* Re: [PATCH] 2.4-ac: migration_thread locking fix
2002-05-11 16:40 [PATCH] 2.4-ac: migration_thread locking fix Robert Love
@ 2002-05-11 16:46 ` Robert Love
0 siblings, 0 replies; 2+ messages in thread
From: Robert Love @ 2002-05-11 16:46 UTC (permalink / raw)
To: alan; +Cc: anton, linux-kernel
[-- Attachment #1: Type: text/plain, Size: 482 bytes --]
On Sat, 2002-05-11 at 09:40, Robert Love wrote:
> Anton Blanchard reported a lock vs. interrupt-off ordering bug in
> migration_thread. The bug is already fixed in 2.5. A patch is attached
> to fix O(1) in your tree.
Oh, this fix was included in the migration_thread cleanup in the latest
batch of fixes I sent you against -ac1. I have attached an all-in-one
patch if you missed it. Otherwise you can get the chunks at the usual
kernel.org location.
Thanks,
Robert Love
[-- Attachment #2: sched-updates-rml-2.4.19-pre8-ac1-1.patch --]
[-- Type: text/x-patch, Size: 11630 bytes --]
diff -urN linux-2.4.19-pre8-ac1/arch/alpha/mm/fault.c linux/arch/alpha/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/alpha/mm/fault.c Wed May 8 12:03:59 2002
+++ linux/arch/alpha/mm/fault.c Wed May 8 13:40:13 2002
@@ -196,8 +196,7 @@
*/
out_of_memory:
if (current->pid == 1) {
- current->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/arm/mm/fault-common.c linux/arch/arm/mm/fault-common.c
--- linux-2.4.19-pre8-ac1/arch/arm/mm/fault-common.c Wed May 8 12:04:09 2002
+++ linux/arch/arm/mm/fault-common.c Wed May 8 13:40:13 2002
@@ -225,8 +225,7 @@
* If we are out of memory for pid1,
* sleep for a while and retry
*/
- tsk->policy |= SCHED_YIELD;
- schedule();
+ yield();
goto survive;
check_stack:
diff -urN linux-2.4.19-pre8-ac1/arch/ia64/mm/fault.c linux/arch/ia64/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/ia64/mm/fault.c Wed May 8 12:04:11 2002
+++ linux/arch/ia64/mm/fault.c Wed May 8 13:40:13 2002
@@ -196,8 +196,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (current->pid == 1) {
- current->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/m68k/mm/fault.c linux/arch/m68k/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/m68k/mm/fault.c Wed May 8 12:04:07 2002
+++ linux/arch/m68k/mm/fault.c Wed May 8 13:40:13 2002
@@ -181,8 +181,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (current->pid == 1) {
- current->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/mips/mm/fault.c linux/arch/mips/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/mips/mm/fault.c Wed May 8 12:04:01 2002
+++ linux/arch/mips/mm/fault.c Wed May 8 13:40:13 2002
@@ -211,8 +211,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (tsk->pid == 1) {
- tsk->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/mips64/mm/fault.c linux/arch/mips64/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/mips64/mm/fault.c Wed May 8 12:04:13 2002
+++ linux/arch/mips64/mm/fault.c Wed May 8 13:40:13 2002
@@ -240,8 +240,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (tsk->pid == 1) {
- tsk->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/ppc/mm/fault.c linux/arch/ppc/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/ppc/mm/fault.c Wed May 8 12:04:04 2002
+++ linux/arch/ppc/mm/fault.c Wed May 8 13:40:13 2002
@@ -197,8 +197,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (current->pid == 1) {
- current->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/s390/mm/fault.c linux/arch/s390/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/s390/mm/fault.c Wed May 8 12:04:13 2002
+++ linux/arch/s390/mm/fault.c Wed May 8 13:40:13 2002
@@ -290,8 +290,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (tsk->pid == 1) {
- tsk->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/s390x/mm/fault.c linux/arch/s390x/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/s390x/mm/fault.c Wed May 8 12:04:14 2002
+++ linux/arch/s390x/mm/fault.c Wed May 8 13:40:13 2002
@@ -290,8 +290,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (tsk->pid == 1) {
- tsk->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/arch/sh/mm/fault.c linux/arch/sh/mm/fault.c
--- linux-2.4.19-pre8-ac1/arch/sh/mm/fault.c Wed May 8 12:04:10 2002
+++ linux/arch/sh/mm/fault.c Wed May 8 13:40:13 2002
@@ -207,8 +207,7 @@
out_of_memory:
up_read(&mm->mmap_sem);
if (current->pid == 1) {
- current->policy |= SCHED_YIELD;
- schedule();
+ yield();
down_read(&mm->mmap_sem);
goto survive;
}
diff -urN linux-2.4.19-pre8-ac1/kernel/sched.c linux/kernel/sched.c
--- linux-2.4.19-pre8-ac1/kernel/sched.c Wed May 8 12:03:14 2002
+++ linux/kernel/sched.c Wed May 8 13:40:18 2002
@@ -22,13 +22,17 @@
#include <linux/kernel_stat.h>
/*
- * Priority of a process goes from 0 to MAX_PRIO-1. The
- * 0 to MAX_RT_PRIO-1 priority range is allocated to RT tasks,
- * the MAX_RT_PRIO to MAX_PRIO range is for SCHED_OTHER tasks.
