2009-12-03 18:59:42 +01:00
|
|
|
/*
|
|
|
|
* Common Block IO controller cgroup interface
|
|
|
|
*
|
|
|
|
* Based on ideas and code from CFQ, CFS and BFQ:
|
|
|
|
* Copyright (C) 2003 Jens Axboe <axboe@kernel.dk>
|
|
|
|
*
|
|
|
|
* Copyright (C) 2008 Fabio Checconi <fabio@gandalf.sssup.it>
|
|
|
|
* Paolo Valente <paolo.valente@unimore.it>
|
|
|
|
*
|
|
|
|
* Copyright (C) 2009 Vivek Goyal <vgoyal@redhat.com>
|
|
|
|
* Nauman Rafique <nauman@google.com>
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
*
|
|
|
|
* For policy-specific per-blkcg data:
|
|
|
|
* Copyright (C) 2015 Paolo Valente <paolo.valente@unimore.it>
|
|
|
|
* Arianna Avanzini <avanzini.arianna@gmail.com>
|
2009-12-03 18:59:42 +01:00
|
|
|
*/
|
|
|
|
#include <linux/ioprio.h>
|
2009-12-03 18:59:49 +01:00
|
|
|
#include <linux/kdev_t.h>
|
2009-12-04 16:36:41 +01:00
|
|
|
#include <linux/module.h>
|
2009-12-07 09:29:39 +01:00
|
|
|
#include <linux/err.h>
|
2010-04-02 00:01:41 +02:00
|
|
|
#include <linux/blkdev.h>
|
2015-05-22 23:13:37 +02:00
|
|
|
#include <linux/backing-dev.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 09:04:11 +01:00
|
|
|
#include <linux/slab.h>
|
2010-04-13 10:05:49 +02:00
|
|
|
#include <linux/genhd.h>
|
2012-03-05 22:15:00 +01:00
|
|
|
#include <linux/delay.h>
|
2012-03-19 23:10:56 +01:00
|
|
|
#include <linux/atomic.h>
|
2015-08-18 23:55:31 +02:00
|
|
|
#include <linux/ctype.h>
|
2015-05-22 23:13:17 +02:00
|
|
|
#include <linux/blk-cgroup.h>
|
2012-03-05 22:15:12 +01:00
|
|
|
#include "blk.h"
|
2009-12-04 16:36:42 +01:00
|
|
|
|
2010-04-09 08:31:19 +02:00
|
|
|
#define MAX_KEY_LEN 100
|
|
|
|
|
2015-07-09 22:39:47 +02:00
|
|
|
/*
|
|
|
|
* blkcg_pol_mutex protects blkcg_policy[] and policy [de]activation.
|
|
|
|
* blkcg_pol_register_mutex nests outside of it and synchronizes entire
|
|
|
|
* policy [un]register operations including cgroup file additions /
|
|
|
|
* removals. Putting cgroup file registration outside blkcg_pol_mutex
|
|
|
|
* allows grabbing it from cgroup callbacks.
|
|
|
|
*/
|
|
|
|
static DEFINE_MUTEX(blkcg_pol_register_mutex);
|
2012-04-13 22:11:26 +02:00
|
|
|
static DEFINE_MUTEX(blkcg_pol_mutex);
|
2012-03-05 22:15:13 +01:00
|
|
|
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
struct blkcg blkcg_root;
|
2012-04-16 22:57:25 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkcg_root);
|
2009-12-04 16:36:41 +01:00
|
|
|
|
2015-05-22 23:13:21 +02:00
|
|
|
struct cgroup_subsys_state * const blkcg_root_css = &blkcg_root.css;
|
|
|
|
|
2012-04-16 22:57:25 +02:00
|
|
|
static struct blkcg_policy *blkcg_policy[BLKCG_MAX_POLS];
|
2012-03-05 22:15:04 +01:00
|
|
|
|
2015-07-09 22:39:49 +02:00
|
|
|
static LIST_HEAD(all_blkcgs); /* protected by blkcg_pol_mutex */
|
|
|
|
|
2012-04-13 22:11:33 +02:00
|
|
|
static bool blkcg_policy_enabled(struct request_queue *q,
|
2012-04-16 22:57:25 +02:00
|
|
|
const struct blkcg_policy *pol)
|
2012-04-13 22:11:33 +02:00
|
|
|
{
|
|
|
|
return pol && test_bit(pol->plid, q->blkcg_pols);
|
|
|
|
}
|
|
|
|
|
2012-03-05 22:15:14 +01:00
|
|
|
/**
|
|
|
|
* blkg_free - free a blkg
|
|
|
|
* @blkg: blkg to free
|
|
|
|
*
|
|
|
|
* Free @blkg which may be partially allocated.
|
|
|
|
*/
|
2012-04-16 22:57:25 +02:00
|
|
|
static void blkg_free(struct blkcg_gq *blkg)
|
2012-03-05 22:15:14 +01:00
|
|
|
{
|
2012-03-05 22:15:20 +01:00
|
|
|
int i;
|
2012-03-05 22:15:16 +01:00
|
|
|
|
|
|
|
if (!blkg)
|
|
|
|
return;
|
|
|
|
|
2013-05-14 22:52:31 +02:00
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++)
|
2015-08-18 23:55:11 +02:00
|
|
|
if (blkg->pd[i])
|
|
|
|
blkcg_policy[i]->pd_free_fn(blkg->pd[i]);
|
2012-03-05 22:15:20 +01:00
|
|
|
|
2015-08-18 23:55:07 +02:00
|
|
|
if (blkg->blkcg != &blkcg_root)
|
|
|
|
blk_exit_rl(&blkg->rl);
|
2015-08-18 23:55:24 +02:00
|
|
|
|
|
|
|
blkg_rwstat_exit(&blkg->stat_ios);
|
|
|
|
blkg_rwstat_exit(&blkg->stat_bytes);
|
2012-03-05 22:15:16 +01:00
|
|
|
kfree(blkg);
|
2012-03-05 22:15:14 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkg_alloc - allocate a blkg
|
|
|
|
* @blkcg: block cgroup the new blkg is associated with
|
|
|
|
* @q: request_queue the new blkg is associated with
|
2012-06-05 05:40:52 +02:00
|
|
|
* @gfp_mask: allocation mask to use
|
2012-03-05 22:15:14 +01:00
|
|
|
*
|
2012-03-05 22:15:20 +01:00
|
|
|
* Allocate a new blkg assocating @blkcg and @q.
|
2012-03-05 22:15:14 +01:00
|
|
|
*/
|
2012-06-05 05:40:52 +02:00
|
|
|
static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
|
|
|
|
gfp_t gfp_mask)
|
2012-03-05 22:15:14 +01:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2012-03-05 22:15:20 +01:00
|
|
|
int i;
|
2012-03-05 22:15:14 +01:00
|
|
|
|
|
|
|
/* alloc and init base part */
|
2012-06-05 05:40:52 +02:00
|
|
|
blkg = kzalloc_node(sizeof(*blkg), gfp_mask, q->node);
|
2012-03-05 22:15:14 +01:00
|
|
|
if (!blkg)
|
|
|
|
return NULL;
|
|
|
|
|
2015-08-18 23:55:24 +02:00
|
|
|
if (blkg_rwstat_init(&blkg->stat_bytes, gfp_mask) ||
|
|
|
|
blkg_rwstat_init(&blkg->stat_ios, gfp_mask))
|
|
|
|
goto err_free;
|
|
|
|
|
2012-03-05 22:15:22 +01:00
|
|
|
blkg->q = q;
|
2012-03-05 22:15:20 +01:00
|
|
|
INIT_LIST_HEAD(&blkg->q_node);
|
2012-03-05 22:15:14 +01:00
|
|
|
blkg->blkcg = blkcg;
|
blkcg: fix use-after-free in __blkg_release_rcu() by making blkcg_gq refcnt an atomic_t
Hello,
So, this patch should do. Joe, Vivek, can one of you guys please
verify that the oops goes away with this patch?
Jens, the original thread can be read at
http://thread.gmane.org/gmane.linux.kernel/1720729
The fix converts blkg->refcnt from int to atomic_t. It does some
overhead but it should be minute compared to everything else which is
going on and the involved cacheline bouncing, so I think it's highly
unlikely to cause any noticeable difference. Also, the refcnt in
question should be converted to a perpcu_ref for blk-mq anyway, so the
atomic_t is likely to go away pretty soon anyway.
Thanks.
------- 8< -------
__blkg_release_rcu() may be invoked after the associated request_queue
is released with a RCU grace period inbetween. As such, the function
and callbacks invoked from it must not dereference the associated
request_queue. This is clearly indicated in the comment above the
function.
Unfortunately, while trying to fix a different issue, 2a4fd070ee85
("blkcg: move bulk of blkcg_gq release operations to the RCU
callback") ignored this and added [un]locking of @blkg->q->queue_lock
to __blkg_release_rcu(). This of course can cause oops as the
request_queue may be long gone by the time this code gets executed.
general protection fault: 0000 [#1] SMP
CPU: 21 PID: 30 Comm: rcuos/21 Not tainted 3.15.0 #1
Hardware name: Stratus ftServer 6400/G7LAZ, BIOS BIOS Version 6.3:57 12/25/2013
task: ffff880854021de0 ti: ffff88085403c000 task.ti: ffff88085403c000
RIP: 0010:[<ffffffff8162e9e5>] [<ffffffff8162e9e5>] _raw_spin_lock_irq+0x15/0x60
RSP: 0018:ffff88085403fdf0 EFLAGS: 00010086
RAX: 0000000000020000 RBX: 0000000000000010 RCX: 0000000000000000
RDX: 000060ef80008248 RSI: 0000000000000286 RDI: 6b6b6b6b6b6b6b6b
RBP: ffff88085403fdf0 R08: 0000000000000286 R09: 0000000000009f39
R10: 0000000000020001 R11: 0000000000020001 R12: ffff88103c17a130
R13: ffff88103c17a080 R14: 0000000000000000 R15: 0000000000000000
FS: 0000000000000000(0000) GS:ffff88107fca0000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000006e5ab8 CR3: 000000000193d000 CR4: 00000000000407e0
Stack:
ffff88085403fe18 ffffffff812cbfc2 ffff88103c17a130 0000000000000000
ffff88103c17a130 ffff88085403fec0 ffffffff810d1d28 ffff880854021de0
ffff880854021de0 ffff88107fcaec58 ffff88085403fe80 ffff88107fcaec30
Call Trace:
[<ffffffff812cbfc2>] __blkg_release_rcu+0x72/0x150
[<ffffffff810d1d28>] rcu_nocb_kthread+0x1e8/0x300
[<ffffffff81091d81>] kthread+0xe1/0x100
[<ffffffff8163813c>] ret_from_fork+0x7c/0xb0
Code: ff 47 04 48 8b 7d 08 be 00 02 00 00 e8 55 48 a4 ff 5d c3 0f 1f 00 66 66 66 66 90 55 48 89 e5
+fa 66 66 90 66 66 90 b8 00 00 02 00 <f0> 0f c1 07 89 c2 c1 ea 10 66 39 c2 75 02 5d c3 83 e2 fe 0f
+b7
RIP [<ffffffff8162e9e5>] _raw_spin_lock_irq+0x15/0x60
RSP <ffff88085403fdf0>
The request_queue locking was added because blkcg_gq->refcnt is an int
protected with the queue lock and __blkg_release_rcu() needs to put
the parent. Let's fix it by making blkcg_gq->refcnt an atomic_t and
dropping queue locking in the function.
Given the general heavy weight of the current request_queue and blkcg
operations, this is unlikely to cause any noticeable overhead.
Moreover, blkcg_gq->refcnt is likely to be converted to percpu_ref in
the near future, so whatever (most likely negligible) overhead it may
add is temporary.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Joe Lawrence <joe.lawrence@stratus.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Link: http://lkml.kernel.org/g/alpine.DEB.2.02.1406081816540.17948@jlaw-desktop.mno.stratus.com
Cc: stable@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
2014-06-19 23:42:57 +02:00
|
|
|
atomic_set(&blkg->refcnt, 1);
|
2012-03-05 22:15:14 +01:00
|
|
|
|
blkcg: implement per-blkg request allocation
Currently, request_queue has one request_list to allocate requests
from regardless of blkcg of the IO being issued. When the unified
request pool is used up, cfq proportional IO limits become meaningless
- whoever grabs the next request being freed wins the race regardless
of the configured weights.
This can be easily demonstrated by creating a blkio cgroup w/ very low
weight, put a program which can issue a lot of random direct IOs there
and running a sequential IO from a different cgroup. As soon as the
request pool is used up, the sequential IO bandwidth crashes.
This patch implements per-blkg request_list. Each blkg has its own
request_list and any IO allocates its request from the matching blkg
making blkcgs completely isolated in terms of request allocation.
* Root blkcg uses the request_list embedded in each request_queue,
which was renamed to @q->root_rl from @q->rq. While making blkcg rl
handling a bit harier, this enables avoiding most overhead for root
blkcg.
* Queue fullness is properly per request_list but bdi isn't blkcg
aware yet, so congestion state currently just follows the root
blkcg. As writeback isn't aware of blkcg yet, this works okay for
async congestion but readahead may get the wrong signals. It's
better than blkcg completely collapsing with shared request_list but
needs to be improved with future changes.
* After this change, each block cgroup gets a full request pool making
resource consumption of each cgroup higher. This makes allowing
non-root users to create cgroups less desirable; however, note that
allowing non-root users to directly manage cgroups is already
severely broken regardless of this patch - each block cgroup
consumes kernel memory and skews IO weight (IO weights are not
hierarchical).
v2: queue-sysfs.txt updated and patch description udpated as suggested
by Vivek.
v3: blk_get_rl() wasn't checking error return from
blkg_lookup_create() and may cause oops on lookup failure. Fix it
by falling back to root_rl on blkg lookup failures. This problem
was spotted by Rakesh Iyer <rni@google.com>.
v4: Updated to accomodate 458f27a982 "block: Avoid missed wakeup in
request waitqueue". blk_drain_queue() now wakes up waiters on all
blkg->rl on the target queue.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2012-06-27 00:05:44 +02:00
|
|
|
/* root blkg uses @q->root_rl, init rl only for !root blkgs */
|
|
|
|
if (blkcg != &blkcg_root) {
|
|
|
|
if (blk_init_rl(&blkg->rl, q, gfp_mask))
|
|
|
|
goto err_free;
|
|
|
|
blkg->rl.blkg = blkg;
|
|
|
|
}
|
|
|
|
|
2012-04-13 22:11:28 +02:00
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++) {
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_policy *pol = blkcg_policy[i];
|
2012-03-05 22:15:20 +01:00
|
|
|
struct blkg_policy_data *pd;
|
2012-03-05 22:15:14 +01:00
|
|
|
|
2012-04-13 22:11:33 +02:00
|
|
|
if (!blkcg_policy_enabled(q, pol))
|
2012-03-05 22:15:20 +01:00
|
|
|
continue;
|
|
|
|
|
|
|
|
/* alloc per-policy data and attach it to blkg */
|
2015-08-18 23:55:11 +02:00
|
|
|
pd = pol->pd_alloc_fn(gfp_mask, q->node);
|
blkcg: implement per-blkg request allocation
Currently, request_queue has one request_list to allocate requests
from regardless of blkcg of the IO being issued. When the unified
request pool is used up, cfq proportional IO limits become meaningless
- whoever grabs the next request being freed wins the race regardless
of the configured weights.
