cgroup migration logic has been updated accordingly but the
controller migration methods haven't and still assume that all tasks
migrate to a single target css; furthermore, the methods were fed the
css in which subtree_control was updated which is the parent of the
target csses. pids controller depends on the migration methods to
move charges and this made the controller attribute charges to the
wrong csses often triggering the following warning by driving a
counter negative.
WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40()
Modules linked in:
CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29
...
ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000
ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00
ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8
Call Trace:
[<ffffffff81551ffc>] dump_stack+0x4e/0x82
[<ffffffff810de202>] warn_slowpath_common+0x82/0xc0
[<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20
[<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40
[<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0
[<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330
[<ffffffff81188e05>] cgroup_migrate+0xf5/0x190
[<ffffffff81189016>] cgroup_attach_task+0x176/0x200
[<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460
[<ffffffff81189684>] cgroup_procs_write+0x14/0x20
[<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0
[<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190
[<ffffffff81265f88>] __vfs_write+0x28/0xe0
[<ffffffff812666fc>] vfs_write+0xac/0x1a0
[<ffffffff81267019>] SyS_write+0x49/0xb0
[<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76
This patch fixes the bug by removing @css parameter from the three
migration methods, ->can_attach, ->cancel_attach() and ->attach() and
updating cgroup_taskset iteration helpers also return the destination
css in addition to the task being migrated. All controllers are
updated accordingly.
* Controllers which don't care whether there are one or multiple
target csses can be converted trivially. cpu, io, freezer, perf,
netclassid and netprio fall in this category.
* cpuset's current implementation assumes that there's single source
and destination and thus doesn't support v2 hierarchy already. The
only change made by this patchset is how that single destination css
is obtained.
* memory migration path already doesn't do anything on v2. How the
single destination css is obtained is updated and the prep stage of
mem_cgroup_can_attach() is reordered to accomodate the change.
* pids is the only controller which was affected by this bug. It now
correctly handles multi-destination migrations and no longer causes
counter underflow from incorrect accounting.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 08:18:21 -07:00
|
|
|
struct cgroup_subsys_state *new_css;
|
2008-10-18 20:27:22 -07:00
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_lock(&freezer_mutex);
|
2012-10-16 15:03:14 -07:00
|
|
|
|
2008-10-29 14:00:52 -07:00
|
|
|
/*
|
2013-08-08 17:11:23 -07:00
|
|
|
* Make the new tasks conform to the current state of @new_css.
|
2012-10-16 15:03:14 -07:00
|
|
|
* For simplicity, when migrating any task to a FROZEN cgroup, we
|
|
|
|
* revert it to FREEZING and let update_if_frozen() determine the
|
|
|
|
* correct state later.
|
|
|
|
*
|
2013-08-08 17:11:23 -07:00
|
|
|
* Tasks in @tset are on @new_css but may not conform to its
|
2012-10-16 15:03:14 -07:00
|
|
|
* current state before executing the following - !frozen tasks may
|
|
|
|
* be visible in a FROZEN cgroup and frozen tasks in a THAWED one.
|
2008-10-29 14:00:52 -07:00
|
|
|
*/
|
cgroup: fix handling of multi-destination migration from subtree_control enabling
Consider the following v2 hierarchy.
P0 (+memory) --- P1 (-memory) --- A
\- B
P0 has memory enabled in its subtree_control while P1 doesn't. If
both A and B contain processes, they would belong to the memory css of
P1. Now if memory is enabled on P1's subtree_control, memory csses
should be created on both A and B and A's processes should be moved to
the former and B's processes the latter. IOW, enabling controllers
can cause atomic migrations into different csses.
The core cgroup migration logic has been updated accordingly but the
controller migration methods haven't and still assume that all tasks
migrate to a single target css; furthermore, the methods were fed the
css in which subtree_control was updated which is the parent of the
target csses. pids controller depends on the migration methods to
move charges and this made the controller attribute charges to the
wrong csses often triggering the following warning by driving a
counter negative.
WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40()
Modules linked in:
CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29
...
ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000
ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00
ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8
Call Trace:
[<ffffffff81551ffc>] dump_stack+0x4e/0x82
[<ffffffff810de202>] warn_slowpath_common+0x82/0xc0
[<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20
[<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40
[<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0
[<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330
[<ffffffff81188e05>] cgroup_migrate+0xf5/0x190
[<ffffffff81189016>] cgroup_attach_task+0x176/0x200
[<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460
[<ffffffff81189684>] cgroup_procs_write+0x14/0x20
[<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0
[<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190
[<ffffffff81265f88>] __vfs_write+0x28/0xe0
[<ffffffff812666fc>] vfs_write+0xac/0x1a0
[<ffffffff81267019>] SyS_write+0x49/0xb0
[<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76
This patch fixes the bug by removing @css parameter from the three
migration methods, ->can_attach, ->cancel_attach() and ->attach() and
updating cgroup_taskset iteration helpers also return the destination
css in addition to the task being migrated. All controllers are
updated accordingly.
* Controllers which don't care whether there are one or multiple
target csses can be converted trivially. cpu, io, freezer, perf,
netclassid and netprio fall in this category.
* cpuset's current implementation assumes that there's single source
and destination and thus doesn't support v2 hierarchy already. The
only change made by this patchset is how that single destination css
is obtained.
* memory migration path already doesn't do anything on v2. How the
single destination css is obtained is updated and the prep stage of
mem_cgroup_can_attach() is reordered to accomodate the change.
* pids is the only controller which was affected by this bug. It now
correctly handles multi-destination migrations and no longer causes
counter underflow from incorrect accounting.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 08:18:21 -07:00
|
|
|
cgroup_taskset_for_each(task, new_css, tset) {
|
2015-12-03 08:18:21 -07:00
|
|
|
struct freezer *freezer = css_freezer(new_css);
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
if (!(freezer->state & CGROUP_FREEZING)) {
|
2012-10-16 15:03:14 -07:00
|
|
|
__thaw_task(task);
|
|
|
|
} else {
|
|
|
|
freeze_task(task);
|
freezer,sched: Rewrite core freezer logic
Rewrite the core freezer to behave better wrt thawing and be simpler
in general.
By replacing PF_FROZEN with TASK_FROZEN, a special block state, it is
ensured frozen tasks stay frozen until thawed and don't randomly wake
up early, as is currently possible.
As such, it does away with PF_FROZEN and PF_FREEZER_SKIP, freeing up
two PF_flags (yay!).
Specifically; the current scheme works a little like:
freezer_do_not_count();
schedule();
freezer_count();
And either the task is blocked, or it lands in try_to_freezer()
through freezer_count(). Now, when it is blocked, the freezer
considers it frozen and continues.
However, on thawing, once pm_freezing is cleared, freezer_count()
stops working, and any random/spurious wakeup will let a task run
before its time.
That is, thawing tries to thaw things in explicit order; kernel
threads and workqueues before doing bringing SMP back before userspace
etc.. However due to the above mentioned races it is entirely possible
for userspace tasks to thaw (by accident) before SMP is back.
This can be a fatal problem in asymmetric ISA architectures (eg ARMv9)
where the userspace task requires a special CPU to run.
As said; replace this with a special task state TASK_FROZEN and add
the following state transitions:
TASK_FREEZABLE -> TASK_FROZEN
__TASK_STOPPED -> TASK_FROZEN
__TASK_TRACED -> TASK_FROZEN
The new TASK_FREEZABLE can be set on any state part of TASK_NORMAL
(IOW. TASK_INTERRUPTIBLE and TASK_UNINTERRUPTIBLE) -- any such state
is already required to deal with spurious wakeups and the freezer
causes one such when thawing the task (since the original state is
lost).
The special __TASK_{STOPPED,TRACED} states *can* be restored since
their canonical state is in ->jobctl.
