Hello, Viresh.
On Tue, Nov 06, 2012 at 04:08:45PM +0530, Viresh Kumar wrote:
> Workqueues queues work on current cpu, if the caller haven't passed a
> preferred
> cpu. This may wake up an idle CPU, which is actually not required.
>
> This work can be processed by any CPU and so we must select a
Hello, guys.
I'm wondering why cpuset_cpus_allowed_fallback() is necessary. This
is called from, e.g., try_to_wake_up()->select_task_rq() when none of
the cpus in ->cpus_allowed is useable. The cpuset callback invokes
do_set_cpus_allowed() w/ the cpuset's cpus_allowed. This was added by
the fol
On Wed, Nov 28, 2012 at 07:17:18AM +0100, Mike Galbraith wrote:
>
> A rescue thread exiting TASK_INTERRUPTIBLE can lead to a task scheduling
> off, never to be seen again. In the case where this occurred, an exiting
> thread hit reiserfs homebrew conditional resched while holding a mutex,
> bring
Hey, Greg.
On Wed, Nov 28, 2012 at 10:26:32AM -0800, Greg Thelen wrote:
> Before this patch init_cgroup_root() created a degenerate list by
> first inserting a element into allcg_list and then initializing the
> inserted list element. The initialization reset the element's
> prev/next fields form
Hello, Greg.
On Wed, Nov 28, 2012 at 12:15:42PM -0800, Greg Thelen wrote:
> @@ -4276,6 +4276,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)
> DEFINE_WAIT(wait);
> struct cgroup_event *event, *tmp;
> struct cgroup_subsys *ss;
> + struct list_head tmp_list;
LIST_HE
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 11 ---
1 file changed, 11 deletions(-)
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index b017887..a423774 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -2412,17 +2412,6 @@ int __cpuset_node_allowed_hardwall(int node, gfp_t
Add CS_ONLINE which is set from css_online() and cleared from
css_offline(). This will enable using generic cgroup iterator while
allowing decoupling cpuset from cgroup internal locking.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 11 ++-
1 file changed, 10 insertions(+), 1 deletion
n't visible to all the iterations and this patch currently doesn't
make any functional difference. This will be used to de-couple cpuset
locking from cgroup core.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 85 -
1 fil
own+0x36/0x50
[] store_online+0x5d/0xe0
[] dev_attr_store+0x18/0x30
[] sysfs_write_file+0xe0/0x150
[] vfs_write+0xa8/0x160
[] sys_write+0x52/0xa0
[] system_call_fastpath+0x16/0x1b
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 28
1 file changed, 24 insertions(+)
e() treats cpusets w/
non-zero ->attach_in_progress like cpusets w/ tasks and refuses to
remove all cpus or mems from it.
This currently doesn't make any functional difference as everything is
protected by cgroup_mutex but enables decoupling the locking.
Signed-off-by: Tejun Heo
.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 186
1 file changed, 106 insertions(+), 80 deletions(-)
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 79be3f0..2ee0e03 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -208,23
g.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 14 ++
1 file changed, 14 insertions(+)
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 68a0906..79be3f0 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -266,6 +266,7 @@ static struct workqueue_struct *cpuset_propagate_hotplug_w
while holding cgroup_mutex and waits for
completion without cgroup_mutex. Each in-flight propagation holds a
reference to the cpuset->css.
This patch doesn't cause any functional difference.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 54 -
The function isn't that hot, the overhead of missing the fast exit is
low, the test itself depends heavily on cgroup internals, and it's
gonna be a hindrance when trying to decouple cpuset locking from
cgroup core. Remove the fast exit path.
Signed-off-by: Tejun Heo
---
kernel/cp
cpuset_attach()
and make the global variables static ones inside cpuset_attach().
While at it, convert cpus_attach to cpumask_t from cpumask_var_t.
There's no reason to mess with dynamic allocation on a static buffer.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c
le_hotplug() can
handle multiple resources going up and down. These properties will
allow async operation.
The reorganization, while drastic, is equivalent and shouldn't cause
any behavior difference. This will enable making hotplug handling
async and remove get_online_cpus() -> cgroup_mut
e CPU / memory hotplug path still grabs the two locks in
the reverse order and thus this is a deadlock hazard; however, the two
locks are already deadlock-prone and the hotplug path will be updated
by further patches.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c
() frees.
