[RFC PATCH 0/9] nohz: Nohz full kick fixes

[RFC PATCH 0/9] nohz: Nohz full kick fixes

[Frederic Weisbecker]

Hi,

There have been several reports of warnings and lockup lately due to
recent or longer term issues.

This patchset is a proposed solution to fix them.

git://git.kernel.org/pub/scm/linux/kernel/git/frederic/linux-dynticks.git
	nohz/fixes

Thanks,
	Frederic
---

Frederic Weisbecker (9):
      nohz: Restore NMI safe local irq work for local nohz kick
      nohz: Move nohz full init call to tick init
      irq_work: Force raised irq work to run on irq work interrupt
      x86: Build irq work only if local apic support
      x86: Tell irq work about self IPI support
      arm: Tell irq work about self IPI support
      arm64: Tell irq work about self IPI support
      nohz: Consolidate nohz full init code
      nohz: nohz full depends on irq work self IPI support


 arch/arm/kernel/smp.c             |  5 +++
 arch/arm64/kernel/smp.c           |  5 +++
 arch/x86/Kconfig                  |  4 +++
 arch/x86/include/asm/entry_arch.h |  2 +-
 arch/x86/kernel/Makefile          |  2 +-
 arch/x86/kernel/entry_64.S        |  2 +-
 arch/x86/kernel/irq_work.c        |  7 +++--
 arch/x86/kernel/irqinit.c         |  2 +-
 include/linux/irq_work.h          |  4 +++
 include/linux/tick.h              |  9 +-----
 init/main.c                       |  2 +-
 kernel/irq_work.c                 | 21 ++++++++++++-
 kernel/time/tick-common.c         |  1 +
 kernel/time/tick-internal.h       |  7 +++++
 kernel/time/tick-sched.c          | 66 ++++++++++++++++++++++++++++-----------
 kernel/time/timer.c               |  2 +-
 16 files changed, 105 insertions(+), 36 deletions(-)

[RFC PATCH 1/9] nohz: Restore NMI safe local irq work for local nohz kick

[Frederic Weisbecker]

The local nohz kick is currently used by perf which needs it to be
NMI-safe. Recent commit though (7d1311b93e58ed55f3a31cc8f94c4b8fe988a2b9)
changed its implementation to fire the local kick using the remote kick
API. It was convenient to make the code more generic but the remote kick
isn't NMI-safe.

As a result:

	WARNING: CPU: 3 PID: 18062 at kernel/irq_work.c:72 irq_work_queue_on+0x11e/0x140()
	CPU: 3 PID: 18062 Comm: trinity-subchil Not tainted 3.16.0+ #34
	0000000000000009 00000000903774d1 ffff880244e06c00 ffffffff9a7f1e37
	0000000000000000 ffff880244e06c38 ffffffff9a0791dd ffff880244fce180
	0000000000000003 ffff880244e06d58 ffff880244e06ef8 0000000000000000
	Call Trace:
	<NMI>  [<ffffffff9a7f1e37>] dump_stack+0x4e/0x7a
	[<ffffffff9a0791dd>] warn_slowpath_common+0x7d/0xa0
	[<ffffffff9a07930a>] warn_slowpath_null+0x1a/0x20
	[<ffffffff9a17ca1e>] irq_work_queue_on+0x11e/0x140
	[<ffffffff9a10a2c7>] tick_nohz_full_kick_cpu+0x57/0x90
	[<ffffffff9a186cd5>] __perf_event_overflow+0x275/0x350
	[<ffffffff9a184f80>] ? perf_event_task_disable+0xa0/0xa0
	[<ffffffff9a01a4cf>] ? x86_perf_event_set_period+0xbf/0x150
	[<ffffffff9a187934>] perf_event_overflow+0x14/0x20
	[<ffffffff9a020386>] intel_pmu_handle_irq+0x206/0x410
	[<ffffffff9a0b54d3>] ? arch_vtime_task_switch+0x63/0x130
	[<ffffffff9a01937b>] perf_event_nmi_handler+0x2b/0x50
	[<ffffffff9a007b72>] nmi_handle+0xd2/0x390
	[<ffffffff9a007aa5>] ? nmi_handle+0x5/0x390
	[<ffffffff9a0d131b>] ? lock_release+0xab/0x330
	[<ffffffff9a008062>] default_do_nmi+0x72/0x1c0
	[<ffffffff9a0c925f>] ? cpuacct_account_field+0xcf/0x200
	[<ffffffff9a008268>] do_nmi+0xb8/0x100

Lets fix this by restoring the use of local irq work for the nohz local
kick.

Reported-by: Dave Jones <davej@redhat.com>
Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 include/linux/tick.h     |  7 +------
 kernel/time/tick-sched.c | 14 ++++++++++++++
 2 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/include/linux/tick.h b/include/linux/tick.h
index 0590523..9a82c7d 100644
--- a/include/linux/tick.h
+++ b/include/linux/tick.h
@@ -183,13 +183,8 @@ static inline bool tick_nohz_full_cpu(int cpu)
 
 extern void tick_nohz_init(void);
 extern void __tick_nohz_full_check(void);
+extern void tick_nohz_full_kick(void);
 extern void tick_nohz_full_kick_cpu(int cpu);
-
-static inline void tick_nohz_full_kick(void)
-{
-	tick_nohz_full_kick_cpu(smp_processor_id());
-}
-
 extern void tick_nohz_full_kick_all(void);
 extern void __tick_nohz_task_switch(struct task_struct *tsk);
 #else
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 99aa6ee..f654a8a 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -225,6 +225,20 @@ static DEFINE_PER_CPU(struct irq_work, nohz_full_kick_work) = {
 };
 
 /*
+ * Kick this CPU if it's full dynticks in order to force it to
+ * re-evaluate its dependency on the tick and restart it if necessary.
+ * This kick, unlike tick_nohz_full_kick_cpu() and tick_nohz_full_kick_all(),
+ * is NMI safe.
+ */
+void tick_nohz_full_kick(void)
+{
+	if (!tick_nohz_full_cpu(smp_processor_id()))
+		return;
+
+	irq_work_queue(&__get_cpu_var(nohz_full_kick_work));
+}
+
+/*
  * Kick the CPU if it's full dynticks in order to force it to
  * re-evaluate its dependency on the tick and restart it if necessary.
  */
-- 
1.8.3.1

[RFC PATCH 2/9] nohz: Move nohz full init call to tick init

[Frederic Weisbecker]

This way we unbloat a bit main.c and more importantly we initialize
nohz full after init_IRQ(). This dependency will be needed in further
patches because nohz full depends on irq work to raise its own IRQ.
This ability, on ARM64, is obtained only after init_IRQ() which initialize
the pointer to __smp_call_function.

Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 include/linux/tick.h        | 2 --
 init/main.c                 | 1 -
 kernel/time/tick-common.c   | 1 +
 kernel/time/tick-internal.h | 7 +++++++
 4 files changed, 8 insertions(+), 3 deletions(-)

diff --git a/include/linux/tick.h b/include/linux/tick.h
index 9a82c7d..595ee86 100644
--- a/include/linux/tick.h
+++ b/include/linux/tick.h
@@ -181,14 +181,12 @@ static inline bool tick_nohz_full_cpu(int cpu)
 	return cpumask_test_cpu(cpu, tick_nohz_full_mask);
 }
 
-extern void tick_nohz_init(void);
 extern void __tick_nohz_full_check(void);
 extern void tick_nohz_full_kick(void);
 extern void tick_nohz_full_kick_cpu(int cpu);
 extern void tick_nohz_full_kick_all(void);
 extern void __tick_nohz_task_switch(struct task_struct *tsk);
 #else
-static inline void tick_nohz_init(void) { }
 static inline bool tick_nohz_full_enabled(void) { return false; }
 static inline bool tick_nohz_full_cpu(int cpu) { return false; }
 static inline void __tick_nohz_full_check(void) { }
diff --git a/init/main.c b/init/main.c
index bb1aed9..8af2f1a 100644
--- a/init/main.c
+++ b/init/main.c
@@ -577,7 +577,6 @@ asmlinkage __visible void __init start_kernel(void)
 		local_irq_disable();
 	idr_init_cache();
 	rcu_init();
-	tick_nohz_init();
 	context_tracking_init();
 	radix_tree_init();
 	/* init some links before init_ISA_irqs() */
diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c
index 0a0608e..052b4b5 100644
--- a/kernel/time/tick-common.c
+++ b/kernel/time/tick-common.c
@@ -400,4 +400,5 @@ void tick_resume(void)
 void __init tick_init(void)
 {
 	tick_broadcast_init();
+	tick_nohz_init();
 }
diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h
index c19c1d8..366aeb4 100644
--- a/kernel/time/tick-internal.h
+++ b/kernel/time/tick-internal.h
@@ -99,6 +99,13 @@ static inline int tick_broadcast_oneshot_active(void) { return 0; }
 static inline bool tick_broadcast_oneshot_available(void) { return false; }
 #endif /* !TICK_ONESHOT */
 
+/* NO_HZ_FULL internal */
+#ifdef CONFIG_NO_HZ_FULL
+extern void tick_nohz_init(void);
+# else
+static inline void tick_nohz_init(void) { }
+#endif
+
 /*
  * Broadcasting support
  */
-- 
1.8.3.1

[RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Frederic Weisbecker]

The nohz full kick, which restarts the tick after some new state depend
on it, can't be called from any place given the operation it does on
timers. If it is called from the scheduler or timers code, chances are that
we run into a deadlock.

