From a127d2bcf1fbc8c8e0b5cf0dab54f7d3ff50ce47 Mon Sep 17 00:00:00 2001 From: Preeti U Murthy Date: Wed, 18 Mar 2015 16:19:27 +0530 Subject: timers/tick/broadcast-hrtimer: Fix suspicious RCU usage in idle loop The hrtimer mode of broadcast queues hrtimers in the idle entry path so as to wakeup cpus in deep idle states. The associated call graph is : cpuidle_idle_call() |____ clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ENTER, ....)) |_____tick_broadcast_set_event() |____clockevents_program_event() |____bc_set_next() The hrtimer_{start/cancel} functions call into tracing which uses RCU. But it is not legal to call into RCU in cpuidle because it is one of the quiescent states. Hence protect this region with RCU_NONIDLE which informs RCU that the cpu is momentarily non-idle. As an aside it is helpful to point out that the clock event device that is programmed here is not a per-cpu clock device; it is a pseudo clock device, used by the broadcast framework alone. The per-cpu clock device programming never goes through bc_set_next(). Signed-off-by: Preeti U Murthy Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Paul E. McKenney Cc: linuxppc-dev@ozlabs.org Cc: mpe@ellerman.id.au Cc: tglx@linutronix.de Link: http://lkml.kernel.org/r/20150318104705.17763.56668.stgit@preeti.in.ibm.com Signed-off-by: Ingo Molnar --- kernel/time/tick-broadcast-hrtimer.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/kernel/time/tick-broadcast-hrtimer.c b/kernel/time/tick-broadcast-hrtimer.c index eb682d5c697c..6aac4beedbbe 100644 --- a/kernel/time/tick-broadcast-hrtimer.c +++ b/kernel/time/tick-broadcast-hrtimer.c @@ -49,6 +49,7 @@ static void bc_set_mode(enum clock_event_mode mode, */ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) { + int bc_moved; /* * We try to cancel the timer first. If the callback is on * flight on some other cpu then we let it handle it. If we @@ -60,9 +61,15 @@ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) * restart the timer because we are in the callback, but we * can set the expiry time and let the callback return * HRTIMER_RESTART. + * + * Since we are in the idle loop at this point and because + * hrtimer_{start/cancel} functions call into tracing, + * calls to these functions must be bound within RCU_NONIDLE. */ - if (hrtimer_try_to_cancel(&bctimer) >= 0) { - hrtimer_start(&bctimer, expires, HRTIMER_MODE_ABS_PINNED); + RCU_NONIDLE(bc_moved = (hrtimer_try_to_cancel(&bctimer) >= 0) ? + !hrtimer_start(&bctimer, expires, HRTIMER_MODE_ABS_PINNED) : + 0); + if (bc_moved) { /* Bind the "device" to the cpu */ bc->bound_on = smp_processor_id(); } else if (bc->bound_on == smp_processor_id()) { -- cgit v1.2.3 From 11bc26fe372fa6da81c82c68f755d2795838a640 Mon Sep 17 00:00:00 2001 From: Richard Weinberger Date: Thu, 26 Mar 2015 10:27:06 +0100 Subject: clocksource/drivers: Fix various !CONFIG_HAS_IOMEM build errors MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix !CONFIG_HAS_IOMEM related build failures in three clocksource drivers. The build failures have the pattern of: drivers/clocksource/sh_cmt.c: In function ‘sh_cmt_map_memory’: drivers/clocksource/sh_cmt.c:920:2: error: implicit declaration of function ‘ioremap_nocache’ [-Werror=implicit-function-declaration] cmt->mapbase = ioremap_nocache(mem->start, resource_size(mem)); Signed-off-by: Richard Weinberger Signed-off-by: Daniel Lezcano Acked-by: Geert Uytterhoeven Cc: maxime.ripard@free-electrons.com Link: http://lkml.kernel.org/r/1427362029-6511-1-git-send-email-daniel.lezcano@linaro.org Signed-off-by: Ingo Molnar --- drivers/clocksource/Kconfig | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig index 68161f7a07d6..a0b036ccb118 100644 --- a/drivers/clocksource/Kconfig +++ b/drivers/clocksource/Kconfig @@ -192,6 +192,7 @@ config SYS_SUPPORTS_EM_STI config SH_TIMER_CMT bool "Renesas CMT timer driver" if COMPILE_TEST depends on GENERIC_CLOCKEVENTS + depends on HAS_IOMEM default SYS_SUPPORTS_SH_CMT help This enables build of a clocksource and clockevent driver for @@ -201,6 +202,7 @@ config SH_TIMER_CMT config SH_TIMER_MTU2 bool "Renesas MTU2 timer driver" if COMPILE_TEST depends on GENERIC_CLOCKEVENTS + depends on HAS_IOMEM default SYS_SUPPORTS_SH_MTU2 help This enables build of a clockevent driver for the Multi-Function @@ -210,6 +212,7 @@ config SH_TIMER_MTU2 config SH_TIMER_TMU bool "Renesas TMU timer driver" if COMPILE_TEST depends on GENERIC_CLOCKEVENTS + depends on HAS_IOMEM default SYS_SUPPORTS_SH_TMU help This enables build of a clocksource and clockevent driver for -- cgit v1.2.3 From 6e206020324c50a95486f6b279a53512febed92d Mon Sep 17 00:00:00 2001 From: Maxime Ripard Date: Thu, 26 Mar 2015 10:27:09 +0100 Subject: clocksource/drivers/sun5i: Fix cpufreq interaction with sched_clock() The sun5i timer is used as the sched-clock on certain systems, and ever since we started using cpufreq, the cpu clock (that is one of the timer's clock indirect parent) now changes as well, along with the actual sched_clock() rate. This is not accurate and not desirable. We can safely remove the sun5i sched-clock on those systems, since we have other reliable sched_clock() sources in the system. Tested-by: Hans de Goede Signed-off-by: Maxime Ripard Signed-off-by: Daniel Lezcano [ Improved the changelog. ] Cc: richard@nod.at Link: http://lkml.kernel.org/r/1427362029-6511-4-git-send-email-daniel.lezcano@linaro.org Signed-off-by: Ingo Molnar --- drivers/clocksource/timer-sun5i.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/drivers/clocksource/timer-sun5i.c b/drivers/clocksource/timer-sun5i.c index 5dcbf90b8015..58597fbcc046 100644 --- a/drivers/clocksource/timer-sun5i.c +++ b/drivers/clocksource/timer-sun5i.c @@ -17,7 +17,6 @@ #include #include #include -#include #include #include #include @@ -137,11 +136,6 @@ static struct irqaction sun5i_timer_irq = { .dev_id = &sun5i_clockevent, }; -static u64 sun5i_timer_sched_read(void) -{ - return ~readl(timer_base + TIMER_CNTVAL_LO_REG(1)); -} - static void __init sun5i_timer_init(struct device_node *node) { struct reset_control *rstc; @@ -172,7 +166,6 @@ static void __init sun5i_timer_init(struct device_node *node) writel(TIMER_CTL_ENABLE | TIMER_CTL_RELOAD, timer_base + TIMER_CTL_REG(1)); - sched_clock_register(sun5i_timer_sched_read, 32, rate); clocksource_mmio_init(timer_base + TIMER_CNTVAL_LO_REG(1), node->name, rate, 340, 32, clocksource_mmio_readl_down); -- cgit v1.2.3