With this, frozen tasks need an explicit TASK_FROZEN wakeup and are
free of undue (early / spurious) wakeups.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20220822114649.055452969@infradead.org
2022-08-22 04:18:22 -07:00
|
|
|
|
2015-12-03 08:18:21 -07:00
|
|
|
/* clear FROZEN and propagate upwards */
|
|
|
|
while (freezer && (freezer->state & CGROUP_FROZEN)) {
|
|
|
|
freezer->state &= ~CGROUP_FROZEN;
|
|
|
|
freezer = parent_freezer(freezer);
|
|
|
|
}
|
2012-10-16 15:03:14 -07:00
|
|
|
}
|
|
|
|
}
|
2008-10-18 20:27:21 -07:00
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_unlock(&freezer_mutex);
|
2011-05-26 16:25:19 -07:00
|
|
|
}
|
|
|
|
|
2014-02-12 14:07:59 -07:00
|
|
|
/**
|
|
|
|
* freezer_fork - cgroup post fork callback
|
|
|
|
* @task: a task which has just been forked
|
|
|
|
*
|
|
|
|
* @task has just been created and should conform to the current state of
|
|
|
|
* the cgroup_freezer it belongs to. This function may race against
|
|
|
|
* freezer_attach(). Losing to freezer_attach() means that we don't have
|
|
|
|
* to do anything as freezer_attach() will put @task into the appropriate
|
|
|
|
* state.
|
|
|
|
*/
|
2015-12-03 08:24:08 -07:00
|
|
|
static void freezer_fork(struct task_struct *task)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
|
|
|
struct freezer *freezer;
|
|
|
|
|
2008-11-12 14:26:50 -07:00
|
|
|
/*
|
2014-02-12 14:07:59 -07:00
|
|
|
* The root cgroup is non-freezable, so we can skip locking the
|
|
|
|
* freezer. This is safe regardless of race with task migration.
|
|
|
|
* If we didn't race or won, skipping is obviously the right thing
|
|
|
|
* to do. If we lost and root is the new cgroup, noop is still the
|
|
|
|
* right thing to do.
|
2008-11-12 14:26:50 -07:00
|
|
|
*/
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
if (task_css_is_root(task, freezer_cgrp_id))
|
|
|
|
return;
|
2008-11-12 14:26:50 -07:00
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_lock(&freezer_mutex);
|
|
|
|
rcu_read_lock();
|
|
|
|
|
|
|
|
freezer = task_freezer(task);
|
|
|
|
if (freezer->state & CGROUP_FREEZING)
|
2011-11-21 13:32:26 -07:00
|
|
|
freeze_task(task);
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
|
2012-10-16 15:03:14 -07:00
|
|
|
rcu_read_unlock();
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_unlock(&freezer_mutex);
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
/**
|
|
|
|
* update_if_frozen - update whether a cgroup finished freezing
|
2013-08-08 17:11:24 -07:00
|
|
|
* @css: css of interest
|
2012-11-09 10:12:30 -07:00
|
|
|
*
|
|
|
|
* Once FREEZING is initiated, transition to FROZEN is lazily updated by
|
|
|
|
* calling this function. If the current state is FREEZING but not FROZEN,
|
|
|
|
* this function checks whether all tasks of this cgroup and the descendant
|
|
|
|
* cgroups finished freezing and, if so, sets FROZEN.
|
|
|
|
*
|
|
|
|
* The caller is responsible for grabbing RCU read lock and calling
|
|
|
|
* update_if_frozen() on all descendants prior to invoking this function.
|
2012-10-16 15:03:14 -07:00
|
|
|
*
|
|
|
|
* Task states and freezer state might disagree while tasks are being
|
2013-08-08 17:11:24 -07:00
|
|
|
* migrated into or out of @css, so we can't verify task states against
|
2012-10-16 15:03:15 -07:00
|
|
|
* @freezer state here. See freezer_attach() for details.