This is why we run the nohz full kick from an irq work. That way we make
sure that the kick runs on a virgin context.

However if that's the case when irq work runs in its own dedicated
self-ipi, things are different for the big bunch of archs that don't
support the self triggered way. In order to handle them, irq works are
also handled by the timer interrupt as fallback.

Now when irq works run on the timer interrupt, the context isn't blank.
More precisely, they can run in the context of the hrtimer that runs the
tick. But the nohz kick cancels and restarts this hrtimer. If it does
the cancellation from the hrtimer itself, we run in an endless loop:

	Kernel panic - not syncing: Watchdog detected hard LOCKUP on cpu 2
	CPU: 2 PID: 7538 Comm: kworker/u8:8 Not tainted 3.16.0+ #34
	Workqueue: btrfs-endio-write normal_work_helper [btrfs]
	 ffff880244c06c88 000000001b486fe1 ffff880244c06bf0 ffffffff8a7f1e37
	 ffffffff8ac52a18 ffff880244c06c78 ffffffff8a7ef928 0000000000000010
	 ffff880244c06c88 ffff880244c06c20 000000001b486fe1 0000000000000000
	Call Trace:
	 <NMI[<ffffffff8a7f1e37>] dump_stack+0x4e/0x7a
	 [<ffffffff8a7ef928>] panic+0xd4/0x207
	 [<ffffffff8a1450e8>] watchdog_overflow_callback+0x118/0x120
	 [<ffffffff8a186b0e>] __perf_event_overflow+0xae/0x350
	 [<ffffffff8a184f80>] ? perf_event_task_disable+0xa0/0xa0
	 [<ffffffff8a01a4cf>] ? x86_perf_event_set_period+0xbf/0x150
	 [<ffffffff8a187934>] perf_event_overflow+0x14/0x20
	 [<ffffffff8a020386>] intel_pmu_handle_irq+0x206/0x410
	 [<ffffffff8a01937b>] perf_event_nmi_handler+0x2b/0x50
	 [<ffffffff8a007b72>] nmi_handle+0xd2/0x390
	 [<ffffffff8a007aa5>] ? nmi_handle+0x5/0x390
	 [<ffffffff8a0cb7f8>] ? match_held_lock+0x8/0x1b0
	 [<ffffffff8a008062>] default_do_nmi+0x72/0x1c0
	 [<ffffffff8a008268>] do_nmi+0xb8/0x100
	 [<ffffffff8a7ff66a>] end_repeat_nmi+0x1e/0x2e
	 [<ffffffff8a0cb7f8>] ? match_held_lock+0x8/0x1b0
	 [<ffffffff8a0cb7f8>] ? match_held_lock+0x8/0x1b0
	 [<ffffffff8a0cb7f8>] ? match_held_lock+0x8/0x1b0
	 <<EOE><IRQ[<ffffffff8a0ccd2f>] lock_acquired+0xaf/0x450
	 [<ffffffff8a0f74c5>] ? lock_hrtimer_base.isra.20+0x25/0x50
	 [<ffffffff8a7fc678>] _raw_spin_lock_irqsave+0x78/0x90
	 [<ffffffff8a0f74c5>] ? lock_hrtimer_base.isra.20+0x25/0x50
	 [<ffffffff8a0f74c5>] lock_hrtimer_base.isra.20+0x25/0x50
	 [<ffffffff8a0f7723>] hrtimer_try_to_cancel+0x33/0x1e0
	 [<ffffffff8a0f78ea>] hrtimer_cancel+0x1a/0x30
	 [<ffffffff8a109237>] tick_nohz_restart+0x17/0x90
	 [<ffffffff8a10a213>] __tick_nohz_full_check+0xc3/0x100
	 [<ffffffff8a10a25e>] nohz_full_kick_work_func+0xe/0x10
	 [<ffffffff8a17c884>] irq_work_run_list+0x44/0x70
	 [<ffffffff8a17c8da>] irq_work_run+0x2a/0x50
	 [<ffffffff8a0f700b>] update_process_times+0x5b/0x70
	 [<ffffffff8a109005>] tick_sched_handle.isra.21+0x25/0x60
	 [<ffffffff8a109b81>] tick_sched_timer+0x41/0x60
	 [<ffffffff8a0f7aa2>] __run_hrtimer+0x72/0x470
	 [<ffffffff8a109b40>] ? tick_sched_do_timer+0xb0/0xb0
	 [<ffffffff8a0f8707>] hrtimer_interrupt+0x117/0x270
	 [<ffffffff8a034357>] local_apic_timer_interrupt+0x37/0x60
	 [<ffffffff8a80010f>] smp_apic_timer_interrupt+0x3f/0x50
	 [<ffffffff8a7fe52f>] apic_timer_interrupt+0x6f/0x80

Two possible solutions to fix this: either we make the kick self-aware
of the context it runs on (tick hrtimer or irq work IPI) and adapt the
action on top of it or we force non-lazy irq works to run on irq work
IPIs when available.

The second solution needs the irq work subsystem to know if the arch
supports irq work IPIs. This information will have to be available anyway
because nohz full can't work without that arch support and we must
initialize irq work properly with that dependency in mind.

So lets provide a way for archs to tell about that irq work IPI support.
On top of that we can force irq works on that IPI to avoid lockups like
above.

Reported-by: Dave Jones <davej@redhat.com>
Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 include/linux/irq_work.h |  3 +++
 init/main.c              |  1 +
 kernel/irq_work.c        | 21 ++++++++++++++++++++-
 kernel/time/timer.c      |  2 +-
 4 files changed, 25 insertions(+), 2 deletions(-)

diff --git a/include/linux/irq_work.h b/include/linux/irq_work.h
index bf9422c..708c895 100644
--- a/include/linux/irq_work.h
+++ b/include/linux/irq_work.h
@@ -39,11 +39,14 @@ bool irq_work_queue_on(struct irq_work *work, int cpu);
 #endif
 
 void irq_work_run(void);
+void irq_work_tick(void);
 void irq_work_sync(struct irq_work *work);
 
 #ifdef CONFIG_IRQ_WORK
+void irq_work_init(void);
 bool irq_work_needs_cpu(void);
 #else
+static inline void irq_work_init(void) { }
 static inline bool irq_work_needs_cpu(void) { return false; }
 #endif
 
diff --git a/init/main.c b/init/main.c
index 8af2f1a..0c0015f 100644
--- a/init/main.c
+++ b/init/main.c
@@ -582,6 +582,7 @@ asmlinkage __visible void __init start_kernel(void)
 	/* init some links before init_ISA_irqs() */
 	early_irq_init();
 	init_IRQ();
+	irq_work_init();
 	tick_init();
 	init_timers();
 	hrtimers_init();
diff --git a/kernel/irq_work.c b/kernel/irq_work.c
index e6bcbe7..17bd203 100644
--- a/kernel/irq_work.c
+++ b/kernel/irq_work.c
@@ -22,6 +22,7 @@
 
 static DEFINE_PER_CPU(struct llist_head, raised_list);
 static DEFINE_PER_CPU(struct llist_head, lazy_list);
+static bool has_own_interrupt;
 
 /*
  * Claim the entry so that no one else will poke at it.
@@ -56,6 +57,11 @@ void __weak arch_irq_work_raise(void)
 	 */
 }
 