This doesn't introduce any visible behavior changes. This will help
cleaning up locking.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 66 ++---
1 file changed, 44 insertions(+), 22 deletions(-)
diff --git a/kernel/cpuset.c b/k
Hello, guys.
Depending on cgroup core locking - cgroup_mutex - is messy and makes
cgroup prone to locking dependency problems. The current code already
has lock dependency loop - memcg nests get_online_cpus() inside
cgroup_mutex. cpuset the other way around.
Regardless of the locking details, w
On Wed, Nov 28, 2012 at 01:50:45PM -0800, Greg Thelen wrote:
> Use list_del_init() rather than list_del() to remove events from
> cgrp->event_list. No functional change. This is just defensive
> coding.
>
> Signed-off-by: Greg Thelen
Applied 1-2 to cgroup/for-3.8. Thanks!
--
tejun
--
To uns
Hello, guys.
cpuset implements its own descendant iteration using
cpuset->stack_list and has its own ->parent pointer. There's nothing
cpuset specific about descendant walking or finding the parent. This
patchset makes cpuset use cgroup generic API instead.
0001-cpuset-implement-cgroup_rightmo
Implement cpuset_for_each_descendant_pre() and replace the
cpuset-specific tree walking using cpuset->stack_list with it.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 123 ++--
1 file changed, 48 insertions(+), 75 deletions(-)
diff --gi
cgroup already tracks the hierarchy. Follow cgroup->parent to find
the parent and drop cpuset->parent.
Signed-off-by: Tejun Heo
---
kernel/cpuset.c | 28 +---
1 file changed, 17 insertions(+), 11 deletions(-)
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 3
Implement cgroup_rightmost_descendant() which returns the right most
descendant of the specified cgroup. This can be used to skip the
cgroup's subtree while iterating with
cgroup_for_each_descendant_pre().
Signed-off-by: Tejun Heo
Cc: Michal Hocko
---
include/linux/cgroup.h | 1 +
k
Hello, Glauber.
On Thu, Nov 29, 2012 at 03:14:41PM +0400, Glauber Costa wrote:
> On 11/29/2012 01:34 AM, Tejun Heo wrote:
> > This patchset decouples cpuset locking from cgroup_mutex. After the
> > patchset, cpuset uses cpuset-specific cpuset_mutex instead of
> > cgroup_mute
On Thu, Nov 29, 2012 at 06:26:46AM -0800, Tejun Heo wrote:
> > What I'll try to do, is to come with another specialized lock in cgroup
> > just for this case. So after taking the cgroup lock, we would also take
> > an extra lock if we are adding another entry - be it task o
Hello, Steven.
On Mon, Mar 18, 2013 at 10:36:23AM -0400, Steven Rostedt wrote:
> kernel BUG at kernel/sched/core.c:1731!
> invalid opcode: [#1] PREEMPT SMP
> CPU 5
> Pid: 16637, comm: kworker/5:0 Not tainted 3.6.11-rt30.25.el6rt.x86_64 #1 HP
> ProLiant DL580 G7
...
> static void try_to_wak
Hey, Steven.
On Mon, Mar 18, 2013 at 12:23:19PM -0400, Steven Rostedt wrote:
> > Maybe I'm confused but I can't really see how the above would be a
> > problem to workqueue in itself. Both rq->lock and gcwq->lock are
> > irq-safe, so spin_lock() not disabling preemption shouldn't be a
> > problem
Hello, Steven.
On Mon, Mar 18, 2013 at 12:30:43PM -0400, Steven Rostedt wrote:
> If you happen to know the critical areas that require preemption to be
> disabled for real, we can encapsulate them with:
>
> preempt_disable_rt();
>
> preempt_enable_rt();
>
> These are currently only
On Mon, Mar 18, 2013 at 12:41:23PM -0400, Steven Rostedt wrote:
> But, I'm worried about the loops that are done while holding this lock.
> Just looking at is_chained_work() that does for_each_busy_worker(), how
> big can that list be? If it's bound by # of CPUs then that may be fine,
> but if it c
Hello,
On Mon, Mar 18, 2013 at 10:41:40AM +0100, Ingo Molnar wrote:
> > This patch replaces PF_THREAD_BOUND with PF_NO_SETAFFINITY.
