From: Sam Ravnborg <sam@ravnborg.org>
To: "Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com>,
"David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>,
LKML <linux-kernel@vger.kernel.org>,
linux-arch@vger.kernel.org,
Linus Torvalds <torvalds@linux-foundation.org>,
Andrew Morton <akpm@linux-foundation.org>,
Rusty Russell <rusty@rustcorp.com.au>,
Paul McKenney <paulmck@linux.vnet.ibm.com>,
Ingo Molnar <mingo@elte.hu>,
Peter Zijlstra <peterz@infradead.org>,
Magnus Damm <magnus.damm@gmail.com>,
"David S. Miller" <davem@davemloft.net>
Subject: Re: [PATCH] sparc: Use generic idle loop
Date: Mon, 1 Apr 2013 11:06:20 +0200 [thread overview]
Message-ID: <20130401090620.GA24861@merkur.ravnborg.org> (raw)
In-Reply-To: <51592EF1.4090302@linux.vnet.ibm.com>
Hi Srivatsa,
thanks for the feedback!
@davem - I need you to look at this again. I am not sure about the changes
I do in sparc64_yield().
> > +/* the idle loop on a Sparc... ;) */
> > +void arch_cpu_idle(void)
> > {
> > + if (sparc_idle)
> > + (*sparc_idle)();
> > + else
> > + cpu_relax();
> > }
>
> You need to enable interrupts when coming out in the 'else' case,
> because we enter with interrupts disabled and expect to come out
> of arch_cpu_idle() with interrupts enabled. (Otherwise you'll hit
> that WARN_ON() in the generic code). Thomas has handled a similar
> case in the mips patch.
Something like this should do it then:
/* Idle loop support */
void arch_cpu_idle(void)
{
if (sparc_idle)
(*sparc_idle)();
else
local_irq_enable();
}
I dropped cpu_relax() as this is a simple barrier() which is
part of the generic code (using rmb()).
> > -static void sparc64_yield(int cpu)
> > +/* The idle loop on sparc64. */
> > +void arch_cpu_idle_enter()
> > {
> > + int cpu;
> > +
> > + cpu = smp_processor_id();
> > +
> > if (tlb_type != hypervisor) {
> > touch_nmi_watchdog();
> > return;
> > @@ -88,31 +93,10 @@ static void sparc64_yield(int cpu)
> > set_thread_flag(TIF_POLLING_NRFLAG);
> > }
> >
>
> Hmm, this looks weird. The contents of this function should have
> been inside arch_cpu_idle() (not arch_cpu_idle_enter).
Yep - I see now.
> And it is
> unnecessary to set/clear POLLING flags here, since they are handled
> in the generic code.
OK
> (Same is the case with the need_resched and
> the cpu_is_offline check).
There are some comments about race-conditions in the commit where they were added.
I think the out-most pair can be dropped as this is done in generic code but
that the inner ones needs to stay.
I simplified arch_cpu_idle() - former sparc64_yield() based on your comments.
- dropped TIF_POLLING_NRFLAG handling
- dropped smp_mb__after_clear_bit() - this is done in generic code using rmb()
- dropped the out-most while (!need_resched() && !cpu_is_offline(cpu))
- always enable irq on exit (but I am not sure if this is correct)
But I need davem to have a look at this - as I am fooling around in
code that I do not know much about..
So this is *NOT* to be applied.