- * Priority values are inverted: lower p->prio value means higher
- * priority.
+ * Priority of a process goes from 0 to 139. The 0-99
+ * priority range is allocated to RT tasks, the 100-139
+ * range is for SCHED_OTHER tasks. Priority values are
+ * inverted: lower p->prio value means higher priority.
+ *
+ * MAX_USER_RT_PRIO allows the actual maximum RT priority
+ * to be separate from the value exported to user-space.
+ * NOTE: MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
*/
#define MAX_RT_PRIO 100
+#define MAX_USER_RT_PRIO 100
#define MAX_PRIO (MAX_RT_PRIO + 40)
/*
@@ -1025,7 +1029,7 @@
*/
int task_prio(task_t *p)
{
- return p->prio - MAX_RT_PRIO;
+ return p->prio - MAX_USER_RT_PRIO;
}
int task_nice(task_t *p)
@@ -1082,11 +1086,11 @@
}
/*
- * Valid priorities for SCHED_FIFO and SCHED_RR are 1..MAX_RT_PRIO-1,
- * valid priority for SCHED_OTHER is 0.
+ * Valid priorities for SCHED_FIFO and SCHED_RR are
+ * 1..MAX_USER_RT_PRIO-1, valid priority for SCHED_OTHER is 0.
*/
retval = -EINVAL;
- if (lp.sched_priority < 0 || lp.sched_priority > MAX_RT_PRIO - 1)
+ if (lp.sched_priority < 0 || lp.sched_priority > MAX_USER_RT_PRIO-1)
goto out_unlock;
if ((policy == SCHED_OTHER) != (lp.sched_priority == 0))
goto out_unlock;
@@ -1106,7 +1110,7 @@
p->policy = policy;
p->rt_priority = lp.sched_priority;
if (policy != SCHED_OTHER)
- p->prio = (MAX_RT_PRIO - 1) - p->rt_priority;
+ p->prio = MAX_USER_RT_PRIO-1 - p->rt_priority;
else
p->prio = p->static_prio;
if (array)
@@ -1229,7 +1233,7 @@
switch (policy) {
case SCHED_FIFO:
case SCHED_RR:
- ret = MAX_RT_PRIO - 1;
+ ret = MAX_USER_RT_PRIO-1;
break;
case SCHED_OTHER:
ret = 0;
@@ -1536,80 +1540,34 @@
down(&req.sem);
}
-/*
- * Treat the bits of migration_mask as lock bits.
- * If the bit corresponding to the cpu a migration_thread is
- * running on then we have failed to claim our cpu and must
- * yield in order to find another.
- */
-static volatile unsigned long migration_mask;
-static atomic_t migration_threads_seeking_cpu;
-static struct completion migration_complete
- = COMPLETION_INITIALIZER(migration_complete);
-
-static int migration_thread(void * unused)
+static int migration_thread(void * bind_cpu)
{
- struct sched_param param = { sched_priority: MAX_RT_PRIO - 1 };
+ int cpu = cpu_logical_map((int) (long) bind_cpu);
+ struct sched_param param = { sched_priority: MAX_RT_PRIO-1 };
runqueue_t *rq;
int ret;
daemonize();
sigfillset(¤t->blocked);
set_fs(KERNEL_DS);
- ret = setscheduler(0, SCHED_FIFO, ¶m);
-
- /*
- * We have to migrate manually - there is no migration thread
- * to do this for us yet :-)
- *
- * We use the following property of the Linux scheduler. At
- * this point no other task is running, so by keeping all
- * migration threads running, the load-balancer will distribute
- * them between all CPUs equally. At that point every migration
- * task binds itself to the current CPU.
