From: Daniel Bristot de Oliveira <bristot@redhat.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Ingo Molnar <mingo@redhat.com>,
Peter Zijlstra <peterz@infradead.org>,
Andy Lutomirski <luto@kernel.org>,
Clark Williams <williams@redhat.com>,
x86@kernel.org
Subject: Re: [PATCH V2 2/2] trace,x86: Add nmi to the irq_vectors class
Date: Mon, 8 Apr 2019 15:57:07 +0200 [thread overview]
Message-ID: <1da7473c-a602-9a49-35a1-6de43bb744b9@redhat.com> (raw)
In-Reply-To: <20190408094633.7ff8dbf4@gandalf.local.home>
On 4/8/19 3:46 PM, Steven Rostedt wrote:
> On Mon, 8 Apr 2019 14:24:47 +0200
> Daniel Bristot de Oliveira <bristot@redhat.com> wrote:
>
>> NMI starts:
>> f-4447 [000] d.Z. 1487.689290: nmi_entry: vector=2
>> f-4447 [000] d.Z. 1487.689291: default_do_nmi <-do_nmi
>> f-4447 [000] d.Z. 1487.689291: nmi_handle <-default_do_nmi
>> f-4447 [000] d.Z. 1487.689291: nmi_cpu_backtrace_handler <-nmi_handle
>>
>> It handles very fast:
>> f-4447 [000] d.Z. 1487.689292: nmi_handler: nmi_cpu_backtrace_handler()
>> delta_ns: 780 handled: 0
>>
>> But some other functions continues running in the NMI context:
>> [Dazed and confused, but trying to continue message ]
>> f-4447 [000] d.Z. 1487.689292: _raw_spin_trylock <-default_do_nmi
>> f-4447 [000] d.Z. 1487.689566: unknown_nmi_error <-default_do_nmi
>> f-4447 [000] d.Z. 1487.689566: nmi_handle <-unknown_nmi_error
>> f-4447 [000] d.Z. 1487.689567: printk <-unknown_nmi_error
>> f-4447 [000] d.Z. 1487.689567: vprintk_func <-printk
>> f-4447 [000] d.Z. 1487.689567: printk_safe_log_store <-vprintk_func
>> f-4447 [000] d.Z. 1487.689569: arch_irq_work_raise <-irq_work_queue
>> f-4447 [000] d.Z. 1487.689569: default_send_IPI_self <-arch_irq_work_raise
>> f-4447 [000] d.Z. 1487.689569: __default_send_IPI_shortcut
>> <-default_send_IPI_self
>
> Just to remove confusion. Your example is to show that the new
> tracepoints would have shown that the NMI was long due to the printk? As
> running printk from NMI (even with the delayed output) isn't a normal
> path.
The example is "to show that the new tracepoints would have shown that the NMI
was longer than what the existing tracepoint pointed." The example was not the
best, I agree.. but... it serves to illustrate the idea.
-- Daniel
next prev parent reply other threads:[~2019-04-08 13:57 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-04-01 18:19 [PATCH V2 1/2] trace,x86: Add external_interrupts to the irq_vectors class Daniel Bristot de Oliveira
2019-04-01 18:19 ` [PATCH V2 2/2] trace,x86: Add nmi " Daniel Bristot de Oliveira
2019-04-05 21:50 ` Thomas Gleixner
2019-04-08 12:24 ` Daniel Bristot de Oliveira
2019-04-08 13:46 ` Steven Rostedt
2019-04-08 13:57 ` Daniel Bristot de Oliveira [this message]
2019-04-08 14:25 ` Steven Rostedt
2019-04-05 21:15 ` [PATCH V2 1/2] trace,x86: Add external_interrupts " Thomas Gleixner
2019-04-08 9:48 ` Daniel Bristot de Oliveira
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=1da7473c-a602-9a49-35a1-6de43bb744b9@redhat.com \
--to=bristot@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=williams@redhat.com \
--cc=x86@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).