From: Ingo Molnar <mingo@elte.hu>
To: Jens Axboe <jens.axboe@oracle.com>, Nick Piggin <npiggin@suse.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
Lennart Sorensen <lsorense@csclub.uwaterloo.ca>,
Andrew Morton <akpm@linux-foundation.org>,
tytso@mit.edu, drees76@gmail.com, jesper@krogh.cc,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Peter Zijlstra <a.p.zijlstra@chello.nl>
Subject: Re: Linux 2.6.29
Date: Fri, 3 Apr 2009 10:15:14 +0200 [thread overview]
Message-ID: <20090403081514.GA21325@elte.hu> (raw)
In-Reply-To: <20090403072507.GO5178@kernel.dk>
* Jens Axboe <jens.axboe@oracle.com> wrote:
> On Thu, Apr 02 2009, Linus Torvalds wrote:
> > On Fri, 3 Apr 2009, Lennart Sorensen wrote:
> > > So so far I would rank anticipatory at about 1000x better than
> > > cfq for my work load. It sure acts a lot more like it used to
> > > back in 2.6.18 times.
[...]
> > Jens - remind us what the problem with AS was wrt CFQ?
>
> CFQ was just faster, plus it supported things like io priorities
> that AS does not.
btw., while pluggable IO schedulers have their upsides:
- They are easier to test during development and deployment.
- The uptick of a new, experimental IO scheduler is faster due to
easier availability.
- Regressions in the primary IO scheduler are easier to prove.
And the technical case for pluggable IO schedulers is much stronger
than the case for pluggable process schedulers:
- Persistent media has persistent workloads - and each workload has
different access patterns.
- The inefficiencies of mixed workloads on the same rotating media
have forced a clear separation of the 'one disk, one workload'
usage model, and has hammered this down people's minds. (Nobody
in their right mind is going to put a big Oracle and SAP
installation on the same [rotating] disk.)
- the 'NOP' scheduler makes sense on media with RAM-like
properties. 90% of CFQ's overhead is useless fluff on such media.
- [ These properties are not there for CPU schedulers: CPUs are
data processors not persistent data storage so they are
fundamentally shared by all workloads and have a lot less
persistent state - so mixing workloads on CPUs is common and
having one good scheduler is paramount. ]
At the risk of restarting the "to plug or not to plug" scheduler
flamewars ;-), the pluggable IO scheduler design has its very clear
downsides as well:
- 99% of users use CFQ, so any bugs in it will hit 99% of the Linux
community and we have not actually won much in terms of helping
real people out in the field.
- We are many years down the road of having replaced AS with the
supposedly better CFQ - and AS is still (or again?) markedly
better for some common tests.
- The 1% of testers/users who find that CFQ sucks and track it down
to CFQ can easily switch back to another IO scheduler: NOP or AS.
This dillutes the quality of _CFQ_, our crown jewel IO scheduler:
as it removes critical participiants from the pool of testers.
They might be only 1% of all Linux users, but they are the 1% who
make things happen upstream.
The result: even if CFQ sucks for some important workloads, the
combined social pressure is IMO never strong enough on upstream
to get our act together. While we might fix the bugs reported
here, the time to realize and address these bugs was way too
long. Power-users configure they way out and go the path of least
resistance and the rest suffers in silence.
- There's not even any feedback in the common case: people think
"hey, what I'm doing must be some oddball thing" and leave it at
that. Even if that oddball thing is not odd at all. Furthermore,
getting feedback _after_ someone has solved their problems by
switching to AS is a lot harder than getting feedback while they
are still hurting and cursing. Yesterday's solved problem is
boring and a lot less worthy to report than today's high-prio
ticket.
- It is _too easy_ to switch to AS, and shops with critical data
will not be as eager to report CFQ problems, and will not be as
eager to test experimental kernel patches that fix CFQ problems,
if they can switch to AS at the flip of a switch.
Ergo, i think pluggable designs for something as critical and as
central as IO scheduling has its clear downsides as it created two
mediocre schedulers:
- CFQ with all the modern features but performance problems on
certain workloads
- Anticipatory with legacy features only but works (much!) better
on some workloads.
... instead of giving us just a single well-working CFQ scheduler.
This, IMHO, in its current form, seems to trump the upsides of IO
schedulers.
So i do think that late during development (i.e. now), _years_ down
the line, we should make it gradually harder for people to use AS.
I'd not remove the AS code per se (it _is_ convenient to test it
without having to patch the kernel - especially now that we _know_
that there is a common problem, and there _are_ genuinely oddball
workloads where it might work better due to luck or design), but
still we should:
- Make it harder to configure in.
- Change the /sys switch-to-AS method to break any existing scripts
that switched CFQ to AS. Add a warning to the syslog if an old
script uses the old method and document the change prominetly but
do _not_ switch the IO scheduler to AS.
- If the user still switched to AS, emit some scary warning about
this being an obsolete IO scheduler, that it is not being tested
as widely as CFQ and hence might have bugs, and that if the user
still feels absolutely compelled to use it, to report his problem
to the appropriate mailing lists so that upstream can fix CFQ
instead.
By splintering the pool of testers and by removing testers from that
pool who are the most important in getting our default IO scheduler
tested we are not doing ourselves any favors.
Btw., my personal opinion is that even such extreme measures dont
work fully right due to social factors, so _my_ preferred choice for
doing such things is well known: to implement one good default
scheduler and to fix all bugs in it ;-)
For IO schedulers i think there's just two sane technical choices
for plugins: one good default scheduler (CFQ) or no IO scheduler at
all (NOP).
The rest is development fuzz or migration fuzz - and such fuzz needs
to be forced to zero after years of stabilization.
What do you think?
