From: "Jason A. Donenfeld" <Jason@zx2c4.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>,
LKML <linux-kernel@vger.kernel.org>,
clemens@ladisch.de, Sultan Alsawaf <sultan@kerneltoast.com>,
Waiman Long <longman@redhat.com>, X86 ML <x86@kernel.org>
Subject: Re: infinite loop in read_hpet from ktime_get_boot_fast_ns
Date: Wed, 12 Jun 2019 11:44:35 +0200 [thread overview]
Message-ID: <CAHmME9obwzZ5x=p3twDfNYux+kg0h4QAGe0ePAkZ2KqvguBK3g@mail.gmail.com> (raw)
In-Reply-To: <20190612090257.GF3436@hirez.programming.kicks-ass.net>
Hey Peter,
On Wed, Jun 12, 2019 at 11:03 AM Peter Zijlstra <peterz@infradead.org> wrote:
> How quasi? Do the comments in kernel/sched/clock.c look like something
> you could use?
>
> As already mentioned in the other tasks, anything ktime will be
> horrifically crap when it ends up using the HPET, the code in
> kernel/sched/clock.c is a best effort to keep using TSC even when it is
> deemed unusable for timekeeping.
Thanks for pointing that out. Indeed the HPET path is a bummer and I'd
like to just escape using ktime all together.
In fact, my accuracy requirements are very lax. I could probably even
deal with an inaccuracy as huge as ~200 milliseconds. But what I do
need is 64-bit, so that it doesn't wrap, allowing me to compare two
stamps taken a long time apart, and for it to take into account sleep
time, like CLOCK_BOOTTIME does, which means get_jiffies_64() doesn't
fit the bill. I was under the impression that I could only get this
with ktime_get_boot & co, because those add the sleep offset.
It looks like, though, kernel/sched/clock.c keeps track of some
offsets too -- __sched_clock_offset and __gtod_offset, and the comment
at the top mentions explicit sleep hooks. I wasn't sure which function
to use from here, though. sched_clock() seems based on jiffies, which
has the 32-bit wraparound issue, and the base implementation doesn't
seem to take into account sleeptime. The x86 implementation seems use
rdtsc and then adds cyc2ns_offset which looks to be based on
cyc2ns_suspend, which I assume is what I want. But there's still the
issue of the 32-bit wraparound on the base implementation.
I guess you know this code better than my quick perusal. Is there some
clock in here that doesn't have a wrap around issue and takes into
account sleeptime, without being super slow like ktime/hpet?
Jason
next prev parent reply other threads:[~2019-06-12 9:44 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-07 14:14 infinite loop in read_hpet from ktime_get_boot_fast_ns Jason A. Donenfeld
2019-06-11 21:09 ` Thomas Gleixner
2019-06-11 21:40 ` Waiman Long
2019-06-12 9:02 ` Peter Zijlstra
2019-06-12 9:44 ` Jason A. Donenfeld [this message]
2019-06-12 12:28 ` Peter Zijlstra
2019-06-12 12:31 ` Peter Zijlstra
2019-06-12 12:58 ` Jason A. Donenfeld
2019-06-12 15:27 ` Peter Zijlstra
2019-06-12 19:46 ` Arnd Bergmann
2019-06-18 17:34 ` Jason A. Donenfeld
2019-06-12 14:01 ` Arnd Bergmann
2019-06-13 15:18 ` Jason A. Donenfeld
2019-06-13 15:40 ` Arnd Bergmann
2019-06-13 16:17 ` Jason A. Donenfeld
2019-06-13 16:26 ` Thomas Gleixner
2019-06-13 16:34 ` Jason A. Donenfeld
2019-06-13 16:41 ` Jason A. Donenfeld
2019-06-13 19:53 ` Thomas Gleixner
2019-06-14 9:14 ` Jason A. Donenfeld
2019-06-14 9:44 ` Thomas Gleixner
2019-06-14 9:56 ` Jason A. Donenfeld
2019-06-14 9:48 ` [PATCH] timekeeping: add get_jiffies_boot_64() for jiffies including sleep Jason A. Donenfeld
2019-06-14 9:55 ` [tip:timers/urgent] timekeeping: Repair ktime_get_coarse*() granularity tip-bot for Thomas Gleixner
2019-06-14 11:18 ` Arnd Bergmann
2019-06-12 9:29 ` infinite loop in read_hpet from ktime_get_boot_fast_ns Peter Zijlstra
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='CAHmME9obwzZ5x=p3twDfNYux+kg0h4QAGe0ePAkZ2KqvguBK3g@mail.gmail.com' \
--to=jason@zx2c4.com \
--cc=clemens@ladisch.de \
--cc=linux-kernel@vger.kernel.org \
--cc=longman@redhat.com \
--cc=peterz@infradead.org \
--cc=sultan@kerneltoast.com \
--cc=tglx@linutronix.de \
--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).