Sam
diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
index 3d361f2..ee5eacc 100644
--- a/arch/sparc/Kconfig
+++ b/arch/sparc/Kconfig
@@ -37,6 +37,7 @@ config SPARC
select GENERIC_SMP_IDLE_THREAD
select GENERIC_CMOS_UPDATE
select GENERIC_CLOCKEVENTS
+ select GENERIC_IDLE_LOOP
select GENERIC_STRNCPY_FROM_USER
select GENERIC_STRNLEN_USER
select MODULES_USE_ELF_RELA
diff --git a/arch/sparc/kernel/hvtramp.S b/arch/sparc/kernel/hvtramp.S
index 9365432..605c960 100644
--- a/arch/sparc/kernel/hvtramp.S
+++ b/arch/sparc/kernel/hvtramp.S
@@ -128,8 +128,7 @@ hv_cpu_startup:
call smp_callin
nop
- call cpu_idle
- mov 0, %o0
+
call cpu_panic
nop
diff --git a/arch/sparc/kernel/process_32.c b/arch/sparc/kernel/process_32.c
index 62eede1..890d7c4 100644
--- a/arch/sparc/kernel/process_32.c
+++ b/arch/sparc/kernel/process_32.c
@@ -64,23 +64,13 @@ extern void fpsave(unsigned long *, unsigned long *, void *, unsigned long *);
struct task_struct *last_task_used_math = NULL;
struct thread_info *current_set[NR_CPUS];
-/*
- * the idle loop on a Sparc... ;)
- */
-void cpu_idle(void)
+/* Idle loop support. */
+void arch_cpu_idle(void)
{
- set_thread_flag(TIF_POLLING_NRFLAG);
-
- /* endless idle loop with no priority at all */
- for (;;) {
- while (!need_resched()) {
- if (sparc_idle)
- (*sparc_idle)();
- else
- cpu_relax();
- }
- schedule_preempt_disabled();
- }
+ if (sparc_idle)
+ (*sparc_idle)();
+ else
+ local_irq_enable();
}
/* XXX cli/sti -> local_irq_xxx here, check this works once SMP is fixed. */
diff --git a/arch/sparc/kernel/process_64.c b/arch/sparc/kernel/process_64.c
index cdb80b2..cecb0d6 100644
--- a/arch/sparc/kernel/process_64.c
+++ b/arch/sparc/kernel/process_64.c
@@ -52,17 +52,12 @@
#include "kstack.h"
-static void sparc64_yield(int cpu)
+/* Idle loop support on sparc64. */
+void arch_cpu_idle(void)
{
if (tlb_type != hypervisor) {
touch_nmi_watchdog();
- return;
- }
-
- clear_thread_flag(TIF_POLLING_NRFLAG);
- smp_mb__after_clear_bit();
-
- while (!need_resched() && !cpu_is_offline(cpu)) {
+ } else {
unsigned long pstate;
/* Disable interrupts. */
@@ -73,7 +68,7 @@ static void sparc64_yield(int cpu)
: "=&r" (pstate)
: "i" (PSTATE_IE));
- if (!need_resched() && !cpu_is_offline(cpu))
+ if (!need_resched() && !cpu_is_offline(smp_processor_id()))
sun4v_cpu_yield();
/* Re-enable interrupts. */
@@ -84,36 +79,16 @@ static void sparc64_yield(int cpu)
: "=&r" (pstate)
: "i" (PSTATE_IE));
}
-
- set_thread_flag(TIF_POLLING_NRFLAG);
+ local_irq_enable();
}
-/* The idle loop on sparc64. */
-void cpu_idle(void)
-{
- int cpu = smp_processor_id();
-
- set_thread_flag(TIF_POLLING_NRFLAG);
-
- while(1) {
- tick_nohz_idle_enter();
- rcu_idle_enter();
-
- while (!need_resched() && !cpu_is_offline(cpu))
- sparc64_yield(cpu);
-
- rcu_idle_exit();
- tick_nohz_idle_exit();
-
#ifdef CONFIG_HOTPLUG_CPU
- if (cpu_is_offline(cpu)) {
- sched_preempt_enable_no_resched();
- cpu_play_dead();
- }
-#endif
- schedule_preempt_disabled();
- }
+void arch_cpu_idle_dead()
+{
+ sched_preempt_enable_no_resched();
+ cpu_play_dead();
}
+#endif
#ifdef CONFIG_COMPAT
static void show_regwindow32(struct pt_regs *regs)
diff --git a/arch/sparc/kernel/smp_32.c b/arch/sparc/kernel/smp_32.c
index 9e7e6d7..e3f2b81 100644
--- a/arch/sparc/kernel/smp_32.c
+++ b/arch/sparc/kernel/smp_32.c
@@ -369,7 +369,7 @@ void __cpuinit sparc_start_secondary(void *arg)
local_irq_enable();
wmb();
- cpu_idle();
+ cpu_startup_entry(CPUHP_ONLINE);
/* We should never reach here! */
BUG();
diff --git a/arch/sparc/kernel/smp_64.c b/arch/sparc/kernel/smp_64.c
index 537eb66..c025ffc 100644
--- a/arch/sparc/kernel/smp_64.c
+++ b/arch/sparc/kernel/smp_64.c
@@ -127,6 +127,8 @@ void __cpuinit smp_callin(void)
/* idle thread is expected to have preempt disabled */
preempt_disable();
+
+ cpu_startup_entry(CPUHP_ONLINE);
}
void cpu_panic(void)
diff --git a/arch/sparc/kernel/trampoline_64.S b/arch/sparc/kernel/trampoline_64.S
