* [patch 0/4] sched: Replace read_lock(&tasklist_lock) with RCU - the easy part
@ 2009-12-09 10:14 Thomas Gleixner
2009-12-09 10:14 ` [patch 1/4] sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam() Thomas Gleixner
` (3 more replies)
0 siblings, 4 replies; 8+ messages in thread
From: Thomas Gleixner @ 2009-12-09 10:14 UTC (permalink / raw)
To: LKML; +Cc: Ingo Molnar, Peter Zijlstra
First batch of patches which replace read_lock(&tasklist_lock) with
RCU.
Thanks,
tglx
^ permalink raw reply [flat|nested] 8+ messages in thread
* [patch 1/4] sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam()
2009-12-09 10:14 [patch 0/4] sched: Replace read_lock(&tasklist_lock) with RCU - the easy part Thomas Gleixner
@ 2009-12-09 10:14 ` Thomas Gleixner
2009-12-14 16:31 ` [tip:sched/urgent] " tip-bot for Thomas Gleixner
2009-12-09 10:15 ` [patch 2/4] sched: Use rcu in sched_get/set_affinity() Thomas Gleixner
` (2 subsequent siblings)
3 siblings, 1 reply; 8+ messages in thread
From: Thomas Gleixner @ 2009-12-09 10:14 UTC (permalink / raw)
To: LKML; +Cc: Ingo Molnar, Peter Zijlstra
[-- Attachment #1: sched-use-rcu-in-sched_getscheduler.patch --]
[-- Type: text/plain, Size: 1844 bytes --]
read_lock(&tasklist_lock) does not protect sys_sched_getscheduler and
sys_sched_getparam() against a concurrent update of the policy or
scheduler parameters as do_sched_setscheduler() does not take the
tasklist_lock. The accessed integers can be retrieved w/o locking and
are snapshots anyway.
Using rcu_read_lock() to protect find_task_by_vpid() and prevent the
task struct from going away is not changing the above situation.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
kernel/sched.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
Index: linux-2.6-tip/kernel/sched.c
===================================================================
--- linux-2.6-tip.orig/kernel/sched.c
+++ linux-2.6-tip/kernel/sched.c
@@ -6477,7 +6477,7 @@ SYSCALL_DEFINE1(sched_getscheduler, pid_
return -EINVAL;
retval = -ESRCH;
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
if (p) {
retval = security_task_getscheduler(p);
@@ -6485,7 +6485,7 @@ SYSCALL_DEFINE1(sched_getscheduler, pid_
retval = p->policy
| (p->sched_reset_on_fork ? SCHED_RESET_ON_FORK : 0);
}
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
return retval;
}
@@ -6503,7 +6503,7 @@ SYSCALL_DEFINE2(sched_getparam, pid_t, p
if (!param || pid < 0)
return -EINVAL;
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
retval = -ESRCH;
if (!p)
@@ -6514,7 +6514,7 @@ SYSCALL_DEFINE2(sched_getparam, pid_t, p
goto out_unlock;
lp.sched_priority = p->rt_priority;
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
/*
* This one might sleep, we cannot do it with a spinlock held ...
@@ -6524,7 +6524,7 @@ SYSCALL_DEFINE2(sched_getparam, pid_t, p
return retval;
out_unlock:
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
return retval;
}
^ permalink raw reply [flat|nested] 8+ messages in thread
* [patch 2/4] sched: Use rcu in sched_get/set_affinity()
2009-12-09 10:14 [patch 0/4] sched: Replace read_lock(&tasklist_lock) with RCU - the easy part Thomas Gleixner
2009-12-09 10:14 ` [patch 1/4] sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam() Thomas Gleixner
@ 2009-12-09 10:15 ` Thomas Gleixner
2009-12-14 16:31 ` [tip:sched/urgent] " tip-bot for Thomas Gleixner
2009-12-09 10:15 ` [patch 3/4] sched: Rename bogus label in sched_setaffinity() Thomas Gleixner
2009-12-09 10:15 ` [patch 4/4] sched: Use rcu in sched_get_rr_param Thomas Gleixner
3 siblings, 1 reply; 8+ messages in thread
From: Thomas Gleixner @ 2009-12-09 10:15 UTC (permalink / raw)
To: LKML; +Cc: Ingo Molnar, Peter Zijlstra
[-- Attachment #1: sched-use-rcu-in-affinity-get-set.patch --]
[-- Type: text/plain, Size: 1675 bytes --]
tasklist_lock is held read locked to protect the find_task_by_vpid()
call and to prevent the task going away. sched_setaffinity acquires a
task struct ref and drops tasklist lock right away. The access to the
cpus_allowed mask is protected by rq->lock.
