All of lore.kernel.org
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: Feng Tang <feng.tang@intel.com>, "Paul E. McKenney" <paulmck@kernel.org>
Cc: linux-kernel@vger.kernel.org, john.stultz@linaro.org,
	sboyd@kernel.org, corbet@lwn.net, Mark.Rutland@arm.com,
	maz@kernel.org, kernel-team@fb.com, neeraju@codeaurora.org,
	ak@linux.intel.com, zhengjun.xing@intel.com,
	Xing Zhengjun <zhengjun.xing@linux.intel.com>
Subject: Re: [PATCH v10 clocksource 6/7] clocksource: Forgive tsc_early pre-calibration drift
Date: Tue, 27 Apr 2021 23:03:22 +0200	[thread overview]
Message-ID: <871ravo2xx.ffs@nanos.tec.linutronix.de> (raw)
In-Reply-To: <20210426150127.GB23119@shbuild999.sh.intel.com>

On Mon, Apr 26 2021 at 23:01, Feng Tang wrote:
> On Sun, Apr 25, 2021 at 03:47:07PM -0700, Paul E. McKenney wrote:
> We've reported one case that tsc can be wrongly judged as 'unstable'
> by 'refined-jiffies' watchdog [1], while reducing the threshold could
> make it easier to be triggered.
>
> It could be reproduced on the a plaform with a 115200 serial console,
> and hpet been disabled (several x86 platforms has this), add 
> 'initcall_debug' cmdline parameter to get more debug message, we can
> see:
>
> [    1.134197] clocksource: timekeeping watchdog on CPU1: Marking clocksource 'tsc-early' as unstable because the skew is too large:
> [    1.134214] clocksource:                       'refined-jiffies' wd_nesc: 500000000 wd_now: ffff8b35 wd_last: ffff8b03 mask: ffffffff

refined-jiffies is the worst of all watchdogs and this obviously cannot
be fixed at all simply because we can lose ticks in that mode. And no,
we cannot compensate for lost ticks via TSC which we in turn "monitor"
via ticks.

Even if we hack around it and make it "work" then the TSC will never
become fully trusted because refined-jiffies cannot support NOHZ/HIGHRES
mode for obvious reasons either. So the system stays in periodic mode
forever.

If there is no independent timer to validate against then the TSC is
better stable and the BIOS/SMM code has to be trusted not to wreckage
TSC. There are no other options.

So TBH, I do not care about this case at all. It's pointless to even
think about it. Either the TSC works on these systems or it doesn't. If
it doesn't, then you have to keep the pieces.

I'm so dead tired of this especially since this is known forever. But
it's obviously better to waste our time than to fix the damned hardware.

Thanks,

        tglx

  parent reply	other threads:[~2021-04-27 21:03 UTC|newest]