|
2008-10-18 20:27:21 -07:00
|
|
|
*/
|
2013-08-08 17:11:24 -07:00
|
|
|
static void update_if_frozen(struct cgroup_subsys_state *css)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2013-08-08 17:11:24 -07:00
|
|
|
struct freezer *freezer = css_freezer(css);
|
2013-08-08 17:11:25 -07:00
|
|
|
struct cgroup_subsys_state *pos;
|
2013-08-08 17:11:26 -07:00
|
|
|
struct css_task_iter it;
|
2008-10-18 20:27:21 -07:00
|
|
|
struct task_struct *task;
|
2012-10-16 15:03:14 -07:00
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
lockdep_assert_held(&freezer_mutex);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
if (!(freezer->state & CGROUP_FREEZING) ||
|
|
|
|
(freezer->state & CGROUP_FROZEN))
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
return;
|
2012-11-09 10:12:30 -07:00
|
|
|
|
|
|
|
/* are all (live) children frozen? */
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
rcu_read_lock();
|
2013-08-08 17:11:25 -07:00
|
|
|
css_for_each_child(pos, css) {
|
|
|
|
struct freezer *child = css_freezer(pos);
|
2008-10-18 20:27:21 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
if ((child->state & CGROUP_FREEZER_ONLINE) &&
|
2014-05-13 08:28:30 -07:00
|
|
|
!(child->state & CGROUP_FROZEN)) {
|
|
|
|
rcu_read_unlock();
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
return;
|
2014-05-13 08:28:30 -07:00
|
|
|
}
|
2012-11-09 10:12:30 -07:00
|
|
|
}
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
rcu_read_unlock();
|
2012-11-09 10:12:30 -07:00
|
|
|
|
|
|
|
/* are all tasks frozen? */
|
2017-05-15 06:34:01 -07:00
|
|
|
css_task_iter_start(css, 0, &it);
|
2012-10-16 15:03:14 -07:00
|
|
|
|
2013-08-08 17:11:26 -07:00
|
|
|
while ((task = css_task_iter_next(&it))) {
|
freezer,sched: Rewrite core freezer logic
Rewrite the core freezer to behave better wrt thawing and be simpler
in general.
By replacing PF_FROZEN with TASK_FROZEN, a special block state, it is
ensured frozen tasks stay frozen until thawed and don't randomly wake
up early, as is currently possible.
As such, it does away with PF_FROZEN and PF_FREEZER_SKIP, freeing up
two PF_flags (yay!).
Specifically; the current scheme works a little like:
freezer_do_not_count();
schedule();
freezer_count();
And either the task is blocked, or it lands in try_to_freezer()
through freezer_count(). Now, when it is blocked, the freezer
considers it frozen and continues.
However, on thawing, once pm_freezing is cleared, freezer_count()
stops working, and any random/spurious wakeup will let a task run
before its time.
That is, thawing tries to thaw things in explicit order; kernel
threads and workqueues before doing bringing SMP back before userspace
etc.. However due to the above mentioned races it is entirely possible
for userspace tasks to thaw (by accident) before SMP is back.
This can be a fatal problem in asymmetric ISA architectures (eg ARMv9)
where the userspace task requires a special CPU to run.
As said; replace this with a special task state TASK_FROZEN and add
the following state transitions:
TASK_FREEZABLE -> TASK_FROZEN
__TASK_STOPPED -> TASK_FROZEN
__TASK_TRACED -> TASK_FROZEN
The new TASK_FREEZABLE can be set on any state part of TASK_NORMAL
(IOW. TASK_INTERRUPTIBLE and TASK_UNINTERRUPTIBLE) -- any such state
is already required to deal with spurious wakeups and the freezer
causes one such when thawing the task (since the original state is
lost).
The special __TASK_{STOPPED,TRACED} states *can* be restored since
their canonical state is in ->jobctl.
With this, frozen tasks need an explicit TASK_FROZEN wakeup and are
free of undue (early / spurious) wakeups.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20220822114649.055452969@infradead.org
2022-08-22 04:18:22 -07:00
|
|
|
if (freezing(task) && !frozen(task))
|
|
|
|
goto out_iter_end;
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
freezer->state |= CGROUP_FROZEN;
|
2012-11-09 10:12:30 -07:00
|
|
|
out_iter_end:
|
2013-08-08 17:11:26 -07:00
|
|
|
css_task_iter_end(&it);
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2013-12-05 10:28:04 -07:00
|
|
|
static int freezer_read(struct seq_file *m, void *v)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2013-12-05 10:28:04 -07:00
|
|
|
struct cgroup_subsys_state *css = seq_css(m), *pos;
|
2008-10-18 20:27:21 -07:00
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_lock(&freezer_mutex);
