From 2e636d5e66c35dfcbaf617aa8fa963f6847478fe Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 16 May 2016 15:01:11 +0200 Subject: sched/preempt: Fix preempt_count manipulations Vikram reported that his ARM64 compiler managed to 'optimize' away the preempt_count manipulations in code like: preempt_enable_no_resched(); put_user(); preempt_disable(); Irrespective of that fact that that is horrible code that should be fixed for many reasons, it does highlight a deficiency in the generic preempt_count manipulators. As it is never right to combine/elide preempt_count manipulations like this. Therefore sprinkle some volatile in the two generic accessors to ensure the compiler is aware of the fact that the preempt_count is observed outside of the regular program-order view and thus cannot be optimized away like this. x86; the only arch not using the generic code is not affected as we do all this in asm in order to use the segment base per-cpu stuff. Reported-by: Vikram Mulukutla Tested-by: Vikram Mulukutla Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: a787870924db ("sched, arch: Create asm/preempt.h") Link: http://lkml.kernel.org/r/20160516131751.GH3205@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- include/asm-generic/preempt.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/asm-generic/preempt.h b/include/asm-generic/preempt.h index 5d8ffa3e6f8c..c1cde3577551 100644 --- a/include/asm-generic/preempt.h +++ b/include/asm-generic/preempt.h @@ -7,10 +7,10 @@ static __always_inline int preempt_count(void) { - return current_thread_info()->preempt_count; + return READ_ONCE(current_thread_info()->preempt_count); } -static __always_inline int *preempt_count_ptr(void) +static __always_inline volatile int *preempt_count_ptr(void) { return ¤t_thread_info()->preempt_count; } -- cgit v1.2.3 From b7e7ade34e6188bee2e3b0d42b51d25137d9e2a5 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 23 May 2016 11:19:07 +0200 Subject: sched/core: Fix remote wakeups Commit: b5179ac70de8 ("sched/fair: Prepare to fix fairness problems on migration") ... introduced a bug: Mike Galbraith found that it introduced a performance regression, while Paul E. McKenney reported lost wakeups and bisected it to this commit. The reason is that I mis-read ttwu_queue() such that I assumed any wakeup that got a remote queue must have had the task migrated. Since this is not so; we need to transfer this information between queueing the wakeup and actually doing the wakeup. Use a new task_struct::sched_flag for this, we already write to sched_contributes_to_load in the wakeup path so this is a hot and modified cacheline. Reported-by: Paul E. McKenney Reported-by: Mike Galbraith Tested-by: Mike Galbraith Signed-off-by: Peter Zijlstra (Intel) Cc: Andrew Hunter Cc: Andy Lutomirski Cc: Ben Segall Cc: Borislav Petkov Cc: Brian Gerst Cc: Dave Hansen Cc: Denys Vlasenko Cc: Fenghua Yu Cc: H. Peter Anvin Cc: Linus Torvalds Cc: Matt Fleming Cc: Morten Rasmussen Cc: Oleg Nesterov Cc: Paul Turner Cc: Pavan Kondeti Cc: Peter Zijlstra Cc: Quentin Casasnovas Cc: Thomas Gleixner Cc: byungchul.park@lge.com Fixes: b5179ac70de8 ("sched/fair: Prepare to fix fairness problems on migration") Link: http://lkml.kernel.org/r/20160523091907.GD15728@worktop.ger.corp.intel.com Signed-off-by: Ingo Molnar --- include/linux/sched.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/sched.h b/include/linux/sched.h index 6cc0df970f1a..e053517a88b6 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1533,6 +1533,7 @@ struct task_struct { unsigned sched_reset_on_fork:1; unsigned sched_contributes_to_load:1; unsigned sched_migrated:1; + unsigned sched_remote_wakeup:1; unsigned :0; /* force alignment to the next boundary */ /* unserialized, strictly 'current' */ -- cgit v1.2.3