Hi Thomas,
On Thu, Mar 21, 2013 at 10:52 PM, Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
> Each architecture implements its own cpu_idle() code, which is more or
> less the same on all architectures (plus/minus a few bugs and a few
> missing extra functionalities, instrumentation ...). That also forces
> everyone who is interested in idle related features to add new
> functionality to every architecture. What a waste.
I noticed there are still prototypes for cpu_idle() in include/linux/cpu.h
and include/linux/smp.h, so I ran a full "git grep -w cpu_idle", which
suggested a few more places that needs updates:
Documentation/s390/Debugging390.txt:0001b304 T cpu_idle
Documentation/s390/Debugging390.txt:is cpu_idle+0x66 ( quiet the cpu
is asleep, don't wake it )
Documentation/scheduler/sched-arch.txt:Your cpu_idle routines need to
obey the following rules:
Documentation/scheduler/sched-arch.txt:3. When cpu_idle finds
(need_resched() == 'true'), it should call
Documentation/trace/events-power.txt:cpu_idle "state=%lu cpu_id=%lu"
Documentation/trace/ftrace.txt: <idle>-0 [002] .N.1
21169.031484: rcu_idle_exit <-cpu_idle
Documentation/trace/ftrace.txt: => cpu_idle
Documentation/trace/ftrace.txt: <idle>-0 3.N.1 11us :
rcu_idle_exit <-cpu_idle
Documentation/trace/ftrace.txt: <idle>-0 3.N.1 11us :
tick_nohz_idle_exit <-cpu_idle
Documentation/trace/ftrace.txt: <idle>-0 3.N.1 25us :
sub_preempt_count <-cpu_idle
Documentation/trace/ftrace.txt: <idle>-0 3.N.. 25us :
schedule <-cpu_idle
Documentation/trace/ftrace.txt: <idle>-0 2.N.2 3us :
cpu_idle: state=4294967295 cpu_id=2
Documentation/virtual/uml/UserModeLinux-HOWTO.txt: #3
0x100a5508 in cpu_idle () at process_kern.c:471
arch/arm/Kconfig: This option adds a write barrier to the
cpu_idle loop so that,
arch/cris/arch-v10/drivers/gpio.c: * from cpu_idle() in kernel/process.c
arch/cris/arch-v10/drivers/gpio.c: * The check in cpu_idle()
reduces latency from ~15 ms to ~6 ms
arch/cris/arch-v32/drivers/mach-fs/gpio.c: * from cpu_idle() in
kernel/process.c
arch/cris/arch-v32/drivers/mach-fs/gpio.c: * The check in
cpu_idle() reduces latency from ~15 ms to ~6 ms
arch/mips/kernel/smtc.c: printk("Dangling IXMT in cpu_idle()\n");
arch/powerpc/kernel/irq.c: * in cpu_idle() will properly re-enable everything.
arch/sparc/kernel/leon_pmc.c:/* leon_pmc.c: LEON Power-down cpu_idle() handler
arch/sparc/kernel/smp_32.c: printk("CPU[%d]: Returns from
cpu_idle!\n", smp_processor_id());
arch/sparc/kernel/smp_64.c: printk("CPU[%d]: Returns from
cpu_idle!\n", smp_processor_id());
arch/tile/include/asm/thread_info.h:/* Enable interrupts racelessly
and nap forever: helper for cpu_idle(). */
drivers/acpi/processor_driver.c: * Set flag to delay
cpu_idle/throttling initialization
include/linux/cpu.h:void cpu_idle(void);
include/linux/smp.h:extern void cpu_idle(void);
include/trace/events/power.h:DEFINE_EVENT(cpu, cpu_idle,
init/main.c: * cpu_idle.
init/main.c: /* Call into cpu_idle with preempt disabled */
init/main.c: * fragile until we cpu_idle() for the first time.
kernel/trace/power-traces.c:EXPORT_TRACEPOINT_SYMBOL_GPL(cpu_idle);
tools/perf/builtin-timechart.c: * Mapping all these
"power:cpu_idle" strings to the tracepoint
tools/perf/builtin-timechart.c: if (strcmp(event_str,
"power:cpu_idle") == 0) {
tools/perf/builtin-timechart.c: "-e", "power:cpu_idle",
tools/perf/builtin-timechart.c: if (!is_valid_tracepoint("power:cpu_idle") &&
tools/perf/builtin-top.c: "cpu_idle",
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
