From: Andi Kleen <andi@firstfloor.org>
To: "Patrick J. LoPresti" <lopresti@gmail.com>
Cc: linux-fsdevel@vger.kernel.org, linux-nfs@vger.kernel.org,
linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: Proposal: Use hi-res clock for file timestamps
Date: Tue, 17 Aug 2010 16:54:03 +0200 [thread overview]
Message-ID: <87aaolwar8.fsf@basil.nowhere.org> (raw)
In-Reply-To: <AANLkTimnyXKahtjaFeSsgcq=xMy-pP3na1jidQhZ-dt2@mail.gmail.com> (Patrick J. LoPresti's message of "Fri, 13 Aug 2010 11:25:49 -0700")
"Patrick J. LoPresti" <lopresti@gmail.com> writes:
>
> 1) Anybody who cares about file system performance is already using
> "noatime" or "relatime", which mitigates the hit greatly.
Consider mtime.
> If the above patch is too slow for some architectures, how about
> making it a configuration option? Call it "CONFIG_1980S_FILE_TICK",
> have it default to YES on the architectures that care and NO on
> anything remotely modern and sane.
>
> OK that's my proposal. Bash away.
I suspect it will be a performance disaster on x86 for VFS intensive
applications on capable file systems. VFS is very performance
critical. These checks lurk on unexpected places too, e.g. on /dev
access.
Even TSC is much slower than just reading the variable.
Also you should check if the file system granuality
even supports it, it's completely wasted on a ext3 for example.
Maybe as a optional sysctl, default to off.
-Andi
--
ak@linux.intel.com -- Speaking for myself only.
WARNING: multiple messages have this Message-ID (diff)
From: Andi Kleen <andi-Vw/NltI1exuRpAAqCnN02g@public.gmane.org>
To: "Patrick J. LoPresti" <lopresti-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: linux-fsdevel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-kernel
<linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
Subject: Re: Proposal: Use hi-res clock for file timestamps
Date: Tue, 17 Aug 2010 16:54:03 +0200 [thread overview]
Message-ID: <87aaolwar8.fsf@basil.nowhere.org> (raw)
In-Reply-To: <AANLkTimnyXKahtjaFeSsgcq=xMy-pP3na1jidQhZ-dt2-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> (Patrick J. LoPresti's message of "Fri, 13 Aug 2010 11:25:49 -0700")
"Patrick J. LoPresti" <lopresti-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> writes:
>
> 1) Anybody who cares about file system performance is already using
> "noatime" or "relatime", which mitigates the hit greatly.
Consider mtime.
> If the above patch is too slow for some architectures, how about
> making it a configuration option? Call it "CONFIG_1980S_FILE_TICK",
> have it default to YES on the architectures that care and NO on
> anything remotely modern and sane.
>
> OK that's my proposal. Bash away.
I suspect it will be a performance disaster on x86 for VFS intensive
applications on capable file systems. VFS is very performance
critical. These checks lurk on unexpected places too, e.g. on /dev
access.
Even TSC is much slower than just reading the variable.
Also you should check if the file system granuality
even supports it, it's completely wasted on a ext3 for example.
Maybe as a optional sysctl, default to off.
-Andi
--
ak-VuQAYsv1563Yd54FQh9/CA@public.gmane.org -- Speaking for myself only.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2010-08-17 14:54 UTC|newest]
Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-08-13 18:25 Proposal: Use hi-res clock for file timestamps Patrick J. LoPresti
2010-08-13 18:45 ` john stultz
2010-08-13 18:45 ` john stultz
2010-08-13 18:57 ` Patrick J. LoPresti
2010-08-13 19:09 ` john stultz
2010-08-13 19:09 ` john stultz
2010-08-13 20:53 ` Patrick J. LoPresti
2010-08-13 20:53 ` Patrick J. LoPresti
2010-08-14 16:45 ` Patrick J. LoPresti
2010-08-15 1:50 ` Bret Towe
2010-08-13 19:57 ` Jim Rees
2010-08-13 20:26 ` john stultz
2010-08-13 20:52 ` Jim Rees
2010-08-17 14:54 ` Andi Kleen [this message]
2010-08-17 14:54 ` Andi Kleen
2010-08-17 17:41 ` J. Bruce Fields
2010-08-17 17:41 ` J. Bruce Fields
2010-08-17 18:29 ` Andi Kleen
2010-08-17 18:29 ` Andi Kleen
2010-08-17 18:50 ` Patrick J. LoPresti
2010-08-17 18:50 ` Patrick J. LoPresti
2010-08-17 19:04 ` J. Bruce Fields
2010-08-17 19:18 ` Patrick J. LoPresti
2010-08-17 19:18 ` Patrick J. LoPresti
2010-08-17 19:39 ` Alan Cox
2010-08-17 19:39 ` Alan Cox
2010-08-17 19:29 ` J. Bruce Fields
2010-08-17 19:29 ` J. Bruce Fields
2010-08-17 19:52 ` Alan Cox
2010-08-17 19:52 ` Alan Cox
2010-08-18 5:53 ` Neil Brown
2010-08-18 5:53 ` Neil Brown
2010-08-18 14:46 ` Patrick J. LoPresti
2010-08-18 14:46 ` Patrick J. LoPresti
2010-08-18 17:32 ` J. Bruce Fields
2010-08-18 17:32 ` J. Bruce Fields
2010-08-18 18:15 ` Chuck Lever
2010-08-18 18:15 ` Chuck Lever
2010-08-18 23:41 ` Neil Brown
2010-08-18 23:41 ` Neil Brown
2010-08-19 0:52 ` Neil Brown
2010-08-19 0:52 ` Neil Brown
2010-08-19 2:08 ` J. Bruce Fields
2010-08-19 2:08 ` J. Bruce Fields
2010-08-19 2:44 ` Neil Brown
2010-08-19 2:44 ` Neil Brown
2010-08-19 22:46 ` J. Bruce Fields
2010-08-19 22:46 ` J. Bruce Fields
2010-08-18 23:47 ` Neil Brown
2010-08-18 23:47 ` Neil Brown
2010-08-18 17:50 ` Andi Kleen
2010-08-18 17:50 ` Andi Kleen
2010-08-18 18:54 ` J. Bruce Fields
2010-08-18 19:25 ` Andi Kleen
2010-08-18 19:30 ` J. Bruce Fields
2010-08-17 19:34 ` Patrick J. LoPresti
2010-08-17 19:34 ` Patrick J. LoPresti
2010-08-17 19:54 ` Alan Cox
2010-08-17 19:43 ` Patrick J. LoPresti
2010-08-17 19:43 ` Patrick J. LoPresti
2010-08-17 19:45 ` J. Bruce Fields
2010-08-17 19:45 ` J. Bruce Fields
2010-08-18 18:12 ` J. Bruce Fields
2010-08-18 18:12 ` J. Bruce Fields
2010-08-19 1:41 ` john stultz
2010-08-19 1:41 ` john stultz
2010-08-19 2:31 ` J. Bruce Fields
2010-08-19 2:31 ` J. Bruce Fields
2010-08-19 3:17 ` john stultz
2010-08-19 3:17 ` john stultz
2010-08-19 22:53 ` J. Bruce Fields
2010-08-18 18:20 ` David Woodhouse
2010-08-18 18:20 ` David Woodhouse
2010-08-18 18:32 ` Patrick J. LoPresti
2010-08-18 18:32 ` Patrick J. LoPresti
2010-08-18 18:53 ` Andi Kleen
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=87aaolwar8.fsf@basil.nowhere.org \
--to=andi@firstfloor.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nfs@vger.kernel.org \
--cc=lopresti@gmail.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.