From 095031052b4c03a7c8ffa51cbab031e442c4f8b7 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 31 Jan 2008 22:45:22 +0100 Subject: RCU: add help text for "RCU implementation type" This patch supplies help text for the "RCU implementation type" kernel configuration choice. Reported-by: Stefan Richter Signed-off-by: Paul E. McKenney Signed-off-by: Ingo Molnar diff --git a/init/Kconfig b/init/Kconfig index 0d0bbf2..dcc96a8 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -775,6 +775,14 @@ config PREEMPT_NOTIFIERS choice prompt "RCU implementation type:" default CLASSIC_RCU + help + This allows you to choose either the classic RCU implementation + that is designed for best read-side performance on non-realtime + systems, or the preemptible RCU implementation for best latency + on realtime systems. Note that some kernel preemption modes + will restrict your choice. + + Select the default if you are unsure. config CLASSIC_RCU bool "Classic RCU" -- cgit v0.10.2 From 296825cbe14d4c95ee9c41ca5824f7487bfb4d9d Mon Sep 17 00:00:00 2001 From: Srivatsa Vaddagiri Date: Thu, 31 Jan 2008 22:45:22 +0100 Subject: sched: fix high wake up latencies with FAIR_USER_SCHED The reason why we are getting better wakeup latencies for !FAIR_USER_SCHED is because of this snippet of code in place_entity(): if (!initial) { /* sleeps upto a single latency don't count. */ if (sched_feat(NEW_FAIR_SLEEPERS) && entity_is_task(se)) ^^^^^^^^^^^^^^^^^^ vruntime -= sysctl_sched_latency; /* ensure we never gain time by being placed backwards. */ vruntime = max_vruntime(se->vruntime, vruntime); } NEW_FAIR_SLEEPERS feature gives credit for sleeping only to tasks and not group-level entities. With the patch attached, I could see that wakeup latencies with FAIR_USER_SCHED are restored to the same level as !FAIR_USER_SCHED. Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 72e25c7..cf958aef 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -520,7 +520,7 @@ place_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int initial) if (!initial) { /* sleeps upto a single latency don't count. */ - if (sched_feat(NEW_FAIR_SLEEPERS) && entity_is_task(se)) + if (sched_feat(NEW_FAIR_SLEEPERS)) vruntime -= sysctl_sched_latency; /* ensure we never gain time by being placed backwards. */ -- cgit v0.10.2 From ef9884e6f29bbe1075204f962a00f7533bf7e8f3 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 31 Jan 2008 22:45:22 +0100 Subject: sched: let +nice tasks have smaller impact MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Michel Dänzr has bisected an interactivity problem with plus-reniced tasks back to this commit: 810e95ccd58d91369191aa4ecc9e6d4a10d8d0c8 is first bad commit commit 810e95ccd58d91369191aa4ecc9e6d4a10d8d0c8 Author: Peter Zijlstra Date: Mon Oct 15 17:00:14 2007 +0200 sched: another wakeup_granularity fix unit mis-match: wakeup_gran was used against a vruntime fix this by assymetrically scaling the vtime of positive reniced tasks. Bisected-by: Michel Dänzer Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index cf958aef..6c091d6 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1106,7 +1106,11 @@ static void check_preempt_wakeup(struct rq *rq, struct task_struct *p) } gran = sysctl_sched_wakeup_granularity; - if (unlikely(se->load.weight != NICE_0_LOAD)) + /* + * More easily preempt - nice tasks, while not making + * it harder for + nice tasks. + */ + if (unlikely(se->load.weight > NICE_0_LOAD)) gran = calc_delta_fair(gran, &se->load); if (pse->vruntime + gran < se->vruntime) -- cgit v0.10.2 From 5aff0531ee2403b319e89bd04055fc41173e95fd Mon Sep 17 00:00:00 2001 From: Gerald Stralko Date: Thu, 31 Jan 2008 22:45:23 +0100 Subject: sched: remove unused params This removes the extra struct task_struct *p parameter in inc_nr_running and dec_nr_running functions. Signed-off by: Jerry Stralko Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index ba4c880..8355e00 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1255,12 +1255,12 @@ static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); #define sched_class_highest (&rt_sched_class) -static void inc_nr_running(struct task_struct *p, struct rq *rq) +static void inc_nr_running(struct rq *rq) { rq->nr_running++; } -static void dec_nr_running(struct task_struct *p, struct rq *rq) +static void dec_nr_running(struct rq *rq) { rq->nr_running--; } @@ -1354,7 +1354,7 @@ static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) rq->nr_uninterruptible--; enqueue_task(rq, p, wakeup); - inc_nr_running(p, rq); + inc_nr_running(rq); } /* @@ -1366,7 +1366,7 @@ static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) rq->nr_uninterruptible++; dequeue_task(rq, p, sleep); - dec_nr_running(p, rq); + dec_nr_running(rq); } /** @@ -2006,7 +2006,7 @@ void fastcall wake_up_new_task(struct task_struct *p, unsigned long clone_flags) * management (if any): */ p->sched_class->task_new(rq, p); - inc_nr_running(p, rq); + inc_nr_running(rq); } check_preempt_curr(rq, p); #ifdef CONFIG_SMP -- cgit v0.10.2 From c4772d99300a9fc13c86aaa370e630c5973664f6 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 31 Jan 2008 22:45:23 +0100 Subject: debug: turn ignore_loglevel into an early param i was debugging early crashes and wondered where all the printks went. The reason: ignore_loglevel_setup() was not called yet ... Signed-off-by: Ingo Molnar diff --git a/kernel/printk.c b/kernel/printk.c index 58bbec6..29ae1e9 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -455,10 +455,10 @@ static int __init ignore_loglevel_setup(char *str) ignore_loglevel = 1; printk(KERN_INFO "debug: ignoring loglevel setting.\n"); - return 1; + return 0; } -__setup("ignore_loglevel", ignore_loglevel_setup); +early_param("ignore_loglevel", ignore_loglevel_setup); /* * Write out chars from start to end - 1 inclusive -- cgit v0.10.2