Ingo
next prev parent reply other threads:[~2009-04-03 8:16 UTC|newest]
Thread overview: 664+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-03-23 23:29 Linux 2.6.29 Linus Torvalds
2009-03-24 6:19 ` Jesper Krogh
2009-03-24 6:46 ` David Rees
2009-03-24 7:32 ` Jesper Krogh
2009-03-24 8:16 ` Ingo Molnar
2009-03-24 11:10 ` Jesper Krogh
2009-03-24 19:00 ` David Rees
2009-03-25 17:42 ` Jesper Krogh
2009-03-25 18:16 ` David Rees
2009-03-25 18:46 ` Jesper Krogh
2009-03-25 18:30 ` Theodore Tso
2009-03-25 18:40 ` Linus Torvalds
2009-03-25 22:05 ` Theodore Tso
2009-03-25 23:23 ` Linus Torvalds
2009-03-25 23:46 ` Bron Gondwana
2009-03-26 0:32 ` Ric Wheeler
2009-03-27 0:11 ` Andrew Morton
2009-03-27 0:27 ` Linus Torvalds
2009-03-27 0:47 ` Andrew Morton
2009-03-27 1:03 ` Linus Torvalds
2009-03-27 1:25 ` Andrew Morton
2009-03-27 2:21 ` David Rees
2009-03-27 3:03 ` Matthew Garrett
2009-03-27 3:36 ` Dave Jones
2009-03-27 3:01 ` Matthew Garrett
2009-03-27 3:38 ` Linus Torvalds
2009-03-27 3:59 ` Linus Torvalds
2009-03-28 23:52 ` david
2009-03-28 5:06 ` Ingo Molnar
2009-04-01 21:03 ` Lennart Sorensen
2009-04-01 21:36 ` Andrew Morton
2009-04-01 22:57 ` Lennart Sorensen
2009-04-03 14:46 ` Mark Lord
2009-04-03 15:16 ` Lennart Sorensen
2009-04-03 15:42 ` Mark Lord
2009-04-03 18:59 ` Jeff Garzik
2009-04-04 8:18 ` Andrew Morton
2009-04-04 12:40 ` Mark Lord
2009-04-05 1:57 ` David Newall
2009-04-05 3:46 ` Mark Lord
2009-04-02 1:00 ` Ingo Molnar
2009-04-03 4:06 ` Lennart Sorensen
2009-04-03 4:13 ` Linus Torvalds
2009-04-03 7:25 ` Jens Axboe
2009-04-03 8:15 ` Ingo Molnar [this message]
2009-04-06 21:46 ` Bill Davidsen
2009-04-03 14:21 ` Lennart Sorensen
2009-04-03 15:05 ` Mark Lord
2009-04-03 15:14 ` Lennart Sorensen
2009-04-03 19:57 ` Jeff Garzik
2009-04-03 21:28 ` Janne Grunau
2009-04-03 21:57 ` Jeff Garzik
2009-04-03 22:32 ` Janne Grunau
2009-04-03 22:57 ` David Rees
2009-04-03 23:29 ` Jeff Garzik
2009-04-03 23:52 ` Linus Torvalds
2009-04-03 22:53 ` David Rees
2009-04-03 23:30 ` Jeff Garzik
2009-04-04 16:29 ` Janne Grunau
2009-04-04 23:02 ` Jeff Garzik
2009-04-05 14:20 ` Janne Grunau
2009-04-06 14:06 ` Lennart Sorensen
2009-04-03 22:28 ` Jeff Moyer
2009-04-06 14:15 ` Lennart Sorensen
2009-04-06 21:27 ` Mark Lord
2009-04-06 21:56 ` Lennart Sorensen
2009-04-02 11:05 ` Janne Grunau
2009-04-02 16:09 ` Andrew Morton
2009-04-02 16:33 ` David Rees
2009-04-02 16:46 ` Linus Torvalds
2009-04-02 16:51 ` Andrew Morton
2009-04-02 22:13 ` Jeff Garzik
2009-04-02 21:56 ` Jeff Garzik
2009-04-02 16:29 ` David Rees
2009-04-02 16:42 ` Andrew Morton
2009-04-02 16:57 ` Linus Torvalds
2009-04-02 17:04 ` Linus Torvalds
2009-04-02 22:09 ` Jeff Garzik
2009-04-02 22:42 ` Linus Torvalds
2009-04-02 22:51 ` Andrew Morton
2009-04-02 23:00 ` Linus Torvalds
2009-04-03 2:01 ` Jeff Garzik
2009-04-03 2:16 ` Linus Torvalds
2009-04-03 3:05 ` Jeff Garzik
2009-04-03 3:34 ` Linus Torvalds
2009-04-03 11:32 ` Chris Mason
2009-04-03 15:07 ` Linus Torvalds
2009-04-03 15:40 ` Chris Mason
2009-04-03 20:05 ` Jeff Garzik
2009-04-03 20:14 ` Linus Torvalds
2009-04-03 21:48 ` Jeff Garzik
2009-04-03 22:06 ` Linus Torvalds
2009-04-03 23:48 ` Jeff Garzik
2009-04-04 12:46 ` Mark Lord
2009-04-04 12:52 ` Huang Yuntao
2009-04-03 23:35 ` Dave Jones
2009-04-04 12:44 ` Mark Lord
2009-04-04 21:10 ` Jeff Garzik
2009-04-03 5:05 ` Nick Piggin
2009-04-03 8:31 ` Jeff Garzik
2009-04-03 8:35 ` Jeff Garzik
2009-04-03 2:38 ` Trenton D. Adams
2009-04-03 2:54 ` Jeff Garzik
2009-04-03 15:14 ` Mark Lord
2009-04-03 15:18 ` Lennart Sorensen
2009-04-03 15:46 ` Mark Lord
2009-04-03 15:28 ` Linus Torvalds
2009-04-02 18:52 ` David Rees
2009-04-02 21:42 ` Theodore Tso