This can be easily demonstrated by creating a blkio cgroup w/ very low
weight, put a program which can issue a lot of random direct IOs there
and running a sequential IO from a different cgroup. As soon as the
request pool is used up, the sequential IO bandwidth crashes.
This patch implements per-blkg request_list. Each blkg has its own
request_list and any IO allocates its request from the matching blkg
making blkcgs completely isolated in terms of request allocation.
* Root blkcg uses the request_list embedded in each request_queue,
which was renamed to @q->root_rl from @q->rq. While making blkcg rl
handling a bit harier, this enables avoiding most overhead for root
blkcg.
* Queue fullness is properly per request_list but bdi isn't blkcg
aware yet, so congestion state currently just follows the root
blkcg. As writeback isn't aware of blkcg yet, this works okay for
async congestion but readahead may get the wrong signals. It's
better than blkcg completely collapsing with shared request_list but
needs to be improved with future changes.
* After this change, each block cgroup gets a full request pool making
resource consumption of each cgroup higher. This makes allowing
non-root users to create cgroups less desirable; however, note that
allowing non-root users to directly manage cgroups is already
severely broken regardless of this patch - each block cgroup
consumes kernel memory and skews IO weight (IO weights are not
hierarchical).
v2: queue-sysfs.txt updated and patch description udpated as suggested
by Vivek.
v3: blk_get_rl() wasn't checking error return from
blkg_lookup_create() and may cause oops on lookup failure. Fix it
by falling back to root_rl on blkg lookup failures. This problem
was spotted by Rakesh Iyer <rni@google.com>.
v4: Updated to accomodate 458f27a982 "block: Avoid missed wakeup in
request waitqueue". blk_drain_queue() now wakes up waiters on all
blkg->rl on the target queue.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2012-06-27 00:05:44 +02:00
|
|
|
if (!pd)
|
|
|
|
goto err_free;
|
2012-03-05 22:15:16 +01:00
|
|
|
|
2012-03-05 22:15:20 +01:00
|
|
|
blkg->pd[i] = pd;
|
|
|
|
pd->blkg = blkg;
|
2013-01-09 17:05:12 +01:00
|
|
|
pd->plid = i;
|
2012-03-05 22:15:20 +01:00
|
|
|
}
|
|
|
|
|
2012-03-05 22:15:14 +01:00
|
|
|
return blkg;
|
blkcg: implement per-blkg request allocation
Currently, request_queue has one request_list to allocate requests
from regardless of blkcg of the IO being issued. When the unified
request pool is used up, cfq proportional IO limits become meaningless
- whoever grabs the next request being freed wins the race regardless
of the configured weights.
This can be easily demonstrated by creating a blkio cgroup w/ very low
weight, put a program which can issue a lot of random direct IOs there
and running a sequential IO from a different cgroup. As soon as the
request pool is used up, the sequential IO bandwidth crashes.
This patch implements per-blkg request_list. Each blkg has its own
request_list and any IO allocates its request from the matching blkg
making blkcgs completely isolated in terms of request allocation.
* Root blkcg uses the request_list embedded in each request_queue,
which was renamed to @q->root_rl from @q->rq. While making blkcg rl
handling a bit harier, this enables avoiding most overhead for root
blkcg.
* Queue fullness is properly per request_list but bdi isn't blkcg
aware yet, so congestion state currently just follows the root
blkcg. As writeback isn't aware of blkcg yet, this works okay for
async congestion but readahead may get the wrong signals. It's
better than blkcg completely collapsing with shared request_list but
needs to be improved with future changes.
* After this change, each block cgroup gets a full request pool making
resource consumption of each cgroup higher. This makes allowing
non-root users to create cgroups less desirable; however, note that
allowing non-root users to directly manage cgroups is already
severely broken regardless of this patch - each block cgroup
consumes kernel memory and skews IO weight (IO weights are not
hierarchical).
v2: queue-sysfs.txt updated and patch description udpated as suggested
by Vivek.
v3: blk_get_rl() wasn't checking error return from
blkg_lookup_create() and may cause oops on lookup failure. Fix it
by falling back to root_rl on blkg lookup failures. This problem
was spotted by Rakesh Iyer <rni@google.com>.
v4: Updated to accomodate 458f27a982 "block: Avoid missed wakeup in
request waitqueue". blk_drain_queue() now wakes up waiters on all
blkg->rl on the target queue.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2012-06-27 00:05:44 +02:00
|
|
|
|
|
|
|
err_free:
|
|
|
|
blkg_free(blkg);
|
|
|
|
return NULL;
|
2012-03-05 22:15:14 +01:00
|
|
|
}
|
|
|
|
|
2015-08-18 23:55:17 +02:00
|
|
|
struct blkcg_gq *blkg_lookup_slowpath(struct blkcg *blkcg,
|
|
|
|
struct request_queue *q, bool update_hint)
|
2012-04-13 23:50:53 +02:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2012-04-13 23:50:53 +02:00
|
|
|
|
2012-04-20 01:29:24 +02:00
|
|
|
/*
|
2013-01-09 17:05:10 +01:00
|
|
|
* Hint didn't match. Look up from the radix tree. Note that the
|
|
|
|
* hint can only be updated under queue_lock as otherwise @blkg
|
|
|
|
* could have already been removed from blkg_tree. The caller is
|
|
|
|
* responsible for grabbing queue_lock if @update_hint.
|
2012-04-20 01:29:24 +02:00
|
|
|
*/
|
|
|
|
blkg = radix_tree_lookup(&blkcg->blkg_tree, q->id);
|
2013-01-09 17:05:10 +01:00
|
|
|
if (blkg && blkg->q == q) {
|
|
|
|
if (update_hint) {
|
|
|
|
lockdep_assert_held(q->queue_lock);
|
|
|
|
rcu_assign_pointer(blkcg->blkg_hint, blkg);
|
|
|
|
}
|
2012-04-20 01:29:24 +02:00
|
|
|
return blkg;
|
2013-01-09 17:05:10 +01:00
|
|
|
}
|
2012-04-20 01:29:24 +02:00
|
|
|
|
2012-04-13 23:50:53 +02:00
|
|
|
return NULL;
|
|
|
|
}
|
2015-08-18 23:55:20 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkg_lookup_slowpath);
|
2012-04-13 23:50:53 +02:00
|
|
|
|
2012-06-05 05:40:52 +02:00
|
|
|
/*
|
|
|
|
* If @new_blkg is %NULL, this function tries to allocate a new one as
|
2015-08-18 23:55:01 +02:00
|
|
|
* necessary using %GFP_NOWAIT. @new_blkg is always consumed on return.
|
2012-06-05 05:40:52 +02:00
|
|
|
*/
|
2013-01-09 17:05:10 +01:00
|
|
|
static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
|
|
|
|
struct request_queue *q,
|
|
|
|
struct blkcg_gq *new_blkg)
|
2011-05-19 21:38:28 +02:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2015-05-22 23:13:38 +02:00
|
|
|
struct bdi_writeback_congested *wb_congested;
|
2013-01-09 17:05:12 +01:00
|
|
|
int i, ret;
|
2011-05-19 21:38:28 +02:00
|
|
|
|
2012-03-05 22:15:06 +01:00
|
|
|
WARN_ON_ONCE(!rcu_read_lock_held());
|
|
|
|
lockdep_assert_held(q->queue_lock);
|
|
|
|
|
2012-03-05 22:15:11 +01:00
|
|
|
/* blkg holds a reference to blkcg */
|
2014-05-13 18:11:01 +02:00
|
|
|
if (!css_tryget_online(&blkcg->css)) {
|
2015-08-18 23:55:28 +02:00
|
|
|
ret = -ENODEV;
|
2013-01-09 17:05:10 +01:00
|
|
|
goto err_free_blkg;
|
2012-06-05 05:40:52 +02:00
|
|
|
}
|
2012-03-05 22:15:06 +01:00
|
|
|
|
2015-05-22 23:13:38 +02:00
|
|
|
wb_congested = wb_congested_get_create(&q->backing_dev_info,
|
2016-11-22 00:03:32 +01:00
|
|
|
blkcg->css.id,
|
|
|
|
GFP_NOWAIT | __GFP_NOWARN);
|
2015-05-22 23:13:38 +02:00
|
|
|
if (!wb_congested) {
|
|
|
|
ret = -ENOMEM;
|
|
|
|
goto err_put_css;
|
|
|
|
}
|
|
|
|
|
2012-04-20 01:29:23 +02:00
|
|
|
/* allocate */
|
2012-06-05 05:40:52 +02:00
|
|
|
if (!new_blkg) {
|
2016-11-22 00:03:32 +01:00
|
|
|
new_blkg = blkg_alloc(blkcg, q, GFP_NOWAIT | __GFP_NOWARN);
|
2012-06-05 05:40:52 +02:00
|
|
|
if (unlikely(!new_blkg)) {
|
2013-01-09 17:05:10 +01:00
|
|
|
ret = -ENOMEM;
|
2015-05-22 23:13:38 +02:00
|
|
|
goto err_put_congested;
|
2012-06-05 05:40:52 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
blkg = new_blkg;
|
2015-05-22 23:13:38 +02:00
|
|
|
blkg->wb_congested = wb_congested;
|
2012-03-05 22:15:06 +01:00
|
|
|
|
2013-05-14 22:52:31 +02:00
|
|
|
/* link parent */
|
2013-01-09 17:05:10 +01:00
|
|
|
if (blkcg_parent(blkcg)) {
|
|
|
|
blkg->parent = __blkg_lookup(blkcg_parent(blkcg), q, false);
|
|
|
|
if (WARN_ON_ONCE(!blkg->parent)) {
|
2015-08-18 23:55:28 +02:00
|
|
|
ret = -ENODEV;
|
2015-05-22 23:13:38 +02:00
|
|
|
goto err_put_congested;
|
2013-01-09 17:05:10 +01:00
|
|
|
}
|
|
|
|
blkg_get(blkg->parent);
|
|
|
|
}
|
|
|
|
|
2013-05-14 22:52:31 +02:00
|
|
|
/* invoke per-policy init */
|
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++) {
|
|
|
|
struct blkcg_policy *pol = blkcg_policy[i];
|
|
|
|
|
|
|
|
if (blkg->pd[i] && pol->pd_init_fn)
|
2015-08-18 23:55:14 +02:00
|
|
|
pol->pd_init_fn(blkg->pd[i]);
|
2013-05-14 22:52:31 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/* insert */
|
2012-03-05 22:15:06 +01:00
|
|
|
spin_lock(&blkcg->lock);
|
2012-04-20 01:29:24 +02:00
|
|
|
ret = radix_tree_insert(&blkcg->blkg_tree, q->id, blkg);
|
|
|
|
if (likely(!ret)) {
|
|
|
|
hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
|
|
|
|
list_add(&blkg->q_node, &q->blkg_list);
|
2013-01-09 17:05:12 +01:00
|
|
|
|
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++) {
|
|
|
|
struct blkcg_policy *pol = blkcg_policy[i];
|
|
|
|
|
|
|
|
if (blkg->pd[i] && pol->pd_online_fn)
|
2015-08-18 23:55:14 +02:00
|
|
|
pol->pd_online_fn(blkg->pd[i]);
|
2013-01-09 17:05:12 +01:00
|
|
|
}
|
2012-04-20 01:29:24 +02:00
|
|
|
}
|
2013-01-09 17:05:12 +01:00
|
|
|
blkg->online = true;
|
2012-03-05 22:15:06 +01:00
|
|
|
spin_unlock(&blkcg->lock);
|
2012-04-20 01:29:23 +02:00
|
|
|
|
2015-05-22 23:13:19 +02:00
|
|
|
if (!ret)
|
2012-04-20 01:29:24 +02:00
|
|
|
return blkg;
|
2012-06-05 05:40:52 +02:00
|
|
|
|
2013-01-09 17:05:10 +01:00
|
|
|
/* @blkg failed fully initialized, use the usual release path */
|
|
|
|
blkg_put(blkg);
|
|
|
|
return ERR_PTR(ret);
|
|
|
|
|
2015-05-22 23:13:38 +02:00
|
|
|
err_put_congested:
|
|
|
|
wb_congested_put(wb_congested);
|
2013-01-09 17:05:10 +01:00
|
|
|
err_put_css:
|
2012-04-20 01:29:23 +02:00
|
|
|
css_put(&blkcg->css);
|
2013-01-09 17:05:10 +01:00
|
|
|
err_free_blkg:
|
2012-06-05 05:40:52 +02:00
|
|
|
blkg_free(new_blkg);
|
2013-01-09 17:05:10 +01:00
|
|
|
return ERR_PTR(ret);
|
2009-12-03 18:59:42 +01:00
|
|
|
}
|
2012-04-13 22:11:34 +02:00
|
|
|
|
2013-01-09 17:05:10 +01:00
|
|
|
/**
|
|
|
|
* blkg_lookup_create - lookup blkg, try to create one if not there
|
|
|
|
* @blkcg: blkcg of interest
|
|
|
|
* @q: request_queue of interest
|
|
|
|
*
|
|
|
|
* Lookup blkg for the @blkcg - @q pair. If it doesn't exist, try to
|
2013-01-09 17:05:10 +01:00
|
|
|
* create one. blkg creation is performed recursively from blkcg_root such
|
|
|
|
* that all non-root blkg's have access to the parent blkg. This function
|
|
|
|
* should be called under RCU read lock and @q->queue_lock.
|
2013-01-09 17:05:10 +01:00
|
|
|
*
|
|
|
|
* Returns pointer to the looked up or created blkg on success, ERR_PTR()
|
|
|
|
* value on error. If @q is dead, returns ERR_PTR(-EINVAL). If @q is not
|
|
|
|
* dead and bypassing, returns ERR_PTR(-EBUSY).
|
|
|
|
*/
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
|
|
|
|
struct request_queue *q)
|
2012-04-13 22:11:34 +02:00
|
|
|
{
|
2013-01-09 17:05:10 +01:00
|
|
|
struct blkcg_gq *blkg;
|
|
|
|
|
|
|
|
WARN_ON_ONCE(!rcu_read_lock_held());
|
|
|
|
lockdep_assert_held(q->queue_lock);
|
|
|
|
|
2012-04-13 22:11:34 +02:00
|
|
|
/*
|
|
|
|
* This could be the first entry point of blkcg implementation and
|
|
|
|
* we shouldn't allow anything to go through for a bypassing queue.
|
|
|
|
*/
|
|
|
|
if (unlikely(blk_queue_bypass(q)))
|
2015-08-18 23:55:28 +02:00
|
|
|
return ERR_PTR(blk_queue_dying(q) ? -ENODEV : -EBUSY);
|
2013-01-09 17:05:10 +01:00
|
|
|
|
|
|
|
blkg = __blkg_lookup(blkcg, q, true);
|
|
|
|
if (blkg)
|
|
|
|
return blkg;
|
|
|
|
|
2013-01-09 17:05:10 +01:00
|
|
|
/*
|
|
|
|
* Create blkgs walking down from blkcg_root to @blkcg, so that all
|
|
|
|
* non-root blkgs have access to their parents.