index da1b781..2e973a2 100644
--- a/arch/sparc/kernel/trampoline_64.S
+++ b/arch/sparc/kernel/trampoline_64.S
@@ -407,8 +407,7 @@ after_lock_tlb:
call smp_callin
nop
- call cpu_idle
- mov 0, %o0
+
call cpu_panic
nop
1: b,a,pt %xcc, 1b
next prev parent reply other threads:[~2013-04-01 9:06 UTC|newest]
Thread overview: 135+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-03-21 21:52 [patch 00/34] idle: Consolidate idle implementations Thomas Gleixner
2013-03-21 21:52 ` [patch 01/34] arch: Cleanup enable/disable_hlt Thomas Gleixner
2013-04-08 20:08 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:52 ` [patch 02/34] arch: Consolidate tsk_is_polling() Thomas Gleixner
2013-03-22 5:01 ` Tony Breeds
2013-03-22 5:20 ` Tony Breeds
2013-03-22 9:26 ` Thomas Gleixner
2013-04-08 20:09 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:52 ` [patch 03/34] idle: Implement set/clr functions for need_resched poll Thomas Gleixner
2013-03-22 9:38 ` James Hogan
2013-04-08 20:10 ` [tip:smp/hotplug] idle: Implement set/ clr " tip-bot for Thomas Gleixner
2013-03-21 21:52 ` [patch 04/34] idle: Provide a generic entry point for the idle code Thomas Gleixner
2013-04-08 20:12 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-04-14 0:42 ` Yinghai Lu
2013-03-21 21:53 ` [patch 06/34] arc: Use generic idle loop Thomas Gleixner
2013-03-22 9:02 ` Vineet Gupta
2013-04-08 20:15 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 05/34] idle: Implement generic idle function Thomas Gleixner
2013-03-23 8:56 ` Heiko Carstens
2013-03-25 10:39 ` Thomas Gleixner
2013-03-28 15:39 ` Srivatsa S. Bhat
2013-04-08 20:14 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-04-15 22:25 ` Tony Luck
2013-04-16 13:28 ` Thomas Gleixner
2013-04-16 18:25 ` Tony Luck
2013-04-16 18:35 ` [PATCH] ia64: Make sure interrupts enabled when we "safe_halt()" Luck, Tony
2013-04-17 9:00 ` [tip:smp/hotplug] ia64: Make sure interrupts enabled when we " safe_halt()" tip-bot for Luck, Tony
2013-11-18 6:05 ` [patch 05/34] idle: Implement generic idle function Viresh Kumar
2013-11-22 22:32 ` Thomas Gleixner
2013-03-21 21:53 ` [patch 07/34] alpha: Use generic idle loop Thomas Gleixner
2013-03-28 15:37 ` Srivatsa S. Bhat
2013-03-29 11:22 ` Thomas Gleixner
2013-04-08 20:16 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 08/34] arm: " Thomas Gleixner
2013-03-22 21:24 ` Kevin Hilman
2013-03-25 11:31 ` Thomas Gleixner
2013-03-25 11:48 ` Russell King - ARM Linux
2013-03-25 14:02 ` Thomas Gleixner
2013-04-08 21:47 ` Russell King - ARM Linux
2013-04-09 9:20 ` Thomas Gleixner
2013-04-09 9:38 ` Russell King - ARM Linux
2013-04-25 20:03 ` Stephen Boyd
2013-04-25 21:01 ` Thomas Gleixner
2013-05-01 0:49 ` Stephen Boyd
2013-05-01 0:55 ` Paul E. McKenney
2013-05-21 0:57 ` [PATCH] ARM: smp: Drop RCU_NONIDLE usage in cpu_die() Stephen Boyd
2013-04-08 20:17 ` [tip:smp/hotplug] arm: Use generic idle loop tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 09/34] arm64: " Thomas Gleixner
2013-03-25 18:06 ` Catalin Marinas
2013-04-08 20:18 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 10/34] avr32: " Thomas Gleixner
2013-04-08 20:20 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-04-09 7:31 ` Hans-Christian Egtvedt
2013-03-21 21:53 ` [patch 12/34] c6x: " Thomas Gleixner
2013-04-08 20:22 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 11/34] bfin: " Thomas Gleixner
2013-04-08 20:21 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 13/34] cris: " Thomas Gleixner
2013-03-27 15:04 ` Jesper Nilsson
2013-03-27 17:10 ` Thomas Gleixner
2013-04-08 20:23 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 14/34] frv: " Thomas Gleixner
2013-04-08 20:24 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 15/34] h8300: " Thomas Gleixner
2013-04-08 20:26 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 16/34] hexagon: " Thomas Gleixner