2009-04-02 21:50 ` Lennart Sorensen
2009-04-03 15:07 ` Mark Lord
2009-04-02 12:17 ` Theodore Tso
2009-04-02 21:54 ` Lennart Sorensen
2009-04-02 23:27 ` Theodore Tso
2009-04-03 0:32 ` Lennart Sorensen
2009-03-27 3:23 ` Theodore Tso
2009-03-27 3:47 ` Matthew Garrett
2009-03-27 5:13 ` Theodore Tso
2009-03-27 5:57 ` Matthew Garrett
2009-03-27 6:21 ` Matthew Garrett
2009-03-27 11:24 ` Theodore Tso
2009-03-27 12:17 ` Linux 2.6.29 - delayed metadata for delayed allocation? Andreas T.Auer
2009-03-27 14:51 ` Linux 2.6.29 Matthew Garrett
2009-03-27 15:08 ` Alan Cox
2009-03-27 15:22 ` Matthew Garrett
2009-03-27 16:15 ` Alan Cox
2009-03-27 16:28 ` Matthew Garrett
2009-03-27 16:51 ` Alan Cox
2009-03-27 17:02 ` Matthew Garrett
2009-03-27 17:19 ` Alan Cox
2009-03-27 18:05 ` Linus Torvalds
2009-03-27 18:35 ` Alan Cox
2009-03-27 19:03 ` Theodore Tso
2009-03-27 19:14 ` Alan Cox
2009-03-27 19:32 ` Theodore Tso
2009-03-27 20:11 ` Andreas T.Auer
2009-03-27 22:01 ` Linus Torvalds
2009-03-31 9:58 ` Neil Brown
2009-03-27 19:19 ` Gene Heskett
2009-03-27 19:48 ` Theodore Tso
2009-03-27 20:02 ` Aaron Cohen
[not found] ` <727e50150903271301l36cff340l33e813bf6f77b4b@mail.gmail.com>
2009-03-27 20:04 ` Theodore Tso
2009-03-27 22:37 ` Gene Heskett
2009-03-27 22:55 ` Theodore Tso
2009-03-28 0:42 ` Gene Heskett
2009-03-27 18:36 ` Hua Zhong
2009-03-27 17:57 ` Linus Torvalds
2009-03-27 18:22 ` Linus Torvalds
2009-03-27 18:32 ` Alan Cox
2009-03-27 18:40 ` Linus Torvalds
2009-03-27 19:00 ` Alan Cox
2009-03-29 9:15 ` Xavier Bestel
2009-03-29 20:16 ` Alan Cox
2009-03-29 21:07 ` Linus Torvalds
2009-03-30 19:37 ` Jeremy Fitzhardinge
2009-03-27 20:27 ` Felipe Contreras
2009-03-27 19:43 ` Jeff Garzik
2009-03-27 20:01 ` Theodore Tso
2009-03-27 22:20 ` Jeff Garzik
2009-03-27 21:46 ` Linus Torvalds
2009-03-27 22:06 ` Jeff Garzik
2009-03-27 22:19 ` Linus Torvalds
2009-03-27 22:25 ` Linus Torvalds
2009-03-28 1:19 ` Jeff Garzik
2009-03-28 1:30 ` David Miller
2009-03-28 2:19 ` Mark Lord
2009-03-28 2:49 ` Jeff Garzik
2009-03-28 13:29 ` Stefan Richter
2009-03-28 14:17 ` Jeff Garzik
2009-03-28 14:35 ` Stefan Richter
2009-03-28 15:17 ` Mark Lord
2009-03-28 16:08 ` Stefan Richter
2009-03-28 16:32 ` Linus Torvalds
2009-03-28 17:22 ` David Hagood
2009-03-29 1:18 ` Jeff Garzik
2009-03-31 18:45 ` Jörn Engel
2009-03-29 23:14 ` Dave Chinner
2009-03-30 0:39 ` Theodore Tso
2009-03-30 1:29 ` Trenton Adams
2009-03-30 3:28 ` Theodore Tso
2009-03-30 3:55 ` Trenton D. Adams
2009-03-30 13:45 ` Theodore Tso
2009-03-30 6:31 ` Dave Chinner
2009-03-30 13:55 ` Theodore Tso
2009-03-30 7:13 ` Andreas T.Auer
2009-03-30 9:05 ` Alan Cox
2009-03-30 10:49 ` Andreas T.Auer
2009-03-30 11:12 ` Alan Cox
2009-03-30 11:17 ` Ric Wheeler
2009-03-30 13:48 ` Mark Lord
2009-03-30 14:00 ` Ric Wheeler
2009-03-30 14:44 ` Mark Lord
2009-03-30 14:58 ` Ric Wheeler
2009-03-30 15:21 ` Mark Lord
2009-03-30 15:27 ` Ric Wheeler
2009-03-30 16:13 ` Linus Torvalds
2009-03-30 16:30 ` Mark Lord
2009-03-30 16:58 ` Linus Torvalds
2009-03-30 17:29 ` Mark Lord
2009-03-30 17:57 ` Chris Mason
2009-03-30 18:39 ` Mark Lord
2009-03-30 18:52 ` Chris Mason
2009-03-30 20:19 ` Mark Lord
2009-03-30 18:54 ` Pasi Kärkkäinen
2009-03-30 15:00 ` Jeff Garzik
2009-03-30 15:34 ` Linus Torvalds
2009-03-30 16:11 ` Ric Wheeler
2009-03-30 16:34 ` Linus Torvalds
2009-03-30 17:11 ` Ric Wheeler
2009-03-30 17:39 ` Mark Lord
2009-03-30 17:51 ` Linus Torvalds
2009-03-30 18:15 ` Ric Wheeler
2009-03-30 19:08 ` Eric Sandeen
2009-03-30 19:22 ` Rik van Riel
2009-03-30 19:41 ` Jeff Garzik
2009-03-30 20:21 ` Michael Tokarev
2009-03-30 20:26 ` Mark Lord
2009-03-30 20:29 ` Mark Lord
2009-03-30 20:35 ` Jeff Garzik
2009-03-30 20:40 ` Mark Lord