> > sched_setaffinity() checks the flag and return -EINVAL if set.
> > set_cpus_allowed_ptr() is no longer affected by the flag.
> >
> > This will allow simplifying
On Mon, Mar 18, 2013 at 02:23:56PM -0400, Steven Rostedt wrote:
> On Mon, 2013-03-18 at 09:43 -0700, Tejun Heo wrote:
> > Hello, Steven.
> >
> > On Mon, Mar 18, 2013 at 12:30:43PM -0400, Steven Rostedt wrote:
> > > If you happen to know the critical areas
On Mon, Mar 18, 2013 at 01:08:07PM -0400, Steven Rostedt wrote:
> On Mon, 2013-03-18 at 09:43 -0700, Tejun Heo wrote:
>
> > Making gcwq locks disable preemption would be much safer / easier, but
> > if that's not desirable, anything touching gcwq->idle_list would be
On Mon, Mar 18, 2013 at 02:57:30PM -0400, Steven Rostedt wrote:
> I like the theory, but it has one flaw. I agree that the update should
> be wrapped in preempt_disable() but since this bug happens on the same
> CPU, the state of the list will be the same when it was preempted to
> when it bugged.
getting unblocked. There's no reason to
trigger BUG while holding rq lock crashing the whole system.
Convert BUG_ON()s in try_to_wake_up_local() to WARN_ON_ONCE()s.
Signed-off-by: Tejun Heo
Cc: Steven Rostedt
---
kernel/sched/core.c |6 --
1 file changed, 4 insertions(+), 2 dele
Hello, Linus.
Please pull from the following branch to receive Lai's patch to fix
highly unlikely but still possible workqueue stall during CPU
hotunplug.
git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git for-3.9-fixes
Thanks.
On Wed, Mar 13, 2013 at 07:54:01PM -0700, Tejun Heo wrote:
> On Wed, Mar 13, 2013 at 04:58:12PM -0700, Tejun Heo wrote:
> > This patch is on top of wq/for-3.10 e626761691 ("workqueue: implement
> > current_is_workqueue_rescuer()").
>
> Oops, forgot the git branch.
On Wed, Mar 13, 2013 at 07:57:18PM -0700, Tejun Heo wrote:
> and available in the following git branch.
>
> git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git review-finer-locking
Applied to wq/for-3.10.
Thanks.
--
tejun
--
To unsubscribe from this list: send the line "uns
Hello, Jan.
On Mon, Mar 18, 2013 at 11:32:44PM +0100, Jan Kara wrote:
> I realized there may be one issue - so far we have a clear identification
> which thread works for which bdi in the thread name (flush-x:y naming).
> That was useful when debugging things. Now with your worker pool this is
>
return NULL;
> > -
> > - t = atomic_cmpxchg(&sd->s_active, v, v + 1);
> > - if (likely(t == v))
> > - break;
> > - if (t < 0)
> > - return NULL;
> > -
> > -
Hello, Jan.
On Tue, Mar 19, 2013 at 8:46 AM, Jan Kara wrote:
> Well, but what you often get is just output of sysrq-w, or sysrq-t, or
> splat from scheduler about stuck task. You often don't have the comfort of
> tracing... Can't we somehow change 'comm' of the task when it starts
> processing
On Mon, Mar 18, 2013 at 09:47:15AM -0700, Tejun Heo wrote:
> Hello,
>
> On Mon, Mar 18, 2013 at 10:41:40AM +0100, Ingo Molnar wrote:
> > > This patch replaces PF_THREAD_BOUND with PF_NO_SETAFFINITY.
> > > sched_setaffinity() checks the flag and return -EINVAL if set.
On Thu, Mar 14, 2013 at 04:01:25PM -0700, Tejun Heo wrote:
> 0001-sched-replace-PF_THREAD_BOUND-with-PF_NO_SETAFFINITY.patch
> 0002-workqueue-convert-worker_pool-worker_ida-to-idr-and-.patch
> 0003-workqueue-relocate-rebind_workers.patch
> 0004-workqueue-directly-restore-CPU
>From a0265a7f5161b6cb55e82b71edb236bbe0d9b3ae Mon Sep 17 00:00:00 2001
From: Tejun Heo
Date: Tue, 19 Mar 2013 13:55:42 -0700
699ce097efe ("workqueue: implement and use pwq_adjust_max_active()")
replaced the only workqueue_freezing usage outside freezer callbacks
with a POOL_FREEZIN
Hey, Stephen.