Thread overview: 72+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-25 22:45 [PATCH v10 clocksource 0/7] Do not mark clocks unstable due to delays for v5.13 Paul E. McKenney
2021-04-25 22:47 ` [PATCH v10 clocksource 1/7] clocksource: Provide module parameters to inject delays in watchdog Paul E. McKenney
2021-04-26  4:07   ` Andi Kleen
2021-04-26  7:13     ` Thomas Gleixner
2021-04-26 15:28     ` Paul E. McKenney
2021-04-26 16:00       ` Andi Kleen
2021-04-26 16:14         ` Paul E. McKenney
2021-04-26 17:56           ` Andi Kleen
2021-04-26 18:24             ` Paul E. McKenney
2021-04-28  4:49               ` Luming Yu
2021-04-28 13:57                 ` Paul E. McKenney
2021-04-28 14:24                   ` Luming Yu
2021-04-28 14:37                     ` Thomas Gleixner
2021-04-25 22:47 ` [PATCH v10 clocksource 2/7] clocksource: Retry clock read if long delays detected Paul E. McKenney
2021-04-27  1:44   ` Feng Tang
2021-04-25 22:47 ` [PATCH v10 clocksource 3/7] clocksource: Check per-CPU clock synchronization when marked unstable Paul E. McKenney
2021-04-26  4:12   ` Andi Kleen
2021-04-26  7:16     ` Thomas Gleixner
2021-04-25 22:47 ` [PATCH v10 clocksource 4/7] clocksource: Provide a module parameter to fuzz per-CPU clock checking Paul E. McKenney
2021-04-25 22:47 ` [PATCH v10 clocksource 5/7] clocksource: Limit number of CPUs checked for clock synchronization Paul E. McKenney
2021-04-25 22:47 ` [PATCH v10 clocksource 6/7] clocksource: Forgive tsc_early pre-calibration drift Paul E. McKenney
2021-04-26 15:01   ` Feng Tang
2021-04-26 15:25     ` Paul E. McKenney
2021-04-26 15:36       ` Feng Tang
2021-04-26 18:26         ` Paul E. McKenney
2021-04-27  1:13           ` Feng Tang
2021-04-27  3:46             ` Paul E. McKenney
2021-04-27  4:16               ` Feng Tang
2021-04-26 15:28     ` Thomas Gleixner
2021-04-27 21:03     ` Thomas Gleixner [this message]
2021-04-27  7:27   ` [clocksource] 8c30ace35d: WARNING:at_kernel/time/clocksource.c:#clocksource_watchdog kernel test robot
2021-04-27  7:27     ` kernel test robot
2021-04-27  8:45     ` Feng Tang
2021-04-27  8:45       ` Feng Tang
2021-04-27 13:37       ` Paul E. McKenney
2021-04-27 13:37         ` Paul E. McKenney
2021-04-27 17:50         ` Paul E. McKenney
2021-04-27 17:50           ` Paul E. McKenney
2021-04-27 21:09           ` Thomas Gleixner
2021-04-27 21:09             ` Thomas Gleixner
2021-04-28  1:48             ` Paul E. McKenney
2021-04-28  1:48               ` Paul E. McKenney
2021-04-28 10:14               ` Thomas Gleixner
2021-04-28 10:14                 ` Thomas Gleixner
2021-04-28 18:31                 ` Paul E. McKenney
2021-04-28 18:31                   ` Paul E. McKenney
2021-04-28 13:34             ` Thomas Gleixner
2021-04-28 13:34               ` Thomas Gleixner
2021-04-28 15:39               ` Peter Zijlstra
2021-04-28 15:39                 ` Peter Zijlstra
2021-04-28 17:00                 ` Thomas Gleixner
2021-04-28 17:00                   ` Thomas Gleixner
2021-04-29  7:38                   ` Feng Tang
2021-04-29  7:38                     ` Feng Tang
2021-04-28 18:31               ` Paul E. McKenney
2021-04-28 18:31                 ` Paul E. McKenney
2021-04-29  8:27                 ` Thomas Gleixner
2021-04-29  8:27                   ` Thomas Gleixner
2021-04-29 14:26                   ` Paul E. McKenney
2021-04-29 14:26                     ` Paul E. McKenney
2021-04-29 17:30                     ` Thomas Gleixner
2021-04-29 17:30                       ` Thomas Gleixner
2021-04-29 23:04                       ` Andi Kleen
2021-04-29 23:04                         ` Andi Kleen
2021-04-30  0:24                         ` Paul E. McKenney
2021-04-30  0:24                           ` Paul E. McKenney
2021-04-30  0:59                           ` Paul E. McKenney
2021-04-30  0:59                             ` Paul E. McKenney
2021-04-30  5:08                       ` Paul E. McKenney
2021-04-30  5:08                         ` Paul E. McKenney
2021-04-25 22:47 ` [PATCH v9 clocksource 6/6] clocksource: Reduce WATCHDOG_THRESHOLD Paul E. McKenney
2021-04-25 22:47 ` [PATCH v10 clocksource 7/7] " Paul E. McKenney

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=871ravo2xx.ffs@nanos.tec.linutronix.de \
    --to=tglx@linutronix.de \
    --cc=Mark.Rutland@arm.com \
    --cc=ak@linux.intel.com \
    --cc=corbet@lwn.net \
    --cc=feng.tang@intel.com \
    --cc=john.stultz@linaro.org \
    --cc=kernel-team@fb.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=maz@kernel.org \
    --cc=neeraju@codeaurora.org \
    --cc=paulmck@kernel.org \
    --cc=sboyd@kernel.org \
    --cc=zhengjun.xing@intel.com \
    --cc=zhengjun.xing@linux.intel.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.