2009-03-30 20:34 ` Jeff Garzik
2009-03-30 20:05 ` Linus Torvalds
2009-03-31 9:27 ` Neil Brown
2009-03-31 21:13 ` Alan Cox
2009-03-31 21:10 ` Alan Cox
2009-03-31 21:55 ` Linus Torvalds
2009-03-30 19:02 ` Bill Davidsen
2009-04-01 1:19 ` david
2009-04-01 16:24 ` Bill Davidsen
2009-04-01 20:15 ` david
2009-04-01 21:33 ` Andreas T.Auer
2009-04-01 22:29 ` david
2009-04-02 2:30 ` Bron Gondwana
2009-04-02 4:55 ` david
2009-04-02 5:29 ` Bron Gondwana
2009-04-02 9:58 ` Andreas T.Auer
2009-04-02 12:30 ` Bill Davidsen
2009-04-01 22:00 ` Harald Arnesen
2009-04-01 22:09 ` Alejandro Riveira Fernández
2009-04-01 22:28 ` david
2009-03-30 3:01 ` Mark Lord
2009-03-30 6:41 ` Andreas T.Auer
2009-03-30 12:55 ` Chris Mason
2009-03-30 17:42 ` Theodore Tso
2009-03-31 23:55 ` Dave Chinner
2009-04-01 12:53 ` Chris Mason
2009-04-01 15:41 ` Andreas T.Auer
2009-04-01 16:02 ` Chris Mason
2009-04-01 18:37 ` Andreas T.Auer
2009-04-01 21:50 ` Theodore Tso
2009-04-01 23:44 ` Matthew Garrett
2009-03-28 16:25 ` Alex Goebel
2009-03-28 21:12 ` Hua Zhong
2009-03-29 8:22 ` Stefan Richter
2009-03-29 0:33 ` david
2009-03-29 1:24 ` Jeff Garzik
2009-03-29 3:43 ` Theodore Tso
2009-03-29 4:53 ` Jeff Garzik
2009-03-31 15:01 ` Thierry Vignaud
2009-03-28 0:18 ` Jeff Garzik
2009-03-28 1:45 ` Linus Torvalds
2009-03-28 2:53 ` Jeff Garzik
2009-03-28 2:56 ` Zid Null
2009-03-28 3:55 ` Gene Heskett
2009-03-28 11:29 ` Alejandro Riveira Fernández
2009-03-28 2:16 ` Mark Lord
2009-03-28 2:38 ` Linus Torvalds
2009-03-28 11:57 ` Andreas T.Auer
2009-03-27 15:20 ` Giacomo A. Catenazzi
2009-03-27 21:11 ` Jeremy Fitzhardinge
2009-03-28 7:45 ` Bojan Smojver
2009-03-28 8:43 ` Bojan Smojver
2009-03-28 21:55 ` Bojan Smojver
2009-03-31 21:51 ` Jeremy Fitzhardinge
2009-03-31 22:30 ` Bojan Smojver
2009-04-01 5:26 ` Bojan Smojver
2009-04-01 6:35 ` Jeremy Fitzhardinge
2009-04-03 12:39 ` Pavel Machek
2009-03-27 0:51 ` Linus Torvalds
2009-03-27 1:03 ` Andrew Morton
2009-03-27 9:58 ` Alan Cox
2009-03-26 2:50 ` Neil Brown
2009-03-26 3:13 ` Theodore Tso
2009-03-24 9:15 ` Alan Cox
2009-03-24 9:32 ` Ingo Molnar
2009-03-24 10:10 ` Alan Cox
2009-03-24 10:31 ` Ingo Molnar
2009-03-24 11:12 ` Andrew Morton
2009-03-24 12:23 ` Alan Cox
2009-03-24 13:37 ` Theodore Tso
2009-03-25 12:37 ` Jan Kara
2009-03-25 15:00 ` Theodore Tso
2009-03-25 17:29 ` Linus Torvalds
2009-03-25 17:57 ` Alan Cox
2009-03-25 18:09 ` David Rees
2009-03-25 18:21 ` Linus Torvalds
2009-03-25 18:26 ` Linus Torvalds
2009-03-25 18:48 ` Ric Wheeler
2009-03-25 18:49 ` Alan Cox
2009-03-25 18:55 ` Ric Wheeler
2009-03-25 18:58 ` Theodore Tso
2009-03-25 19:48 ` Christoph Hellwig
2009-03-25 21:50 ` Theodore Tso
2009-03-26 2:10 ` Matthew Garrett
2009-03-26 2:36 ` Jeff Garzik
2009-03-26 2:42 ` Matthew Garrett
[not found] ` <f73f7ab80903251944s581166bbk31c26db50750814a@mail.gmail.com>
2009-03-26 2:46 ` Kyle Moffett
2009-03-26 2:51 ` Jeff Garzik
2009-03-26 3:03 ` Kyle Moffett
2009-03-26 3:40 ` Linus Torvalds
2009-03-26 3:57 ` David Miller
2009-03-26 4:58 ` Kyle Moffett
2009-03-26 6:24 ` Jeff Garzik
2009-03-26 12:49 ` Kyle Moffett
2009-03-26 2:47 ` Matthew Garrett
2009-03-26 2:54 ` Kyle Moffett
2009-03-25 20:45 ` Linus Torvalds
2009-03-25 21:51 ` Theodore Tso
2009-03-25 23:21 ` Linus Torvalds
2009-03-25 23:50 ` Jan Kara
2009-03-26 0:04 ` Linus Torvalds
2009-03-26 9:06 ` ext3 IO latency measurements (was: Linux 2.6.29) Ingo Molnar
2009-03-26 9:09 ` Ingo Molnar
2009-03-26 11:08 ` Jens Axboe
2009-03-26 14:27 ` Arjan van de Ven
2009-03-26 14:36 ` Jens Axboe
2009-03-26 14:49 ` Arjan van de Ven
2009-03-26 11:37 ` Theodore Tso
2009-03-26 12:44 ` Ingo Molnar
2009-03-26 12:46 ` Ingo Molnar
2009-03-26 14:03 ` Ingo Molnar