On Tue, Mar 19, 2013 at 01:19:38PM +1100, Stephen Rothwell wrote:
> @@@ -456,40 -462,30 +462,30 @@@ static int worker_pool_assign_id(struc
> {
> int ret;
>
> - mutex_lock(&worker_pool_idr_mutex);
> - ret = idr_alloc(&worker_pool_idr, pool, 0, 0, GFP_KERNEL);
> -
On Fri, Feb 15, 2013 at 11:55:43AM -0500, Aristeu Rozanski wrote:
> This patchset implements device cgroup hierarchy. Exceptions will be
> propagated down in the tree and local preferences will be re-evaluated
> everytime a change in its parent occours, reapplying them if it's still
> possible.
Ap
On Wed, Mar 13, 2013 at 09:17:09AM +0800, Li Zefan wrote:
> These two functions share most of the code.
>
> Signed-off-by: Li Zefan
Applied to cgroup/for-3.10.
Thanks.
--
tejun
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vge
rsion.
cgroup_mutex is no longer abused by controllers and can be put outer
to threadgroup_lock. Reverse the locking order in
attach_task_by_pid().
Signed-off-by: Tejun Heo
Cc: Li Zefan
---
Li, can you please ack this?
Thanks!
kernel/cgroup.c | 21 -
1 file changed, 8 inser
On Wed, Mar 20, 2013 at 09:05:40AM +1100, Stephen Rothwell wrote:
> > Anyways, I pulled master into wq/for-next and resolved it there, so it
> > shouldn't cause you any more trouble.
>
> Ah, OK, thanks. One small point, when you do a back merge like that,
> you should always put an explanation i
duled NUMA awareness support. Let's drop the non-essential 'H'
postfix from unbound kworker name.
While at it, restructure kthread_create*() invocation to help future
NUMA related changes.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 15 ---
1 file chang
k contained inside single NUMA node, but this will serve as
foundation for making all unbound pools NUMA-affine.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 18 --
1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
inde
is implemented.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 151ce49..25dab9d 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -230,8 +230,6 @@ struct wq_d
e any functional changes.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 75 +-
1 file changed, 52 insertions(+), 23 deletions(-)
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 3f820a5..bbbfc92 100644
--- a/kernel/workqueue.c
_workqueue as first_pwq() so this patch doesn't make any behavior
changes.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 48 +---
1 file changed, 37 insertions(+), 11 deletions(-)
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 25dab9d..3f
#x27;t part of a pool's attributes. It only
affects how apply_workqueue_attrs() picks which pools to use.
After "pool_ids" change, first_pwq() doesn't have any user left.
Removed.
Signed-off-by: Tejun Heo
---
Documentation/kernel-parameters.txt | 9 +++
include/l
d from workqueue users by the number of concurrently
active work items and this change shouldn't matter much.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 108 ++---
1 file changed, 86 insertions(+), 22 deletions(-)
diff --git a/kernel
Hello,
There are two types of workqueues - per-cpu and unbound. The former
is bound to each CPU and the latter isn't not bound to any by default.
While the recently added attrs support allows unbound workqueues to be
confined to subset of CPUs, it still is quite cumbersome for
applications where
.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 35 ++-
1 file changed, 34 insertions(+), 1 deletion(-)
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 775c2f4..9b096e3 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -44,6 +44,7 @@
#include
the above
assumption will no longer hold. Introduce workqueue->unbound_attrs
which records the current attrs in effect and use it for sysfs instead
of first_pwq()->attrs.
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 36
1 file changed, 24 insertions(
Currently workqueue->name[] is of flexible length. We want to use the
flexible field for something more useful and there isn't much benefit
in allowing arbitrary name length anyway. Make it fixed len capping
at 24 bytes.
Signed-off-by: Tejun Heo
---
kernel/workqueu
Hey,
On Wed, Mar 20, 2013 at 11:01:50PM +0900, JoonSoo Kim wrote:
> 2013/3/19 Tejun Heo :
> > On Wed, Mar 13, 2013 at 07:57:18PM -0700, Tejun Heo wrote:
> >> and available in the following git branch.