rcu_read_lock() provides the same protection here.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
kernel/sched.c | 16 ++++++----------
1 file changed, 6 insertions(+), 10 deletions(-)
Index: linux-2.6-tip/kernel/sched.c
===================================================================
--- linux-2.6-tip.orig/kernel/sched.c
+++ linux-2.6-tip/kernel/sched.c
@@ -6535,22 +6535,18 @@ long sched_setaffinity(pid_t pid, const
int retval;
get_online_cpus();
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
if (!p) {
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
put_online_cpus();
return -ESRCH;
}
- /*
- * It is not safe to call set_cpus_allowed with the
- * tasklist_lock held. We will bump the task_struct's
- * usage count and then drop tasklist_lock.
- */
+ /* Prevent p going away */
get_task_struct(p);
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
if (!alloc_cpumask_var(&cpus_allowed, GFP_KERNEL)) {
retval = -ENOMEM;
@@ -6636,7 +6632,7 @@ long sched_getaffinity(pid_t pid, struct
int retval;
get_online_cpus();
- read_lock(&tasklist_lock);
+ rcu_read_lock();
retval = -ESRCH;
p = find_process_by_pid(pid);
@@ -6652,7 +6648,7 @@ long sched_getaffinity(pid_t pid, struct
task_rq_unlock(rq, &flags);
out_unlock:
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
put_online_cpus();
return retval;
^ permalink raw reply [flat|nested] 8+ messages in thread
* [patch 3/4] sched: Rename bogus label in sched_setaffinity()
2009-12-09 10:14 [patch 0/4] sched: Replace read_lock(&tasklist_lock) with RCU - the easy part Thomas Gleixner
2009-12-09 10:14 ` [patch 1/4] sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam() Thomas Gleixner
2009-12-09 10:15 ` [patch 2/4] sched: Use rcu in sched_get/set_affinity() Thomas Gleixner
@ 2009-12-09 10:15 ` Thomas Gleixner
2009-12-09 10:15 ` [patch 4/4] sched: Use rcu in sched_get_rr_param Thomas Gleixner
3 siblings, 0 replies; 8+ messages in thread
From: Thomas Gleixner @ 2009-12-09 10:15 UTC (permalink / raw)
To: LKML; +Cc: Ingo Molnar, Peter Zijlstra
[-- Attachment #1: sched-fix-bogus-label.patch --]
[-- Type: text/plain, Size: 952 bytes --]
out_unlock: is bogus as there is nothing to unlock. Rename it.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
kernel/sched.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
Index: linux-2.6-tip/kernel/sched.c
===================================================================
--- linux-2.6-tip.orig/kernel/sched.c
+++ linux-2.6-tip/kernel/sched.c
@@ -6558,11 +6558,11 @@ long sched_setaffinity(pid_t pid, const
}
retval = -EPERM;
if (!check_same_owner(p) && !capable(CAP_SYS_NICE))
- goto out_unlock;
+ goto out_free;
retval = security_task_setscheduler(p, 0, NULL);
if (retval)
- goto out_unlock;
+ goto out_free;
cpuset_cpus_allowed(p, cpus_allowed);
cpumask_and(new_mask, in_mask, cpus_allowed);
@@ -6581,7 +6581,7 @@ long sched_setaffinity(pid_t pid, const
goto again;
}
}
-out_unlock:
+out_free:
free_cpumask_var(new_mask);
out_free_cpus_allowed:
free_cpumask_var(cpus_allowed);
^ permalink raw reply [flat|nested] 8+ messages in thread
* [patch 4/4] sched: Use rcu in sched_get_rr_param
2009-12-09 10:14 [patch 0/4] sched: Replace read_lock(&tasklist_lock) with RCU - the easy part Thomas Gleixner
` (2 preceding siblings ...)