2009-03-26 14:13 ` Ingo Molnar
2009-03-26 14:30 ` Andrew Morton
2009-03-26 15:32 ` relatime: update once per day patches (was: ext3 IO latency measurements) Frans Pop
2009-03-26 15:47 ` Andrew Morton
2009-03-26 16:14 ` Linus Torvalds
2009-03-26 16:24 ` Andrew Morton
2009-03-26 17:12 ` Frans Pop
2009-03-26 17:48 ` Andrew Morton
2009-03-26 18:49 ` Matthew Garrett
2009-03-26 19:20 ` Andrew Morton
2009-03-26 19:43 ` Matthew Garrett
2009-03-27 11:25 ` David Hagood
2009-03-26 16:30 ` Theodore Tso
2009-03-26 16:40 ` Jose Celestino
2009-03-26 17:14 ` Frans Pop
2009-03-26 16:53 ` Frans Pop
2009-03-26 16:53 ` Linus Torvalds
2009-03-26 17:32 ` [PATCH] Allow relatime to update atime once a day Matthew Garrett
2009-03-26 17:56 ` Alexey Dobriyan
2009-03-26 18:55 ` Alan Cox
2009-03-26 14:47 ` ext3 IO latency measurements (was: Linux 2.6.29) Theodore Tso
2009-03-26 16:20 ` Linus Torvalds
2009-03-26 17:06 ` Andreas Schwab
2009-03-26 17:07 ` Theodore Tso
2009-03-26 17:16 ` Linus Torvalds
2009-03-26 17:49 ` [PATCH 1/2] Add a strictatime mount option Matthew Garrett
2009-03-26 17:53 ` [PATCH 2/2] Make relatime default Matthew Garrett
2009-03-26 18:48 ` Alan Cox
2009-03-26 22:27 ` Linus Torvalds
2009-03-27 0:15 ` Frans Pop
2009-03-27 0:30 ` Linus Torvalds
2009-03-27 14:06 ` Alan Cox
2009-03-27 2:05 ` Frans Pop
2009-04-09 20:13 ` Pavel Machek
2009-04-09 20:47 ` Linus Torvalds
2009-04-09 21:15 ` Pavel Machek
2009-04-09 21:20 ` Linus Torvalds
2009-04-09 22:00 ` Pavel Machek
2009-03-30 14:42 ` Andrea Arcangeli
2009-03-30 14:52 ` Xavier Bestel
2009-03-30 19:26 ` Bill Davidsen
2009-03-26 18:52 ` [PATCH 1/2] Add a strictatime mount option Alan Cox
2009-03-26 18:59 ` ext3 IO latency measurements (was: Linux 2.6.29) Alan Cox
2009-03-26 20:02 ` Matthew Garrett
2009-03-26 20:42 ` Alan Cox
2009-03-26 20:55 ` Matthew Garrett
2009-03-26 20:58 ` Alan Cox
2009-03-26 23:04 ` Bron Gondwana
2009-03-27 11:22 ` Alan Cox
2009-03-27 12:19 ` Bron Gondwana
2009-03-27 13:56 ` Alan Cox
2009-03-27 12:00 ` ext3 IO latency measurements Giacomo A. Catenazzi
2009-03-26 17:29 ` ext3 IO latency measurements (was: Linux 2.6.29) Frans Pop
2009-03-26 17:32 ` Bill Nottingham
2009-03-26 17:41 ` Linus Torvalds
2009-03-26 18:23 ` Bill Nottingham
2009-03-26 22:24 ` Linus Torvalds
2009-03-27 13:47 ` Bill Nottingham
2009-03-26 18:54 ` Alan Cox
2009-03-26 15:28 ` Theodore Tso
2009-03-26 23:02 ` Ingo Molnar
2009-03-26 23:59 ` Theodore Tso
2009-03-27 0:08 ` Ingo Molnar
2009-03-27 0:40 ` Jesse Barnes
2009-03-31 11:51 ` Neil Brown
2009-03-26 12:22 ` Pekka Enberg
2009-03-26 12:23 ` Pekka Enberg
2009-03-26 14:38 ` Andrew Morton
2009-03-26 18:11 ` Jan Kara
2009-03-26 18:35 ` Andrew Morton
2009-03-27 21:26 ` Jan Kara
2009-03-26 22:39 ` Linus Torvalds
2009-03-26 22:57 ` Andrew Morton
2009-03-27 21:38 ` Jan Kara
2009-03-27 22:10 ` Linus Torvalds
2009-03-28 19:43 ` Andrew Morton
2009-04-09 21:59 ` updated: ext3 IO latency measurements on v2.6.30-rc1 Ingo Molnar
2009-04-10 7:34 ` Heinz Diehl
2009-05-18 16:37 ` Sanjoy Mahajan
2009-03-25 23:57 ` Linux 2.6.29 Linus Torvalds
2009-03-26 0:22 ` Jan Kara
2009-03-26 1:34 ` Linus Torvalds
2009-03-26 2:59 ` Theodore Tso
2009-03-26 16:24 ` Jan Kara
2009-03-24 13:20 ` Theodore Tso
2009-03-24 13:30 ` Ingo Molnar
2009-03-24 13:51 ` Theodore Tso
2009-03-24 16:34 ` Jesper Krogh
2009-03-24 17:32 ` Linus Torvalds
2009-03-24 18:20 ` Mark Lord
2009-03-24 18:41 ` Eric Sandeen
2009-03-24 13:52 ` Alan Cox
2009-03-24 14:28 ` Theodore Tso
2009-03-24 15:18 ` Alan Cox
2009-03-24 17:55 ` Jan Kara
2009-03-24 17:55 ` Linus Torvalds
2009-03-24 18:41 ` Kyle Moffett
2009-03-24 19:17 ` Linus Torvalds
2009-03-24 18:45 ` Theodore Tso
2009-03-24 19:21 ` Linus Torvalds