|
2012-11-09 10:12:30 -07:00
|
|
|
rcu_read_lock();
|
2008-10-18 20:27:21 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
/* update states bottom-up */
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
css_for_each_descendant_post(pos, css) {
|
2014-05-13 09:11:01 -07:00
|
|
|
if (!css_tryget_online(pos))
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
continue;
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
2013-08-08 17:11:25 -07:00
|
|
|
update_if_frozen(pos);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
rcu_read_lock();
|
|
|
|
css_put(pos);
|
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
rcu_read_unlock();
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_unlock(&freezer_mutex);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
2013-08-08 17:11:24 -07:00
|
|
|
seq_puts(m, freezer_state_strs(css_freezer(css)->state));
|
2008-10-18 20:27:21 -07:00
|
|
|
seq_putc(m, '\n');
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:29 -07:00
|
|
|
static void freeze_cgroup(struct freezer *freezer)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2013-08-08 17:11:26 -07:00
|
|
|
struct css_task_iter it;
|
2008-10-18 20:27:21 -07:00
|
|
|
struct task_struct *task;
|
|
|
|
|
2017-05-15 06:34:01 -07:00
|
|
|
css_task_iter_start(&freezer->css, 0, &it);
|
2013-08-08 17:11:26 -07:00
|
|
|
while ((task = css_task_iter_next(&it)))
|
2012-10-16 15:03:14 -07:00
|
|
|
freeze_task(task);
|
2013-08-08 17:11:26 -07:00
|
|
|
css_task_iter_end(&it);
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:29 -07:00
|
|
|
static void unfreeze_cgroup(struct freezer *freezer)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2013-08-08 17:11:26 -07:00
|
|
|
struct css_task_iter it;
|
2008-10-18 20:27:21 -07:00
|
|
|
struct task_struct *task;
|
|
|
|
|
2017-05-15 06:34:01 -07:00
|
|
|
css_task_iter_start(&freezer->css, 0, &it);
|
2013-08-08 17:11:26 -07:00
|
|
|
while ((task = css_task_iter_next(&it)))
|
2011-11-21 13:32:23 -07:00
|
|
|
__thaw_task(task);
|
2013-08-08 17:11:26 -07:00
|
|
|
css_task_iter_end(&it);
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
/**
|
|
|
|
* freezer_apply_state - apply state change to a single cgroup_freezer
|
|
|
|
* @freezer: freezer to apply state change to
|
|
|
|
* @freeze: whether to freeze or unfreeze
|
2012-11-09 10:12:30 -07:00
|
|
|
* @state: CGROUP_FREEZING_* flag to set or clear
|
|
|
|
*
|
|
|
|
* Set or clear @state on @cgroup according to @freeze, and perform
|
|
|
|
* freezing or thawing as necessary.
|
2012-11-09 10:12:30 -07:00
|
|
|
*/
|
2012-11-09 10:12:30 -07:00
|
|
|
static void freezer_apply_state(struct freezer *freezer, bool freeze,
|
|
|
|
unsigned int state)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2012-10-16 15:03:15 -07:00
|
|
|
/* also synchronizes against task migration, see freezer_attach() */
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
lockdep_assert_held(&freezer_mutex);
|
2008-10-29 14:00:54 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
if (!(freezer->state & CGROUP_FREEZER_ONLINE))
|
|
|
|
return;
|
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
if (freeze) {
|
2012-11-09 10:12:30 -07:00
|
|
|
if (!(freezer->state & CGROUP_FREEZING))
|
2023-04-05 06:15:32 -07:00
|
|
|
static_branch_inc_cpuslocked(&freezer_active);
|
2012-11-09 10:12:30 -07:00
|
|
|
freezer->state |= state;
|
2012-11-09 10:12:29 -07:00
|
|
|
freeze_cgroup(freezer);
|
2012-11-09 10:12:30 -07:00
|
|
|
} else {
|
2012-11-09 10:12:30 -07:00
|
|
|
bool was_freezing = freezer->state & CGROUP_FREEZING;
|
|
|
|
|
|
|
|
freezer->state &= ~state;
|
|
|
|
|
|
|
|
if (!(freezer->state & CGROUP_FREEZING)) {
|
|
|
|
freezer->state &= ~CGROUP_FROZEN;
|
freezer,sched: Rewrite core freezer logic
Rewrite the core freezer to behave better wrt thawing and be simpler
in general.
By replacing PF_FROZEN with TASK_FROZEN, a special block state, it is
ensured frozen tasks stay frozen until thawed and don't randomly wake
up early, as is currently possible.
As such, it does away with PF_FROZEN and PF_FREEZER_SKIP, freeing up
two PF_flags (yay!).
Specifically; the current scheme works a little like:
freezer_do_not_count();
schedule();
freezer_count();
And either the task is blocked, or it lands in try_to_freezer()
through freezer_count(). Now, when it is blocked, the freezer
considers it frozen and continues.