2013-04-08 20:27 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 18/34] m32r: " Thomas Gleixner
2013-04-08 20:29 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 17/34] ia64: " Thomas Gleixner
2013-03-28 15:40 ` Srivatsa S. Bhat
2013-04-08 20:28 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 19/34] m68k: " Thomas Gleixner
2013-04-08 20:30 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 20/34] metag: " Thomas Gleixner
2013-03-22 10:16 ` James Hogan
2013-03-25 11:26 ` Thomas Gleixner
2013-04-08 20:32 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 21/34] microblaze: " Thomas Gleixner
2013-04-08 20:33 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 22/34] mips: " Thomas Gleixner
2013-03-28 15:42 ` Srivatsa S. Bhat
2013-04-08 20:34 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 24/34] openrisc: " Thomas Gleixner
2013-04-08 20:36 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 23/34] mn10300: " Thomas Gleixner
2013-04-08 20:35 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 25/34] parisc: " Thomas Gleixner
2013-04-08 20:38 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 26/34] powerpc: " Thomas Gleixner
2013-03-28 15:40 ` Srivatsa S. Bhat
2013-04-01 9:13 ` Deepthi Dharwar
2013-04-08 20:39 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 27/34] s390: " Thomas Gleixner
2013-03-23 9:38 ` Heiko Carstens
2013-03-23 9:39 ` Heiko Carstens
2013-04-08 20:40 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 28/34] score: " Thomas Gleixner
2013-04-08 20:41 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 30/34] tile: Enter idle with preemption disabled Thomas Gleixner
2013-03-22 20:34 ` Chris Metcalf
2013-03-22 20:40 ` Chris Metcalf
2013-03-28 15:43 ` Srivatsa S. Bhat
2013-04-08 20:44 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 29/34] sh: Use generic idle loop Thomas Gleixner
2013-03-28 15:40 ` Srivatsa S. Bhat
2013-04-08 20:42 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 31/34] tile: " Thomas Gleixner
2013-03-28 15:41 ` Srivatsa S. Bhat
2013-03-29 11:24 ` Thomas Gleixner
2013-04-08 20:45 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 32/34] unicore: " Thomas Gleixner
2013-04-08 20:46 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 33/34] x86: " Thomas Gleixner
2013-03-28 15:43 ` Srivatsa S. Bhat
2013-04-08 20:47 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-21 21:53 ` [patch 34/34] xtensa: " Thomas Gleixner
2013-03-22 12:37 ` Max Filippov
2013-04-08 20:48 ` [tip:smp/hotplug] " tip-bot for Thomas Gleixner
2013-03-22 20:09 ` [patch 00/34] idle: Consolidate idle implementations Sam Ravnborg
[not found] ` <alpine.LFD.2.02.1303271940150.22263@ionos>
[not found] ` <5153EC43.7070808@zankel.net>
2013-03-28 9:24 ` Thomas Gleixner
2013-03-28 22:16 ` Chris Zankel
2013-03-29 16:19 ` Sam Ravnborg
2013-03-29 20:29 ` [PATCH] sparc: Use generic idle loop Sam Ravnborg
2013-03-31 23:46 ` David Miller
2013-04-01 6:53 ` Srivatsa S. Bhat
2013-04-01 9:06 ` Sam Ravnborg [this message]
2013-04-08 12:33 ` Srivatsa S. Bhat
2013-04-08 17:10 ` Sam Ravnborg
2013-04-08 19:24 ` David Miller
2013-04-11 19:38 ` [PATCH v2] " Sam Ravnborg
2013-04-12 18:56 ` Thomas Gleixner
2013-04-12 18:58 ` David Miller
2013-04-13 19:53 ` [tip:smp/hotplug] " tip-bot for Sam Ravnborg
2013-05-03 9:47 ` [patch 00/34] idle: Consolidate idle implementations Geert Uytterhoeven
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20130401090620.GA24861@merkur.ravnborg.org \
--to=sam@ravnborg.org \
--cc=akpm@linux-foundation.org \
--cc=davem@davemloft.net \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=magnus.damm@gmail.com \
--cc=mingo@elte.hu \
--cc=paulmck@linux.vnet.ibm.com \
--cc=peterz@infradead.org \
--cc=rusty@rustcorp.com.au \
--cc=srivatsa.bhat@linux.vnet.ibm.com \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®