Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <c60f5634e8605cb4c2ef4646b6e511e6135bea48.camel@linux.intel.com>
Subject: Re: PNP0501 serial driver takes almost 2 seconds to suspend/resume
 (printk issue)
From:   Todd Brandt <todd.e.brandt@linux.intel.com>
Reply-To: todd.e.brandt@linux.intel.com
To:     John Ogness <john.ogness@linutronix.de>,
        Sergey Senozhatsky <senozhatsky@chromium.org>,
        Petr Mladek <pmladek@suse.com>, rafael.j.wysocki@intel.com,
        len.brown@intel.com
Cc:     Andy Shevchenko <andy.shevchenko@gmail.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        "open list:SERIAL DRIVERS" <linux-serial@vger.kernel.org>,
        Steven Rostedt <rostedt@goodmis.org>
Date:   Wed, 13 Jul 2022 10:11:22 -0700
In-Reply-To: <87ilo1wdac.fsf@jogness.linutronix.de>
References: <12fb98fe27e23e3f74a139e5e8eb83a97a343372.camel@linux.intel.com>
         <51b9e2cc3baf61a604bd239b736ec2d12f1f6695.camel@linux.intel.com>
         <87czegxccb.fsf@jogness.linutronix.de>
         <045ebee30af2b80aaeace1dab18ecd113e3f17c7.camel@linux.intel.com>
         <87tu7qvx1q.fsf@jogness.linutronix.de>
         <CAHp75VfyzMNMO2NRwXwSjAmQqBbdRG3+SzyFDG+90dmvmg1xLQ@mail.gmail.com>
         <87o7xwbuoy.fsf@jogness.linutronix.de> <Ysvbp8vz7R9hDNqx@alley>
         <Ysv3JNs4RwE7kAou@google.com> <87ilo1wdac.fsf@jogness.linutronix.de>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

I've updated the bugzilla entry with new data:
https://bugzilla.kernel.org/show_bug.cgi?id=216216

I just added 3 new tests for 5.19.0-rc6 on 3 machines that see this
issue: otcpl-asus-e200-cht (cherry trail), otcpl-aml-y (amber lake),
and otcpl-whl-u (whiskey lake). The kernel has the CONFIG_PRINTK_CALLER
option enabled.

The test is a S2idle and is run thusly:
%> sleepgraph -dev -m freeze -rtcwake 15

I've included the dmesg boot logs for all three. The dmesg
suspend/resume logs are included in the html timelines by clicking the
"dmesg" button in the upper right hand corner of the timeline. There's
a "log" button as well that shows other system into.

These files are attached to the bugzilla entry.
otcpl-aml-y-5.19.0-rc6-boot-dmesg.txt
otcpl-aml-y-5.19.0-rc6-freeze.html
otcpl-asus-e200-cht-5.19.0-rc6-boot-dmesg.txt
otcpl-asus-e200-cht-5.19.0-rc6-freeze.html
otcpl-whl-u-5.19.0-rc6-boot-dmesg.txt
otcpl-whl-u-5.19.0-rc6-freeze.html

If possible can we move this thread to the bugzilla comment section?


On Wed, 2022-07-13 at 11:57 +0206, John Ogness wrote:
> On 2022-07-11, Sergey Senozhatsky <senozhatsky@chromium.org> wrote:
> > > It seems that __pr_flush() does not check whether all consoles
> > > are
> > > suspended. In this case the progress is not possible and it has
> > > to
> > > wait the entire timeout.
> > 
> > But isn't console_suspended set after pr_flush() call?
> 
> There should not be any printing after the suspend_console() message.
> If
> Todd's report is coming from 5.19-rc1, then it is likely a kthread
> issue, where the kthread is not respecting @console_suspended. (This
> would still need to be fixed for the kthreads, but would not be
> relevant
> for 5.19.)
> 
> John