> >>
> >> git://git.kernel.org/pub/scm/linux/kernel/git/
On Wed, Mar 20, 2013 at 11:08:29PM +0900, JoonSoo Kim wrote:
> 2013/3/20 Tejun Heo :
> > Unbound workqueues are going to be NUMA-affine. Add wq_numa_tbl_len
> > and wq_numa_possible_cpumask[] in preparation. The former is the
> > highest NUMA node ID + 1 and the latter is ma
On Wed, Mar 20, 2013 at 03:43:28PM +1100, Stephen Rothwell wrote:
> It has returned today. Please be more careful.
>
> In file included from include/linux/memcontrol.h:22:0,
> from include/linux/swap.h:8,
> from include/linux/suspend.h:4,
> from
On Wed, Mar 20, 2013 at 08:58:08AM +0800, Li Zefan wrote:
> On 2013/3/20 6:02, Tejun Heo wrote:
> > It doesn't make sense to nest cgroup_mutex inside threadgroup_lock
> > when it should be outer to most all locks used by all cgroup
> > controllers. It was nested in
On Wed, Mar 20, 2013 at 11:03:53PM +0800, Lai Jiangshan wrote:
> > +enomem:
> > + free_workqueue_attrs(tmp_attrs);
> > + if (pwq_tbl) {
> > + for_each_node(node)
> > + kfree(pwq_tbl[node]);
>
> It will free the dfl_pwq multi times.
Oops, you're righ
On Wed, Mar 20, 2013 at 8:26 AM, Lai Jiangshan wrote:
>> for_eahc_node(node)
>> if (pwq_tbl[node] != dfl_pwq)
>> kfree(pwq_tbl[node]);
>> kfree(dfl_pwq);
>
> I also missed.
> we still need to put_unbound_pool() before free(pwq)
Yeap, we do.
On Wed, Mar 20, 2013 at 11:43:30PM +0800, Lai Jiangshan wrote:
> > + for_each_node(node)
> > +
> > BUG_ON(!alloc_cpumask_var_node(&wq_numa_possible_cpumask[node],
> > + GFP_KERNEL, node));
> > + for_each_possible_cpu(cpu) {
> >
On Wed, Mar 20, 2013 at 11:52:03PM +0800, Lai Jiangshan wrote:
> > +static struct pool_workqueue *alloc_unbound_pwq(struct workqueue_struct
> > *wq,
> > + const struct workqueue_attrs *attrs)
> > +{
> > + struct worker_pool *pool;
> > + struct pool
Hey,
On Thu, Mar 21, 2013 at 12:38:17AM +0800, Lai Jiangshan wrote:
> I am sorry for replying so late and replied with so huge patchset.
>
> But problem happened now, my patches and your patches are conflict.
> Which patchset should be rebased?
>
> I think my patches need be merged at first. Thu
d from workqueue users by the number of concurrently
active work items and this change shouldn't matter much.
v2: Fixed pwq freeing in apply_workqueue_attrs() error path. Spotted
by Lai.
Signed-off-by: Tejun Heo
Cc: Lai Jiangshan
---
kern
shan.
Signed-off-by: Tejun Heo
Cc: Lai Jiangshan
---
kernel/workqueue.c |6 --
1 file changed, 4 insertions(+), 2 deletions(-)
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -3683,12 +3683,14 @@ static void pwq_adjust_max_active(struct
spin_unlock(&pwq->pool->loc
On Wed, Mar 20, 2013 at 03:28:01AM +0800, Lai Jiangshan wrote:
> When we create a new pool via get_unbound_pool() when freezing.
> the pool->flags' POOL_FREEZING is incorrect.
>
> Fix it by adding POOL_FREEZING if workqueue_freezing.
> (wq_mutex is already held for accessing workqueue_freezing.)
>
ing POOL_FREEZING if workqueue_freezing. wq_mutex is
already held so no further locking is necessary. This also removes
the unused static variable warning when !CONFIG_FREEZER.
tj: Updated commit message.
Signed-off-by: Lai Jiangshan
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 3 +++
1 f
On Wed, Mar 20, 2013 at 03:28:02AM +0800, Lai Jiangshan wrote:
> pool->worker_idr nor worker->id is not protected by RCU.
> don't need to free pool->worker_idr by RCU.
>
> Just free it directly.