2009-03-24 19:40 ` Ric Wheeler
2009-03-24 19:55 ` Jeff Garzik
2009-03-25 9:34 ` Benny Halevy
2009-03-25 9:39 ` Jens Axboe
2009-03-25 19:32 ` Jeff Garzik
2009-03-25 19:43 ` Christoph Hellwig
2009-03-25 19:43 ` Jens Axboe
2009-03-25 19:49 ` Ric Wheeler
2009-03-25 19:57 ` Jens Axboe
2009-03-25 20:41 ` Hugh Dickins
2009-03-26 8:57 ` Jens Axboe
2009-03-26 14:47 ` Hugh Dickins
2009-03-26 15:46 ` Jens Axboe
2009-03-26 18:21 ` Hugh Dickins
2009-03-26 18:32 ` Jens Axboe
2009-03-26 19:00 ` Hugh Dickins
2009-03-26 19:03 ` Jens Axboe
2009-03-25 20:16 ` Jeff Garzik
2009-03-25 20:25 ` Ric Wheeler
2009-03-25 21:22 ` James Bottomley
2009-03-26 8:59 ` Jens Axboe
2009-03-30 19:05 ` range-based cache flushing (was Re: Linux 2.6.29) Jeff Garzik
2009-04-01 0:14 ` James Bottomley
2009-04-01 1:28 ` Jeff Garzik
2009-04-01 21:20 ` James Bottomley
2009-03-25 21:27 ` Linux 2.6.29 Benny Halevy
2009-03-25 20:25 ` Jeff Garzik
2009-03-25 20:40 ` Linus Torvalds
2009-03-25 20:57 ` Ric Wheeler
2009-03-25 23:02 ` Linus Torvalds
2009-03-26 0:28 ` Ric Wheeler
2009-03-26 1:36 ` Linus Torvalds
2009-03-25 21:29 ` [PATCH] issue storage device flush via sync_blockdev() (was Re: Linux 2.6.29) Jeff Garzik
2009-03-25 21:56 ` Eric Sandeen
2009-03-25 23:08 ` Jeff Garzik
2009-03-26 2:31 ` Eric Sandeen
2009-03-26 14:19 ` Ric Wheeler
2009-03-26 0:58 ` Ric Wheeler
2009-03-26 1:26 ` Jeff Garzik
2009-03-26 1:33 ` Jeff Garzik
2009-03-26 1:39 ` Ric Wheeler
2009-03-26 8:24 ` Christoph Hellwig
2009-03-27 7:59 ` Jens Axboe
2009-03-25 22:01 ` Alan Cox
2009-03-25 23:12 ` Jeff Garzik
2009-03-26 3:24 ` [PATCH v2] issue storage device flush via sync_blockdev() Jeff Garzik
2009-03-27 2:50 ` Theodore Tso
2009-03-27 3:17 ` Jeff Garzik
2009-03-27 3:30 ` Theodore Tso
2009-03-27 20:50 ` [PATCH] issue storage dev flush from generic file_fsync helper Jeff Garzik
2009-03-29 8:25 ` Christoph Hellwig
2009-03-30 1:25 ` Fernando Luis Vázquez Cao
2009-03-30 1:36 ` [PATCH 1/5] block: Add block_flush_device() Fernando Luis Vázquez Cao
2009-03-30 1:40 ` [PATCH 2/5] ext3: call blkdev_issue_flush on fsync() Fernando Luis Vázquez Cao
2009-03-30 1:51 ` Jeff Garzik
2009-03-30 2:50 ` Fernando Luis Vázquez Cao
2009-03-30 12:04 ` Fernando Luis Vázquez Cao
2009-03-30 12:09 ` [PATCH 1/7] block: Add block_flush_device() Fernando Luis Vázquez Cao
2009-03-30 15:07 ` Bartlomiej Zolnierkiewicz
2009-03-31 6:09 ` Fernando Luis Vázquez Cao
2009-03-30 17:34 ` Linus Torvalds
2009-03-30 17:50 ` Jeff Garzik
2009-03-30 17:55 ` Jens Axboe
2009-03-30 18:27 ` Linus Torvalds
2009-03-30 18:54 ` Jens Axboe
2009-03-30 19:16 ` Jeff Garzik
2009-03-30 19:24 ` Chris Mason
2009-03-30 20:09 ` Andi Kleen
2009-03-30 20:15 ` Chris Mason
2009-03-30 19:59 ` Linus Torvalds
2009-03-30 20:31 ` Jeff Garzik
2009-03-30 19:45 ` Linus Torvalds
2009-03-30 20:17 ` Jens Axboe
2009-03-30 20:36 ` Linus Torvalds
2009-03-31 2:14 ` Ric Wheeler
2009-03-31 2:47 ` Linus Torvalds
2009-03-31 6:04 ` Jens Axboe
2009-03-31 11:15 ` Ric Wheeler
2009-03-31 14:55 ` Linus Torvalds
2009-03-31 15:22 ` Chris Mason
2009-03-31 15:41 ` Ric Wheeler
2009-03-31 16:15 ` Linus Torvalds
2009-03-31 16:43 ` Jens Axboe
2009-03-31 16:57 ` Linus Torvalds
2009-03-31 17:19 ` Jens Axboe
2009-04-01 0:54 ` Tejun Heo
2009-03-31 17:03 ` Jens Axboe
2009-04-01 0:43 ` Tejun Heo
2009-03-31 17:14 ` Ric Wheeler
2009-03-31 19:25 ` Mark Lord
2009-03-31 15:54 ` Linus Torvalds
2009-03-31 16:29 ` Alan Cox
2009-03-31 6:01 ` Jens Axboe
2009-03-30 20:52 ` Mark Lord
2009-03-30 20:57 ` Jeff Garzik
2009-03-31 13:16 ` Chris Mason
2009-03-31 13:23 ` Mark Lord
2009-03-31 13:28 ` Chris Mason
2009-03-31 15:49 ` Eric Sandeen
2009-03-31 16:37 ` Mark Lord