+int __init __weak arch_irq_work_has_own_interrupt(void)
+{
+	return 0;
+}
+
 #ifdef CONFIG_SMP
 /*
  * Enqueue the irq_work @work on @cpu unless it's already pending
@@ -115,7 +121,8 @@ bool irq_work_needs_cpu(void)
 
 	raised = &__get_cpu_var(raised_list);
 	lazy = &__get_cpu_var(lazy_list);
-	if (llist_empty(raised) && llist_empty(lazy))
+
+	if ((llist_empty(raised) || has_own_interrupt) && llist_empty(lazy))
 		return false;
 
 	/* All work should have been flushed before going offline */
@@ -171,6 +178,13 @@ void irq_work_run(void)
 }
 EXPORT_SYMBOL_GPL(irq_work_run);
 
+void irq_work_tick(void)
+{
+	if (!has_own_interrupt)
+		irq_work_run_list(&__get_cpu_var(raised_list));
+	irq_work_run_list(&__get_cpu_var(lazy_list));
+}
+
 /*
  * Synchronize against the irq_work @entry, ensures the entry is not
  * currently in use.
@@ -183,3 +197,8 @@ void irq_work_sync(struct irq_work *work)
 		cpu_relax();
 }
 EXPORT_SYMBOL_GPL(irq_work_sync);
+
+void __init irq_work_init(void)
+{
+	has_own_interrupt = arch_irq_work_has_own_interrupt();
+}
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index aca5dfe..9bbb834 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1385,7 +1385,7 @@ void update_process_times(int user_tick)
 	rcu_check_callbacks(cpu, user_tick);
 #ifdef CONFIG_IRQ_WORK
 	if (in_irq())
-		irq_work_run();
+		irq_work_tick();
 #endif
 	scheduler_tick();
 	run_posix_cpu_timers(p);
-- 
1.8.3.1

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Peter Zijlstra]

On Thu, Aug 21, 2014 at 04:52:51PM +0200, Frederic Weisbecker wrote:
> diff --git a/kernel/irq_work.c b/kernel/irq_work.c
> index e6bcbe7..17bd203 100644
> --- a/kernel/irq_work.c
> +++ b/kernel/irq_work.c
> @@ -22,6 +22,7 @@
>  
>  static DEFINE_PER_CPU(struct llist_head, raised_list);
>  static DEFINE_PER_CPU(struct llist_head, lazy_list);
> +static bool has_own_interrupt;
>  
>  /*
>   * Claim the entry so that no one else will poke at it.

> +void irq_work_tick(void)
> +{
> +	if (!has_own_interrupt)
> +		irq_work_run_list(&__get_cpu_var(raised_list));
> +	irq_work_run_list(&__get_cpu_var(lazy_list));
> +}
> +
>  /*
>   * Synchronize against the irq_work @entry, ensures the entry is not
>   * currently in use.

Why not something like the below; then it becomes a compile time matter.

---
diff --git a/arch/alpha/include/asm/Kbuild b/arch/alpha/include/asm/Kbuild
index e858aa0..a52cbf1 100644
--- a/arch/alpha/include/asm/Kbuild
+++ b/arch/alpha/include/asm/Kbuild
@@ -4,6 +4,7 @@ generic-y += clkdev.h
 generic-y += cputime.h
 generic-y += exec.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
 generic-y += scatterlist.h
diff --git a/arch/arc/include/asm/Kbuild b/arch/arc/include/asm/Kbuild
index e76fd79..b8fffc1 100644
--- a/arch/arc/include/asm/Kbuild
+++ b/arch/arc/include/asm/Kbuild
@@ -18,6 +18,7 @@ generic-y += ioctl.h
 generic-y += ioctls.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kmap_types.h
 generic-y += kvm_para.h
 generic-y += local.h
diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
index 70cd84e..202905e 100644
--- a/arch/arm/include/asm/Kbuild
+++ b/arch/arm/include/asm/Kbuild
@@ -11,6 +11,7 @@ generic-y += hash.h
 generic-y += ioctl.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += local.h
 generic-y += local64.h
diff --git a/arch/arm64/include/asm/Kbuild b/arch/arm64/include/asm/Kbuild
index 0b3fcf8..d617789 100644
--- a/arch/arm64/include/asm/Kbuild
+++ b/arch/arm64/include/asm/Kbuild
@@ -9,8 +9,8 @@ generic-y += current.h
 generic-y += delay.h
 generic-y += div64.h
 generic-y += dma.h
-generic-y += emergency-restart.h
 generic-y += early_ioremap.h
+generic-y += emergency-restart.h
 generic-y += errno.h
 generic-y += ftrace.h
 generic-y += hash.h
@@ -19,6 +19,7 @@ generic-y += ioctl.h
 generic-y += ioctls.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += kvm_para.h
diff --git a/arch/avr32/include/asm/Kbuild b/arch/avr32/include/asm/Kbuild
index 00a0f3c..2a71b1c 100644
--- a/arch/avr32/include/asm/Kbuild
+++ b/arch/avr32/include/asm/Kbuild
@@ -9,6 +9,7 @@ generic-y += exec.h
 generic-y += futex.h
 generic-y += hash.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += local.h
 generic-y += local64.h
 generic-y += mcs_spinlock.h
diff --git a/arch/blackfin/include/asm/Kbuild b/arch/blackfin/include/asm/Kbuild
index 0d93b9a..46ed6bb 100644
--- a/arch/blackfin/include/asm/Kbuild
+++ b/arch/blackfin/include/asm/Kbuild
@@ -15,6 +15,7 @@ generic-y += hw_irq.h
 generic-y += ioctl.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += kvm_para.h
diff --git a/arch/c6x/include/asm/Kbuild b/arch/c6x/include/asm/Kbuild
index 8dbdce8..e77e0c1 100644
--- a/arch/c6x/include/asm/Kbuild
+++ b/arch/c6x/include/asm/Kbuild
@@ -22,6 +22,7 @@ generic-y += ioctl.h
 generic-y += ioctls.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += local.h
diff --git a/arch/cris/include/asm/Kbuild b/arch/cris/include/asm/Kbuild
index 31742df..802b94c4 100644
--- a/arch/cris/include/asm/Kbuild
+++ b/arch/cris/include/asm/Kbuild
@@ -8,6 +8,7 @@ generic-y += clkdev.h
 generic-y += cputime.h
 generic-y += exec.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += kvm_para.h
 generic-y += linkage.h
 generic-y += mcs_spinlock.h
diff --git a/arch/frv/include/asm/Kbuild b/arch/frv/include/asm/Kbuild
index 5b73921..3caf05c 100644
--- a/arch/frv/include/asm/Kbuild
+++ b/arch/frv/include/asm/Kbuild
@@ -3,6 +3,7 @@ generic-y += clkdev.h
 generic-y += cputime.h
 generic-y += exec.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
 generic-y += scatterlist.h
diff --git a/arch/hexagon/include/asm/Kbuild b/arch/hexagon/include/asm/Kbuild
index 0e69796..5f234a5 100644
--- a/arch/hexagon/include/asm/Kbuild
+++ b/arch/hexagon/include/asm/Kbuild
@@ -23,6 +23,7 @@ generic-y += ioctls.h
 generic-y += iomap.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += local.h
diff --git a/arch/ia64/include/asm/Kbuild b/arch/ia64/include/asm/Kbuild
index e8317d2..747320b 100644
--- a/arch/ia64/include/asm/Kbuild
+++ b/arch/ia64/include/asm/Kbuild
@@ -2,6 +2,7 @@
 generic-y += clkdev.h
 generic-y += exec.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += kvm_para.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
diff --git a/arch/m32r/include/asm/Kbuild b/arch/m32r/include/asm/Kbuild
index accc10a..e02448b 100644
--- a/arch/m32r/include/asm/Kbuild
+++ b/arch/m32r/include/asm/Kbuild
@@ -3,6 +3,7 @@ generic-y += clkdev.h
 generic-y += cputime.h
 generic-y += exec.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += module.h
 generic-y += preempt.h
diff --git a/arch/m68k/include/asm/Kbuild b/arch/m68k/include/asm/Kbuild
index c67c94a..dbaf9f3 100644
--- a/arch/m68k/include/asm/Kbuild
+++ b/arch/m68k/include/asm/Kbuild
@@ -11,6 +11,7 @@ generic-y += hw_irq.h
 generic-y += ioctl.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += kvm_para.h
diff --git a/arch/metag/include/asm/Kbuild b/arch/metag/include/asm/Kbuild
index c29ead8..7b8111c 100644
--- a/arch/metag/include/asm/Kbuild
+++ b/arch/metag/include/asm/Kbuild
@@ -19,6 +19,7 @@ generic-y += ioctl.h
 generic-y += ioctls.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += kvm_para.h
diff --git a/arch/microblaze/include/asm/Kbuild b/arch/microblaze/include/asm/Kbuild
index 27a3acd..448143b 100644
--- a/arch/microblaze/include/asm/Kbuild
+++ b/arch/microblaze/include/asm/Kbuild
@@ -5,6 +5,7 @@ generic-y += cputime.h
 generic-y += device.h
 generic-y += exec.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
 generic-y += scatterlist.h
diff --git a/arch/mips/include/asm/Kbuild b/arch/mips/include/asm/Kbuild
index 335e529..57012ef 100644
--- a/arch/mips/include/asm/Kbuild
+++ b/arch/mips/include/asm/Kbuild
@@ -3,6 +3,7 @@ generic-y += cputime.h
 generic-y += current.h
 generic-y += emergency-restart.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += local64.h
 generic-y += mcs_spinlock.h
 generic-y += mutex.h
diff --git a/arch/mn10300/include/asm/Kbuild b/arch/mn10300/include/asm/Kbuild
index ecbd667..77eb1a6 100644
--- a/arch/mn10300/include/asm/Kbuild
+++ b/arch/mn10300/include/asm/Kbuild
@@ -4,6 +4,7 @@ generic-y += clkdev.h
 generic-y += cputime.h
 generic-y += exec.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
 generic-y += scatterlist.h
diff --git a/arch/openrisc/include/asm/Kbuild b/arch/openrisc/include/asm/Kbuild
index 480af0d..89b61d7 100644
--- a/arch/openrisc/include/asm/Kbuild
+++ b/arch/openrisc/include/asm/Kbuild
@@ -31,6 +31,7 @@ generic-y += ioctl.h
 generic-y += ioctls.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += kvm_para.h
diff --git a/arch/parisc/include/asm/Kbuild b/arch/parisc/include/asm/Kbuild
index ecf25e6..ffb024b 100644
--- a/arch/parisc/include/asm/Kbuild
+++ b/arch/parisc/include/asm/Kbuild
@@ -10,6 +10,7 @@ generic-y += exec.h
 generic-y += hash.h
 generic-y += hw_irq.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kvm_para.h
 generic-y += local.h
diff --git a/arch/powerpc/include/asm/Kbuild b/arch/powerpc/include/asm/Kbuild
index 7f23f16..31e8f59 100644
--- a/arch/powerpc/include/asm/Kbuild
+++ b/arch/powerpc/include/asm/Kbuild
@@ -1,6 +1,7 @@
 