|
|
|
|
*/
|
|
|
|
while (true) {
|
|
|
|
struct blkcg *pos = blkcg;
|
|
|
|
struct blkcg *parent = blkcg_parent(blkcg);
|
|
|
|
|
|
|
|
while (parent && !__blkg_lookup(parent, q, false)) {
|
|
|
|
pos = parent;
|
|
|
|
parent = blkcg_parent(parent);
|
|
|
|
}
|
|
|
|
|
|
|
|
blkg = blkg_create(pos, q, NULL);
|
|
|
|
if (pos == blkcg || IS_ERR(blkg))
|
|
|
|
return blkg;
|
|
|
|
}
|
2012-04-13 22:11:34 +02:00
|
|
|
}
|
2009-12-03 18:59:42 +01:00
|
|
|
|
2012-04-16 22:57:25 +02:00
|
|
|
static void blkg_destroy(struct blkcg_gq *blkg)
|
2012-03-05 22:15:19 +01:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg *blkcg = blkg->blkcg;
|
2015-08-18 23:55:24 +02:00
|
|
|
struct blkcg_gq *parent = blkg->parent;
|
2013-01-09 17:05:12 +01:00
|
|
|
int i;
|
2012-03-05 22:15:19 +01:00
|
|
|
|
2012-06-05 13:36:44 +02:00
|
|
|
lockdep_assert_held(blkg->q->queue_lock);
|
2012-03-05 22:15:21 +01:00
|
|
|
lockdep_assert_held(&blkcg->lock);
|
2012-03-05 22:15:19 +01:00
|
|
|
|
|
|
|
/* Something wrong if we are trying to remove same group twice */
|
2012-03-05 22:15:20 +01:00
|
|
|
WARN_ON_ONCE(list_empty(&blkg->q_node));
|
2012-03-05 22:15:21 +01:00
|
|
|
WARN_ON_ONCE(hlist_unhashed(&blkg->blkcg_node));
|
2012-04-20 01:29:24 +02:00
|
|
|
|
2013-01-09 17:05:12 +01:00
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++) {
|
|
|
|
struct blkcg_policy *pol = blkcg_policy[i];
|
|
|
|
|
|
|
|
if (blkg->pd[i] && pol->pd_offline_fn)
|
2015-08-18 23:55:14 +02:00
|
|
|
pol->pd_offline_fn(blkg->pd[i]);
|
2013-01-09 17:05:12 +01:00
|
|
|
}
|
2015-08-18 23:55:24 +02:00
|
|
|
|
|
|
|
if (parent) {
|
|
|
|
blkg_rwstat_add_aux(&parent->stat_bytes, &blkg->stat_bytes);
|
|
|
|
blkg_rwstat_add_aux(&parent->stat_ios, &blkg->stat_ios);
|
|
|
|
}
|
|
|
|
|
2013-01-09 17:05:12 +01:00
|
|
|
blkg->online = false;
|
|
|
|
|
2012-04-20 01:29:24 +02:00
|
|
|
radix_tree_delete(&blkcg->blkg_tree, blkg->q->id);
|
2012-03-05 22:15:20 +01:00
|
|
|
list_del_init(&blkg->q_node);
|
2012-03-05 22:15:21 +01:00
|
|
|
hlist_del_init_rcu(&blkg->blkcg_node);
|
2012-03-05 22:15:19 +01:00
|
|
|
|
2012-04-20 01:29:24 +02:00
|
|
|
/*
|
|
|
|
* Both setting lookup hint to and clearing it from @blkg are done
|
|
|
|
* under queue_lock. If it's not pointing to @blkg now, it never
|
|
|
|
* will. Hint assignment itself can race safely.
|
|
|
|
*/
|
2014-02-17 22:35:57 +01:00
|
|
|
if (rcu_access_pointer(blkcg->blkg_hint) == blkg)
|
2012-04-20 01:29:24 +02:00
|
|
|
rcu_assign_pointer(blkcg->blkg_hint, NULL);
|
|
|
|
|
2012-03-05 22:15:19 +01:00
|
|
|
/*
|
|
|
|
* Put the reference taken at the time of creation so that when all
|
|
|
|
* queues are gone, group can be destroyed.
|
|
|
|
*/
|
|
|
|
blkg_put(blkg);
|
|
|
|
}
|
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
/**
|
|
|
|
* blkg_destroy_all - destroy all blkgs associated with a request_queue
|
|
|
|
* @q: request_queue of interest
|
|
|
|
*
|
2012-04-13 22:11:34 +02:00
|
|
|
* Destroy all blkgs associated with @q.
|
2012-03-05 22:15:21 +01:00
|
|
|
*/
|
2012-04-13 22:11:34 +02:00
|
|
|
static void blkg_destroy_all(struct request_queue *q)
|
2012-03-05 22:15:00 +01:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg, *n;
|
2012-03-05 22:15:00 +01:00
|
|
|
|
2012-04-13 22:11:35 +02:00
|
|
|
lockdep_assert_held(q->queue_lock);
|
2012-03-05 22:15:00 +01:00
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
list_for_each_entry_safe(blkg, n, &q->blkg_list, q_node) {
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg *blkcg = blkg->blkcg;
|
2012-03-05 22:15:00 +01:00
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
spin_lock(&blkcg->lock);
|
|
|
|
blkg_destroy(blkg);
|
|
|
|
spin_unlock(&blkcg->lock);
|
2012-03-05 22:15:00 +01:00
|
|
|
}
|
2015-09-05 21:47:36 +02:00
|
|
|
|
|
|
|
q->root_blkg = NULL;
|
|
|
|
q->root_rl.blkg = NULL;
|
2012-03-05 22:15:00 +01:00
|
|
|
}
|
|
|
|
|
2013-05-14 22:52:31 +02:00
|
|
|
/*
|
|
|
|
* A group is RCU protected, but having an rcu lock does not mean that one
|
|
|
|
* can access all the fields of blkg and assume these are valid. For
|
|
|
|
* example, don't try to follow throtl_data and request queue links.
|
|
|
|
*
|
|
|
|
* Having a reference to blkg under an rcu allows accesses to only values
|
|
|
|
* local to groups like group stats and group rate limits.
|
|
|
|
*/
|
|
|
|
void __blkg_release_rcu(struct rcu_head *rcu_head)
|
2012-03-05 22:15:15 +01:00
|
|
|
{
|
2013-05-14 22:52:31 +02:00
|
|
|
struct blkcg_gq *blkg = container_of(rcu_head, struct blkcg_gq, rcu_head);
|
2013-05-14 22:52:31 +02:00
|
|
|
|
2013-01-09 17:05:10 +01:00
|
|
|
/* release the blkcg and parent blkg refs this blkg has been holding */
|
2012-03-05 22:15:15 +01:00
|
|
|
css_put(&blkg->blkcg->css);
|
blkcg: fix use-after-free in __blkg_release_rcu() by making blkcg_gq refcnt an atomic_t
Hello,
So, this patch should do. Joe, Vivek, can one of you guys please
verify that the oops goes away with this patch?
Jens, the original thread can be read at
http://thread.gmane.org/gmane.linux.kernel/1720729
The fix converts blkg->refcnt from int to atomic_t. It does some
overhead but it should be minute compared to everything else which is
going on and the involved cacheline bouncing, so I think it's highly
unlikely to cause any noticeable difference. Also, the refcnt in
question should be converted to a perpcu_ref for blk-mq anyway, so the
atomic_t is likely to go away pretty soon anyway.
Thanks.
------- 8< -------
__blkg_release_rcu() may be invoked after the associated request_queue
is released with a RCU grace period inbetween. As such, the function
and callbacks invoked from it must not dereference the associated
request_queue. This is clearly indicated in the comment above the
function.
Unfortunately, while trying to fix a different issue, 2a4fd070ee85
("blkcg: move bulk of blkcg_gq release operations to the RCU
callback") ignored this and added [un]locking of @blkg->q->queue_lock
to __blkg_release_rcu(). This of course can cause oops as the
request_queue may be long gone by the time this code gets executed.
general protection fault: 0000 [#1] SMP
CPU: 21 PID: 30 Comm: rcuos/21 Not tainted 3.15.0 #1
Hardware name: Stratus ftServer 6400/G7LAZ, BIOS BIOS Version 6.3:57 12/25/2013
task: ffff880854021de0 ti: ffff88085403c000 task.ti: ffff88085403c000
RIP: 0010:[<ffffffff8162e9e5>] [<ffffffff8162e9e5>] _raw_spin_lock_irq+0x15/0x60
RSP: 0018:ffff88085403fdf0 EFLAGS: 00010086
RAX: 0000000000020000 RBX: 0000000000000010 RCX: 0000000000000000
RDX: 000060ef80008248 RSI: 0000000000000286 RDI: 6b6b6b6b6b6b6b6b
RBP: ffff88085403fdf0 R08: 0000000000000286 R09: 0000000000009f39
R10: 0000000000020001 R11: 0000000000020001 R12: ffff88103c17a130
R13: ffff88103c17a080 R14: 0000000000000000 R15: 0000000000000000
FS: 0000000000000000(0000) GS:ffff88107fca0000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000006e5ab8 CR3: 000000000193d000 CR4: 00000000000407e0
Stack:
ffff88085403fe18 ffffffff812cbfc2 ffff88103c17a130 0000000000000000
ffff88103c17a130 ffff88085403fec0 ffffffff810d1d28 ffff880854021de0
ffff880854021de0 ffff88107fcaec58 ffff88085403fe80 ffff88107fcaec30
Call Trace:
[<ffffffff812cbfc2>] __blkg_release_rcu+0x72/0x150
[<ffffffff810d1d28>] rcu_nocb_kthread+0x1e8/0x300
[<ffffffff81091d81>] kthread+0xe1/0x100
[<ffffffff8163813c>] ret_from_fork+0x7c/0xb0
Code: ff 47 04 48 8b 7d 08 be 00 02 00 00 e8 55 48 a4 ff 5d c3 0f 1f 00 66 66 66 66 90 55 48 89 e5
+fa 66 66 90 66 66 90 b8 00 00 02 00 <f0> 0f c1 07 89 c2 c1 ea 10 66 39 c2 75 02 5d c3 83 e2 fe 0f
+b7
RIP [<ffffffff8162e9e5>] _raw_spin_lock_irq+0x15/0x60
RSP <ffff88085403fdf0>
The request_queue locking was added because blkcg_gq->refcnt is an int
protected with the queue lock and __blkg_release_rcu() needs to put
the parent. Let's fix it by making blkcg_gq->refcnt an atomic_t and
dropping queue locking in the function.
Given the general heavy weight of the current request_queue and blkcg
operations, this is unlikely to cause any noticeable overhead.
Moreover, blkcg_gq->refcnt is likely to be converted to percpu_ref in
the near future, so whatever (most likely negligible) overhead it may
add is temporary.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Joe Lawrence <joe.lawrence@stratus.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Link: http://lkml.kernel.org/g/alpine.DEB.2.02.1406081816540.17948@jlaw-desktop.mno.stratus.com
Cc: stable@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
2014-06-19 23:42:57 +02:00
|
|
|
if (blkg->parent)
|
2013-01-09 17:05:10 +01:00
|
|
|
blkg_put(blkg->parent);
|
2012-03-05 22:15:15 +01:00
|
|
|
|
2015-05-22 23:13:38 +02:00
|
|
|
wb_congested_put(blkg->wb_congested);
|
|
|
|
|
2013-05-14 22:52:31 +02:00
|
|
|
blkg_free(blkg);
|
2012-03-05 22:15:15 +01:00
|
|
|
}
|
2013-05-14 22:52:31 +02:00
|
|
|
EXPORT_SYMBOL_GPL(__blkg_release_rcu);
|
2012-03-05 22:15:15 +01:00
|
|
|
|
blkcg: implement per-blkg request allocation
Currently, request_queue has one request_list to allocate requests
from regardless of blkcg of the IO being issued. When the unified
request pool is used up, cfq proportional IO limits become meaningless
- whoever grabs the next request being freed wins the race regardless
of the configured weights.
This can be easily demonstrated by creating a blkio cgroup w/ very low
weight, put a program which can issue a lot of random direct IOs there
and running a sequential IO from a different cgroup. As soon as the
request pool is used up, the sequential IO bandwidth crashes.
This patch implements per-blkg request_list. Each blkg has its own
request_list and any IO allocates its request from the matching blkg
making blkcgs completely isolated in terms of request allocation.
* Root blkcg uses the request_list embedded in each request_queue,
which was renamed to @q->root_rl from @q->rq. While making blkcg rl
handling a bit harier, this enables avoiding most overhead for root
blkcg.
* Queue fullness is properly per request_list but bdi isn't blkcg
aware yet, so congestion state currently just follows the root
blkcg. As writeback isn't aware of blkcg yet, this works okay for
async congestion but readahead may get the wrong signals. It's
better than blkcg completely collapsing with shared request_list but
needs to be improved with future changes.
* After this change, each block cgroup gets a full request pool making
resource consumption of each cgroup higher. This makes allowing
non-root users to create cgroups less desirable; however, note that
allowing non-root users to directly manage cgroups is already
severely broken regardless of this patch - each block cgroup
consumes kernel memory and skews IO weight (IO weights are not
hierarchical).
v2: queue-sysfs.txt updated and patch description udpated as suggested
by Vivek.
v3: blk_get_rl() wasn't checking error return from
blkg_lookup_create() and may cause oops on lookup failure. Fix it
by falling back to root_rl on blkg lookup failures. This problem
was spotted by Rakesh Iyer <rni@google.com>.
v4: Updated to accomodate 458f27a982 "block: Avoid missed wakeup in
request waitqueue". blk_drain_queue() now wakes up waiters on all
blkg->rl on the target queue.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2012-06-27 00:05:44 +02:00
|
|
|
/*
|
|
|
|
* The next function used by blk_queue_for_each_rl(). It's a bit tricky
|
|
|
|
* because the root blkg uses @q->root_rl instead of its own rl.
|
|
|
|
*/
|
|
|
|
struct request_list *__blk_queue_next_rl(struct request_list *rl,
|
|
|
|
struct request_queue *q)
|
|
|
|
{
|
|
|
|
struct list_head *ent;
|
|
|
|
struct blkcg_gq *blkg;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Determine the current blkg list_head. The first entry is
|
|
|
|
* root_rl which is off @q->blkg_list and mapped to the head.
|
|
|
|
*/
|
|
|
|
if (rl == &q->root_rl) {
|
|
|
|
ent = &q->blkg_list;
|
2012-10-22 03:15:37 +02:00
|
|
|
/* There are no more block groups, hence no request lists */
|
|
|
|
if (list_empty(ent))
|
|
|
|
return NULL;
|
blkcg: implement per-blkg request allocation
Currently, request_queue has one request_list to allocate requests
from regardless of blkcg of the IO being issued. When the unified
request pool is used up, cfq proportional IO limits become meaningless
- whoever grabs the next request being freed wins the race regardless
of the configured weights.
This can be easily demonstrated by creating a blkio cgroup w/ very low
weight, put a program which can issue a lot of random direct IOs there
and running a sequential IO from a different cgroup. As soon as the
request pool is used up, the sequential IO bandwidth crashes.