>
> Signed-off-by: Lai Jiangshan
...
> @@ -3462,6 +3461,7 @@ static void put_unbound_pool(struct work
On Wed, Mar 20, 2013 at 03:28:03AM +0800, Lai Jiangshan wrote:
> current_is_workqueue_rescuer() <-> current is worker and worker->rescue_wq
>
> Signed-off-by: Lai Jiangshan
Applied to wq/for-3.10.
--
tejun
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body o
On Wed, Mar 20, 2013 at 03:28:04AM +0800, Lai Jiangshan wrote:
> "if (!freezable || !(pwq->pool->flags & POOL_FREEZING))" is hard to read.
>
> Swap the two branches. it becomes
> "if (freezable && (pwq->pool->flags & POOL_FREEZING))", it is better.
>
> Signed-off-by: Lai Jiangshan
> ---
> kerne
both of the above two
cases. This also makes thaw_workqueues() simpler.
tj: Updated comments and description.
Signed-off-by: Lai Jiangshan
Signed-off-by: Tejun Heo
---
kernel/workqueue.c | 13 ++---
1 file changed, 6 insertions(+), 7 deletions(-)
diff --git a/kernel/workqueue.c
On Wed, Mar 20, 2013 at 03:28:06AM +0800, Lai Jiangshan wrote:
> currently wq_mutext protects:
>
> * worker_pool_idr and unbound_pool_hash
> * pool->refcnt
> * workqueues list
> * workqueue->flags, ->nr_drainers
> * workqueue_freezing
>
> We can see that it protects very different things.
> So we
On Wed, Mar 20, 2013 at 03:28:10AM +0800, Lai Jiangshan wrote:
> rcu_read_lock_sched() is better than preempt_disable() if the code is
> protected by RCU_SCHED.
>
> Signed-off-by: Lai Jiangshan
Applied to wq/for-3.10.
Thanks.
--
tejun
--
To unsubscribe from this list: send the line "unsubscri
On Wed, Mar 20, 2013 at 03:28:15AM +0800, Lai Jiangshan wrote:
> static struct worker *alloc_worker(void)
> {
> struct worker *worker;
> @@ -2326,7 +2262,8 @@ repeat:
> spin_unlock_irq(&wq_mayday_lock);
>
> /* migrate to the target cpu if possible */
> -
On Wed, Mar 20, 2013 at 03:28:17AM +0800, Lai Jiangshan wrote:
> If we have 4096 CPUs, workqueue_cpu_up_callback() will travel too much CPUs,
> to avoid it, we use for_each_cpu_worker_pool for the cpu pools and
> use unbound_pool_hash for unbound pools.
>
> After it, for_each_pool() becomes unused
On Wed, Mar 20, 2013 at 03:28:18AM +0800, Lai Jiangshan wrote:
> Simply move it to pool->lock C.S.
Patch description should explain *why* in addition to what. Also,
please don't use abbreviations like C.S. Yeah, I know it's critical
section but it might as well be counter strike or context switc
On Wed, Mar 20, 2013 at 03:28:21AM +0800, Lai Jiangshan wrote:
> If lockdep complains something for other subsystem, lockdep_is_held() can be
> false negative. so we need to also test debug_locks before do assert.
>
> Signed-off-by: Lai Jiangshan
Applied to wq/for-3.10.
Thanks.
--
tejun
--
To
So, overall,
On Wed, Mar 20, 2013 at 03:28:00AM +0800, Lai Jiangshan wrote:
...
> In this list, we can find that:
> 1) wq_mutex protects too much different kind of things.
I don't agree with this and unless you can come up with a much better
reason, won't be splitting wq_mutex further. Also, I'm
On Wed, Mar 20, 2013 at 11:35 AM, Oleg Nesterov wrote:
> then we need "do not abuse ->cred_guard_mutex in threadgroup_lock()"
> acked by you and Li. Please let me know if I should resend it.
Yeah, we want that one regardless of this one. Please feel free to add
my Acked-by (if I hadn't acked alre
d from workqueue users by the number of concurrently
active work items and this change shouldn't matter much.
v2: Fixed pwq freeing in apply_workqueue_attrs() error path. Spotted
by Lai.
Signed-off-by: Tejun Heo
Cc: Lai Jiangshan
---
Please forget about the previous posting. It was fre
On Tue, Mar 19, 2013 at 05:00:19PM -0700, Tejun Heo wrote:
> and also available in the following git branch.