 generic-y += clkdev.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
 generic-y += rwsem.h
diff --git a/arch/s390/include/asm/Kbuild b/arch/s390/include/asm/Kbuild
index b3fea07..773f866 100644
--- a/arch/s390/include/asm/Kbuild
+++ b/arch/s390/include/asm/Kbuild
@@ -2,6 +2,7 @@
 
 generic-y += clkdev.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
 generic-y += scatterlist.h
diff --git a/arch/score/include/asm/Kbuild b/arch/score/include/asm/Kbuild
index aad2091..1909d2a 100644
--- a/arch/score/include/asm/Kbuild
+++ b/arch/score/include/asm/Kbuild
@@ -6,6 +6,7 @@ generic-y += barrier.h
 generic-y += clkdev.h
 generic-y += cputime.h
 generic-y += hash.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += preempt.h
 generic-y += scatterlist.h
diff --git a/arch/sh/include/asm/Kbuild b/arch/sh/include/asm/Kbuild
index c19e47d..5a6c9ac 100644
--- a/arch/sh/include/asm/Kbuild
+++ b/arch/sh/include/asm/Kbuild
@@ -12,6 +12,7 @@ generic-y += hash.h
 generic-y += ioctl.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kvm_para.h
 generic-y += local.h
 generic-y += local64.h
diff --git a/arch/sparc/include/asm/Kbuild b/arch/sparc/include/asm/Kbuild
index cdd1b44..f5f94ce 100644
--- a/arch/sparc/include/asm/Kbuild
+++ b/arch/sparc/include/asm/Kbuild
@@ -8,6 +8,7 @@ generic-y += emergency-restart.h
 generic-y += exec.h
 generic-y += hash.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += linkage.h
 generic-y += local.h
 generic-y += local64.h
diff --git a/arch/tile/include/asm/Kbuild b/arch/tile/include/asm/Kbuild
index 0aa5675..e6462b8 100644
--- a/arch/tile/include/asm/Kbuild
+++ b/arch/tile/include/asm/Kbuild
@@ -17,6 +17,7 @@ generic-y += ioctl.h
 generic-y += ioctls.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += local.h
 generic-y += local64.h
 generic-y += mcs_spinlock.h
diff --git a/arch/um/include/asm/Kbuild b/arch/um/include/asm/Kbuild
index 7bd64aa..244b12c 100644
--- a/arch/um/include/asm/Kbuild
+++ b/arch/um/include/asm/Kbuild
@@ -14,6 +14,7 @@ generic-y += hash.h
 generic-y += hw_irq.h
 generic-y += io.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += mcs_spinlock.h
 generic-y += mutex.h
diff --git a/arch/unicore32/include/asm/Kbuild b/arch/unicore32/include/asm/Kbuild
index 1e5fb87..5a2bb53 100644
--- a/arch/unicore32/include/asm/Kbuild
+++ b/arch/unicore32/include/asm/Kbuild
@@ -22,6 +22,7 @@ generic-y += ioctl.h
 generic-y += ioctls.h
 generic-y += ipcbuf.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += local.h
diff --git a/arch/x86/include/asm/Kbuild b/arch/x86/include/asm/Kbuild
index 3bf000f..8fa909c 100644
--- a/arch/x86/include/asm/Kbuild
+++ b/arch/x86/include/asm/Kbuild
@@ -7,5 +7,6 @@ genhdr-y += unistd_x32.h
 generic-y += clkdev.h
 generic-y += cputime.h
 generic-y += early_ioremap.h
+generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += scatterlist.h
diff --git a/arch/xtensa/include/asm/Kbuild b/arch/xtensa/include/asm/Kbuild
index c3d20ba..105d389 100644
--- a/arch/xtensa/include/asm/Kbuild
+++ b/arch/xtensa/include/asm/Kbuild
@@ -12,6 +12,7 @@ generic-y += hardirq.h
 generic-y += hash.h
 generic-y += ioctl.h
 generic-y += irq_regs.h
+generic-y += irq_work.h
 generic-y += kdebug.h
 generic-y += kmap_types.h
 generic-y += kvm_para.h
diff --git a/include/asm-generic/irq_work.h b/include/asm-generic/irq_work.h
index e69de29..e962b41 100644
--- a/include/asm-generic/irq_work.h
+++ b/include/asm-generic/irq_work.h
@@ -0,0 +1,10 @@
+#ifndef __ASM_IRQ_WORK_H
+#define __ASM_IRQ_WORK_H
+
+static inline bool arch_irq_work_interrupt(void)
+{
+	return false;
+}
+
+#undef /* __ASM_IRQ_WORK_H */
+

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Frederic Weisbecker]

On Thu, Sep 04, 2014 at 08:11:37AM +0200, Peter Zijlstra wrote:
> On Thu, Aug 21, 2014 at 04:52:51PM +0200, Frederic Weisbecker wrote:
> > diff --git a/kernel/irq_work.c b/kernel/irq_work.c
> > index e6bcbe7..17bd203 100644
> > --- a/kernel/irq_work.c
> > +++ b/kernel/irq_work.c
> > @@ -22,6 +22,7 @@
> >  
> >  static DEFINE_PER_CPU(struct llist_head, raised_list);
> >  static DEFINE_PER_CPU(struct llist_head, lazy_list);
> > +static bool has_own_interrupt;
> >  
> >  /*
> >   * Claim the entry so that no one else will poke at it.
> 
> > +void irq_work_tick(void)
> > +{
> > +	if (!has_own_interrupt)
> > +		irq_work_run_list(&__get_cpu_var(raised_list));
> > +	irq_work_run_list(&__get_cpu_var(lazy_list));
> > +}
> > +
> >  /*
> >   * Synchronize against the irq_work @entry, ensures the entry is not
> >   * currently in use.
> 
> Why not something like the below; then it becomes a compile time matter.

Problem is that some archs only have that information at runtime.
See x86 for example which tests cpu_has_apic.