2009-03-30 12:11 ` [PATCH 2/7] ext3: call blkdev_issue_flush() on fsync() Fernando Luis Vázquez Cao
2009-03-30 14:04 ` Theodore Tso
2009-03-30 14:15 ` Chris Mason
2009-03-30 14:33 ` Theodore Tso
2009-03-31 1:26 ` Tejun Heo
2009-03-31 1:58 ` Theodore Tso
2009-03-31 2:14 ` Tejun Heo
2009-03-31 11:18 ` Jens Axboe
2009-03-31 21:29 ` Jeff Garzik
2009-04-01 1:03 ` Tejun Heo
2009-03-30 12:15 ` [PATCH 3/7] ext4: " Fernando Luis Vázquez Cao
2009-03-30 12:18 ` [PATCH 4/7] vfs: call blkdev_issue_flush() from generic file_fsync() helper Fernando Luis Vázquez Cao
2009-03-30 12:22 ` [PATCH 5/7] vfs: Add wbcflush sysfs knob to disable storage device writeback cache flushes Fernando Luis Vázquez Cao
2009-03-30 12:36 ` Jens Axboe
2009-03-30 14:18 ` Fernando Luis Vázquez Cao
2009-03-30 14:35 ` Jens Axboe
2009-03-31 6:49 ` Fernando Luis Vázquez Cao
2009-03-31 10:38 ` Jens Axboe
2009-03-31 11:56 ` Fernando Luis Vázquez Cao
2009-03-30 15:14 ` Bartlomiej Zolnierkiewicz
2009-03-30 17:51 ` Jens Axboe
2009-03-30 17:55 ` Jeff Garzik
2009-03-30 17:59 ` Jens Axboe
2009-03-30 19:09 ` Jeff Garzik
2009-03-30 20:56 ` Bartlomiej Zolnierkiewicz
2009-03-30 22:01 ` Jeff Garzik
2009-03-30 12:33 ` [PATCH 6/7] xfs: propagate issue-flush error code Fernando Luis Vázquez Cao
2009-03-30 15:20 ` Bartlomiej Zolnierkiewicz
2009-03-31 23:37 ` Dave Chinner
2009-04-01 3:52 ` Fernando Luis Vázquez Cao
2009-03-30 12:36 ` [PATCH 7/7] reiserfs: " Fernando Luis Vázquez Cao
2009-03-30 15:25 ` Bartlomiej Zolnierkiewicz
2009-03-30 1:43 ` [PATCH 3/5] ext4: call blkdev_issue_flush on fsync Fernando Luis Vázquez Cao
2009-03-30 1:53 ` [PATCH 4/5] vfs: call blkdev_issue_flush from generic file_fsync helper Fernando Luis Vázquez Cao
2009-03-30 1:59 ` [PATCH 5/5] vfs: Add wbcflush sysfs knob to disable storage device writeback cache flushes Fernando Luis Vázquez Cao
2009-03-25 21:33 ` Linux 2.6.29 Jeff Garzik
2009-03-27 7:57 ` Jens Axboe
2009-03-27 14:13 ` Theodore Tso
2009-03-27 14:35 ` Christoph Hellwig
2009-03-27 15:03 ` Ric Wheeler
2009-03-27 20:38 ` Jeff Garzik
2009-03-28 0:14 ` Alan Cox
2009-03-29 8:25 ` Christoph Hellwig
2009-03-27 19:14 ` Chris Mason
2009-03-27 7:46 ` Jens Axboe
2009-03-31 20:49 ` Jeff Garzik
2009-03-31 22:02 ` Ric Wheeler
2009-03-31 22:22 ` Jeff Garzik
2009-04-01 18:34 ` Mark Lord
2009-03-24 20:24 ` David Rees
2009-03-25 7:30 ` David Rees
2009-03-24 23:03 ` Jesse Barnes
2009-03-25 0:05 ` Arjan van de Ven
2009-03-25 17:59 ` David Rees
2009-03-25 18:40 ` Stephen Clark
2009-03-26 23:53 ` Mark Lord
2009-03-25 2:09 ` Theodore Tso
2009-03-25 3:57 ` Jesse Barnes
2009-03-27 11:27 ` Martin Steigerwald
2009-03-24 12:27 ` Andi Kleen
2009-04-02 14:00 ` Mathieu Desnoyers
2009-03-24 13:02 ` Revert "gro: Fix legacy path napi_complete crash", (was: Re: Linux 2.6.29) Ingo Molnar
2009-03-24 13:12 ` Ingo Molnar
2009-03-24 13:12 ` Ingo Molnar
2009-03-24 13:35 ` Herbert Xu
2009-03-24 14:06 ` Ingo Molnar
2009-03-24 14:33 ` Robert Schwebel
2009-03-24 14:39 ` Ingo Molnar
2009-03-24 15:09 ` Herbert Xu
2009-03-24 15:29 ` Sascha Hauer
2009-03-24 15:36 ` Ingo Molnar
2009-03-24 15:47 ` Ingo Molnar
2009-03-24 15:59 ` Herbert Xu
2009-03-24 16:02 ` Ingo Molnar
2009-03-24 19:19 ` Ingo Molnar
2009-03-24 20:54 ` Ingo Molnar
2009-03-24 21:17 ` Revert "gro: Fix legacy path napi_complete crash", David Miller
2009-03-24 22:01 ` Ingo Molnar
2009-03-25 0:33 ` Revert "gro: Fix legacy path napi_complete crash", (was: Re: Linux 2.6.29) Herbert Xu
2009-03-25 0:32 ` Herbert Xu
2009-03-25 2:09 ` Revert "gro: Fix legacy path napi_complete crash", David Miller
2009-03-24 21:36 ` David Miller
2009-03-24 22:47 ` David Miller
2009-03-25 0:24 ` Herbert Xu
2009-03-25 0:23 ` Herbert Xu
