From: Shaohua Li <shaohua.li@intel.com>
To: lkml <linux-kernel@vger.kernel.org>
Cc: akpm <akpm@osdl.org>
Subject: [PATCH]CPU hotplug breaks wake_up_new_task
Date: Tue, 31 May 2005 15:35:09 +0800 [thread overview]
Message-ID: <1117524909.3820.11.camel@linux-hp.sh.intel.com> (raw)
Hi,
There is a race condition at wake_up_new_task at CPU hotplug case.
Say do_fork
copy_process (which sets new forked task's current cpu, cpu_allowed)
<-------- the new forked task's current cpu is offline
wake_up_new_task
wake_up_new_task will put the forked task into a dead cpu.
Thanks,
Shaohua
Signed-off-by: Shaohua Li<shaohua.li@intel.com>
---
linux-2.6.11-rc5-mm1-root/kernel/sched.c | 25 +++++++++++++++++++++++--
1 files changed, 23 insertions(+), 2 deletions(-)
diff -puN kernel/sched.c~wake_up_new_task_to_online_cpu kernel/sched.c
--- linux-2.6.11-rc5-mm1/kernel/sched.c~wake_up_new_task_to_online_cpu 2005-05-31 13:39:43.888682784 +0800
+++ linux-2.6.11-rc5-mm1-root/kernel/sched.c 2005-05-31 14:49:58.537958704 +0800
@@ -1412,6 +1412,10 @@ void fastcall sched_fork(task_t *p, int
put_cpu();
}
+#ifdef CONFIG_HOTPLUG_CPU
+static int task_select_online_cpu(int dead_cpu, struct task_struct *tsk);
+#endif
+
/*
* wake_up_new_task - wake up a newly created task for the first time.
*
@@ -1426,9 +1430,20 @@ void fastcall wake_up_new_task(task_t *
runqueue_t *rq, *this_rq;
rq = task_rq_lock(p, &flags);
- BUG_ON(p->state != TASK_RUNNING);
this_cpu = smp_processor_id();
cpu = task_cpu(p);
+#ifdef CONFIG_HOTPLUG_CPU
+ while (!cpu_online(cpu)) {
+ cpu = task_select_online_cpu(cpu, p);
+ set_task_cpu(p, cpu);
+ task_rq_unlock(rq, &flags);
+ /* CPU hotplug might occur here */
+ rq = task_rq_lock(p, &flags);
+ this_cpu = smp_processor_id();
+ cpu = task_cpu(p);
+ }
+#endif
+ BUG_ON(p->state != TASK_RUNNING);
/*
* We decrease the sleep average of forking parents
@@ -4457,7 +4472,7 @@ wait_to_die:
#ifdef CONFIG_HOTPLUG_CPU
/* Figure out where task on dead CPU should go, use force if neccessary. */
-static void move_task_off_dead_cpu(int dead_cpu, struct task_struct *tsk)
+static int task_select_online_cpu(int dead_cpu, struct task_struct *tsk)
{
int dest_cpu;
cpumask_t mask;
@@ -4486,6 +4501,12 @@ static void move_task_off_dead_cpu(int d
"longer affine to cpu%d\n",
tsk->pid, tsk->comm, dead_cpu);
}
+ return dest_cpu;
+}
+
+static void move_task_off_dead_cpu(int dead_cpu, struct task_struct *tsk)
+{
+ int dest_cpu = task_select_online_cpu(dead_cpu, tsk);
__migrate_task(tsk, dead_cpu, dest_cpu);
}
_
next reply other threads:[~2005-05-31 7:28 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-05-31 7:35 Shaohua Li [this message]
2005-05-31 8:00 ` [PATCH]CPU hotplug breaks wake_up_new_task Ashok Raj
2005-05-31 8:35 ` Shaohua Li
2005-05-31 8:50 ` Nick Piggin
2005-05-31 9:11 ` Shaohua Li
2005-05-31 9:08 ` Nick Piggin
2005-05-31 9:40 ` Srivatsa Vaddagiri
2005-05-31 9:46 ` Nick Piggin
2005-05-31 10:40 ` Srivatsa Vaddagiri
2005-05-31 10:49 ` Nick Piggin
-- strict thread matches above, loose matches on Subject: below --
2005-05-31 11:41 Li, Shaohua
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1117524909.3820.11.camel@linux-hp.sh.intel.com \
--to=shaohua.li@intel.com \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.