2009-12-09 10:15 ` [patch 3/4] sched: Rename bogus label in sched_setaffinity() Thomas Gleixner
@ 2009-12-09 10:15 ` Thomas Gleixner
2009-12-14 16:31 ` [tip:sched/urgent] sched: Use rcu in sched_get_rr_param() tip-bot for Thomas Gleixner
3 siblings, 1 reply; 8+ messages in thread
From: Thomas Gleixner @ 2009-12-09 10:15 UTC (permalink / raw)
To: LKML; +Cc: Ingo Molnar, Peter Zijlstra
[-- Attachment #1: sched-use-rcu-in-sched-get-rr-param.patch --]
[-- Type: text/plain, Size: 1314 bytes --]
read_lock(&tasklist_lock) does not protect sys_sched_get_rr_param()
against a concurrent update of the policy or scheduler parameters as
do_sched_scheduler() does not take the tasklist_lock.
The access to task->sched_class->get_rr_interval is protected by
task_rq_lock(task).
Use rcu_read_lock() to protect find_task_by_vpid() and prevent the
task struct from going away.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
kernel/sched.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
Index: linux-2.6-tip/kernel/sched.c
===================================================================
--- linux-2.6-tip.orig/kernel/sched.c
+++ linux-2.6-tip/kernel/sched.c
@@ -6892,7 +6892,7 @@ SYSCALL_DEFINE2(sched_rr_get_interval, p
return -EINVAL;
retval = -ESRCH;
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
if (!p)
goto out_unlock;
@@ -6905,13 +6905,13 @@ SYSCALL_DEFINE2(sched_rr_get_interval, p
time_slice = p->sched_class->get_rr_interval(rq, p);
task_rq_unlock(rq, &flags);
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
jiffies_to_timespec(time_slice, &t);
retval = copy_to_user(interval, &t, sizeof(t)) ? -EFAULT : 0;
return retval;
out_unlock:
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
return retval;
}
^ permalink raw reply [flat|nested] 8+ messages in thread
* [tip:sched/urgent] sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam()
2009-12-09 10:14 ` [patch 1/4] sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam() Thomas Gleixner
@ 2009-12-14 16:31 ` tip-bot for Thomas Gleixner
0 siblings, 0 replies; 8+ messages in thread
From: tip-bot for Thomas Gleixner @ 2009-12-14 16:31 UTC (permalink / raw)
To: linux-tip-commits; +Cc: linux-kernel, hpa, mingo, peterz, tglx, mingo
Commit-ID: 5fe85be081edf0ac92d83f9c39e0ab5c1371eb82
Gitweb: http://git.kernel.org/tip/5fe85be081edf0ac92d83f9c39e0ab5c1371eb82
Author: Thomas Gleixner <tglx@linutronix.de>
AuthorDate: Wed, 9 Dec 2009 10:14:58 +0000
Committer: Ingo Molnar <mingo@elte.hu>
CommitDate: Mon, 14 Dec 2009 17:11:34 +0100
sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam()
read_lock(&tasklist_lock) does not protect
sys_sched_getscheduler and sys_sched_getparam() against a
concurrent update of the policy or scheduler parameters as
do_sched_setscheduler() does not take the tasklist_lock. The
accessed integers can be retrieved w/o locking and are snapshots
anyway.