However, on thawing, once pm_freezing is cleared, freezer_count()
stops working, and any random/spurious wakeup will let a task run
before its time.
That is, thawing tries to thaw things in explicit order; kernel
threads and workqueues before doing bringing SMP back before userspace
etc.. However due to the above mentioned races it is entirely possible
for userspace tasks to thaw (by accident) before SMP is back.
This can be a fatal problem in asymmetric ISA architectures (eg ARMv9)
where the userspace task requires a special CPU to run.
As said; replace this with a special task state TASK_FROZEN and add
the following state transitions:
TASK_FREEZABLE -> TASK_FROZEN
__TASK_STOPPED -> TASK_FROZEN
__TASK_TRACED -> TASK_FROZEN
The new TASK_FREEZABLE can be set on any state part of TASK_NORMAL
(IOW. TASK_INTERRUPTIBLE and TASK_UNINTERRUPTIBLE) -- any such state
is already required to deal with spurious wakeups and the freezer
causes one such when thawing the task (since the original state is
lost).
The special __TASK_{STOPPED,TRACED} states *can* be restored since
their canonical state is in ->jobctl.
With this, frozen tasks need an explicit TASK_FROZEN wakeup and are
free of undue (early / spurious) wakeups.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20220822114649.055452969@infradead.org
2022-08-22 04:18:22 -07:00
|
|
|
if (was_freezing)
|
2023-04-05 06:15:32 -07:00
|
|
|
static_branch_dec_cpuslocked(&freezer_active);
|
2012-11-09 10:12:30 -07:00
|
|
|
unfreeze_cgroup(freezer);
|
|
|
|
}
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
2012-11-09 10:12:30 -07:00
|
|
|
}
|
2011-11-21 13:32:25 -07:00
|
|
|
|
2012-11-09 10:12:30 -07:00
|
|
|
/**
|
|
|
|
* freezer_change_state - change the freezing state of a cgroup_freezer
|
|
|
|
* @freezer: freezer of interest
|
|
|
|
* @freeze: whether to freeze or thaw
|
|
|
|
*
|
2012-11-09 10:12:30 -07:00
|
|
|
* Freeze or thaw @freezer according to @freeze. The operations are
|
|
|
|
* recursive - all descendants of @freezer will be affected.
|
2012-11-09 10:12:30 -07:00
|
|
|
*/
|
|
|
|
static void freezer_change_state(struct freezer *freezer, bool freeze)
|
|
|
|
{
|
2013-08-08 17:11:25 -07:00
|
|
|
struct cgroup_subsys_state *pos;
|
2012-11-09 10:12:30 -07:00
|
|
|
|
2023-04-05 06:15:32 -07:00
|
|
|
cpus_read_lock();
|
2012-11-09 10:12:30 -07:00
|
|
|
/*
|
|
|
|
* Update all its descendants in pre-order traversal. Each
|
|
|
|
* descendant will try to inherit its parent's FREEZING state as
|
|
|
|
* CGROUP_FREEZING_PARENT.
|
|
|
|
*/
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_lock(&freezer_mutex);
|
2012-11-09 10:12:30 -07:00
|
|
|
rcu_read_lock();
|
2013-08-08 17:11:25 -07:00
|
|
|
css_for_each_descendant_pre(pos, &freezer->css) {
|
|
|
|
struct freezer *pos_f = css_freezer(pos);
|
2012-11-09 10:12:30 -07:00
|
|
|
struct freezer *parent = parent_freezer(pos_f);
|
|
|
|
|
2014-05-13 09:11:01 -07:00
|
|
|
if (!css_tryget_online(pos))
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
continue;
|
|
|
|
rcu_read_unlock();
|
2013-08-08 17:11:27 -07:00
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
if (pos_f == freezer)
|
2013-08-08 17:11:27 -07:00
|
|
|
freezer_apply_state(pos_f, freeze,
|
|
|
|
CGROUP_FREEZING_SELF);
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
else
|
2013-08-08 17:11:27 -07:00
|
|
|
freezer_apply_state(pos_f,
|
|
|
|
parent->state & CGROUP_FREEZING,
|
|
|
|
CGROUP_FREEZING_PARENT);
|
|
|
|
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
rcu_read_lock();
|
|
|
|
css_put(pos);
|
2012-11-09 10:12:30 -07:00
|
|
|
}
|
|
|
|
rcu_read_unlock();
|
cgroup_freezer: replace freezer->lock with freezer_mutex
After 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it
to css_set_rwsem"), css task iterators requires sleepable context as
it may block on css_set_rwsem. I missed that cgroup_freezer was
iterating tasks under IRQ-safe spinlock freezer->lock. This leads to
errors like the following on freezer state reads and transitions.