This patch implements per-blkg request_list. Each blkg has its own
request_list and any IO allocates its request from the matching blkg
making blkcgs completely isolated in terms of request allocation.
* Root blkcg uses the request_list embedded in each request_queue,
which was renamed to @q->root_rl from @q->rq. While making blkcg rl
handling a bit harier, this enables avoiding most overhead for root
blkcg.
* Queue fullness is properly per request_list but bdi isn't blkcg
aware yet, so congestion state currently just follows the root
blkcg. As writeback isn't aware of blkcg yet, this works okay for
async congestion but readahead may get the wrong signals. It's
better than blkcg completely collapsing with shared request_list but
needs to be improved with future changes.
* After this change, each block cgroup gets a full request pool making
resource consumption of each cgroup higher. This makes allowing
non-root users to create cgroups less desirable; however, note that
allowing non-root users to directly manage cgroups is already
severely broken regardless of this patch - each block cgroup
consumes kernel memory and skews IO weight (IO weights are not
hierarchical).
v2: queue-sysfs.txt updated and patch description udpated as suggested
by Vivek.
v3: blk_get_rl() wasn't checking error return from
blkg_lookup_create() and may cause oops on lookup failure. Fix it
by falling back to root_rl on blkg lookup failures. This problem
was spotted by Rakesh Iyer <rni@google.com>.
v4: Updated to accomodate 458f27a982 "block: Avoid missed wakeup in
request waitqueue". blk_drain_queue() now wakes up waiters on all
blkg->rl on the target queue.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2012-06-27 00:05:44 +02:00
|
|
|
} else {
|
|
|
|
blkg = container_of(rl, struct blkcg_gq, rl);
|
|
|
|
ent = &blkg->q_node;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* walk to the next list_head, skip root blkcg */
|
|
|
|
ent = ent->next;
|
|
|
|
if (ent == &q->root_blkg->q_node)
|
|
|
|
ent = ent->next;
|
|
|
|
if (ent == &q->blkg_list)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
blkg = container_of(ent, struct blkcg_gq, q_node);
|
|
|
|
return &blkg->rl;
|
|
|
|
}
|
|
|
|
|
2013-08-09 02:11:24 +02:00
|
|
|
static int blkcg_reset_stats(struct cgroup_subsys_state *css,
|
|
|
|
struct cftype *cftype, u64 val)
|
2010-04-02 00:01:24 +02:00
|
|
|
{
|
2013-08-09 02:11:24 +02:00
|
|
|
struct blkcg *blkcg = css_to_blkcg(css);
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2012-04-13 22:11:26 +02:00
|
|
|
int i;
|
2010-04-02 00:01:24 +02:00
|
|
|
|
2015-07-09 22:39:47 +02:00
|
|
|
mutex_lock(&blkcg_pol_mutex);
|
2010-04-02 00:01:24 +02:00
|
|
|
spin_lock_irq(&blkcg->lock);
|
2012-03-08 19:53:58 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Note that stat reset is racy - it doesn't synchronize against
|
|
|
|
* stat updates. This is a debug feature which shouldn't exist
|
|
|
|
* anyway. If you get hit by a race, retry.
|
|
|
|
*/
|
hlist: drop the node parameter from iterators
I'm not sure why, but the hlist for each entry iterators were conceived
list_for_each_entry(pos, head, member)
The hlist ones were greedy and wanted an extra parameter:
hlist_for_each_entry(tpos, pos, head, member)
Why did they need an extra pos parameter? I'm not quite sure. Not only
they don't really need it, it also prevents the iterator from looking
exactly like the list iterator, which is unfortunate.
Besides the semantic patch, there was some manual work required:
- Fix up the actual hlist iterators in linux/list.h
- Fix up the declaration of other iterators based on the hlist ones.
- A very small amount of places were using the 'node' parameter, this
was modified to use 'obj->member' instead.
- Coccinelle didn't handle the hlist_for_each_entry_safe iterator
properly, so those had to be fixed up manually.
The semantic patch which is mostly the work of Peter Senna Tschudin is here:
@@
iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host;
type T;
expression a,c,d,e;
identifier b;
statement S;
@@
-T b;
<+... when != b
(
hlist_for_each_entry(a,
- b,
c, d) S
|
hlist_for_each_entry_continue(a,
- b,
c) S
|
hlist_for_each_entry_from(a,
- b,
c) S
|
hlist_for_each_entry_rcu(a,
- b,
c, d) S
|
hlist_for_each_entry_rcu_bh(a,
- b,
c, d) S
|
hlist_for_each_entry_continue_rcu_bh(a,
- b,
c) S
|
for_each_busy_worker(a, c,
- b,
d) S
|
ax25_uid_for_each(a,
- b,
c) S
|
ax25_for_each(a,
- b,
c) S
|
inet_bind_bucket_for_each(a,
- b,
c) S
|
sctp_for_each_hentry(a,
- b,
c) S
|
sk_for_each(a,
- b,
c) S
|
sk_for_each_rcu(a,
- b,
c) S
|
sk_for_each_from
-(a, b)
+(a)
S
+ sk_for_each_from(a) S
|
sk_for_each_safe(a,
- b,
c, d) S
|
sk_for_each_bound(a,
- b,
c) S
|
hlist_for_each_entry_safe(a,
- b,
c, d, e) S
|
hlist_for_each_entry_continue_rcu(a,
- b,
c) S
|
nr_neigh_for_each(a,
- b,
c) S
|
nr_neigh_for_each_safe(a,
- b,
c, d) S
|
nr_node_for_each(a,
- b,
c) S
|
nr_node_for_each_safe(a,
- b,
c, d) S
|
- for_each_gfn_sp(a, c, d, b) S
+ for_each_gfn_sp(a, c, d) S
|
- for_each_gfn_indirect_valid_sp(a, c, d, b) S
+ for_each_gfn_indirect_valid_sp(a, c, d) S
|
for_each_host(a,
- b,
c) S
|
for_each_host_safe(a,
- b,
c, d) S
|
for_each_mesh_entry(a,
- b,
c, d) S
)
...+>
[akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c]
[akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c]
[akpm@linux-foundation.org: checkpatch fixes]
[akpm@linux-foundation.org: fix warnings]
[akpm@linux-foudnation.org: redo intrusive kvm changes]
Tested-by: Peter Senna Tschudin <peter.senna@gmail.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 02:06:00 +01:00
|
|
|
hlist_for_each_entry(blkg, &blkcg->blkg_list, blkcg_node) {
|
2015-08-18 23:55:24 +02:00
|
|
|
blkg_rwstat_reset(&blkg->stat_bytes);
|
|
|
|
blkg_rwstat_reset(&blkg->stat_ios);
|
|
|
|
|
2012-04-13 22:11:28 +02:00
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++) {
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_policy *pol = blkcg_policy[i];
|
2012-03-05 22:15:16 +01:00
|
|
|
|
2015-08-18 23:55:14 +02:00
|
|
|
if (blkg->pd[i] && pol->pd_reset_stats_fn)
|
|
|
|
pol->pd_reset_stats_fn(blkg->pd[i]);
|
2012-04-13 22:11:26 +02:00
|
|
|
}
|
2010-04-02 00:01:24 +02:00
|
|
|
}
|
2011-05-19 21:38:30 +02:00
|
|
|
|
2010-04-02 00:01:24 +02:00
|
|
|
spin_unlock_irq(&blkcg->lock);
|
2012-04-13 22:11:26 +02:00
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
2010-04-02 00:01:24 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-08-18 23:55:33 +02:00
|
|
|
const char *blkg_dev_name(struct blkcg_gq *blkg)
|
2010-04-02 00:01:24 +02:00
|
|
|
{
|
2012-04-01 23:38:42 +02:00
|
|
|
/* some drivers (floppy) instantiate a queue w/o disk registered */
|
|
|
|
if (blkg->q->backing_dev_info.dev)
|
|
|
|
return dev_name(blkg->q->backing_dev_info.dev);
|
|
|
|
return NULL;
|
2010-04-02 00:01:24 +02:00
|
|
|
}
|
2015-08-18 23:55:33 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkg_dev_name);
|
2010-04-02 00:01:24 +02:00
|
|
|
|
2012-04-01 23:38:42 +02:00
|
|
|
/**
|
|
|
|
* blkcg_print_blkgs - helper for printing per-blkg data
|
|
|
|
* @sf: seq_file to print to
|
|
|
|
* @blkcg: blkcg of interest
|
|
|
|
* @prfill: fill function to print out a blkg
|
|
|
|
* @pol: policy in question
|
|
|
|
* @data: data to be passed to @prfill
|
|
|
|
* @show_total: to print out sum of prfill return values or not
|
|
|
|
*
|
|
|
|
* This function invokes @prfill on each blkg of @blkcg if pd for the
|
|
|
|
* policy specified by @pol exists. @prfill is invoked with @sf, the
|
2013-01-09 17:05:13 +01:00
|
|
|
* policy data and @data and the matching queue lock held. If @show_total
|
|
|
|
* is %true, the sum of the return values from @prfill is printed with
|
|
|
|
* "Total" label at the end.
|
2012-04-01 23:38:42 +02:00
|
|
|
*
|
|
|
|
* This is to be used to construct print functions for
|
|
|
|
* cftype->read_seq_string method.
|
|
|
|
*/
|
2012-04-16 22:57:25 +02:00
|
|
|
void blkcg_print_blkgs(struct seq_file *sf, struct blkcg *blkcg,
|
2012-04-16 22:57:26 +02:00
|
|
|
u64 (*prfill)(struct seq_file *,
|
|
|
|
struct blkg_policy_data *, int),
|
2012-04-16 22:57:25 +02:00
|
|
|
const struct blkcg_policy *pol, int data,
|
2012-04-13 22:11:27 +02:00
|
|
|
bool show_total)
|
2011-05-19 21:38:28 +02:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2012-04-01 23:38:42 +02:00
|
|
|
u64 total = 0;
|
2011-05-19 21:38:28 +02:00
|
|
|
|
2013-01-09 17:05:13 +01:00
|
|
|
rcu_read_lock();
|
2013-02-28 21:52:24 +01:00
|
|
|
hlist_for_each_entry_rcu(blkg, &blkcg->blkg_list, blkcg_node) {
|
2013-01-09 17:05:13 +01:00
|
|
|
spin_lock_irq(blkg->q->queue_lock);
|
2012-04-13 22:11:33 +02:00
|
|
|
if (blkcg_policy_enabled(blkg->q, pol))
|
2012-04-16 22:57:26 +02:00
|
|
|
total += prfill(sf, blkg->pd[pol->plid], data);
|
2013-01-09 17:05:13 +01:00
|
|
|
spin_unlock_irq(blkg->q->queue_lock);
|
|
|
|
}
|
|
|
|
rcu_read_unlock();
|
2012-04-01 23:38:42 +02:00
|
|
|
|
|
|
|
if (show_total)
|
|
|
|
seq_printf(sf, "Total %llu\n", (unsigned long long)total);
|
|
|
|
}
|
2012-04-01 23:38:43 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkcg_print_blkgs);
|
2012-04-01 23:38:42 +02:00
|
|
|
|
|
|
|
/**
|
|
|
|
* __blkg_prfill_u64 - prfill helper for a single u64 value
|
|
|
|
* @sf: seq_file to print to
|
2012-04-16 22:57:26 +02:00
|
|
|
* @pd: policy private data of interest
|
2012-04-01 23:38:42 +02:00
|
|
|
* @v: value to print
|
|
|
|
*
|
2012-04-16 22:57:26 +02:00
|
|
|
* Print @v to @sf for the device assocaited with @pd.
|
2012-04-01 23:38:42 +02:00
|
|
|
*/
|
2012-04-16 22:57:26 +02:00
|
|
|
u64 __blkg_prfill_u64(struct seq_file *sf, struct blkg_policy_data *pd, u64 v)
|
2012-04-01 23:38:42 +02:00
|
|
|
{
|
2012-04-16 22:57:26 +02:00
|
|
|
const char *dname = blkg_dev_name(pd->blkg);
|
2012-04-01 23:38:42 +02:00
|
|
|
|
|
|
|
if (!dname)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
seq_printf(sf, "%s %llu\n", dname, (unsigned long long)v);
|
|
|
|
return v;
|
|
|
|
}
|
2012-04-01 23:38:43 +02:00
|
|
|
EXPORT_SYMBOL_GPL(__blkg_prfill_u64);
|
2012-04-01 23:38:42 +02:00
|
|
|
|
|
|
|
/**
|
|
|
|
* __blkg_prfill_rwstat - prfill helper for a blkg_rwstat
|
|
|
|
* @sf: seq_file to print to
|
2012-04-16 22:57:26 +02:00
|
|
|
* @pd: policy private data of interest
|
2012-04-01 23:38:42 +02:00
|
|
|
* @rwstat: rwstat to print
|
|
|
|
*
|
2012-04-16 22:57:26 +02:00
|
|
|
* Print @rwstat to @sf for the device assocaited with @pd.
|
2012-04-01 23:38:42 +02:00
|
|
|
*/
|
2012-04-16 22:57:26 +02:00
|
|
|
u64 __blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
|
2012-04-01 23:38:43 +02:00
|
|
|
const struct blkg_rwstat *rwstat)
|
2012-04-01 23:38:42 +02:00
|
|
|
{
|
|
|
|
static const char *rwstr[] = {
|
|
|
|
[BLKG_RWSTAT_READ] = "Read",
|
|
|
|
[BLKG_RWSTAT_WRITE] = "Write",
|
|
|
|
[BLKG_RWSTAT_SYNC] = "Sync",
|
|
|
|
[BLKG_RWSTAT_ASYNC] = "Async",
|
|
|
|
};
|
2012-04-16 22:57:26 +02:00
|
|
|
const char *dname = blkg_dev_name(pd->blkg);
|
2012-04-01 23:38:42 +02:00
|
|
|
u64 v;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (!dname)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
for (i = 0; i < BLKG_RWSTAT_NR; i++)
|
|
|
|
seq_printf(sf, "%s %s %llu\n", dname, rwstr[i],
|
2015-08-18 23:55:22 +02:00
|
|
|
(unsigned long long)atomic64_read(&rwstat->aux_cnt[i]));
|
2012-04-01 23:38:42 +02:00
|
|
|
|
2015-08-18 23:55:22 +02:00
|
|
|
v = atomic64_read(&rwstat->aux_cnt[BLKG_RWSTAT_READ]) +
|
|
|
|
atomic64_read(&rwstat->aux_cnt[BLKG_RWSTAT_WRITE]);
|
2012-04-01 23:38:42 +02:00
|
|
|
seq_printf(sf, "%s Total %llu\n", dname, (unsigned long long)v);
|
|
|
|
return v;
|
|
|
|
}
|
2013-01-09 17:05:12 +01:00
|
|
|
EXPORT_SYMBOL_GPL(__blkg_prfill_rwstat);
|
2012-04-01 23:38:42 +02:00
|
|
|
|
2012-04-01 23:38:45 +02:00
|
|
|
/**
|
|
|
|
* blkg_prfill_stat - prfill callback for blkg_stat
|
|
|
|
* @sf: seq_file to print to
|
2012-04-16 22:57:26 +02:00
|
|
|
* @pd: policy private data of interest
|
|
|
|
* @off: offset to the blkg_stat in @pd
|
2012-04-01 23:38:45 +02:00
|
|
|
*
|
|
|
|
* prfill callback for printing a blkg_stat.