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Peter Zijlstra]

On Thu, Sep 04, 2014 at 03:33:40PM +0200, Frederic Weisbecker wrote:
> > Why not something like the below; then it becomes a compile time matter.
> 
> Problem is that some archs only have that information at runtime.
> See x86 for example which tests cpu_has_apic.

So the other archs can still benefit, no? Also, on x86 cpu features can
be used with asm_goto/jump_labels, your approach wrecks that too by
copying it into some static variable, see static_cpu_has().

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Frederic Weisbecker]

On Thu, Sep 04, 2014 at 05:40:20PM +0200, Peter Zijlstra wrote:
> On Thu, Sep 04, 2014 at 03:33:40PM +0200, Frederic Weisbecker wrote:
> > > Why not something like the below; then it becomes a compile time matter.
> > 
> > Problem is that some archs only have that information at runtime.
> > See x86 for example which tests cpu_has_apic.
> 
> So the other archs can still benefit, no? Also, on x86 cpu features can
> be used with asm_goto/jump_labels, your approach wrecks that too by
> copying it into some static variable, see static_cpu_has().

Ah I didn't know about that static_cpu_has().
I'm going to try the way you suggest then.

Thanks.

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Frederic Weisbecker]

On Thu, Sep 04, 2014 at 08:11:37AM +0200, Peter Zijlstra wrote:
> On Thu, Aug 21, 2014 at 04:52:51PM +0200, Frederic Weisbecker wrote:
> > diff --git a/kernel/irq_work.c b/kernel/irq_work.c
> > index e6bcbe7..17bd203 100644
> > --- a/kernel/irq_work.c
> > +++ b/kernel/irq_work.c
> > @@ -22,6 +22,7 @@
> >  
> >  static DEFINE_PER_CPU(struct llist_head, raised_list);
> >  static DEFINE_PER_CPU(struct llist_head, lazy_list);
> > +static bool has_own_interrupt;
> >  
> >  /*
> >   * Claim the entry so that no one else will poke at it.
> 
> > +void irq_work_tick(void)
> > +{
> > +	if (!has_own_interrupt)
> > +		irq_work_run_list(&__get_cpu_var(raised_list));
> > +	irq_work_run_list(&__get_cpu_var(lazy_list));
> > +}
> > +
> >  /*
> >   * Synchronize against the irq_work @entry, ensures the entry is not
> >   * currently in use.
> 
> Why not something like the below; then it becomes a compile time matter.

You have a script that does that arch/*/include/asm/Kbuild edit for you right?
Is this something in scripts/ ?

Thanks.
 
