[lkml]   [2004]   [Sep]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
Patch in this message
Subject[Patch 2/4] cpusets simplify cpus_allowed setting in attach
When reattaching a task to a different cpuset, I had
code that attempted to preserve the cpuset relative
placement of the task (such as might have been done
using sched_setaffinity, mbind or set_mempolicy) by
rebinding the task to the intersection of its old
and new placement, if not empty.

This resulted in strange and puzzling behaviour, such
as having a tasks cpus_allowed not change if the task
was reattached to the root cpuset.

Simplify the code - when attaching a task to a cpuset,
simply set its cpus_allowed to that of the cpuset.
This is much less surprising in practice.

Signed-off-by: Paul Jackson <>

Index: 2.6.9-rc1-mm4/kernel/cpuset.c
--- 2.6.9-rc1-mm4.orig/kernel/cpuset.c 2004-09-08 15:22:59.000000000 -0700
+++ 2.6.9-rc1-mm4/kernel/cpuset.c 2004-09-08 18:44:33.000000000 -0700
@@ -605,7 +605,6 @@ static int attach_task(struct cpuset *cs
pid_t pid;
struct task_struct *tsk;
struct cpuset *oldcs;
- cpumask_t cpus;

if (sscanf(buf, "%d", &pid) != 1)
return -EIO;
@@ -645,16 +644,7 @@ static int attach_task(struct cpuset *cs
tsk->cpuset = cs;

- /*
- * If the tasks current CPU placement overlaps with its new cpuset,
- * then let it run in that overlap. Otherwise fallback to simply
- * letting it have the run of the CPUs in the new cpuset.
- */
- if (cpus_intersects(tsk->cpus_allowed, cs->cpus_allowed))
- cpus_and(cpus, tsk->cpus_allowed, cs->cpus_allowed);
- else
- cpus = cs->cpus_allowed;
- set_cpus_allowed(tsk, cpus);
+ set_cpus_allowed(tsk, cs->cpus_allowed);

if (atomic_dec_and_test(&oldcs->count))
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <> 1.650.933.1373
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 14:06    [W:0.064 / U:9.356 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site