BUG: sleeping function called from invalid context at /work
/os/work/kernel/locking/rwsem.c:20
in_atomic(): 0, irqs_disabled(): 0, pid: 462, name: bash
5 locks held by bash/462:
#0: (sb_writers#7){.+.+.+}, at: [<ffffffff811f0843>] vfs_write+0x1a3/0x1c0
#1: (&of->mutex){+.+.+.}, at: [<ffffffff8126d78b>] kernfs_fop_write+0xbb/0x170
#2: (s_active#70){.+.+.+}, at: [<ffffffff8126d793>] kernfs_fop_write+0xc3/0x170
#3: (freezer_mutex){+.+...}, at: [<ffffffff81135981>] freezer_write+0x61/0x1e0
#4: (rcu_read_lock){......}, at: [<ffffffff81135973>] freezer_write+0x53/0x1e0
Preemption disabled at:[<ffffffff81104404>] console_unlock+0x1e4/0x460
CPU: 3 PID: 462 Comm: bash Not tainted 3.15.0-rc1-work+ #10
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
ffff88000916a6d0 ffff88000e0a3da0 ffffffff81cf8c96 0000000000000000
ffff88000e0a3dc8 ffffffff810cf4f2 ffffffff82388040 ffff880013aaf740
0000000000000002 ffff88000e0a3de8 ffffffff81d05974 0000000000000246
Call Trace:
[<ffffffff81cf8c96>] dump_stack+0x4e/0x7a
[<ffffffff810cf4f2>] __might_sleep+0x162/0x260
[<ffffffff81d05974>] down_read+0x24/0x60
[<ffffffff81133e87>] css_task_iter_start+0x27/0x70
[<ffffffff8113584d>] freezer_apply_state+0x5d/0x130
[<ffffffff81135a16>] freezer_write+0xf6/0x1e0
[<ffffffff8112eb88>] cgroup_file_write+0xd8/0x230
[<ffffffff8126d7b7>] kernfs_fop_write+0xe7/0x170
[<ffffffff811f0756>] vfs_write+0xb6/0x1c0
[<ffffffff811f121d>] SyS_write+0x4d/0xc0
[<ffffffff81d08292>] system_call_fastpath+0x16/0x1b
freezer->lock used to be used in hot paths but that time is long gone
and there's no reason for the lock to be IRQ-safe spinlock or even
per-cgroup. In fact, given the fact that a cgroup may contain large
number of tasks, it's not a good idea to iterate over them while
holding IRQ-safe spinlock.
Let's simplify locking by replacing per-cgroup freezer->lock with
global freezer_mutex. This also makes the comments explaining the
intricacies of policy inheritance and the locking around it as the
states are protected by a common mutex.
The conversion is mostly straight-forward. The followings are worth
mentioning.
* freezer_css_online() no longer needs double locking.
* freezer_attach() now performs propagation simply while holding
freezer_mutex. update_if_frozen() race no longer exists and the
comment is removed.
* freezer_fork() now tests whether the task is in root cgroup using
the new task_css_is_root() without doing rcu_read_lock/unlock(). If
not, it grabs freezer_mutex and performs the operation.
* freezer_read() and freezer_change_state() grab freezer_mutex across
the whole operation and pin the css while iterating so that each
descendant processing happens in sleepable context.