|
|
|
|
*/
|
2012-04-16 22:57:26 +02:00
|
|
|
u64 blkg_prfill_stat(struct seq_file *sf, struct blkg_policy_data *pd, int off)
|
2012-04-01 23:38:42 +02:00
|
|
|
{
|
2012-04-16 22:57:26 +02:00
|
|
|
return __blkg_prfill_u64(sf, pd, blkg_stat_read((void *)pd + off));
|
2012-04-01 23:38:42 +02:00
|
|
|
}
|
2012-04-01 23:38:45 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkg_prfill_stat);
|
2012-04-01 23:38:42 +02:00
|
|
|
|
2012-04-01 23:38:45 +02:00
|
|
|
/**
|
|
|
|
* blkg_prfill_rwstat - prfill callback for blkg_rwstat
|
|
|
|
* @sf: seq_file to print to
|
2012-04-16 22:57:26 +02:00
|
|
|
* @pd: policy private data of interest
|
|
|
|
* @off: offset to the blkg_rwstat in @pd
|
2012-04-01 23:38:45 +02:00
|
|
|
*
|
|
|
|
* prfill callback for printing a blkg_rwstat.
|
|
|
|
*/
|
2012-04-16 22:57:26 +02:00
|
|
|
u64 blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
|
|
|
|
int off)
|
2012-04-01 23:38:42 +02:00
|
|
|
{
|
2012-04-16 22:57:26 +02:00
|
|
|
struct blkg_rwstat rwstat = blkg_rwstat_read((void *)pd + off);
|
2012-04-01 23:38:42 +02:00
|
|
|
|
2012-04-16 22:57:26 +02:00
|
|
|
return __blkg_prfill_rwstat(sf, pd, &rwstat);
|
2012-04-01 23:38:42 +02:00
|
|
|
}
|
2012-04-01 23:38:45 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkg_prfill_rwstat);
|
2012-04-01 23:38:42 +02:00
|
|
|
|
2015-08-18 23:55:24 +02:00
|
|
|
static u64 blkg_prfill_rwstat_field(struct seq_file *sf,
|
|
|
|
struct blkg_policy_data *pd, int off)
|
|
|
|
{
|
|
|
|
struct blkg_rwstat rwstat = blkg_rwstat_read((void *)pd->blkg + off);
|
|
|
|
|
|
|
|
return __blkg_prfill_rwstat(sf, pd, &rwstat);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkg_print_stat_bytes - seq_show callback for blkg->stat_bytes
|
|
|
|
* @sf: seq_file to print to
|
|
|
|
* @v: unused
|
|
|
|
*
|
|
|
|
* To be used as cftype->seq_show to print blkg->stat_bytes.
|
|
|
|
* cftype->private must be set to the blkcg_policy.
|
|
|
|
*/
|
|
|
|
int blkg_print_stat_bytes(struct seq_file *sf, void *v)
|
|
|
|
{
|
|
|
|
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
|
|
|
|
blkg_prfill_rwstat_field, (void *)seq_cft(sf)->private,
|
|
|
|
offsetof(struct blkcg_gq, stat_bytes), true);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkg_print_stat_bytes);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkg_print_stat_bytes - seq_show callback for blkg->stat_ios
|
|
|
|
* @sf: seq_file to print to
|
|
|
|
* @v: unused
|
|
|
|
*
|
|
|
|
* To be used as cftype->seq_show to print blkg->stat_ios. cftype->private
|
|
|
|
* must be set to the blkcg_policy.
|
|
|
|
*/
|
|
|
|
int blkg_print_stat_ios(struct seq_file *sf, void *v)
|
|
|
|
{
|
|
|
|
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
|
|
|
|
blkg_prfill_rwstat_field, (void *)seq_cft(sf)->private,
|
|
|
|
offsetof(struct blkcg_gq, stat_ios), true);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkg_print_stat_ios);
|
|
|
|
|
|
|
|
static u64 blkg_prfill_rwstat_field_recursive(struct seq_file *sf,
|
|
|
|
struct blkg_policy_data *pd,
|
|
|
|
int off)
|
|
|
|
{
|
|
|
|
struct blkg_rwstat rwstat = blkg_rwstat_recursive_sum(pd->blkg,
|
|
|
|
NULL, off);
|
|
|
|
return __blkg_prfill_rwstat(sf, pd, &rwstat);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkg_print_stat_bytes_recursive - recursive version of blkg_print_stat_bytes
|
|
|
|
* @sf: seq_file to print to
|
|
|
|
* @v: unused
|
|
|
|
*/
|
|
|
|
int blkg_print_stat_bytes_recursive(struct seq_file *sf, void *v)
|
|
|
|
{
|
|
|
|
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
|
|
|
|
blkg_prfill_rwstat_field_recursive,
|
|
|
|
(void *)seq_cft(sf)->private,
|
|
|
|
offsetof(struct blkcg_gq, stat_bytes), true);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkg_print_stat_bytes_recursive);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkg_print_stat_ios_recursive - recursive version of blkg_print_stat_ios
|
|
|
|
* @sf: seq_file to print to
|
|
|
|
* @v: unused
|
|
|
|
*/
|
|
|
|
int blkg_print_stat_ios_recursive(struct seq_file *sf, void *v)
|
|
|
|
{
|
|
|
|
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
|
|
|
|
blkg_prfill_rwstat_field_recursive,
|
|
|
|
(void *)seq_cft(sf)->private,
|
|
|
|
offsetof(struct blkcg_gq, stat_ios), true);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkg_print_stat_ios_recursive);
|
|
|
|
|
2013-01-09 17:05:12 +01:00
|
|
|
/**
|
|
|
|
* blkg_stat_recursive_sum - collect hierarchical blkg_stat
|
2015-08-18 23:55:23 +02:00
|
|
|
* @blkg: blkg of interest
|
|
|
|
* @pol: blkcg_policy which contains the blkg_stat
|
|
|
|
* @off: offset to the blkg_stat in blkg_policy_data or @blkg
|
2013-01-09 17:05:12 +01:00
|
|
|
*
|
2015-08-18 23:55:23 +02:00
|
|
|
* Collect the blkg_stat specified by @blkg, @pol and @off and all its
|
|
|
|
* online descendants and their aux counts. The caller must be holding the
|
|
|
|
* queue lock for online tests.
|
|
|
|
*
|
|
|
|
* If @pol is NULL, blkg_stat is at @off bytes into @blkg; otherwise, it is
|
|
|
|
* at @off bytes into @blkg's blkg_policy_data of the policy.
|
2013-01-09 17:05:12 +01:00
|
|
|
*/
|
2015-08-18 23:55:23 +02:00
|
|
|
u64 blkg_stat_recursive_sum(struct blkcg_gq *blkg,
|
|
|
|
struct blkcg_policy *pol, int off)
|
2013-01-09 17:05:12 +01:00
|
|
|
{
|
|
|
|
struct blkcg_gq *pos_blkg;
|
2013-08-09 02:11:25 +02:00
|
|
|
struct cgroup_subsys_state *pos_css;
|
2013-08-09 02:11:27 +02:00
|
|
|
u64 sum = 0;
|
2013-01-09 17:05:12 +01:00
|
|
|
|
2015-08-18 23:55:23 +02:00
|
|
|
lockdep_assert_held(blkg->q->queue_lock);
|
2013-01-09 17:05:12 +01:00
|
|
|
|
|
|
|
rcu_read_lock();
|
2015-08-18 23:55:23 +02:00
|
|
|
blkg_for_each_descendant_pre(pos_blkg, pos_css, blkg) {
|
|
|
|
struct blkg_stat *stat;
|
|
|
|
|
|
|
|
if (!pos_blkg->online)
|
|
|
|
continue;
|
2013-01-09 17:05:12 +01:00
|
|
|
|
2015-08-18 23:55:23 +02:00
|
|
|
if (pol)
|
|
|
|
stat = (void *)blkg_to_pd(pos_blkg, pol) + off;
|
|
|
|
else
|
|
|
|
stat = (void *)blkg + off;
|
|
|
|
|
|
|
|
sum += blkg_stat_read(stat) + atomic64_read(&stat->aux_cnt);
|
2013-01-09 17:05:12 +01:00
|
|
|
}
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
return sum;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkg_stat_recursive_sum);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkg_rwstat_recursive_sum - collect hierarchical blkg_rwstat
|
2015-08-18 23:55:23 +02:00
|
|
|
* @blkg: blkg of interest
|
|
|
|
* @pol: blkcg_policy which contains the blkg_rwstat
|
|
|
|
* @off: offset to the blkg_rwstat in blkg_policy_data or @blkg
|
2013-01-09 17:05:12 +01:00
|
|
|
*
|
2015-08-18 23:55:23 +02:00
|
|
|
* Collect the blkg_rwstat specified by @blkg, @pol and @off and all its
|
|
|
|
* online descendants and their aux counts. The caller must be holding the
|
|
|
|
* queue lock for online tests.
|
|
|
|
*
|
|
|
|
* If @pol is NULL, blkg_rwstat is at @off bytes into @blkg; otherwise, it
|
|
|
|
* is at @off bytes into @blkg's blkg_policy_data of the policy.
|
2013-01-09 17:05:12 +01:00
|
|
|
*/
|
2015-08-18 23:55:23 +02:00
|
|
|
struct blkg_rwstat blkg_rwstat_recursive_sum(struct blkcg_gq *blkg,
|
|
|
|
struct blkcg_policy *pol, int off)
|
2013-01-09 17:05:12 +01:00
|
|
|
{
|
|
|
|
struct blkcg_gq *pos_blkg;
|
2013-08-09 02:11:25 +02:00
|
|
|
struct cgroup_subsys_state *pos_css;
|
2013-08-09 02:11:27 +02:00
|
|
|
struct blkg_rwstat sum = { };
|
2013-01-09 17:05:12 +01:00
|
|
|
int i;
|
|
|
|
|
2015-08-18 23:55:23 +02:00
|
|
|
lockdep_assert_held(blkg->q->queue_lock);
|
2013-01-09 17:05:12 +01:00
|
|
|
|
|
|
|
rcu_read_lock();
|
2015-08-18 23:55:23 +02:00
|
|
|
blkg_for_each_descendant_pre(pos_blkg, pos_css, blkg) {
|
2015-08-18 23:55:26 +02:00
|
|
|
struct blkg_rwstat *rwstat;
|
2013-01-09 17:05:12 +01:00
|
|
|
|
|
|
|
if (!pos_blkg->online)
|
|
|
|
continue;
|
|
|
|
|
2015-08-18 23:55:23 +02:00
|
|
|
if (pol)
|
|
|
|
rwstat = (void *)blkg_to_pd(pos_blkg, pol) + off;
|
|
|
|
else
|
|
|
|
rwstat = (void *)pos_blkg + off;
|
|
|
|
|
2013-01-09 17:05:12 +01:00
|
|
|
for (i = 0; i < BLKG_RWSTAT_NR; i++)
|
2015-08-18 23:55:26 +02:00
|
|
|
atomic64_add(atomic64_read(&rwstat->aux_cnt[i]) +
|
|
|
|
percpu_counter_sum_positive(&rwstat->cpu_cnt[i]),
|
|
|
|
&sum.aux_cnt[i]);
|
2013-01-09 17:05:12 +01:00
|
|
|
}
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
return sum;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkg_rwstat_recursive_sum);
|
|
|
|
|
2012-04-01 23:38:43 +02:00
|
|
|
/**
|
|
|
|
* blkg_conf_prep - parse and prepare for per-blkg config update
|
|
|
|
* @blkcg: target block cgroup
|
2012-04-13 22:11:29 +02:00
|
|
|
* @pol: target policy
|
2012-04-01 23:38:43 +02:00
|
|
|
* @input: input string
|
|
|
|
* @ctx: blkg_conf_ctx to be filled
|
|
|
|
*
|
|
|
|
* Parse per-blkg config update from @input and initialize @ctx with the
|
2015-08-18 23:55:31 +02:00
|
|
|
* result. @ctx->blkg points to the blkg to be updated and @ctx->body the
|
|
|
|
* part of @input following MAJ:MIN. This function returns with RCU read
|
|
|
|
* lock and queue lock held and must be paired with blkg_conf_finish().
|
2012-04-01 23:38:43 +02:00
|
|
|
*/
|
2012-04-16 22:57:25 +02:00
|
|
|
int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
|
2015-08-18 23:55:31 +02:00
|
|
|
char *input, struct blkg_conf_ctx *ctx)
|
2012-04-13 22:11:29 +02:00
|
|
|
__acquires(rcu) __acquires(disk->queue->queue_lock)
|
2010-04-13 10:05:49 +02:00
|
|
|
{
|
2012-04-01 23:38:43 +02:00
|
|
|
struct gendisk *disk;
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2016-02-09 20:33:35 +01:00
|
|
|
struct module *owner;
|
2012-04-01 23:38:43 +02:00
|
|
|
unsigned int major, minor;
|
2015-08-18 23:55:31 +02:00
|
|
|
int key_len, part, ret;
|
|
|
|
char *body;
|
2010-04-13 10:05:49 +02:00
|
|
|
|
2015-08-18 23:55:31 +02:00
|
|
|
if (sscanf(input, "%u:%u%n", &major, &minor, &key_len) != 2)
|
2012-04-01 23:38:43 +02:00
|
|
|
return -EINVAL;
|
2012-04-01 23:38:43 +02:00
|
|
|
|
2015-08-18 23:55:31 +02:00
|
|
|
body = input + key_len;
|
|
|
|
if (!isspace(*body))
|
|
|
|
return -EINVAL;
|
|
|
|
body = skip_spaces(body);
|
|
|
|
|
2012-04-01 23:38:43 +02:00
|
|
|
disk = get_gendisk(MKDEV(major, minor), &part);
|
2015-07-23 00:05:53 +02:00
|
|
|
if (!disk)
|
2015-08-18 23:55:28 +02:00
|
|
|
return -ENODEV;
|
2015-07-23 00:05:53 +02:00
|
|
|
if (part) {
|
2016-02-09 20:33:35 +01:00
|
|
|
owner = disk->fops->owner;
|
2015-07-23 00:05:53 +02:00
|
|
|
put_disk(disk);
|
2016-02-09 20:33:35 +01:00
|
|
|
module_put(owner);
|
2015-08-18 23:55:28 +02:00
|
|
|
return -ENODEV;
|
2015-07-23 00:05:53 +02:00
|
|
|
}
|
2012-03-05 22:15:07 +01:00
|
|
|
|
|
|
|
rcu_read_lock();
|
2012-03-05 22:15:08 +01:00
|
|
|
spin_lock_irq(disk->queue->queue_lock);
|
2012-04-13 22:11:29 +02:00
|
|
|
|
2012-04-13 22:11:33 +02:00
|
|
|
if (blkcg_policy_enabled(disk->queue, pol))
|
2012-04-13 22:11:34 +02:00
|
|
|
blkg = blkg_lookup_create(blkcg, disk->queue);
|
2012-04-13 22:11:33 +02:00
|
|
|
else
|
2015-08-18 23:55:28 +02:00
|
|
|
blkg = ERR_PTR(-EOPNOTSUPP);
|
2012-03-05 22:15:07 +01:00
|
|
|
|
2012-03-05 22:15:08 +01:00
|
|
|
if (IS_ERR(blkg)) {
|
|
|
|
ret = PTR_ERR(blkg);
|
2012-04-01 23:38:43 +02:00
|
|
|
rcu_read_unlock();
|
2012-04-13 22:11:29 +02:00
|
|
|
spin_unlock_irq(disk->queue->queue_lock);
|
2016-02-09 20:33:35 +01:00
|
|
|
owner = disk->fops->owner;
|
2012-04-01 23:38:43 +02:00
|
|
|
put_disk(disk);
|
2016-02-09 20:33:35 +01:00
|
|
|
module_put(owner);
|
2012-04-01 23:38:43 +02:00
|
|
|
/*
|
|
|
|
* If queue was bypassing, we should retry. Do so after a
|
|
|
|
* short msleep(). It isn't strictly necessary but queue
|
|
|
|
* can be bypassing for some time and it's always nice to
|
|
|
|
* avoid busy looping.
