[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 4 Apr 2014 12:09:21 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Andy Lutomirski <luto@...capital.net>
Cc: "Theodore Ts'o" <tytso@....edu>, Joerg Roedel <joro@...tes.org>,
Jiri Kosina <jkosina@...e.cz>,
Andrew Morton <akpm@...ux-foundation.org>,
Mateusz Guzik <mguzik@...hat.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Steven Rostedt <rostedt@...dmis.org>,
LKML <linux-kernel@...r.kernel.org>,
Thomas Gleixner <tglx@...utronix.de>,
"H. Peter Anvin" <hpa@...or.com>, Borislav Petkov <bp@...en8.de>,
Ingo Molnar <mingo@...nel.org>, Mel Gorman <mgorman@...e.de>,
Kay Sievers <kay@...y.org>
Subject: Re: [RFC PATCH] cmdline: Hide "debug" from /proc/cmdline
On Fri, Apr 4, 2014 at 11:57 AM, Andy Lutomirski <luto@...capital.net> wrote:
>
> What would break if we used CLOCK_BOOTTIME or CLOCK_MONOTONIC?
I wouldn't object to trying - I thought you meant changing the format
itself. If we keep the semantics (seconds since boot) but just change
the clock source in other ways, I wouldn't worry.
Of course, one issue is locking and performance. Most clock sources
are nasty as hell to read. "sched_clock()" is special.
But I'm absolutely ok with trying things out, so if you have an idea
for a patch...
> It would be really neat if we even went back and edited the queued up
> messages once we have a clocksource.
Sounds great, I'd love to see it. Patches welcome ;)
> There are also log message metadata fields now; we could stick more
> than one timestamp in there, and existing tools either won't see them
> at all (if they use klogctl) or they already know how to ignore
> unknown fields (/dev/kmsg).
That sounds like overkill, but trying to aim for something that is
actually accurate to within a second is fine.
That said, I *also* suspect that we could easily just have a simple
heartbeat message that gives accurate real time. Adding a kernel
heartbeat to dmesg is something I've often wanted to have myself.
Having something that prints out real wall-clock time every 30 minutes
or so in dmesg does not sound like a bad idea, and could be used to
correct for time skew with the existing one even if we decided that
CLOCK_BOOTTIME has too many locking/performance issues.
Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists