[PATCH RT 19/25][RFC 3.0.23-rt39-rc1] timer: Fix hotplug for -rt
From: Steven Rostedt <rostedt@goodmis.org>
Date: 2012-03-06 16:21:56
Also in:
lkml
Subsystem:
the rest · Maintainer:
Linus Torvalds
From: Steven Rostedt <rostedt@goodmis.org>
Revert the RT patch:
Author: Ingo Molnar [off-list ref]
Date: Fri Jul 3 08:30:32 2009 -0500
timers: fix timer hotplug on -rt
Here we are in the CPU_DEAD notifier, and we must not sleep nor
enable interrupts.
There's no problem with sleeping in this notifier.
But the get_cpu_var() had to be converted to a get_local_var().
Replace the previous fix with the get_local_var() convert.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Carsten Emde <redacted>
Cc: John Kacur <jkacur@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Clark Williams <redacted>
Cc: stable-rt@vger.kernel.org
Link: http://lkml.kernel.org/r/20120301190344.948157137@goodmis.org
Signed-off-by: Thomas Gleixner <redacted>
---
kernel/timer.c | 16 +++++-----------
1 files changed, 5 insertions(+), 11 deletions(-)
diff --git a/kernel/timer.c b/kernel/timer.c
index 937799f..a319c6a 100644
--- a/kernel/timer.c
+++ b/kernel/timer.c@@ -1743,21 +1743,17 @@ static void __cpuinit migrate_timers(int cpu) { struct tvec_base *old_base; struct tvec_base *new_base; - unsigned long flags; int i; BUG_ON(cpu_online(cpu)); old_base = per_cpu(tvec_bases, cpu); - new_base = get_cpu_var(tvec_bases); + new_base = get_local_var(tvec_bases); /* * The caller is globally serialized and nobody else * takes two locks at once, deadlock is not possible. */ - local_irq_save(flags); - while (!spin_trylock(&new_base->lock)) - cpu_relax(); - while (!spin_trylock(&old_base->lock)) - cpu_relax(); + spin_lock_irq(&new_base->lock); + spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING); BUG_ON(old_base->running_timer);
@@ -1771,10 +1767,8 @@ static void __cpuinit migrate_timers(int cpu) } spin_unlock(&old_base->lock); - spin_unlock(&new_base->lock); - local_irq_restore(flags);