Using rcu_read_lock() to protect find_task_by_vpid() and prevent
the task struct from going away is not changing the above
situation.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20091209100706.753790977@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
kernel/sched.c | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/kernel/sched.c b/kernel/sched.c
index 258c73c..1782bee 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -6458,7 +6458,7 @@ SYSCALL_DEFINE1(sched_getscheduler, pid_t, pid)
return -EINVAL;
retval = -ESRCH;
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
if (p) {
retval = security_task_getscheduler(p);
@@ -6466,7 +6466,7 @@ SYSCALL_DEFINE1(sched_getscheduler, pid_t, pid)
retval = p->policy
| (p->sched_reset_on_fork ? SCHED_RESET_ON_FORK : 0);
}
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
return retval;
}
@@ -6484,7 +6484,7 @@ SYSCALL_DEFINE2(sched_getparam, pid_t, pid, struct sched_param __user *, param)
if (!param || pid < 0)
return -EINVAL;
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
retval = -ESRCH;
if (!p)
@@ -6495,7 +6495,7 @@ SYSCALL_DEFINE2(sched_getparam, pid_t, pid, struct sched_param __user *, param)
goto out_unlock;
lp.sched_priority = p->rt_priority;
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
/*
* This one might sleep, we cannot do it with a spinlock held ...
@@ -6505,7 +6505,7 @@ SYSCALL_DEFINE2(sched_getparam, pid_t, pid, struct sched_param __user *, param)
return retval;
out_unlock:
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
return retval;
}
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [tip:sched/urgent] sched: Use rcu in sched_get/set_affinity()
2009-12-09 10:15 ` [patch 2/4] sched: Use rcu in sched_get/set_affinity() Thomas Gleixner
@ 2009-12-14 16:31 ` tip-bot for Thomas Gleixner
0 siblings, 0 replies; 8+ messages in thread
From: tip-bot for Thomas Gleixner @ 2009-12-14 16:31 UTC (permalink / raw)
To: linux-tip-commits; +Cc: linux-kernel, hpa, mingo, peterz, tglx, mingo
Commit-ID: 23f5d142519621b16cf2b378cf8adf4dcf01a616
Gitweb: http://git.kernel.org/tip/23f5d142519621b16cf2b378cf8adf4dcf01a616
Author: Thomas Gleixner <tglx@linutronix.de>
AuthorDate: Wed, 9 Dec 2009 10:15:01 +0000
Committer: Ingo Molnar <mingo@elte.hu>
CommitDate: Mon, 14 Dec 2009 17:11:35 +0100
sched: Use rcu in sched_get/set_affinity()
tasklist_lock is held read locked to protect the
find_task_by_vpid() call and to prevent the task going away.
sched_setaffinity acquires a task struct ref and drops tasklist
lock right away. The access to the cpus_allowed mask is
protected by rq->lock.
rcu_read_lock() provides the same protection here.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20091209100706.789059966@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
kernel/sched.c | 16 ++++++----------
1 files changed, 6 insertions(+), 10 deletions(-)
diff --git a/kernel/sched.c b/kernel/sched.c
index 1782bee..7989312 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -6516,22 +6516,18 @@ long sched_setaffinity(pid_t pid, const struct cpumask *in_mask)
int retval;
get_online_cpus();
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
if (!p) {
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
put_online_cpus();
return -ESRCH;
}
- /*
- * It is not safe to call set_cpus_allowed with the
- * tasklist_lock held. We will bump the task_struct's
- * usage count and then drop tasklist_lock.
