public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: qianjun.kernel@gmail.com
To: tglx@linutronix.de, peterz@infradead.org, will@kernel.org,
	luto@kernel.org, linux-kernel@vger.kernel.org
Cc: laoar.shao@gmail.com, qais.yousef@arm.com, urezki@gmail.com,
	jun qian <qianjun.kernel@gmail.com>
Subject: [PATCH V7 2/4] softirq: Factor loop termination condition
Date: Tue, 15 Sep 2020 19:56:07 +0800	[thread overview]
Message-ID: <20200915115609.85106-3-qianjun.kernel@gmail.com> (raw)
In-Reply-To: <20200915115609.85106-1-qianjun.kernel@gmail.com>

From: jun qian <qianjun.kernel@gmail.com>

Add a function interface to end softirq process

Signed-off-by: jun qian <qianjun.kernel@gmail.com>
---
 kernel/softirq.c | 44 +++++++++++++++++++++++++-------------------
 1 file changed, 25 insertions(+), 19 deletions(-)

diff --git a/kernel/softirq.c b/kernel/softirq.c
index 0db77ab..5a627cd 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -198,22 +198,6 @@ void __local_bh_enable_ip(unsigned long ip, unsigned int cnt)
 }
 EXPORT_SYMBOL(__local_bh_enable_ip);
 
-/*
- * We restart softirq processing for at most MAX_SOFTIRQ_RESTART times,
- * but break the loop if need_resched() is set or after 2 ms.
- * The MAX_SOFTIRQ_TIME provides a nice upper bound in most cases, but in
- * certain cases, such as stop_machine(), jiffies may cease to
- * increment and so we need the MAX_SOFTIRQ_RESTART limit as
- * well to make sure we eventually return from this method.
- *
- * These limits have been established via experimentation.
- * The two things to balance is latency against fairness -
- * we want to handle softirqs as soon as possible, but they
- * should not be able to lock up the box.
- */
-#define MAX_SOFTIRQ_TIME	(2 * NSEC_PER_MSEC)
-#define MAX_SOFTIRQ_RESTART 10
-
 #ifdef CONFIG_TRACE_IRQFLAGS
 /*
  * When we run softirqs from irq_exit() and thus on the hardirq stack we need
@@ -247,11 +231,34 @@ static inline void lockdep_softirq_end(bool in_hardirq)
 static inline void lockdep_softirq_end(bool in_hardirq) { }
 #endif
 
+/*
+ * We restart softirq processing but break the loop if need_resched() is
+ * set or after 2 ms. The MAX_SOFTIRQ_RESTART guarantees a loop termination
+ * if sched_clock() were ever to stall.
+ *
+ * These limits have been established via experimentation.  The two things
+ * to balance is latency against fairness - we want to handle softirqs as
+ * soon as possible, but they should not be able to lock up the box.
+ */
+#define MAX_SOFTIRQ_TIME	(2 * NSEC_PER_MSEC)
+#define MAX_SOFTIRQ_RESTART    10
+
+static inline bool __softirq_needs_break(u64 start)
+{
+	if (need_resched())
+		return true;
+
+	if (sched_clock() - start >= MAX_SOFTIRQ_TIME)
+		return true;
+
+	return false;
+}
+
 asmlinkage __visible void __softirq_entry __do_softirq(void)
 {
 	u64 start = sched_clock();
 	unsigned long old_flags = current->flags;
-	int max_restart = MAX_SOFTIRQ_RESTART;
+	unsigned int max_restart = MAX_SOFTIRQ_RESTART;
 	struct softirq_action *h;
 	bool in_hardirq;
 	__u32 pending;
@@ -308,8 +315,7 @@ asmlinkage __visible void __softirq_entry __do_softirq(void)
 
 	pending = local_softirq_pending();
 	if (pending) {
-		if (sched_clock() - start < MAX_SOFTIRQ_TIME && !need_resched() &&
-		    --max_restart)
+		if (!__softirq_needs_break(start) && --max_restart)
 			goto restart;
 
 		wakeup_softirqd();
-- 
1.8.3.1


  parent reply	other threads:[~2020-09-15 12:12 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-15 11:56 [PATCH V7 0/4] Softirq:avoid large sched delay from the pending softirqs qianjun.kernel
2020-09-15 11:56 ` [PATCH V7 1/4] softirq: Use sched_clock() based timeout qianjun.kernel
2020-09-24  8:34   ` Thomas Gleixner
2020-09-15 11:56 ` qianjun.kernel [this message]
2020-09-24  8:36   ` [PATCH V7 2/4] softirq: Factor loop termination condition Thomas Gleixner
2020-09-24 12:31     ` Thomas Gleixner
2020-09-15 11:56 ` [PATCH V7 3/4] softirq: Rewrite softirq processing loop qianjun.kernel
2020-09-15 11:56 ` [PATCH V7 4/4] softirq: Allow early break the " qianjun.kernel
2020-09-24 15:37   ` Thomas Gleixner
2020-09-24 23:08     ` Frederic Weisbecker
2020-09-24 23:10       ` Frederic Weisbecker
2020-09-25 22:37       ` Thomas Gleixner
2020-09-25  0:42     ` Frederic Weisbecker
2020-09-25 22:42       ` Thomas Gleixner
2020-09-26 12:22         ` Frederic Weisbecker
2020-09-28 10:51       ` jun qian
2020-09-29 11:44         ` Frederic Weisbecker
2020-10-09 15:01           ` Qais Yousef
2020-10-13 10:43             ` Frederic Weisbecker
2020-10-13 12:40               ` Qais Yousef
2020-09-26  2:00     ` jun qian
2020-09-27  1:05   ` [softirq] 56c21abbe6: will-it-scale.per_process_ops -9.1% regression kernel test robot
2020-09-28  9:20   ` [PATCH V7 4/4] softirq: Allow early break the softirq processing loop Peter Zijlstra
2020-09-28 11:15     ` jun qian
2020-09-28  9:22   ` Peter Zijlstra
2020-09-28 10:09     ` Thomas Gleixner

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=20200915115609.85106-3-qianjun.kernel@gmail.com \
    --to=qianjun.kernel@gmail.com \
    --cc=laoar.shao@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luto@kernel.org \
    --cc=peterz@infradead.org \
    --cc=qais.yousef@arm.com \
    --cc=tglx@linutronix.de \
    --cc=urezki@gmail.com \
    --cc=will@kernel.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