> ---
> diff --git a/arch/alpha/include/asm/Kbuild b/arch/alpha/include/asm/Kbuild
> index e858aa0..a52cbf1 100644
> --- a/arch/alpha/include/asm/Kbuild
> +++ b/arch/alpha/include/asm/Kbuild
> @@ -4,6 +4,7 @@ generic-y += clkdev.h
>  generic-y += cputime.h
>  generic-y += exec.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
>  generic-y += scatterlist.h
> diff --git a/arch/arc/include/asm/Kbuild b/arch/arc/include/asm/Kbuild
> index e76fd79..b8fffc1 100644
> --- a/arch/arc/include/asm/Kbuild
> +++ b/arch/arc/include/asm/Kbuild
> @@ -18,6 +18,7 @@ generic-y += ioctl.h
>  generic-y += ioctls.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kmap_types.h
>  generic-y += kvm_para.h
>  generic-y += local.h
> diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
> index 70cd84e..202905e 100644
> --- a/arch/arm/include/asm/Kbuild
> +++ b/arch/arm/include/asm/Kbuild
> @@ -11,6 +11,7 @@ generic-y += hash.h
>  generic-y += ioctl.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += local.h
>  generic-y += local64.h
> diff --git a/arch/arm64/include/asm/Kbuild b/arch/arm64/include/asm/Kbuild
> index 0b3fcf8..d617789 100644
> --- a/arch/arm64/include/asm/Kbuild
> +++ b/arch/arm64/include/asm/Kbuild
> @@ -9,8 +9,8 @@ generic-y += current.h
>  generic-y += delay.h
>  generic-y += div64.h
>  generic-y += dma.h
> -generic-y += emergency-restart.h
>  generic-y += early_ioremap.h
> +generic-y += emergency-restart.h
>  generic-y += errno.h
>  generic-y += ftrace.h
>  generic-y += hash.h
> @@ -19,6 +19,7 @@ generic-y += ioctl.h
>  generic-y += ioctls.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += kvm_para.h
> diff --git a/arch/avr32/include/asm/Kbuild b/arch/avr32/include/asm/Kbuild
> index 00a0f3c..2a71b1c 100644
> --- a/arch/avr32/include/asm/Kbuild
> +++ b/arch/avr32/include/asm/Kbuild
> @@ -9,6 +9,7 @@ generic-y += exec.h
>  generic-y += futex.h
>  generic-y += hash.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += local.h
>  generic-y += local64.h
>  generic-y += mcs_spinlock.h
> diff --git a/arch/blackfin/include/asm/Kbuild b/arch/blackfin/include/asm/Kbuild
> index 0d93b9a..46ed6bb 100644
> --- a/arch/blackfin/include/asm/Kbuild
> +++ b/arch/blackfin/include/asm/Kbuild
> @@ -15,6 +15,7 @@ generic-y += hw_irq.h
>  generic-y += ioctl.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += kvm_para.h
> diff --git a/arch/c6x/include/asm/Kbuild b/arch/c6x/include/asm/Kbuild
> index 8dbdce8..e77e0c1 100644
> --- a/arch/c6x/include/asm/Kbuild
> +++ b/arch/c6x/include/asm/Kbuild
> @@ -22,6 +22,7 @@ generic-y += ioctl.h
>  generic-y += ioctls.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += local.h
> diff --git a/arch/cris/include/asm/Kbuild b/arch/cris/include/asm/Kbuild
> index 31742df..802b94c4 100644
> --- a/arch/cris/include/asm/Kbuild
> +++ b/arch/cris/include/asm/Kbuild
> @@ -8,6 +8,7 @@ generic-y += clkdev.h
>  generic-y += cputime.h
>  generic-y += exec.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += kvm_para.h
>  generic-y += linkage.h
>  generic-y += mcs_spinlock.h
> diff --git a/arch/frv/include/asm/Kbuild b/arch/frv/include/asm/Kbuild
> index 5b73921..3caf05c 100644
> --- a/arch/frv/include/asm/Kbuild
> +++ b/arch/frv/include/asm/Kbuild
> @@ -3,6 +3,7 @@ generic-y += clkdev.h
>  generic-y += cputime.h
>  generic-y += exec.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
>  generic-y += scatterlist.h
> diff --git a/arch/hexagon/include/asm/Kbuild b/arch/hexagon/include/asm/Kbuild
> index 0e69796..5f234a5 100644
> --- a/arch/hexagon/include/asm/Kbuild
> +++ b/arch/hexagon/include/asm/Kbuild
> @@ -23,6 +23,7 @@ generic-y += ioctls.h
>  generic-y += iomap.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += local.h
> diff --git a/arch/ia64/include/asm/Kbuild b/arch/ia64/include/asm/Kbuild
> index e8317d2..747320b 100644
> --- a/arch/ia64/include/asm/Kbuild
> +++ b/arch/ia64/include/asm/Kbuild
> @@ -2,6 +2,7 @@
>  generic-y += clkdev.h
>  generic-y += exec.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += kvm_para.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
> diff --git a/arch/m32r/include/asm/Kbuild b/arch/m32r/include/asm/Kbuild
> index accc10a..e02448b 100644
> --- a/arch/m32r/include/asm/Kbuild
> +++ b/arch/m32r/include/asm/Kbuild
> @@ -3,6 +3,7 @@ generic-y += clkdev.h
>  generic-y += cputime.h
>  generic-y += exec.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += module.h
>  generic-y += preempt.h
> diff --git a/arch/m68k/include/asm/Kbuild b/arch/m68k/include/asm/Kbuild
> index c67c94a..dbaf9f3 100644
> --- a/arch/m68k/include/asm/Kbuild
> +++ b/arch/m68k/include/asm/Kbuild
> @@ -11,6 +11,7 @@ generic-y += hw_irq.h
>  generic-y += ioctl.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += kvm_para.h
> diff --git a/arch/metag/include/asm/Kbuild b/arch/metag/include/asm/Kbuild
> index c29ead8..7b8111c 100644
> --- a/arch/metag/include/asm/Kbuild
> +++ b/arch/metag/include/asm/Kbuild
> @@ -19,6 +19,7 @@ generic-y += ioctl.h
>  generic-y += ioctls.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += kvm_para.h
> diff --git a/arch/microblaze/include/asm/Kbuild b/arch/microblaze/include/asm/Kbuild
> index 27a3acd..448143b 100644
> --- a/arch/microblaze/include/asm/Kbuild
> +++ b/arch/microblaze/include/asm/Kbuild
> @@ -5,6 +5,7 @@ generic-y += cputime.h
>  generic-y += device.h
>  generic-y += exec.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
>  generic-y += scatterlist.h
> diff --git a/arch/mips/include/asm/Kbuild b/arch/mips/include/asm/Kbuild
> index 335e529..57012ef 100644
> --- a/arch/mips/include/asm/Kbuild
> +++ b/arch/mips/include/asm/Kbuild
> @@ -3,6 +3,7 @@ generic-y += cputime.h
>  generic-y += current.h
>  generic-y += emergency-restart.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += local64.h
>  generic-y += mcs_spinlock.h
>  generic-y += mutex.h
> diff --git a/arch/mn10300/include/asm/Kbuild b/arch/mn10300/include/asm/Kbuild
> index ecbd667..77eb1a6 100644
> --- a/arch/mn10300/include/asm/Kbuild
> +++ b/arch/mn10300/include/asm/Kbuild
> @@ -4,6 +4,7 @@ generic-y += clkdev.h
>  generic-y += cputime.h
>  generic-y += exec.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
>  generic-y += scatterlist.h
> diff --git a/arch/openrisc/include/asm/Kbuild b/arch/openrisc/include/asm/Kbuild
> index 480af0d..89b61d7 100644
> --- a/arch/openrisc/include/asm/Kbuild
> +++ b/arch/openrisc/include/asm/Kbuild
> @@ -31,6 +31,7 @@ generic-y += ioctl.h
>  generic-y += ioctls.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += kvm_para.h
> diff --git a/arch/parisc/include/asm/Kbuild b/arch/parisc/include/asm/Kbuild
> index ecf25e6..ffb024b 100644
> --- a/arch/parisc/include/asm/Kbuild
> +++ b/arch/parisc/include/asm/Kbuild
> @@ -10,6 +10,7 @@ generic-y += exec.h
>  generic-y += hash.h
>  generic-y += hw_irq.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kvm_para.h
>  generic-y += local.h
> diff --git a/arch/powerpc/include/asm/Kbuild b/arch/powerpc/include/asm/Kbuild
> index 7f23f16..31e8f59 100644
> --- a/arch/powerpc/include/asm/Kbuild
> +++ b/arch/powerpc/include/asm/Kbuild
> @@ -1,6 +1,7 @@
>  
>  generic-y += clkdev.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
>  generic-y += rwsem.h
> diff --git a/arch/s390/include/asm/Kbuild b/arch/s390/include/asm/Kbuild
> index b3fea07..773f866 100644
> --- a/arch/s390/include/asm/Kbuild
> +++ b/arch/s390/include/asm/Kbuild
> @@ -2,6 +2,7 @@
>  
>  generic-y += clkdev.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
>  generic-y += scatterlist.h
> diff --git a/arch/score/include/asm/Kbuild b/arch/score/include/asm/Kbuild
> index aad2091..1909d2a 100644
> --- a/arch/score/include/asm/Kbuild
> +++ b/arch/score/include/asm/Kbuild
> @@ -6,6 +6,7 @@ generic-y += barrier.h
>  generic-y += clkdev.h
>  generic-y += cputime.h
>  generic-y += hash.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += preempt.h
>  generic-y += scatterlist.h
> diff --git a/arch/sh/include/asm/Kbuild b/arch/sh/include/asm/Kbuild
> index c19e47d..5a6c9ac 100644
> --- a/arch/sh/include/asm/Kbuild
> +++ b/arch/sh/include/asm/Kbuild
> @@ -12,6 +12,7 @@ generic-y += hash.h
>  generic-y += ioctl.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kvm_para.h
>  generic-y += local.h
>  generic-y += local64.h
> diff --git a/arch/sparc/include/asm/Kbuild b/arch/sparc/include/asm/Kbuild
> index cdd1b44..f5f94ce 100644
> --- a/arch/sparc/include/asm/Kbuild
> +++ b/arch/sparc/include/asm/Kbuild
> @@ -8,6 +8,7 @@ generic-y += emergency-restart.h
>  generic-y += exec.h
>  generic-y += hash.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += linkage.h
>  generic-y += local.h
>  generic-y += local64.h
> diff --git a/arch/tile/include/asm/Kbuild b/arch/tile/include/asm/Kbuild
> index 0aa5675..e6462b8 100644
> --- a/arch/tile/include/asm/Kbuild
> +++ b/arch/tile/include/asm/Kbuild
> @@ -17,6 +17,7 @@ generic-y += ioctl.h
>  generic-y += ioctls.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += local.h
>  generic-y += local64.h
>  generic-y += mcs_spinlock.h
> diff --git a/arch/um/include/asm/Kbuild b/arch/um/include/asm/Kbuild
> index 7bd64aa..244b12c 100644
> --- a/arch/um/include/asm/Kbuild
> +++ b/arch/um/include/asm/Kbuild
> @@ -14,6 +14,7 @@ generic-y += hash.h
>  generic-y += hw_irq.h
>  generic-y += io.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += mcs_spinlock.h
>  generic-y += mutex.h
> diff --git a/arch/unicore32/include/asm/Kbuild b/arch/unicore32/include/asm/Kbuild
> index 1e5fb87..5a2bb53 100644
> --- a/arch/unicore32/include/asm/Kbuild
> +++ b/arch/unicore32/include/asm/Kbuild
> @@ -22,6 +22,7 @@ generic-y += ioctl.h
>  generic-y += ioctls.h
>  generic-y += ipcbuf.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += local.h
> diff --git a/arch/x86/include/asm/Kbuild b/arch/x86/include/asm/Kbuild
> index 3bf000f..8fa909c 100644
> --- a/arch/x86/include/asm/Kbuild
> +++ b/arch/x86/include/asm/Kbuild
> @@ -7,5 +7,6 @@ genhdr-y += unistd_x32.h
>  generic-y += clkdev.h
>  generic-y += cputime.h
>  generic-y += early_ioremap.h
> +generic-y += irq_work.h
>  generic-y += mcs_spinlock.h
>  generic-y += scatterlist.h
> diff --git a/arch/xtensa/include/asm/Kbuild b/arch/xtensa/include/asm/Kbuild
> index c3d20ba..105d389 100644
> --- a/arch/xtensa/include/asm/Kbuild
> +++ b/arch/xtensa/include/asm/Kbuild
> @@ -12,6 +12,7 @@ generic-y += hardirq.h
>  generic-y += hash.h
>  generic-y += ioctl.h
>  generic-y += irq_regs.h
> +generic-y += irq_work.h
>  generic-y += kdebug.h
>  generic-y += kmap_types.h
>  generic-y += kvm_para.h
> diff --git a/include/asm-generic/irq_work.h b/include/asm-generic/irq_work.h
> index e69de29..e962b41 100644
> --- a/include/asm-generic/irq_work.h
> +++ b/include/asm-generic/irq_work.h
> @@ -0,0 +1,10 @@
> +#ifndef __ASM_IRQ_WORK_H
> +#define __ASM_IRQ_WORK_H
> +
> +static inline bool arch_irq_work_interrupt(void)
> +{
> +	return false;
> +}
> +
> +#undef /* __ASM_IRQ_WORK_H */
> +

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Peter Zijlstra]

On Sat, Sep 06, 2014 at 03:35:15PM +0200, Frederic Weisbecker wrote:
> You have a script that does that arch/*/include/asm/Kbuild edit for you right?
> Is this something in scripts/ ?

See commit b119fa61d440 ("locking/mcs: Order the header files in Kbuild
of each architecture in alphabetical order")

That script is a little buggy, but should work in the current case. So
first do something like:

	for i in arch/*/include/asm/Kbuild; do
		echo "generic-y += foo.h" >> $i;
	done

Then run that script.

IIRC the bug was where a single line contained multiple headers and/or a
continuation. I should have a better version somewhere, but this one was
easier to find.