- */
+ /* Prevent p going away */
get_task_struct(p);
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
if (!alloc_cpumask_var(&cpus_allowed, GFP_KERNEL)) {
retval = -ENOMEM;
@@ -6617,7 +6613,7 @@ long sched_getaffinity(pid_t pid, struct cpumask *mask)
int retval;
get_online_cpus();
- read_lock(&tasklist_lock);
+ rcu_read_lock();
retval = -ESRCH;
p = find_process_by_pid(pid);
@@ -6633,7 +6629,7 @@ long sched_getaffinity(pid_t pid, struct cpumask *mask)
task_rq_unlock(rq, &flags);
out_unlock:
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
put_online_cpus();
return retval;
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [tip:sched/urgent] sched: Use rcu in sched_get_rr_param()
2009-12-09 10:15 ` [patch 4/4] sched: Use rcu in sched_get_rr_param Thomas Gleixner
@ 2009-12-14 16:31 ` tip-bot for Thomas Gleixner
0 siblings, 0 replies; 8+ messages in thread
From: tip-bot for Thomas Gleixner @ 2009-12-14 16:31 UTC (permalink / raw)
To: linux-tip-commits; +Cc: linux-kernel, hpa, mingo, peterz, tglx, mingo
Commit-ID: 1a551ae715825bb2a2107a2dd68de024a1fa4e32
Gitweb: http://git.kernel.org/tip/1a551ae715825bb2a2107a2dd68de024a1fa4e32
Author: Thomas Gleixner <tglx@linutronix.de>
AuthorDate: Wed, 9 Dec 2009 10:15:11 +0000
Committer: Ingo Molnar <mingo@elte.hu>
CommitDate: Mon, 14 Dec 2009 17:11:35 +0100
sched: Use rcu in sched_get_rr_param()
read_lock(&tasklist_lock) does not protect
sys_sched_get_rr_param() against a concurrent update of the
policy or scheduler parameters as do_sched_scheduler() does not
take the tasklist_lock.
The access to task->sched_class->get_rr_interval is protected by
task_rq_lock(task).
Use rcu_read_lock() to protect find_task_by_vpid() and prevent
the task struct from going away.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20091209100706.862897167@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
kernel/sched.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/kernel/sched.c b/kernel/sched.c
index 7989312..db5c266 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -6873,7 +6873,7 @@ SYSCALL_DEFINE2(sched_rr_get_interval, pid_t, pid,
return -EINVAL;
retval = -ESRCH;
- read_lock(&tasklist_lock);
+ rcu_read_lock();
p = find_process_by_pid(pid);
if (!p)
goto out_unlock;
@@ -6886,13 +6886,13 @@ SYSCALL_DEFINE2(sched_rr_get_interval, pid_t, pid,
time_slice = p->sched_class->get_rr_interval(rq, p);
task_rq_unlock(rq, &flags);
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
jiffies_to_timespec(time_slice, &t);
retval = copy_to_user(interval, &t, sizeof(t)) ? -EFAULT : 0;
return retval;
out_unlock:
- read_unlock(&tasklist_lock);
+ rcu_read_unlock();
return retval;
}
^ permalink raw reply related [flat|nested] 8+ messages in thread
end of thread, other threads:[~2009-12-14 16:32 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-12-09 10:14 [patch 0/4] sched: Replace read_lock(&tasklist_lock) with RCU - the easy part Thomas Gleixner
2009-12-09 10:14 ` [patch 1/4] sched: Use rcu in sys_sched_getscheduler/sys_sched_getparam() Thomas Gleixner
2009-12-14 16:31 ` [tip:sched/urgent] " tip-bot for Thomas Gleixner
2009-12-09 10:15 ` [patch 2/4] sched: Use rcu in sched_get/set_affinity() Thomas Gleixner
2009-12-14 16:31 ` [tip:sched/urgent] " tip-bot for Thomas Gleixner
2009-12-09 10:15 ` [patch 3/4] sched: Rename bogus label in sched_setaffinity() Thomas Gleixner
2009-12-09 10:15 ` [patch 4/4] sched: Use rcu in sched_get_rr_param Thomas Gleixner
2009-12-14 16:31 ` [tip:sched/urgent] sched: Use rcu in sched_get_rr_param() tip-bot for Thomas Gleixner
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).