- */
/*
- * Enter the loop with preemption disabled so that
- * smp_processor_id() remains valid through the check. The
- * interior of the wait loop re-enables preemption in an
- * attempt to get scheduled off the current cpu. When the
- * loop is exited the lock bit in migration_mask is acquired
- * and preemption is disabled on the way out. This way the
- * cpu acquired remains valid when ->cpus_allowed is set.
+ * The first migration thread is started on CPU #0. This one can
+ * migrate the other migration threads to their destination CPUs.
*/
- while (test_and_set_bit(smp_processor_id(), &migration_mask))
- yield();
+ if (cpu != 0) {
+ while (!cpu_rq(cpu_logical_map(0))->migration_thread)
+ yield();
+ set_cpus_allowed(current, 1UL << cpu);
+ }
+ printk("migration_task %d on cpu=%d\n", cpu, smp_processor_id());
+ ret = setscheduler(0, SCHED_FIFO, ¶m);
- current->cpus_allowed = 1 << smp_processor_id();
rq = this_rq();
rq->migration_thread = current;
- /*
- * Now that we've bound ourselves to a cpu, post to
- * migration_threads_seeking_cpu and wait for everyone else.
- * Preemption should remain disabled and the cpu should remain
- * in busywait. Yielding the cpu will allow the livelock
- * where where a timing pattern causes an idle task seeking a
- * migration_thread to always find the unbound migration_thread
- * running on the cpu's it tries to steal tasks from.
- */
- atomic_dec(&migration_threads_seeking_cpu);
- while (atomic_read(&migration_threads_seeking_cpu))
- cpu_relax();
-
sprintf(current->comm, "migration_CPU%d", smp_processor_id());
- /*
- * Everyone's found their cpu, so now wake migration_init().
- * Multiple wakeups are harmless; removal from the waitqueue
- * has locking built-in, and waking an empty queue is valid.
- */
- complete(&migration_complete);
-
- /*
- * Initiate the event loop.
- */
for (;;) {
runqueue_t *rq_src, *rq_dest;
struct list_head *head;
@@ -1634,25 +1592,25 @@
cpu_dest = __ffs(p->cpus_allowed);
rq_dest = cpu_rq(cpu_dest);
repeat:
- cpu_src = p->cpu;
+ cpu_src = p->thread_info->cpu;
rq_src = cpu_rq(cpu_src);
local_irq_save(flags);
double_rq_lock(rq_src, rq_dest);
- if (p->cpu != cpu_src) {
- local_irq_restore(flags);
+ if (p->thread_info->cpu != cpu_src) {
double_rq_unlock(rq_src, rq_dest);
+ local_irq_restore(flags);
goto repeat;
}
if (rq_src == rq) {
- p->cpu = cpu_dest;
+ p->thread_info->cpu = cpu_dest;
if (p->array) {
deactivate_task(p, rq_src);
activate_task(p, rq_dest);
}
}
- local_irq_restore(flags);
double_rq_unlock(rq_src, rq_dest);
+ local_irq_restore(flags);
up(&req->sem);
}
@@ -1660,28 +1618,19 @@
void __init migration_init(void)
{
- unsigned long orig_cache_decay_ticks;
int cpu;
- atomic_set(&migration_threads_seeking_cpu, smp_num_cpus);
-
- orig_cache_decay_ticks = cache_decay_ticks;
- cache_decay_ticks = 0;
-
- for (cpu = 0; cpu < smp_num_cpus; cpu++)
- if (kernel_thread(migration_thread, NULL,
+ current->cpus_allowed = 1UL << cpu_logical_map(0);
+ for (cpu = 0; cpu < smp_num_cpus; cpu++) {
+ if (kernel_thread(migration_thread, (void *) (long) cpu,
CLONE_FS | CLONE_FILES | CLONE_SIGNAL) < 0)
BUG();
+ }
+ current->cpus_allowed = -1L;
- /*
- * We cannot have missed the wakeup for the migration_thread
- * bound for the cpu migration_init() is running on cannot
- * acquire this cpu until migration_init() has yielded it by
- * means of wait_for_completion().
- */
- wait_for_completion(&migration_complete);
-
- cache_decay_ticks = orig_cache_decay_ticks;
+ for (cpu = 0; cpu < smp_num_cpus; cpu++)
+ while (!cpu_rq(cpu_logical_map(cpu))->migration_thread)
+ schedule_timeout(2);
}
#endif /* CONFIG_SMP */
^ permalink raw reply [flat|nested] 2+ messages in thread