From 5c0342ca7ef17220d8dd2da68d0d349c26ab19df Mon Sep 17 00:00:00 2001 From: Claudio Scordino Date: Tue, 14 Nov 2017 12:19:26 +0100 Subject: [PATCH 1/3] sched/deadline: Fix the description of runtime accounting in the documentation Signed-off-by: Claudio Scordino Signed-off-by: Luca Abeni Acked-by: Daniel Bristot de Oliveira Acked-by: Peter Zijlstra Cc: Jonathan Corbet Cc: Linus Torvalds Cc: Mathieu Poirier Cc: Thomas Gleixner Cc: Tommaso Cucinotta Cc: linux-doc@vger.kernel.org Link: http://lkml.kernel.org/r/1510658366-28995-1-git-send-email-claudio@evidence.eu.com Signed-off-by: Ingo Molnar --- Documentation/scheduler/sched-deadline.txt | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/Documentation/scheduler/sched-deadline.txt b/Documentation/scheduler/sched-deadline.txt index e89e36ec15a5..8ce78f82ae23 100644 --- a/Documentation/scheduler/sched-deadline.txt +++ b/Documentation/scheduler/sched-deadline.txt @@ -204,10 +204,17 @@ CONTENTS It does so by decrementing the runtime of the executing task Ti at a pace equal to - dq = -max{ Ui, (1 - Uinact) } dt + dq = -max{ Ui / Umax, (1 - Uinact - Uextra) } dt - where Uinact is the inactive utilization, computed as (this_bq - running_bw), - and Ui is the bandwidth of task Ti. + where: + + - Ui is the bandwidth of task Ti; + - Umax is the maximum reclaimable utilization (subjected to RT throttling + limits); + - Uinact is the (per runqueue) inactive utilization, computed as + (this_bq - running_bw); + - Uextra is the (per runqueue) extra reclaimable utilization + (subjected to RT throttling limits). Let's now see a trivial example of two deadline tasks with runtime equal From aa5222e92f8000ed3c1c38dddf11c83222aadfb3 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 13 Oct 2017 10:01:22 +0300 Subject: [PATCH 2/3] sched/deadline: Don't use dubious signed bitfields It doesn't cause a run-time bug, but these bitfields should be unsigned. When it's signed ->dl_throttled is set to either 0 or -1, instead of 0 and 1 as expected. The sched.h file is included into tons of places so Sparse generates a flood of warnings like this: ./include/linux/sched.h:477:54: error: dubious one-bit signed bitfield Reported-by: Matthew Wilcox Reported-by: Xin Long Signed-off-by: Dan Carpenter Reviewed-by: Luca Abeni Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: kernel-janitors@vger.kernel.org Cc: luca abeni Link: http://lkml.kernel.org/r/20171013070121.dzcncojuj2f4utij@mwanda Signed-off-by: Ingo Molnar --- include/linux/sched.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index a5dc7c98b0a2..21991d668d35 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -473,10 +473,10 @@ struct sched_dl_entity { * conditions between the inactive timer handler and the wakeup * code. */ - int dl_throttled : 1; - int dl_boosted : 1; - int dl_yielded : 1; - int dl_non_contending : 1; + unsigned int dl_throttled : 1; + unsigned int dl_boosted : 1; + unsigned int dl_yielded : 1; + unsigned int dl_non_contending : 1; /* * Bandwidth enforcement timer. Each -deadline task has its From 3f5fe9fef5b2da06b6319fab8123056da5217c3f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 22 Nov 2017 13:05:48 +0100 Subject: [PATCH 3/3] sched/debug: Fix task state recording/printout The recent conversion of the task state recording to use task_state_index() broke the sched_switch tracepoint task state output. task_state_index() returns surprisingly an index (0-7) which is then printed with __print_flags() applying bitmasks. Not really working and resulting in weird states like 'prev_state=t' instead of 'prev_state=I'. Use TASK_REPORT_MAX instead of TASK_STATE_MAX to report preemption. Build a bitmask from the return value of task_state_index() and store it in entry->prev_state, which makes __print_flags() work as expected. Signed-off-by: Thomas Gleixner Cc: Linus Torvalds Cc: Paul E. McKenney Cc: Peter Zijlstra Cc: Steven Rostedt Cc: stable@vger.kernel.org Fixes: efb40f588b43 ("sched/tracing: Fix trace_sched_switch task-state printing") Link: http://lkml.kernel.org/r/alpine.DEB.2.20.1711221304180.1751@nanos Signed-off-by: Ingo Molnar --- include/trace/events/sched.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h index 306b31de5194..bc01e06bc716 100644 --- a/include/trace/events/sched.h +++ b/include/trace/events/sched.h @@ -116,9 +116,9 @@ static inline long __trace_sched_switch_state(bool preempt, struct task_struct * * RUNNING (we will not have dequeued if state != RUNNING). */ if (preempt) - return TASK_STATE_MAX; + return TASK_REPORT_MAX; - return task_state_index(p); + return 1 << task_state_index(p); } #endif /* CREATE_TRACE_POINTS */ @@ -164,7 +164,7 @@ TRACE_EVENT(sched_switch, { 0x40, "P" }, { 0x80, "I" }) : "R", - __entry->prev_state & TASK_STATE_MAX ? "+" : "", + __entry->prev_state & TASK_REPORT_MAX ? "+" : "", __entry->next_comm, __entry->next_pid, __entry->next_prio) );