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Frederic Weisbecker]

On Sat, Sep 06, 2014 at 05:45:56PM +0200, Peter Zijlstra wrote:
> On Sat, Sep 06, 2014 at 03:35:15PM +0200, Frederic Weisbecker wrote:
> > You have a script that does that arch/*/include/asm/Kbuild edit for you right?
> > Is this something in scripts/ ?
> 
> See commit b119fa61d440 ("locking/mcs: Order the header files in Kbuild
> of each architecture in alphabetical order")
> 
> That script is a little buggy, but should work in the current case. So
> first do something like:
> 
> 	for i in arch/*/include/asm/Kbuild; do
> 		echo "generic-y += foo.h" >> $i;
> 	done
> 
> Then run that script.
> 
> IIRC the bug was where a single line contained multiple headers and/or a
> continuation. I should have a better version somewhere, but this one was
> easier to find.

I see. Meanwhile I've applied your suggestion as a separate patch in
the series. May I use your SOB?

Thanks!

Re: [RFC PATCH 3/9] irq_work: Force raised irq work to run on irq work interrupt

[Peter Zijlstra]

[-- Attachment #1: Type: text/plain, Size: 1019 bytes --]

On Sat, Sep 06, 2014 at 07:12:26PM +0200, Frederic Weisbecker wrote:
> On Sat, Sep 06, 2014 at 05:45:56PM +0200, Peter Zijlstra wrote:
> > On Sat, Sep 06, 2014 at 03:35:15PM +0200, Frederic Weisbecker wrote:
> > > You have a script that does that arch/*/include/asm/Kbuild edit for you right?
> > > Is this something in scripts/ ?
> > 
> > See commit b119fa61d440 ("locking/mcs: Order the header files in Kbuild
> > of each architecture in alphabetical order")
> > 
> > That script is a little buggy, but should work in the current case. So
> > first do something like:
> > 
> > 	for i in arch/*/include/asm/Kbuild; do
> > 		echo "generic-y += foo.h" >> $i;
> > 	done
> > 
> > Then run that script.
> > 
> > IIRC the bug was where a single line contained multiple headers and/or a
> > continuation. I should have a better version somewhere, but this one was
> > easier to find.
> 
> I see. Meanwhile I've applied your suggestion as a separate patch in
> the series. May I use your SOB?

Sure.

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

[RFC PATCH 4/9] x86: Build irq work only if local apic support

[Frederic Weisbecker]

Irq work won't run without local apic support. This check is currently
hacked-up through ifdeffery. Lets do conditional build instead. That
way we avoid further ifdefs with the implementation of
arch_irq_work_has_own_interrupt().

Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 arch/x86/Kconfig                  | 4 ++++
 arch/x86/include/asm/entry_arch.h | 2 +-
 arch/x86/kernel/Makefile          | 2 +-
 arch/x86/kernel/entry_64.S        | 2 +-
 arch/x86/kernel/irq_work.c        | 2 --
 arch/x86/kernel/irqinit.c         | 2 +-
 6 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 5d0bf1a..147772f 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -837,6 +837,10 @@ config X86_LOCAL_APIC
 	def_bool y
 	depends on X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_APIC || PCI_MSI
 
+config X86_IRQ_WORK
+	def_bool y
+	depends on X86_LOCAL_APIC && IRQ_WORK
+
 config X86_IO_APIC
 	def_bool y
 	depends on X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC || PCI_MSI
diff --git a/arch/x86/include/asm/entry_arch.h b/arch/x86/include/asm/entry_arch.h
index dc5fa66..f8344d2 100644
--- a/arch/x86/include/asm/entry_arch.h
+++ b/arch/x86/include/asm/entry_arch.h
@@ -38,7 +38,7 @@ BUILD_INTERRUPT(apic_timer_interrupt,LOCAL_TIMER_VECTOR)
 BUILD_INTERRUPT(error_interrupt,ERROR_APIC_VECTOR)
 BUILD_INTERRUPT(spurious_interrupt,SPURIOUS_APIC_VECTOR)
 
-#ifdef CONFIG_IRQ_WORK
+#ifdef CONFIG_X86_IRQ_WORK
 BUILD_INTERRUPT(irq_work_interrupt, IRQ_WORK_VECTOR)
 #endif
 
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index b5ea75c..ddb4994 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -22,7 +22,7 @@ obj-y			:= process_$(BITS).o signal.o entry_$(BITS).o
 obj-y			+= traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o
 obj-y			+= time.o ioport.o ldt.o dumpstack.o nmi.o
 obj-y			+= setup.o x86_init.o i8259.o irqinit.o jump_label.o
-obj-$(CONFIG_IRQ_WORK)  += irq_work.o
+obj-$(CONFIG_X86_IRQ_WORK)  += irq_work.o
 obj-y			+= probe_roms.o
 obj-$(CONFIG_X86_32)	+= i386_ksyms_32.o
 obj-$(CONFIG_X86_64)	+= sys_x86_64.o x8664_ksyms_64.o
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 2fac134..85056bd7 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -1043,7 +1043,7 @@ apicinterrupt ERROR_APIC_VECTOR \
 apicinterrupt SPURIOUS_APIC_VECTOR \
 	spurious_interrupt smp_spurious_interrupt
 
-#ifdef CONFIG_IRQ_WORK
+#ifdef CONFIG_X86_IRQ_WORK
 apicinterrupt IRQ_WORK_VECTOR \
 	irq_work_interrupt smp_irq_work_interrupt
 #endif
diff --git a/arch/x86/kernel/irq_work.c b/arch/x86/kernel/irq_work.c
index 1de84e3..54f60a2 100644
--- a/arch/x86/kernel/irq_work.c
+++ b/arch/x86/kernel/irq_work.c
@@ -40,11 +40,9 @@ __visible void smp_trace_irq_work_interrupt(struct pt_regs *regs)
 
 void arch_irq_work_raise(void)
 {
-#ifdef CONFIG_X86_LOCAL_APIC
 	if (!cpu_has_apic)
 		return;
 
 	apic->send_IPI_self(IRQ_WORK_VECTOR);
 	apic_wait_icr_idle();
-#endif
 }
diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c
index 1e6cff5..566d377 100644
--- a/arch/x86/kernel/irqinit.c
+++ b/arch/x86/kernel/irqinit.c
@@ -176,7 +176,7 @@ static void __init apic_intr_init(void)
 	alloc_intr_gate(ERROR_APIC_VECTOR, error_interrupt);
 
 	/* IRQ work interrupts: */
-# ifdef CONFIG_IRQ_WORK
+# ifdef CONFIG_X86_IRQ_WORK
 	alloc_intr_gate(IRQ_WORK_VECTOR, irq_work_interrupt);
 # endif
 
-- 
1.8.3.1

[RFC PATCH 5/9] x86: Tell irq work about self IPI support

[Frederic Weisbecker]

x86 supports irq work when local apic is available. This is partly known
on runtime so lets implement arch_irq_work_has_own_interrupt()
accordingly.

This should be safely called after setup_arch().

Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 arch/x86/kernel/irq_work.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/x86/kernel/irq_work.c b/arch/x86/kernel/irq_work.c
index 54f60a2..b5f8fff 100644
--- a/arch/x86/kernel/irq_work.c
+++ b/arch/x86/kernel/irq_work.c
@@ -38,6 +38,11 @@ __visible void smp_trace_irq_work_interrupt(struct pt_regs *regs)
 	exiting_irq();
 }
 
+int __init arch_irq_work_has_own_interrupt(void)
+{
+	return cpu_has_apic;
+}
+
 void arch_irq_work_raise(void)
 {
 	if (!cpu_has_apic)
-- 
1.8.3.1

[RFC PATCH 6/9] arm: Tell irq work about self IPI support

[Frederic Weisbecker]

ARM irq work IPI support depends on SMP support. That information is
partly known at early boottime. Lets implement
arch_irq_work_has_own_interrupt() accordingly.

Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Stephen Boyd <sboyd@codeaurora.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 arch/arm/kernel/smp.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c
index 9388a3d..560c52e 100644
--- a/arch/arm/kernel/smp.c
+++ b/arch/arm/kernel/smp.c
@@ -506,6 +506,11 @@ void arch_irq_work_raise(void)
 	if (is_smp())
 		smp_cross_call(cpumask_of(smp_processor_id()), IPI_IRQ_WORK);
 }
+
+int __init arch_irq_work_has_own_interrupt(void)
+{
+	return is_smp();
+}
 #endif
 
 #ifdef CONFIG_GENERIC_CLOCKEVENTS_BROADCAST
-- 
1.8.3.1

Re: [RFC PATCH 6/9] arm: Tell irq work about self IPI support

[Russell King - ARM Linux]

On Thu, Aug 21, 2014 at 04:52:54PM +0200, Frederic Weisbecker wrote:
> ARM irq work IPI support depends on SMP support. That information is
> partly known at early boottime. Lets implement
> arch_irq_work_has_own_interrupt() accordingly.