Fixes: 96d365e0b86e ("cgroup: make css_set_lock a rwsem and rename it to css_set_rwsem")
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
2014-05-07 18:31:17 -07:00
|
|
|
mutex_unlock(&freezer_mutex);
|
2023-04-05 06:15:32 -07:00
|
|
|
cpus_read_unlock();
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2014-05-13 09:16:21 -07:00
|
|
|
static ssize_t freezer_write(struct kernfs_open_file *of,
|
|
|
|
char *buf, size_t nbytes, loff_t off)
|
2008-10-18 20:27:21 -07:00
|
|
|
{
|
2012-11-09 10:12:30 -07:00
|
|
|
bool freeze;
|
2008-10-18 20:27:21 -07:00
|
|
|
|
2014-05-13 09:16:21 -07:00
|
|
|
buf = strstrip(buf);
|
|
|
|
|
|
|
|
if (strcmp(buf, freezer_state_strs(0)) == 0)
|
2012-11-09 10:12:30 -07:00
|
|
|
freeze = false;
|
2014-05-13 09:16:21 -07:00
|
|
|
else if (strcmp(buf, freezer_state_strs(CGROUP_FROZEN)) == 0)
|
2012-11-09 10:12:30 -07:00
|
|
|
freeze = true;
|
2008-10-18 20:27:21 -07:00
|
|
|
else
|
2008-11-12 14:26:50 -07:00
|
|
|
return -EINVAL;
|
2008-10-18 20:27:21 -07:00
|
|
|
|
2014-05-13 09:16:21 -07:00
|
|
|
freezer_change_state(css_freezer(of_css(of)), freeze);
|
|
|
|
return nbytes;
|
2008-10-18 20:27:21 -07:00
|
|
|
}
|
|
|
|
|
2013-08-08 17:11:24 -07:00
|
|
|
static u64 freezer_self_freezing_read(struct cgroup_subsys_state *css,
|
|
|
|
struct cftype *cft)
|
2012-11-09 10:12:30 -07:00
|
|
|
{
|
2013-08-08 17:11:24 -07:00
|
|
|
struct freezer *freezer = css_freezer(css);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
|
|
|
return (bool)(freezer->state & CGROUP_FREEZING_SELF);
|
|
|
|
}
|
|
|
|
|
2013-08-08 17:11:24 -07:00
|
|
|
static u64 freezer_parent_freezing_read(struct cgroup_subsys_state *css,
|
|
|
|
struct cftype *cft)
|
2012-11-09 10:12:30 -07:00
|
|
|
{
|
2013-08-08 17:11:24 -07:00
|
|
|
struct freezer *freezer = css_freezer(css);
|
2012-11-09 10:12:30 -07:00
|
|
|
|
|
|
|
return (bool)(freezer->state & CGROUP_FREEZING_PARENT);
|
|
|
|
}
|
|
|
|
|
2008-10-18 20:27:21 -07:00
|
|
|
static struct cftype files[] = {
|
|
|
|
{
|
|
|
|
.name = "state",
|
2012-04-01 12:09:55 -07:00
|
|
|
.flags = CFTYPE_NOT_ON_ROOT,
|
2013-12-05 10:28:04 -07:00
|
|
|
.seq_show = freezer_read,
|
2014-05-13 09:16:21 -07:00
|
|
|
.write = freezer_write,
|
2008-10-18 20:27:21 -07:00
|
|
|
},
|
2012-11-09 10:12:30 -07:00
|
|
|
{
|
|
|
|
.name = "self_freezing",
|
|
|
|
.flags = CFTYPE_NOT_ON_ROOT,
|
|
|
|
.read_u64 = freezer_self_freezing_read,
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "parent_freezing",
|
|
|
|
.flags = CFTYPE_NOT_ON_ROOT,
|
|
|
|
.read_u64 = freezer_parent_freezing_read,
|
|
|
|
},
|
2012-04-01 12:09:55 -07:00
|
|
|
{ } /* terminate */
|
2008-10-18 20:27:21 -07:00
|
|
|
};
|
|
|
|
|
2014-02-08 08:36:58 -07:00
|
|
|
struct cgroup_subsys freezer_cgrp_subsys = {
|
2012-11-19 09:13:38 -07:00
|
|
|
.css_alloc = freezer_css_alloc,
|
|
|
|
.css_online = freezer_css_online,
|
|
|
|
.css_offline = freezer_css_offline,
|
|
|
|
.css_free = freezer_css_free,
|
2012-10-16 15:03:14 -07:00
|
|
|
.attach = freezer_attach,
|
2008-10-18 20:27:21 -07:00
|
|
|
.fork = freezer_fork,
|
2014-07-15 08:05:09 -07:00
|
|
|
.legacy_cftypes = files,
|
2008-10-18 20:27:21 -07:00
|
|
|
};
|