|
|
|
|
*/
|
|
|
|
if (ret == -EBUSY) {
|
|
|
|
msleep(10);
|
|
|
|
ret = restart_syscall();
|
2010-09-15 23:06:36 +02:00
|
|
|
}
|
2012-04-01 23:38:43 +02:00
|
|
|
return ret;
|
2010-09-15 23:06:33 +02:00
|
|
|
}
|
2012-04-01 23:38:43 +02:00
|
|
|
|
|
|
|
ctx->disk = disk;
|
|
|
|
ctx->blkg = blkg;
|
2015-08-18 23:55:31 +02:00
|
|
|
ctx->body = body;
|
2012-04-01 23:38:43 +02:00
|
|
|
return 0;
|
2010-04-13 10:05:49 +02:00
|
|
|
}
|
2012-04-01 23:38:43 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkg_conf_prep);
|
2010-04-13 10:05:49 +02:00
|
|
|
|
2012-04-01 23:38:43 +02:00
|
|
|
/**
|
|
|
|
* blkg_conf_finish - finish up per-blkg config update
|
|
|
|
* @ctx: blkg_conf_ctx intiailized by blkg_conf_prep()
|
|
|
|
*
|
|
|
|
* Finish up after per-blkg config update. This function must be paired
|
|
|
|
* with blkg_conf_prep().
|
|
|
|
*/
|
2012-04-01 23:38:43 +02:00
|
|
|
void blkg_conf_finish(struct blkg_conf_ctx *ctx)
|
2012-04-13 22:11:29 +02:00
|
|
|
__releases(ctx->disk->queue->queue_lock) __releases(rcu)
|
2010-04-13 10:05:49 +02:00
|
|
|
{
|
2016-02-09 20:33:35 +01:00
|
|
|
struct module *owner;
|
|
|
|
|
2012-04-13 22:11:29 +02:00
|
|
|
spin_unlock_irq(ctx->disk->queue->queue_lock);
|
2012-04-01 23:38:43 +02:00
|
|
|
rcu_read_unlock();
|
2016-02-09 20:33:35 +01:00
|
|
|
owner = ctx->disk->fops->owner;
|
2012-04-01 23:38:43 +02:00
|
|
|
put_disk(ctx->disk);
|
2016-02-09 20:33:35 +01:00
|
|
|
module_put(owner);
|
2010-04-13 10:05:49 +02:00
|
|
|
}
|
2012-04-01 23:38:43 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkg_conf_finish);
|
2010-04-13 10:05:49 +02:00
|
|
|
|
2015-08-18 23:55:34 +02:00
|
|
|
static int blkcg_print_stat(struct seq_file *sf, void *v)
|
|
|
|
{
|
|
|
|
struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
|
|
|
|
struct blkcg_gq *blkg;
|
|
|
|
|
|
|
|
rcu_read_lock();
|
|
|
|
|
|
|
|
hlist_for_each_entry_rcu(blkg, &blkcg->blkg_list, blkcg_node) {
|
|
|
|
const char *dname;
|
|
|
|
struct blkg_rwstat rwstat;
|
|
|
|
u64 rbytes, wbytes, rios, wios;
|
|
|
|
|
|
|
|
dname = blkg_dev_name(blkg);
|
|
|
|
if (!dname)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
spin_lock_irq(blkg->q->queue_lock);
|
|
|
|
|
|
|
|
rwstat = blkg_rwstat_recursive_sum(blkg, NULL,
|
|
|
|
offsetof(struct blkcg_gq, stat_bytes));
|
|
|
|
rbytes = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_READ]);
|
|
|
|
wbytes = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_WRITE]);
|
|
|
|
|
|
|
|
rwstat = blkg_rwstat_recursive_sum(blkg, NULL,
|
|
|
|
offsetof(struct blkcg_gq, stat_ios));
|
|
|
|
rios = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_READ]);
|
|
|
|
wios = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_WRITE]);
|
|
|
|
|
|
|
|
spin_unlock_irq(blkg->q->queue_lock);
|
|
|
|
|
|
|
|
if (rbytes || wbytes || rios || wios)
|
|
|
|
seq_printf(sf, "%s rbytes=%llu wbytes=%llu rios=%llu wios=%llu\n",
|
|
|
|
dname, rbytes, wbytes, rios, wios);
|
|
|
|
}
|
|
|
|
|
|
|
|
rcu_read_unlock();
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2016-06-14 17:04:32 +02:00
|
|
|
static struct cftype blkcg_files[] = {
|
2015-08-18 23:55:34 +02:00
|
|
|
{
|
|
|
|
.name = "stat",
|
2015-10-22 02:48:37 +02:00
|
|
|
.flags = CFTYPE_NOT_ON_ROOT,
|
2015-08-18 23:55:34 +02:00
|
|
|
.seq_show = blkcg_print_stat,
|
|
|
|
},
|
|
|
|
{ } /* terminate */
|
|
|
|
};
|
|
|
|
|
2016-06-14 17:04:32 +02:00
|
|
|
static struct cftype blkcg_legacy_files[] = {
|
2010-04-09 08:31:19 +02:00
|
|
|
{
|
|
|
|
.name = "reset_stats",
|
2012-04-16 22:57:25 +02:00
|
|
|
.write_u64 = blkcg_reset_stats,
|
2009-12-03 18:59:49 +01:00
|
|
|
},
|
2012-04-01 21:09:55 +02:00
|
|
|
{ } /* terminate */
|
2009-12-03 18:59:42 +01:00
|
|
|
};
|
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
/**
|
2012-11-19 17:13:38 +01:00
|
|
|
* blkcg_css_offline - cgroup css_offline callback
|
2013-08-09 02:11:23 +02:00
|
|
|
* @css: css of interest
|
2012-03-05 22:15:21 +01:00
|
|
|
*
|
2013-08-09 02:11:23 +02:00
|
|
|
* This function is called when @css is about to go away and responsible
|
|
|
|
* for shooting down all blkgs associated with @css. blkgs should be
|
2012-03-05 22:15:21 +01:00
|
|
|
* removed while holding both q and blkcg locks. As blkcg lock is nested
|
|
|
|
* inside q lock, this function performs reverse double lock dancing.
|
|
|
|
*
|
|
|
|
* This is the blkcg counterpart of ioc_release_fn().
|
|
|
|
*/
|
2013-08-09 02:11:23 +02:00
|
|
|
static void blkcg_css_offline(struct cgroup_subsys_state *css)
|
2009-12-03 18:59:42 +01:00
|
|
|
{
|
2013-08-09 02:11:23 +02:00
|
|
|
struct blkcg *blkcg = css_to_blkcg(css);
|
2009-12-03 18:59:47 +01:00
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
spin_lock_irq(&blkcg->lock);
|
2012-03-05 22:15:11 +01:00
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
while (!hlist_empty(&blkcg->blkg_list)) {
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg = hlist_entry(blkcg->blkg_list.first,
|
|
|
|
struct blkcg_gq, blkcg_node);
|
2012-03-05 22:15:22 +01:00
|
|
|
struct request_queue *q = blkg->q;
|
2009-12-03 18:59:47 +01:00
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
if (spin_trylock(q->queue_lock)) {
|
|
|
|
blkg_destroy(blkg);
|
|
|
|
spin_unlock(q->queue_lock);
|
|
|
|
} else {
|
|
|
|
spin_unlock_irq(&blkcg->lock);
|
|
|
|
cpu_relax();
|
2012-03-29 20:57:08 +02:00
|
|
|
spin_lock_irq(&blkcg->lock);
|
2010-05-03 14:28:55 +02:00
|
|
|
}
|
2012-03-05 22:15:21 +01:00
|
|
|
}
|
2009-12-03 18:59:47 +01:00
|
|
|
|
2012-03-05 22:15:21 +01:00
|
|
|
spin_unlock_irq(&blkcg->lock);
|
2015-05-22 23:13:37 +02:00
|
|
|
|
|
|
|
wb_blkcg_offline(blkcg);
|
2012-03-05 22:15:11 +01:00
|
|
|
}
|
|
|
|
|
2013-08-09 02:11:23 +02:00
|
|
|
static void blkcg_css_free(struct cgroup_subsys_state *css)
|
2012-03-05 22:15:11 +01:00
|
|
|
{
|
2013-08-09 02:11:23 +02:00
|
|
|
struct blkcg *blkcg = css_to_blkcg(css);
|
2015-08-18 23:55:08 +02:00
|
|
|
int i;
|
2012-03-05 22:15:11 +01:00
|
|
|
|
2015-07-09 22:39:49 +02:00
|
|
|
mutex_lock(&blkcg_pol_mutex);
|
2015-08-18 23:55:16 +02:00
|
|
|
|
2015-07-09 22:39:49 +02:00
|
|
|
list_del(&blkcg->all_blkcgs_node);
|
|
|
|
|
2015-08-18 23:55:08 +02:00
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++)
|
2015-08-18 23:55:16 +02:00
|
|
|
if (blkcg->cpd[i])
|
|
|
|
blkcg_policy[i]->cpd_free_fn(blkcg->cpd[i]);
|
|
|
|
|
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
|
|
|
|
2015-08-18 23:55:08 +02:00
|
|
|
kfree(blkcg);
|
2009-12-03 18:59:42 +01:00
|
|
|
}
|
|
|
|
|
2013-08-09 02:11:23 +02:00
|
|
|
static struct cgroup_subsys_state *
|
|
|
|
blkcg_css_alloc(struct cgroup_subsys_state *parent_css)
|
2009-12-03 18:59:42 +01:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg *blkcg;
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
struct cgroup_subsys_state *ret;
|
|
|
|
int i;
|
2009-12-03 18:59:42 +01:00
|
|
|
|
2015-07-09 22:39:49 +02:00
|
|
|
mutex_lock(&blkcg_pol_mutex);
|
|
|
|
|
2013-08-09 02:11:23 +02:00
|
|
|
if (!parent_css) {
|
2012-04-16 22:57:25 +02:00
|
|
|
blkcg = &blkcg_root;
|
2015-08-18 23:55:08 +02:00
|
|
|
} else {
|
|
|
|
blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
|
|
|
|
if (!blkcg) {
|
|
|
|
ret = ERR_PTR(-ENOMEM);
|
|
|
|
goto free_blkcg;
|
|
|
|
}
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < BLKCG_MAX_POLS ; i++) {
|
|
|
|
struct blkcg_policy *pol = blkcg_policy[i];
|
|
|
|
struct blkcg_policy_data *cpd;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If the policy hasn't been attached yet, wait for it
|
|
|
|
* to be attached before doing anything else. Otherwise,
|
|
|
|
* check if the policy requires any specific per-cgroup
|
|
|
|
* data: if it does, allocate and initialize it.
|
|
|
|
*/
|
2015-08-18 23:55:16 +02:00
|
|
|
if (!pol || !pol->cpd_alloc_fn)
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
continue;
|
|
|
|
|
2015-08-18 23:55:16 +02:00
|
|
|
cpd = pol->cpd_alloc_fn(GFP_KERNEL);
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
if (!cpd) {
|
|
|
|
ret = ERR_PTR(-ENOMEM);
|
|
|
|
goto free_pd_blkcg;
|
|
|
|
}
|
2015-08-18 23:55:15 +02:00
|
|
|
blkcg->cpd[i] = cpd;
|
|
|
|
cpd->blkcg = blkcg;
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
cpd->plid = i;
|
2015-08-18 23:55:16 +02:00
|
|
|
if (pol->cpd_init_fn)
|
|
|
|
pol->cpd_init_fn(cpd);
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
}
|
2009-12-03 18:59:42 +01:00
|
|
|
|
|
|
|
spin_lock_init(&blkcg->lock);
|
2016-11-22 00:03:32 +01:00
|
|
|
INIT_RADIX_TREE(&blkcg->blkg_tree, GFP_NOWAIT | __GFP_NOWARN);
|
2009-12-03 18:59:42 +01:00
|
|
|
INIT_HLIST_HEAD(&blkcg->blkg_list);
|
2015-05-22 23:13:37 +02:00
|
|
|
#ifdef CONFIG_CGROUP_WRITEBACK
|
|
|
|
INIT_LIST_HEAD(&blkcg->cgwb_list);
|
|
|
|
#endif
|
2015-07-09 22:39:49 +02:00
|
|
|
list_add_tail(&blkcg->all_blkcgs_node, &all_blkcgs);
|
|
|
|
|
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
2009-12-03 18:59:42 +01:00
|
|
|
return &blkcg->css;
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
|
|
|
|
free_pd_blkcg:
|
|
|
|
for (i--; i >= 0; i--)
|
2015-08-18 23:55:16 +02:00
|
|
|
if (blkcg->cpd[i])
|
|
|
|
blkcg_policy[i]->cpd_free_fn(blkcg->cpd[i]);
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
free_blkcg:
|
|
|
|
kfree(blkcg);
|
2015-07-09 22:39:49 +02:00
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
block, cgroup: implement policy-specific per-blkcg data
The block IO (blkio) controller enables the block layer to provide service
guarantees in a hierarchical fashion. Specifically, service guarantees
are provided by registered request-accounting policies. As of now, a
proportional-share and a throttling policy are available. They are
implemented, respectively, by the CFQ I/O scheduler and the blk-throttle
subsystem. Unfortunately, as for adding new policies, the current
implementation of the block IO controller is only halfway ready to allow
new policies to be plugged in. This commit provides a solution to make
the block IO controller fully ready to handle new policies.
In what follows, we first describe briefly the current state, and then
list the changes made by this commit.
The throttling policy does not need any per-cgroup information to perform
its task. In contrast, the proportional share policy uses, for each cgroup,
both the weight assigned by the user to the cgroup, and a set of dynamically-
computed weights, one for each device.
The first, user-defined weight is stored in the blkcg data structure: the
block IO controller allocates a private blkcg data structure for each
cgroup in the blkio cgroups hierarchy (regardless of which policy is active).
In other words, the block IO controller internally mirrors the blkio cgroups
with private blkcg data structures.
