lkml.org 
[lkml]   [2012]   [Mar]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[tip:irq/core] genirq: Get rid of unnecessary IRQTF_DIED flag
Commit-ID:  5234ffb9f74cfa8993d174782bc861dd9b7b5bfb
Gitweb: http://git.kernel.org/tip/5234ffb9f74cfa8993d174782bc861dd9b7b5bfb
Author: Alexander Gordeev <agordeev@redhat.com>
AuthorDate: Fri, 9 Mar 2012 14:59:59 +0100
Committer: Thomas Gleixner <tglx@linutronix.de>
CommitDate: Fri, 9 Mar 2012 17:19:09 +0100

genirq: Get rid of unnecessary IRQTF_DIED flag

Currently IRQTF_DIED flag is set when a IRQ thread handler calls do_exit()
But also PF_EXITING per process flag gets set when a thread exits. This
fix eliminates the duplicate by using PF_EXITING flag.

Also, there is a race condition in exit_irq_thread(). In case a thread's
bit is cleared in desc->threads_oneshot (and the IRQ line gets unmasked),
but before IRQTF_DIED flag is set, a new interrupt might come in and set
just cleared bit again, this time forever. This fix throws IRQTF_DIED flag
away, eliminating the race as a result.

[ tglx: Test THREAD_EXITING first as suggested by Oleg ]

Reported-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Link: http://lkml.kernel.org/r/20120309135958.GD2114@dhcp-26-207.brq.redhat.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
kernel/exit.c | 4 ++--
kernel/irq/handle.c | 2 +-
kernel/irq/internals.h | 2 --
kernel/irq/manage.c | 11 +----------
4 files changed, 4 insertions(+), 15 deletions(-)
diff --git a/kernel/exit.c b/kernel/exit.c
index 4b4042f..752d2c0 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -935,8 +935,6 @@ void do_exit(long code)
schedule();
}

- exit_irq_thread();
-
exit_signals(tsk); /* sets PF_EXITING */
/*
* tsk->flags are checked in the futex code to protect against
@@ -945,6 +943,8 @@ void do_exit(long code)
smp_mb();
raw_spin_unlock_wait(&tsk->pi_lock);

+ exit_irq_thread();
+
if (unlikely(in_atomic()))
printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
current->comm, task_pid_nr(current),
diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c
index 470d08c..500aaf6 100644
--- a/kernel/irq/handle.c
+++ b/kernel/irq/handle.c
@@ -60,7 +60,7 @@ static void irq_wake_thread(struct irq_desc *desc, struct irqaction *action)
* device interrupt, so no irq storm is lurking. If the
* RUNTHREAD bit is already set, nothing to do.
*/
- if (test_bit(IRQTF_DIED, &action->thread_flags) ||
+ if ((action->thread->flags & PF_EXITING) ||
test_and_set_bit(IRQTF_RUNTHREAD, &action->thread_flags))
return;

diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h
index b795231..5c233e0 100644
--- a/kernel/irq/internals.h
+++ b/kernel/irq/internals.h
@@ -20,14 +20,12 @@ extern bool noirqdebug;
/*
* Bits used by threaded handlers:
* IRQTF_RUNTHREAD - signals that the interrupt handler thread should run
- * IRQTF_DIED - handler thread died
* IRQTF_WARNED - warning "IRQ_WAKE_THREAD w/o thread_fn" has been printed
* IRQTF_AFFINITY - irq thread is requested to adjust affinity
* IRQTF_FORCED_THREAD - irq action is force threaded
*/
enum {
IRQTF_RUNTHREAD,
- IRQTF_DIED,
IRQTF_WARNED,
IRQTF_AFFINITY,
IRQTF_FORCED_THREAD,
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 3feab4a..a94466d 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -845,17 +845,8 @@ void exit_irq_thread(void)

desc = irq_to_desc(action->irq);

- /*
- * Prevent a stale desc->threads_oneshot. Must be called
- * before setting the IRQTF_DIED flag.
- */
+ /* Prevent a stale desc->threads_oneshot */
irq_finalize_oneshot(desc, action, true);
-
- /*
- * Set the THREAD DIED flag to prevent further wakeups of the
- * soon to be gone threaded handler.
- */
- set_bit(IRQTF_DIED, &action->flags);
}

static void irq_setup_forced_threading(struct irqaction *new)

\
 
 \ /
  Last update: 2012-03-09 22:17    [from the cache]
©2003-2011 Jasper Spaans