2009-03-25 2:11 ` David Miller
2009-03-25 7:33 ` Ingo Molnar
2009-03-25 8:04 ` David Miller
2009-03-25 12:08 ` Herbert Xu
2009-03-25 12:20 ` Ingo Molnar
2009-03-25 12:26 ` Herbert Xu
2009-03-25 22:01 ` Ingo Molnar
2009-03-25 22:20 ` Ken Witherow
2009-03-26 9:07 ` Herbert Xu
2009-03-26 9:25 ` Ingo Molnar
2009-03-25 22:54 ` Jarek Poplawski
2009-03-26 0:03 ` David Miller
2009-03-26 0:10 ` David Miller
2009-03-26 6:43 ` Jarek Poplawski
2009-03-26 7:52 ` David Miller
2009-03-26 7:59 ` Jarek Poplawski
2009-03-26 2:41 ` Herbert Xu
2009-03-26 3:20 ` David Miller
2009-03-26 3:40 ` Herbert Xu
2009-03-26 9:18 ` Jarek Poplawski
2009-03-26 7:39 ` Jarek Poplawski
2009-03-26 7:59 ` David Miller
2009-03-25 9:34 ` Ingo Molnar
2009-03-24 15:22 ` Revert "gro: Fix legacy path napi_complete crash", (was: Re: Linux 2.6.29) Sascha Hauer
2009-03-27 13:35 ` Linux 2.6.29 Hans-Peter Jansen
2009-03-27 14:53 ` Geert Uytterhoeven
2009-03-27 15:46 ` Mike Galbraith
2009-03-27 16:02 ` Linus Torvalds
2009-03-28 7:50 ` Mike Galbraith
2009-03-30 22:00 ` Hans-Peter Jansen
2009-03-30 22:07 ` Arjan van de Ven
2009-03-30 10:18 ` Pavel Machek
2009-03-31 13:33 ` Rafael J. Wysocki
2009-03-31 15:30 ` Hans-Peter Jansen
2009-03-31 19:37 ` Jeff Garzik
2009-03-31 19:47 ` Arjan van de Ven
2009-04-02 19:01 ` Andreas T.Auer
2009-03-27 16:49 ` Frans Pop
[not found] <cj1Ut-1i2-7@gated-at.bofh.it>
[not found] ` <cj8ji-35a-5@gated-at.bofh.it>
[not found] ` <cj8CC-3uO-5@gated-at.bofh.it>
[not found] ` <cjaXP-7vB-17@gated-at.bofh.it>
[not found] ` <cjbhc-7VP-13@gated-at.bofh.it>
[not found] ` <cjbTS-AR-11@gated-at.bofh.it>
[not found] ` <cjcdl-10E-31@gated-at.bofh.it>
[not found] ` <cjeS0-5nC-33@gated-at.bofh.it>
2009-03-25 15:19 ` Bodo Eggert
[not found] ` <cj9oW-4PO-1@gated-at.bofh.it>
[not found] ` <cjkaL-5xg-15@gated-at.bofh.it>
[not found] ` <cjGbn-6Xx-35@gated-at.bofh.it>
[not found] ` <cjGl4-7ak-61@gated-at.bofh.it>
[not found] ` <cjJsv-3Uu-5@gated-at.bofh.it>
[not found] ` <cjKHX-5MF-19@gated-at.bofh.it>
[not found] ` <ck7XP-JA-13@gated-at.bofh.it>
[not found] ` <ck8h7-18H-1@gated-at.bofh.it>
[not found] ` <ck8AA-1y0-7@gated-at.bofh.it>
[not found] ` <ck8Kd-20o-7@gated-at.bofh.it>
[not found] ` <ckaVR-5o8-9@gated-at.bofh.it>
[not found] ` <ckbf4-5M8-7@gated-at.bofh.it>
[not found] ` <ckcE8-845-9@gated-at.bofh.it>
2009-03-27 21:53 ` Bodo Eggert
2009-03-28 6:51 ` Mike Galbraith
2009-03-28 12:12 ` Theodore Tso
[not found] ` <ckdgW-uA-1@gated-at.bofh.it>
[not found] ` <ckdK0-1om-1@gated-at.bofh.it>
[not found] ` <ckiqk-uZ-13@gated-at.bofh.it>
[not found] ` <cklHL-5IW-23@gated-at.bofh.it>
[not found] ` <ckm0Q-6o4-3@gated-at.bofh.it>
[not found] ` <ckmaG-6AC-5@gated-at.bofh.it>
[not found] ` <ckmWO-7Ro-5@gated-at.bofh.it>
[not found] ` <ckn6L-84h-19@gated-at.bofh.it>
[not found] ` <cknzG-ff-7@gated-at.bofh.it>
[not found] ` <cknJk-IB-9@gated-at.bofh.it>
[not found] ` <ckoFm-2eg-3@gated-at.bofh.it>
[not found] ` <ckoYP-2DC-13@gated-at.bofh.it>
2009-03-28 11:53 ` Bodo Eggert
2009-03-29 14:45 ` Pavel Machek
2009-03-29 15:47 ` Linus Torvalds
2009-03-29 19:15 ` Pavel Machek
2009-03-30 14:22 ` Morten P.D. Stevens
[not found] ` <ck93x-2oJ-3@gated-at.bofh.it>
2009-03-27 23:22 ` Bodo Eggert
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=20090403081514.GA21325@elte.hu \
--to=mingo@elte.hu \
--cc=a.p.zijlstra@chello.nl \
--cc=akpm@linux-foundation.org \
--cc=drees76@gmail.com \
--cc=jens.axboe@oracle.com \
--cc=jesper@krogh.cc \
--cc=linux-kernel@vger.kernel.org \
--cc=lsorense@csclub.uwaterloo.ca \
--cc=npiggin@suse.de \
--cc=torvalds@linux-foundation.org \
--cc=tytso@mit.edu \
/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.