On the other hand, for each cgroup and device, the corresponding dynamically-
computed weight is maintained in the following, different way. For each device,
the block IO controller keeps a private blkcg_gq structure for each cgroup in
blkio. In other words, block IO also keeps one private mirror copy of the blkio
cgroups hierarchy for each device, made of blkcg_gq structures.
Each blkcg_gq structure keeps per-policy information in a generic array of
dynamically-allocated 'dedicated' data structures, one for each registered
policy (so currently the array contains two elements). To be inserted into the
generic array, each dedicated data structure embeds a generic blkg_policy_data
structure. Consider now the array contained in the blkcg_gq structure
corresponding to a given pair of cgroup and device: one of the elements
of the array contains the dedicated data structure for the proportional-share
policy, and this dedicated data structure contains the dynamically-computed
weight for that pair of cgroup and device.
The generic strategy adopted for storing per-policy data in blkcg_gq structures
is already capable of handling new policies, whereas the one adopted with blkcg
structures is not, because per-policy data are hard-coded in the blkcg
structures themselves (currently only data related to the proportional-
share policy).
This commit addresses the above issues through the following changes:
. It generalizes blkcg structures so that per-policy data are stored in the same
way as in blkcg_gq structures.
Specifically, it lets also the blkcg structure store per-policy data in a
generic array of dynamically-allocated dedicated data structures. We will
refer to these data structures as blkcg dedicated data structures, to
distinguish them from the dedicated data structures inserted in the generic
arrays kept by blkcg_gq structures.
To allow blkcg dedicated data structures to be inserted in the generic array
inside a blkcg structure, this commit also introduces a new blkcg_policy_data
structure, which is the equivalent of blkg_policy_data for blkcg dedicated
data structures.
. It adds to the blkcg_policy structure, i.e., to the descriptor of a policy, a
cpd_size field and a cpd_init field, to be initialized by the policy with,
respectively, the size of the blkcg dedicated data structures, and the
address of a constructor function for blkcg dedicated data structures.
. It moves the CFQ-specific fields embedded in the blkcg data structure (i.e.,
the fields related to the proportional-share policy), into a new blkcg
dedicated data structure called cfq_group_data.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini.arianna@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-06-05 23:38:42 +02:00
|
|
|
return ret;
|
2009-12-03 18:59:42 +01:00
|
|
|
}
|
|
|
|
|
2012-03-05 22:15:12 +01:00
|
|
|
/**
|
|
|
|
* blkcg_init_queue - initialize blkcg part of request queue
|
|
|
|
* @q: request_queue to initialize
|
|
|
|
*
|
|
|
|
* Called from blk_alloc_queue_node(). Responsible for initializing blkcg
|
|
|
|
* part of new request_queue @q.
|
|
|
|
*
|
|
|
|
* RETURNS:
|
|
|
|
* 0 on success, -errno on failure.
|
|
|
|
*/
|
|
|
|
int blkcg_init_queue(struct request_queue *q)
|
|
|
|
{
|
2015-05-22 23:13:19 +02:00
|
|
|
struct blkcg_gq *new_blkg, *blkg;
|
|
|
|
bool preloaded;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
new_blkg = blkg_alloc(&blkcg_root, q, GFP_KERNEL);
|
|
|
|
if (!new_blkg)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
preloaded = !radix_tree_preload(GFP_KERNEL);
|
2012-03-05 22:15:12 +01:00
|
|
|
|
2015-05-22 23:13:19 +02:00
|
|
|
/*
|
|
|
|
* Make sure the root blkg exists and count the existing blkgs. As
|
|
|
|
* @q is bypassing at this point, blkg_lookup_create() can't be
|
|
|
|
* used. Open code insertion.
|
|
|
|
*/
|
|
|
|
rcu_read_lock();
|
|
|
|
spin_lock_irq(q->queue_lock);
|
|
|
|
blkg = blkg_create(&blkcg_root, q, new_blkg);
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
if (preloaded)
|
|
|
|
radix_tree_preload_end();
|
|
|
|
|
|
|
|
if (IS_ERR(blkg)) {
|
2015-08-18 23:55:07 +02:00
|
|
|
blkg_free(new_blkg);
|
2015-05-22 23:13:19 +02:00
|
|
|
return PTR_ERR(blkg);
|
|
|
|
}
|
|
|
|
|
|
|
|
q->root_blkg = blkg;
|
|
|
|
q->root_rl.blkg = blkg;
|
2012-03-05 22:15:12 +01:00
|
|
|
|
2015-05-22 23:13:19 +02:00
|
|
|
ret = blk_throtl_init(q);
|
|
|
|
if (ret) {
|
|
|
|
spin_lock_irq(q->queue_lock);
|
|
|
|
blkg_destroy_all(q);
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
}
|
|
|
|
return ret;
|
2012-03-05 22:15:12 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkcg_drain_queue - drain blkcg part of request_queue
|
|
|
|
* @q: request_queue to drain
|
|
|
|
*
|
|
|
|
* Called from blk_drain_queue(). Responsible for draining blkcg part.
|
|
|
|
*/
|
|
|
|
void blkcg_drain_queue(struct request_queue *q)
|
|
|
|
{
|
|
|
|
lockdep_assert_held(q->queue_lock);
|
|
|
|
|
2014-07-06 00:43:21 +02:00
|
|
|
/*
|
|
|
|
* @q could be exiting and already have destroyed all blkgs as
|
|
|
|
* indicated by NULL root_blkg. If so, don't confuse policies.
|
|
|
|
*/
|
|
|
|
if (!q->root_blkg)
|
|
|
|
return;
|
|
|
|
|
2012-03-05 22:15:12 +01:00
|
|
|
blk_throtl_drain(q);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkcg_exit_queue - exit and release blkcg part of request_queue
|
|
|
|
* @q: request_queue being released
|
|
|
|
*
|
|
|
|
* Called from blk_release_queue(). Responsible for exiting blkcg part.
|
|
|
|
*/
|
|
|
|
void blkcg_exit_queue(struct request_queue *q)
|
|
|
|
{
|
2012-04-13 22:11:35 +02:00
|
|
|
spin_lock_irq(q->queue_lock);
|
2012-04-13 22:11:34 +02:00
|
|
|
blkg_destroy_all(q);
|
2012-04-13 22:11:35 +02:00
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
|
2012-03-05 22:15:12 +01:00
|
|
|
blk_throtl_exit(q);
|
|
|
|
}
|
|
|
|
|
2009-12-03 18:59:42 +01:00
|
|
|
/*
|
|
|
|
* We cannot support shared io contexts, as we have no mean to support
|
|
|
|
* two tasks with the same ioc in two different groups without major rework
|
|
|
|
* of the main cic data structures. For now we allow a task to change
|
|
|
|
* its cgroup only if it's the only owner of its ioc.
|
|
|
|
*/
|
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 16:18:21 +01:00
|
|
|
static int blkcg_can_attach(struct cgroup_taskset *tset)
|
2009-12-03 18:59:42 +01:00
|
|
|
{
|
2011-12-13 03:12:21 +01:00
|
|
|
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 16:18:21 +01:00
|
|
|
struct cgroup_subsys_state *dst_css;
|
2009-12-03 18:59:42 +01:00
|
|
|
struct io_context *ioc;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
/* task_lock() is needed to avoid races with exit_io_context() */
|
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 16:18:21 +01:00
|
|
|
cgroup_taskset_for_each(task, dst_css, tset) {
|
2011-12-13 03:12:21 +01:00
|
|
|
task_lock(task);
|
|
|
|
ioc = task->io_context;
|
|
|
|
if (ioc && atomic_read(&ioc->nr_tasks) > 1)
|
|
|
|
ret = -EINVAL;
|
|
|
|
task_unlock(task);
|
|
|
|
if (ret)
|
|
|
|
break;
|
|
|
|
}
|
2009-12-03 18:59:42 +01:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2015-08-18 23:55:36 +02:00
|
|
|
static void blkcg_bind(struct cgroup_subsys_state *root_css)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
mutex_lock(&blkcg_pol_mutex);
|
|
|
|
|
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++) {
|
|
|
|
struct blkcg_policy *pol = blkcg_policy[i];
|
|
|
|
struct blkcg *blkcg;
|
|
|
|
|
|
|
|
if (!pol || !pol->cpd_bind_fn)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node)
|
|
|
|
if (blkcg->cpd[pol->plid])
|
|
|
|
pol->cpd_bind_fn(blkcg->cpd[pol->plid]);
|
|
|
|
}
|
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
|
|
|
}
|
|
|
|
|
blkcg: rename subsystem name from blkio to io
blkio interface has become messy over time and is currently the
largest. In addition to the inconsistent naming scheme, it has
multiple stat files which report more or less the same thing, a number
of debug stat files which expose internal details which shouldn't have
been part of the public interface in the first place, recursive and
non-recursive stats and leaf and non-leaf knobs.
Both recursive vs. non-recursive and leaf vs. non-leaf distinctions
don't make any sense on the unified hierarchy as only leaf cgroups can
contain processes. cgroups is going through a major interface
revision with the unified hierarchy involving significant fundamental
usage changes and given that a significant portion of the interface
doesn't make sense anymore, it's a good time to reorganize the
interface.
As the first step, this patch renames the external visible subsystem
name from "blkio" to "io". This is more concise, matches the other
two major subsystem names, "cpu" and "memory", and better suited as
blkcg will be involved in anything writeback related too whether an
actual block device is involved or not.
As the subsystem legacy_name is set to "blkio", the only userland
visible change outside the unified hierarchy is that blkcg is reported
as "io" instead of "blkio" in the subsystem initialized message during
boot. On the unified hierarchy, blkcg now appears as "io".
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: cgroups@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-18 23:55:29 +02:00
|
|
|
struct cgroup_subsys io_cgrp_subsys = {
|
2012-11-19 17:13:38 +01:00
|
|
|
.css_alloc = blkcg_css_alloc,
|
|
|
|
.css_offline = blkcg_css_offline,
|
|
|
|
.css_free = blkcg_css_free,
|
2012-04-16 22:57:25 +02:00
|
|
|
.can_attach = blkcg_can_attach,
|
2015-08-18 23:55:36 +02:00
|
|
|
.bind = blkcg_bind,
|
2015-08-18 23:55:34 +02:00
|
|
|
.dfl_cftypes = blkcg_files,
|
2015-08-18 23:55:30 +02:00
|
|
|
.legacy_cftypes = blkcg_legacy_files,
|
blkcg: rename subsystem name from blkio to io
blkio interface has become messy over time and is currently the
largest. In addition to the inconsistent naming scheme, it has
multiple stat files which report more or less the same thing, a number
of debug stat files which expose internal details which shouldn't have
been part of the public interface in the first place, recursive and
non-recursive stats and leaf and non-leaf knobs.
Both recursive vs. non-recursive and leaf vs. non-leaf distinctions
don't make any sense on the unified hierarchy as only leaf cgroups can
contain processes. cgroups is going through a major interface
revision with the unified hierarchy involving significant fundamental
usage changes and given that a significant portion of the interface
doesn't make sense anymore, it's a good time to reorganize the
interface.
As the first step, this patch renames the external visible subsystem
name from "blkio" to "io". This is more concise, matches the other
two major subsystem names, "cpu" and "memory", and better suited as
blkcg will be involved in anything writeback related too whether an
actual block device is involved or not.
As the subsystem legacy_name is set to "blkio", the only userland
visible change outside the unified hierarchy is that blkcg is reported
as "io" instead of "blkio" in the subsystem initialized message during
boot. On the unified hierarchy, blkcg now appears as "io".
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: cgroups@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-18 23:55:29 +02:00
|
|
|
.legacy_name = "blkio",
|
2014-07-09 00:02:57 +02:00
|
|
|
#ifdef CONFIG_MEMCG
|
|
|
|
/*
|
|
|
|
* This ensures that, if available, memcg is automatically enabled
|
|
|
|
* together on the default hierarchy so that the owner cgroup can
|
|
|
|
* be retrieved from writeback pages.
|
|
|
|
*/
|
|
|
|
.depends_on = 1 << memory_cgrp_id,
|
|
|
|
#endif
|
2012-04-01 21:09:55 +02:00
|
|
|
};
|
blkcg: rename subsystem name from blkio to io
blkio interface has become messy over time and is currently the
largest. In addition to the inconsistent naming scheme, it has
multiple stat files which report more or less the same thing, a number
of debug stat files which expose internal details which shouldn't have
been part of the public interface in the first place, recursive and
non-recursive stats and leaf and non-leaf knobs.
Both recursive vs. non-recursive and leaf vs. non-leaf distinctions
don't make any sense on the unified hierarchy as only leaf cgroups can
contain processes. cgroups is going through a major interface
revision with the unified hierarchy involving significant fundamental
usage changes and given that a significant portion of the interface
doesn't make sense anymore, it's a good time to reorganize the
interface.
As the first step, this patch renames the external visible subsystem
name from "blkio" to "io". This is more concise, matches the other
two major subsystem names, "cpu" and "memory", and better suited as
blkcg will be involved in anything writeback related too whether an
actual block device is involved or not.
As the subsystem legacy_name is set to "blkio", the only userland
visible change outside the unified hierarchy is that blkcg is reported
as "io" instead of "blkio" in the subsystem initialized message during
boot. On the unified hierarchy, blkcg now appears as "io".