Sorry, this makes no sense on its own.  As there's no context for this
to know whether it's an appropriate change or not, it would be
irresponsible fo me to provide an ack for something I don't understand.

-- 
FTTC broadband for 0.8mile line: currently at 9.5Mbps down 400kbps up
according to speedtest.net.

Re: [RFC PATCH 6/9] arm: Tell irq work about self IPI support

[Frederic Weisbecker]

On Thu, Aug 21, 2014 at 04:21:23PM +0100, Russell King - ARM Linux wrote:
> On Thu, Aug 21, 2014 at 04:52:54PM +0200, Frederic Weisbecker wrote:
> > ARM irq work IPI support depends on SMP support. That information is
> > partly known at early boottime. Lets implement
> > arch_irq_work_has_own_interrupt() accordingly.
> 
> Sorry, this makes no sense on its own.  As there's no context for this
> to know whether it's an appropriate change or not, it would be
> irresponsible fo me to provide an ack for something I don't understand.

Ah you're right, I should have Cc'ed you for all patches in the series.
I didn't because I wanted to avoid spaming you but sending that patch alone
doesn't make much sense.

Let me bounce the rest to you.

[RFC PATCH 7/9] arm64: Tell irq work about self IPI support

[Frederic Weisbecker]

ARM64 irq work support depends on __smp_cross_call to point to some
relevant IRQ controller operations. This information should be available
after the call to init_IRQ().

Lets implement arch_irq_work_has_own_interrupt() accordingly.

Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Larry Bassel <larry.bassel@linaro.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 arch/arm64/kernel/smp.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c
index 4743397..b3e8b9e 100644
--- a/arch/arm64/kernel/smp.c
+++ b/arch/arm64/kernel/smp.c
@@ -534,6 +534,11 @@ void arch_irq_work_raise(void)
 	if (__smp_cross_call)
 		smp_cross_call(cpumask_of(smp_processor_id()), IPI_IRQ_WORK);
 }
+
+int __init arch_irq_work_has_own_interrupt(void)
+{
+	return !!__smp_cross_call;
+}
 #endif
 
 static DEFINE_RAW_SPINLOCK(stop_lock);
-- 
1.8.3.1

[RFC PATCH 8/9] nohz: Consolidate nohz full init code

[Frederic Weisbecker]

The supports for CONFIG_NO_HZ_FULL_ALL=y and the nohz_full= kernel
parameter both have their own way to do the same thing: allocate
full dynticks cpumasks, fill them and initialize some state variables.

Lets consolidate that all in the same place.

While at it, convert some regular printk message to warnings when
fundamental allocations fail.

Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 kernel/time/tick-sched.c | 38 +++++++++++++++++++-------------------
 1 file changed, 19 insertions(+), 19 deletions(-)

diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index f654a8a..eb4af01 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -295,22 +295,12 @@ out:
 /* Parse the boot-time nohz CPU list from the kernel parameters. */
 static int __init tick_nohz_full_setup(char *str)
 {
-	int cpu;
-
 	alloc_bootmem_cpumask_var(&tick_nohz_full_mask);
-	alloc_bootmem_cpumask_var(&housekeeping_mask);
 	if (cpulist_parse(str, tick_nohz_full_mask) < 0) {
 		pr_warning("NOHZ: Incorrect nohz_full cpumask\n");
+		free_bootmem_cpumask_var(tick_nohz_full_mask);
 		return 1;
 	}
-
-	cpu = smp_processor_id();
-	if (cpumask_test_cpu(cpu, tick_nohz_full_mask)) {
-		pr_warning("NO_HZ: Clearing %d from nohz_full range for timekeeping\n", cpu);
-		cpumask_clear_cpu(cpu, tick_nohz_full_mask);
-	}
-	cpumask_andnot(housekeeping_mask,
-		       cpu_possible_mask, tick_nohz_full_mask);
 	tick_nohz_full_running = true;
 
 	return 1;
@@ -349,18 +339,11 @@ static int tick_nohz_init_all(void)
 
 #ifdef CONFIG_NO_HZ_FULL_ALL
 	if (!alloc_cpumask_var(&tick_nohz_full_mask, GFP_KERNEL)) {
-		pr_err("NO_HZ: Can't allocate full dynticks cpumask\n");
-		return err;
-	}
-	if (!alloc_cpumask_var(&housekeeping_mask, GFP_KERNEL)) {
-		pr_err("NO_HZ: Can't allocate not-full dynticks cpumask\n");
+		WARN(1, "NO_HZ: Can't allocate full dynticks cpumask\n");
 		return err;
 	}
 	err = 0;
 	cpumask_setall(tick_nohz_full_mask);
-	cpumask_clear_cpu(smp_processor_id(), tick_nohz_full_mask);
-	cpumask_clear(housekeeping_mask);
-	cpumask_set_cpu(smp_processor_id(), housekeeping_mask);
 	tick_nohz_full_running = true;
 #endif
 	return err;
@@ -375,6 +358,23 @@ void __init tick_nohz_init(void)
 			return;
 	}
 
+	if (!alloc_cpumask_var(&housekeeping_mask, GFP_KERNEL)) {
+		WARN(1, "NO_HZ: Can't allocate not-full dynticks cpumask\n");
+		cpumask_clear(tick_nohz_full_mask);
+		tick_nohz_full_running = false;
+		return;
+	}
+
+	cpu = smp_processor_id();
+
+	if (cpumask_test_cpu(cpu, tick_nohz_full_mask)) {
+		pr_warning("NO_HZ: Clearing %d from nohz_full range for timekeeping\n", cpu);
+		cpumask_clear_cpu(cpu, tick_nohz_full_mask);
+	}
+
+	cpumask_andnot(housekeeping_mask,
+		       cpu_possible_mask, tick_nohz_full_mask);
+
 	for_each_cpu(cpu, tick_nohz_full_mask)
 		context_tracking_cpu_set(cpu);
 
-- 
1.8.3.1

[RFC PATCH 9/9] nohz: nohz full depends on irq work self IPI support

[Frederic Weisbecker]

The nohz full functionality depends on IRQ work to trigger its own
interrupts. As it's used to restart the tick, we can't rely on the tick
fallback for irq work callbacks, ie: we can't use the tick to restart
the tick itself as it may well be completely stopped.

Lets reject the full dynticks initialization if that arch support isn't
available.

As a side effect, this makes sure that nohz kick is never called from
the tick. Some lockup have been reported due to the kick trying to cancel
the tick hrtimer from the tick itself, which of course doesn't work.

Cc: Catalin Iacob <iacobcatalin@gmail.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 include/linux/irq_work.h |  1 +
 kernel/time/tick-sched.c | 14 ++++++++++++++
 2 files changed, 15 insertions(+)

diff --git a/include/linux/irq_work.h b/include/linux/irq_work.h
index 708c895..cffdf47 100644
--- a/include/linux/irq_work.h
+++ b/include/linux/irq_work.h
@@ -33,6 +33,7 @@ void init_irq_work(struct irq_work *work, void (*func)(struct irq_work *))
 #define DEFINE_IRQ_WORK(name, _f) struct irq_work name = { .func = (_f), }
 
 bool irq_work_queue(struct irq_work *work);
+int arch_irq_work_has_own_interrupt(void);
 
 #ifdef CONFIG_SMP
 bool irq_work_queue_on(struct irq_work *work, int cpu);
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index eb4af01..767a081 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -365,6 +365,20 @@ void __init tick_nohz_init(void)
 		return;
 	}
 
+	/*
+	 * Full dynticks uses irq work to drive the tick rescheduling on safe
+	 * locking contexts. But then we need irq work to raise its own
+	 * interrupts to avoid circular dependency on the tick
+	 */
+	if (!arch_irq_work_has_own_interrupt()) {
+		pr_warning("NO_HZ: Can't run full dynticks because arch doesn't "
+			   "support irq work self-IPIs\n");
+		cpumask_clear(tick_nohz_full_mask);
+		cpumask_copy(housekeeping_mask, cpu_possible_mask);
+		tick_nohz_full_running = false;
+		return;
+	}
+
 	cpu = smp_processor_id();
 
 	if (cpumask_test_cpu(cpu, tick_nohz_full_mask)) {
-- 
1.8.3.1