linux-rt-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Joel Fernandes <joelaf@lab126.com>
To: <linux-rt-users@vger.kernel.org>
Subject: why does this irqsoff tracer output show interrupts disabled between irq to softirq transition?
Date: Tue, 1 Mar 2016 21:37:59 -0800	[thread overview]
Message-ID: <56D67C37.10207@lab126.com> (raw)

Hi,

I am using the irqsoff tracer on a (slightly old) kernel and I am seeing something like this. 
As you can see, after servicing the hrtimer_interrupt, and calling irq_exit(), a softirq runs.

But the softirq (which happens to be run_timer_softirq) is being called with interrupts 
disabled, - but from my understanding, interrupts are to remain enabled during softirq 
operation (unless ofcourse a spinlock is shared between a softirq and an irq handler and 
someone calls spinlock_irq_save).

So considering this, does the below makes sense?  Thanks for any insight!


  mmcqd/0-1435    0d.h1  182us : preempt_count_add <-_raw_spin_lock
  mmcqd/0-1435    0d.h2  182us : enqueue_hrtimer <-__run_hrtimer
  mmcqd/0-1435    0d.h2  183us : _raw_spin_unlock <-hrtimer_interrupt
  mmcqd/0-1435    0d.h2  183us : preempt_count_sub <-_raw_spin_unlock
  mmcqd/0-1435    0d.h1  184us : tick_program_event <-hrtimer_interrupt
  mmcqd/0-1435    0d.h1  184us : clockevents_program_event <-tick_program_event
  mmcqd/0-1435    0d.h1  184us : ktime_get <-clockevents_program_event
  mmcqd/0-1435    0d.h1  185us : lapic_next_deadline <-clockevents_program_event
  mmcqd/0-1435    0d.h1  186us : irq_exit <-smp_apic_timer_interrupt
  mmcqd/0-1435    0d.h1  186us : irqtime_account_irq <-irq_exit
  mmcqd/0-1435    0d.h1  187us : preempt_count_sub <-irq_exit
  mmcqd/0-1435    0d..1  187us : __do_softirq <-irq_exit
  mmcqd/0-1435    0d..1  188us : msecs_to_jiffies <-__do_softirq
  mmcqd/0-1435    0d..1  188us : irqtime_account_irq <-__do_softirq
  mmcqd/0-1435    0d..1  188us : __local_bh_disable_ip <-__do_softirq
  mmcqd/0-1435    0d.s1  190us : preempt_count_add <-_raw_spin_lock_irq
  mmcqd/0-1435    0d.s2  191us+: _raw_spin_unlock_irq <-run_timer_softirq
  mmcqd/0-1435    0d.s1  194us : preempt_count_add <-_raw_read_lock_irqsave
  mmcqd/0-1435    0d.s2  195us+: _raw_read_unlock_irqrestore <-cpufreq_cpu_get
  mmcqd/0-1435    0d.s1  201us : preempt_count_add <-_raw_spin_lock_irqsave
  mmcqd/0-1435    0d.s2  202us : _raw_spin_unlock_irqrestore <-cpufreq_stats_update
  mmcqd/0-1435    0d.s1  203us : preempt_count_add <-_raw_spin_lock_irqsave
  mmcqd/0-1435    0d.s2  203us+: _raw_spin_unlock_irqrestore <-cpufreq_stat_notifier_trans
  mmcqd/0-1435    0d.s1  205us : preempt_count_add <-_raw_spin_lock_irqsave
  mmcqd/0-1435    0d.s2  205us : detach_if_pending <-mod_timer_pinned
  mmcqd/0-1435    0d.s2  206us : internal_add_timer <-mod_timer_pinned
  mmcqd/0-1435    0d.s2  207us : __internal_add_timer <-internal_add_timer

             reply	other threads:[~2016-03-02  5:38 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-03-02  5:37 Joel Fernandes [this message]
2016-03-05  7:19 ` why does this irqsoff tracer output show interrupts disabled between irq to softirq transition? Joel Fernandes

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=56D67C37.10207@lab126.com \
    --to=joelaf@lab126.com \
    --cc=linux-rt-users@vger.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;
as well as URLs for NNTP newsgroup(s).