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: cgroups@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-18 23:55:29 +02:00
|
|
|
EXPORT_SYMBOL_GPL(io_cgrp_subsys);
|
2012-04-01 21:09:55 +02:00
|
|
|
|
2012-04-13 22:11:33 +02:00
|
|
|
/**
|
|
|
|
* blkcg_activate_policy - activate a blkcg policy on a request_queue
|
|
|
|
* @q: request_queue of interest
|
|
|
|
* @pol: blkcg policy to activate
|
|
|
|
*
|
|
|
|
* Activate @pol on @q. Requires %GFP_KERNEL context. @q goes through
|
|
|
|
* bypass mode to populate its blkgs with policy_data for @pol.
|
|
|
|
*
|
|
|
|
* Activation happens with @q bypassed, so nobody would be accessing blkgs
|
|
|
|
* from IO path. Update of each blkg is protected by both queue and blkcg
|
|
|
|
* locks so that holding either lock and testing blkcg_policy_enabled() is
|
|
|
|
* always enough for dereferencing policy data.
|
|
|
|
*
|
|
|
|
* The caller is responsible for synchronizing [de]activations and policy
|
|
|
|
* [un]registerations. Returns 0 on success, -errno on failure.
|
|
|
|
*/
|
|
|
|
int blkcg_activate_policy(struct request_queue *q,
|
2012-04-16 22:57:25 +02:00
|
|
|
const struct blkcg_policy *pol)
|
2012-04-13 22:11:33 +02:00
|
|
|
{
|
2015-08-18 23:55:09 +02:00
|
|
|
struct blkg_policy_data *pd_prealloc = NULL;
|
2015-05-22 23:13:19 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2015-08-18 23:55:09 +02:00
|
|
|
int ret;
|
2012-04-13 22:11:33 +02:00
|
|
|
|
|
|
|
if (blkcg_policy_enabled(q, pol))
|
|
|
|
return 0;
|
|
|
|
|
2017-01-18 23:37:27 +01:00
|
|
|
if (q->mq_ops)
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_mq_freeze_queue(q);
|
2017-01-18 23:37:27 +01:00
|
|
|
else
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_queue_bypass_start(q);
|
2015-08-18 23:55:09 +02:00
|
|
|
pd_prealloc:
|
|
|
|
if (!pd_prealloc) {
|
2015-08-18 23:55:11 +02:00
|
|
|
pd_prealloc = pol->pd_alloc_fn(GFP_KERNEL, q->node);
|
2015-08-18 23:55:09 +02:00
|
|
|
if (!pd_prealloc) {
|
2012-04-13 22:11:33 +02:00
|
|
|
ret = -ENOMEM;
|
2015-08-18 23:55:09 +02:00
|
|
|
goto out_bypass_end;
|
2012-04-13 22:11:33 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_lock_irq(q->queue_lock);
|
|
|
|
|
|
|
|
list_for_each_entry(blkg, &q->blkg_list, q_node) {
|
2015-08-18 23:55:09 +02:00
|
|
|
struct blkg_policy_data *pd;
|
|
|
|
|
|
|
|
if (blkg->pd[pol->plid])
|
|
|
|
continue;
|
2012-04-13 22:11:33 +02:00
|
|
|
|
2016-11-22 00:03:32 +01:00
|
|
|
pd = pol->pd_alloc_fn(GFP_NOWAIT | __GFP_NOWARN, q->node);
|
2015-08-18 23:55:09 +02:00
|
|
|
if (!pd)
|
|
|
|
swap(pd, pd_prealloc);
|
|
|
|
if (!pd) {
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
goto pd_prealloc;
|
|
|
|
}
|
2012-04-13 22:11:33 +02:00
|
|
|
|
|
|
|
blkg->pd[pol->plid] = pd;
|
|
|
|
pd->blkg = blkg;
|
2013-01-09 17:05:12 +01:00
|
|
|
pd->plid = pol->plid;
|
2015-08-18 23:55:10 +02:00
|
|
|
if (pol->pd_init_fn)
|
2015-08-18 23:55:14 +02:00
|
|
|
pol->pd_init_fn(pd);
|
2012-04-13 22:11:33 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
__set_bit(pol->plid, q->blkcg_pols);
|
|
|
|
ret = 0;
|
2015-08-18 23:55:09 +02:00
|
|
|
|
2012-04-13 22:11:33 +02:00
|
|
|
spin_unlock_irq(q->queue_lock);
|
2015-08-18 23:55:09 +02:00
|
|
|
out_bypass_end:
|
2017-01-18 23:37:27 +01:00
|
|
|
if (q->mq_ops)
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_mq_unfreeze_queue(q);
|
2017-01-18 23:37:27 +01:00
|
|
|
else
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_queue_bypass_end(q);
|
2015-08-18 23:55:11 +02:00
|
|
|
if (pd_prealloc)
|
|
|
|
pol->pd_free_fn(pd_prealloc);
|
2012-04-13 22:11:33 +02:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkcg_activate_policy);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blkcg_deactivate_policy - deactivate a blkcg policy on a request_queue
|
|
|
|
* @q: request_queue of interest
|
|
|
|
* @pol: blkcg policy to deactivate
|
|
|
|
*
|
|
|
|
* Deactivate @pol on @q. Follows the same synchronization rules as
|
|
|
|
* blkcg_activate_policy().
|
|
|
|
*/
|
|
|
|
void blkcg_deactivate_policy(struct request_queue *q,
|
2012-04-16 22:57:25 +02:00
|
|
|
const struct blkcg_policy *pol)
|
2012-04-13 22:11:33 +02:00
|
|
|
{
|
2012-04-16 22:57:25 +02:00
|
|
|
struct blkcg_gq *blkg;
|
2012-04-13 22:11:33 +02:00
|
|
|
|
|
|
|
if (!blkcg_policy_enabled(q, pol))
|
|
|
|
return;
|
|
|
|
|
2017-01-18 23:37:27 +01:00
|
|
|
if (q->mq_ops)
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_mq_freeze_queue(q);
|
2017-01-18 23:37:27 +01:00
|
|
|
else
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_queue_bypass_start(q);
|
|
|
|
|
2012-04-13 22:11:33 +02:00
|
|
|
spin_lock_irq(q->queue_lock);
|
|
|
|
|
|
|
|
__clear_bit(pol->plid, q->blkcg_pols);
|
|
|
|
|
|
|
|
list_for_each_entry(blkg, &q->blkg_list, q_node) {
|
|
|
|
/* grab blkcg lock too while removing @pd from @blkg */
|
|
|
|
spin_lock(&blkg->blkcg->lock);
|
|
|
|
|
2015-08-18 23:55:11 +02:00
|
|
|
if (blkg->pd[pol->plid]) {
|
2015-08-18 23:55:14 +02:00
|
|
|
if (pol->pd_offline_fn)
|
|
|
|
pol->pd_offline_fn(blkg->pd[pol->plid]);
|
2015-08-18 23:55:11 +02:00
|
|
|
pol->pd_free_fn(blkg->pd[pol->plid]);
|
|
|
|
blkg->pd[pol->plid] = NULL;
|
|
|
|
}
|
2012-04-13 22:11:33 +02:00
|
|
|
|
|
|
|
spin_unlock(&blkg->blkcg->lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
2017-01-17 14:03:22 +01:00
|
|
|
|
2017-01-18 23:37:27 +01:00
|
|
|
if (q->mq_ops)
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_mq_unfreeze_queue(q);
|
2017-01-18 23:37:27 +01:00
|
|
|
else
|
2017-01-17 14:03:22 +01:00
|
|
|
blk_queue_bypass_end(q);
|
2012-04-13 22:11:33 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blkcg_deactivate_policy);
|
|
|
|
|
2012-04-13 22:11:28 +02:00
|
|
|
/**
|
2012-04-16 22:57:25 +02:00
|
|
|
* blkcg_policy_register - register a blkcg policy
|
|
|
|
* @pol: blkcg policy to register
|
2012-04-13 22:11:28 +02:00
|
|
|
*
|
2012-04-16 22:57:25 +02:00
|
|
|
* Register @pol with blkcg core. Might sleep and @pol may be modified on
|
|
|
|
* successful registration. Returns 0 on success and -errno on failure.
|
2012-04-13 22:11:28 +02:00
|
|
|
*/
|
2014-06-23 00:31:56 +02:00
|
|
|
int blkcg_policy_register(struct blkcg_policy *pol)
|
2009-12-04 16:36:42 +01:00
|
|
|
{
|
2015-07-09 22:39:50 +02:00
|
|
|
struct blkcg *blkcg;
|
2012-04-13 22:11:28 +02:00
|
|
|
int i, ret;
|
2012-03-05 22:15:20 +01:00
|
|
|
|
2015-07-09 22:39:47 +02:00
|
|
|
mutex_lock(&blkcg_pol_register_mutex);
|
2012-04-13 22:11:26 +02:00
|
|
|
mutex_lock(&blkcg_pol_mutex);
|
|
|
|
|
2012-04-13 22:11:28 +02:00
|
|
|
/* find an empty slot */
|
|
|
|
ret = -ENOSPC;
|
|
|
|
for (i = 0; i < BLKCG_MAX_POLS; i++)
|
2012-04-16 22:57:25 +02:00
|
|
|
if (!blkcg_policy[i])
|
2012-04-13 22:11:28 +02:00
|
|
|
break;
|
|
|
|
if (i >= BLKCG_MAX_POLS)
|
2015-07-09 22:39:47 +02:00
|
|
|
goto err_unlock;
|
2012-03-05 22:15:04 +01:00
|
|
|
|
2015-07-09 22:39:50 +02:00
|
|
|
/* register @pol */
|
2012-04-16 22:57:25 +02:00
|
|
|
pol->plid = i;
|
2015-07-09 22:39:50 +02:00
|
|
|
blkcg_policy[pol->plid] = pol;
|
|
|
|
|
|
|
|
/* allocate and install cpd's */
|
2015-08-18 23:55:16 +02:00
|
|
|
if (pol->cpd_alloc_fn) {
|
2015-07-09 22:39:50 +02:00
|
|
|
list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
|
|
|
|
struct blkcg_policy_data *cpd;
|
|
|
|
|
2015-08-18 23:55:16 +02:00
|
|
|
cpd = pol->cpd_alloc_fn(GFP_KERNEL);
|
2016-09-29 17:33:30 +02:00
|
|
|
if (!cpd)
|
2015-07-09 22:39:50 +02:00
|
|
|
goto err_free_cpds;
|
|
|
|
|
2015-08-18 23:55:15 +02:00
|
|
|
blkcg->cpd[pol->plid] = cpd;
|
|
|
|
cpd->blkcg = blkcg;
|
2015-07-09 22:39:50 +02:00
|
|
|
cpd->plid = pol->plid;
|
2015-08-18 23:55:15 +02:00
|
|
|
pol->cpd_init_fn(cpd);
|
2015-07-09 22:39:50 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-07-09 22:39:47 +02:00
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
2012-04-13 22:11:28 +02:00
|
|
|
|
|
|
|
/* everything is in place, add intf files for the new policy */
|
2015-08-18 23:55:34 +02:00
|
|
|
if (pol->dfl_cftypes)
|
|
|
|
WARN_ON(cgroup_add_dfl_cftypes(&io_cgrp_subsys,
|
|
|
|
pol->dfl_cftypes));
|
2015-08-18 23:55:30 +02:00
|
|
|
if (pol->legacy_cftypes)
|
blkcg: rename subsystem name from blkio to io
blkio interface has become messy over time and is currently the
largest. In addition to the inconsistent naming scheme, it has
multiple stat files which report more or less the same thing, a number
of debug stat files which expose internal details which shouldn't have
been part of the public interface in the first place, recursive and
non-recursive stats and leaf and non-leaf knobs.
Both recursive vs. non-recursive and leaf vs. non-leaf distinctions
don't make any sense on the unified hierarchy as only leaf cgroups can
contain processes. cgroups is going through a major interface
revision with the unified hierarchy involving significant fundamental
usage changes and given that a significant portion of the interface
doesn't make sense anymore, it's a good time to reorganize the
interface.
As the first step, this patch renames the external visible subsystem
name from "blkio" to "io". This is more concise, matches the other
two major subsystem names, "cpu" and "memory", and better suited as
blkcg will be involved in anything writeback related too whether an
actual block device is involved or not.
As the subsystem legacy_name is set to "blkio", the only userland
visible change outside the unified hierarchy is that blkcg is reported
as "io" instead of "blkio" in the subsystem initialized message during
boot. On the unified hierarchy, blkcg now appears as "io".
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Li Zefan <lizefan@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: cgroups@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-18 23:55:29 +02:00
|
|
|
WARN_ON(cgroup_add_legacy_cftypes(&io_cgrp_subsys,
|
2015-08-18 23:55:30 +02:00
|
|
|
pol->legacy_cftypes));
|
2015-07-09 22:39:47 +02:00
|
|
|
mutex_unlock(&blkcg_pol_register_mutex);
|
|
|
|
return 0;
|
|
|
|
|
2015-07-09 22:39:50 +02:00
|
|
|
err_free_cpds:
|
2015-08-18 23:55:16 +02:00
|
|
|
if (pol->cpd_alloc_fn) {
|
2015-07-09 22:39:50 +02:00
|
|
|
list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
|
2015-08-18 23:55:16 +02:00
|
|
|
if (blkcg->cpd[pol->plid]) {
|
|
|
|
pol->cpd_free_fn(blkcg->cpd[pol->plid]);
|
|
|
|
blkcg->cpd[pol->plid] = NULL;
|
|
|
|
}
|
2015-07-09 22:39:50 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
blkcg_policy[pol->plid] = NULL;
|
2015-07-09 22:39:47 +02:00
|
|
|
err_unlock:
|
2012-04-13 22:11:26 +02:00
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
2015-07-09 22:39:47 +02:00
|
|
|
mutex_unlock(&blkcg_pol_register_mutex);
|
2012-04-13 22:11:28 +02:00
|
|
|
return ret;
|
2009-12-04 16:36:42 +01:00
|
|
|
}
|
2012-04-16 22:57:25 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkcg_policy_register);
|
2009-12-04 16:36:42 +01:00
|
|
|
|
2012-04-13 22:11:28 +02:00
|
|
|
/**
|
2012-04-16 22:57:25 +02:00
|
|
|
* blkcg_policy_unregister - unregister a blkcg policy
|
|
|
|
* @pol: blkcg policy to unregister
|
2012-04-13 22:11:28 +02:00
|
|
|
*
|
2012-04-16 22:57:25 +02:00
|
|
|
* Undo blkcg_policy_register(@pol). Might sleep.
|
2012-04-13 22:11:28 +02:00
|
|
|
*/
|
2012-04-16 22:57:25 +02:00
|
|
|
void blkcg_policy_unregister(struct blkcg_policy *pol)
|
2009-12-04 16:36:42 +01:00
|
|
|
{
|
2015-07-09 22:39:50 +02:00
|
|
|
struct blkcg *blkcg;
|
|
|
|
|
2015-07-09 22:39:47 +02:00
|
|
|
mutex_lock(&blkcg_pol_register_mutex);
|
2012-04-13 22:11:26 +02:00
|
|
|
|
2012-04-16 22:57:25 +02:00
|
|
|
if (WARN_ON(blkcg_policy[pol->plid] != pol))
|
2012-04-13 22:11:28 +02:00
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
/* kill the intf files first */
|
2015-08-18 23:55:34 +02:00
|
|
|
if (pol->dfl_cftypes)
|
|
|
|
cgroup_rm_cftypes(pol->dfl_cftypes);
|
2015-08-18 23:55:30 +02:00
|
|
|
if (pol->legacy_cftypes)
|
|
|
|
cgroup_rm_cftypes(pol->legacy_cftypes);
|
2012-04-01 23:38:43 +02:00
|
|
|
|
2015-07-09 22:39:50 +02:00
|
|
|
/* remove cpds and unregister */
|
2015-07-09 22:39:47 +02:00
|
|
|
mutex_lock(&blkcg_pol_mutex);
|
2015-07-09 22:39:50 +02:00
|
|
|
|
2015-08-18 23:55:16 +02:00
|
|
|
if (pol->cpd_alloc_fn) {
|
2015-07-09 22:39:50 +02:00
|
|
|
list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
|
2015-08-18 23:55:16 +02:00
|
|
|
if (blkcg->cpd[pol->plid]) {
|
|
|
|
pol->cpd_free_fn(blkcg->cpd[pol->plid]);
|
|
|
|
blkcg->cpd[pol->plid] = NULL;
|
|
|
|
}
|
2015-07-09 22:39:50 +02:00
|
|
|
}
|
|
|
|
}
|
2012-04-16 22:57:25 +02:00
|
|
|
blkcg_policy[pol->plid] = NULL;
|
2015-07-09 22:39:50 +02:00
|
|
|
|
2012-04-13 22:11:26 +02:00
|
|
|
mutex_unlock(&blkcg_pol_mutex);
|
2015-07-09 22:39:47 +02:00
|
|
|
out_unlock:
|
|
|
|
mutex_unlock(&blkcg_pol_register_mutex);
|
2009-12-04 16:36:42 +01:00
|
|
|
}
|
2012-04-16 22:57:25 +02:00
|
|
|
EXPORT_SYMBOL_GPL(blkcg_policy_unregister);
|