>
> git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git review-numa
Branch rebased on top of the current wq/for-3.10 with updated patches.
The new co
On Wed, Mar 20, 2013 at 3:09 PM, Stephen Rothwell wrote:
> There has been no change to the cgroup tree
> (git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git#for-next).
> Forgot to push?
Yeah, I'm on a roll today. My apologies. Just pushed out.
Thanks.
--
tejun
--
To unsubscribe from t
Hello, Viresh.
On Mon, Mar 18, 2013 at 08:53:25PM +0530, Viresh Kumar wrote:
> queue_work() queues work on current cpu. This may wake up an idle CPU, which
> is
> actually not required.
>
> Some of these works can be processed by any CPU and so we must select a
> non-idle
> CPU here. The initia
Hello, Dave.
On Thu, Mar 21, 2013 at 12:57:21PM +1100, Dave Chinner wrote:
> When you have a system that has 50+ active filesystems (pretty
> common in the distributed storage environments were every disk has
> it's own filesystem), knowing which filesystem(s) are getting stuck
> in writeback from
igned offset - ie. use round_up(id + 1, slot_distance) instead of
id += slot_distance.
Signed-off-by: Tejun Heo
Reported-by: David Teigland
Cc: KAMEZAWA Hiroyuki
---
lib/idr.c |9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/lib/idr.c b/lib/idr.c
index 6482390..ca
On Sat, Feb 02, 2013 at 03:10:48PM -0800, Tejun Heo wrote:
> Fix it by ensuring proceeding to the next slot doesn't carry over the
> unaligned offset - ie. use round_up(id + 1, slot_distance) instead of
> id += slot_distance.
>
> Signed-off-by: Tejun Heo
> Reported-b
Hello,
* Andrew, I think this one is best routed through -mm together with
the previous series. Please read on.
* Bruce, I couldn't convert nfsd. Can you please help? More on it
later.
* Stanislav, Eric, James, can you please take a look at ipc/util.c
conversion? I moved allocation ins
Move slot filling to idr_fill_slot() from idr_get_new_above_int() and
make idr_get_new_above() directly call it. idr_get_new_above_int() is
no longer needed and removed.
This will be used to implement a new ID allocation interface.
Signed-off-by: Tejun Heo
---
lib/idr.c | 30
idr allocation in blk_alloc_devt() wasn't synchronized against lookup
and removal, and its limit check was off by one - 1 << MINORBITS is
the number of minors allowed, not the maximum allowed minor.
Add locking and rename MAX_EXT_DEVT to NR_EXT_DEVT and fix limit
checking.
Signed-of
Convert to the much saner new idr interface.
Only compile tested.
* drm_ctxbitmap_next() error handling in drm_addctx() seems broken.
drm_ctxbitmap_next() return -errno on failure not -1.
Signed-off-by: Tejun Heo
Cc: David Airlie
Cc: dri-de...@lists.freedesktop.org
---
This patch depends on
Convert to the much saner new idr interface.
Only compile tested.
Signed-off-by: Tejun Heo
Cc: Roland Dreier
Cc: Sean Hefty
Cc: Hal Rosenstock
Cc: linux-r...@vger.kernel.org
---
This patch depends on an earlier idr changes and I think it would be
best to route these together through -mm
Convert to the much saner new idr interface.
Only compile tested.
Signed-off-by: Tejun Heo
Cc: Jack Morgenstein
Cc: Or Gerlitz
Cc: Roland Dreier
Cc: linux-r...@vger.kernel.org
---
This patch depends on an earlier idr changes and I think it would be
best to route these together through -mm
Convert to the much saner new idr interface.
Only compile tested.
Signed-off-by: Tejun Heo
Cc: Samuel Ortiz
---
This patch depends on an earlier idr changes and I think it would be
best to route these together through -mm. Please holler if there's
any objection. Thanks.
driver
Convert to the much saner new idr interface.
Only compile tested.
Signed-off-by: Tejun Heo
Cc: Rodolfo Giometti
---
This patch depends on an earlier idr changes and I think it would be
best to route these together through -mm. Please holler if there's
any objection. Thanks.
driver
1 - 100 of 11394 matches
Mail list logo
