2009/1/14 Chris Mason chris.ma...@oracle.com:
On Tue, 2009-01-13 at 18:21 +0100, Peter Zijlstra wrote:
On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
So do a v10, and ask people to test.
---
Subject: mutex: implement adaptive spinning
From: Peter Zijlstra a.p.zijls...@chello.nl
* Chris Mason chris.ma...@oracle.com wrote:
v10 is better that not spinning, but its in the 5-10% range. So, I've
been trying to find ways to close the gap, just to understand exactly
where it is different.
If I take out:
/*
* If there are pending waiters, join them.
On Wed, 14 Jan 2009, Ingo Molnar wrote:
* Chris Mason chris.ma...@oracle.com wrote:
v10 is better that not spinning, but its in the 5-10% range. So, I've
been trying to find ways to close the gap, just to understand exactly
where it is different.
If I take out:
/*
On Wed, 2009-01-14 at 07:43 -0800, Linus Torvalds wrote:
On Wed, 14 Jan 2009, Ingo Molnar wrote:
* Chris Mason chris.ma...@oracle.com wrote:
v10 is better that not spinning, but its in the 5-10% range. So, I've
been trying to find ways to close the gap, just to understand
On Mon, 2009-01-12 at 19:32 +0200, Avi Kivity wrote:
Peter Zijlstra wrote:
Spinlocks can use 'pure' MCS locks.
How about this, then. In mutex_lock(), keep wait_lock locked and only
release it when scheduling out. Waiter spinning naturally follows. If
spinlocks are cache friendly
Full series, including changelogs available at:
http://programming.kicks-ass.net/kernel-patches/mutex-adaptive-spin/
and should shortly appear in a git tree near Ingo :-)
mutex: small cleanup
mutex: preemption fixes
mutex: implement adaptive spinning
mutex: set owner in mutex_lock() only
* Linus Torvalds torva...@linux-foundation.org wrote:
If I take out:
/*
* If there are pending waiters, join them.
*/
if (!list_empty(lock-wait_list))
break;
v10 pops dbench 50 up to 1800MB/s. The other tests soundly beat my
spinning and
On Wed, 2009-01-14 at 18:18 +0100, Nick Piggin wrote:
@@ -173,21 +237,21 @@ __mutex_lock_common(struct mutex *lock,
spin_unlock_mutex(lock-wait_lock, flags);
debug_mutex_free_waiter(waiter);
+ preempt_enable();
Nick Piggin wrote:
(no they're not, Nick's ticket locks still spin on a shared cacheline
IIRC -- the MCS locks mentioned could fix this)
It reminds me. I wrote a basic variation of MCS spinlocks a while back. And
converted dcache lock to use it, which showed large dbench improvements on
a
2009/1/14 Chris Mason chris.ma...@oracle.com:
On Wed, 2009-01-14 at 12:18 +0100, Dmitry Adamushko wrote:
2009/1/14 Chris Mason chris.ma...@oracle.com:
On Tue, 2009-01-13 at 18:21 +0100, Peter Zijlstra wrote:
On Tue, 2009-01-13 at 08:49 -0800, Linus Torvalds wrote:
So do a v10, and ask
* Peter Zijlstra pet...@infradead.org wrote:
Full series, including changelogs available at:
http://programming.kicks-ass.net/kernel-patches/mutex-adaptive-spin/
and should shortly appear in a git tree near Ingo :-)
Linus,
Please pull the adaptive-mutexes-for-linus git tree from:
On Wed, 2009-01-14 at 19:33 +0100, Ingo Molnar wrote:
* Peter Zijlstra pet...@infradead.org wrote:
Full series, including changelogs available at:
http://programming.kicks-ass.net/kernel-patches/mutex-adaptive-spin/
and should shortly appear in a git tree near Ingo :-)
Linus,
* Ingo Molnar mi...@elte.hu wrote:
Linus,
Please pull the adaptive-mutexes-for-linus git tree from:
git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git
adaptive-mutexes-for-linus
We dropped two fresh patches from v11 for the time being: the two debug
patches,
* Peter Zijlstra a.p.zijls...@chello.nl wrote:
On Wed, 2009-01-14 at 10:53 -0800, Andrew Morton wrote:
On Wed, 14 Jan 2009 19:33:19 +0100 Ingo Molnar mi...@elte.hu wrote:
Please pull the adaptive-mutexes-for-linus git tree
fear
- It seems a major shortcoming that the feature
On Wed, 2009-01-14 at 11:36 -0800, Andrew Morton wrote:
Do people enable CONFIG_SCHED_DEBUG?
Well, I have it always enabled, but I've honestly no idea if that makes
me weird.
CONFIG_DEBUG_MUTEXES=n, CONFIG_SCHED_DEBUG=y is getting to be a pretty
small subset?
Could be, do you fancy me doing
* Andrew Morton a...@linux-foundation.org wrote:
On Wed, 14 Jan 2009 20:00:08 +0100
Ingo Molnar mi...@elte.hu wrote:
* Andrew Morton a...@linux-foundation.org wrote:
On Wed, 14 Jan 2009 19:33:19 +0100 Ingo Molnar mi...@elte.hu wrote:
Please pull the
* Peter Zijlstra a.p.zijls...@chello.nl wrote:
On Wed, 2009-01-14 at 11:36 -0800, Andrew Morton wrote:
Do people enable CONFIG_SCHED_DEBUG?
Well, I have it always enabled, but I've honestly no idea if that makes
me weird.
CONFIG_DEBUG_MUTEXES=n, CONFIG_SCHED_DEBUG=y is getting to be
On Wed, 14 Jan 2009 21:14:35 +0100
Ingo Molnar mi...@elte.hu wrote:
* Andrew Morton a...@linux-foundation.org wrote:
On Wed, 14 Jan 2009 20:00:08 +0100
Ingo Molnar mi...@elte.hu wrote:
* Andrew Morton a...@linux-foundation.org wrote:
On Wed, 14 Jan 2009 19:33:19 +0100
On Wed, 14 Jan 2009 21:27:36 +0100
Ingo Molnar mi...@elte.hu wrote:
* Peter Zijlstra a.p.zijls...@chello.nl wrote:
On Wed, 2009-01-14 at 11:36 -0800, Andrew Morton wrote:
Do people enable CONFIG_SCHED_DEBUG?
Well, I have it always enabled, but I've honestly no idea if that makes
* Andrew Morton a...@linux-foundation.org wrote:
Do people enable CONFIG_SCHED_DEBUG?
If they suspect performance problems and want to analyze them?
The vast majority of users do not and usually cannot compile their own
kernels.
... which they derive from distro kernels or some old
On Wed, 14 Jan 2009 21:51:22 +0100
Ingo Molnar mi...@elte.hu wrote:
* Andrew Morton a...@linux-foundation.org wrote:
Do people enable CONFIG_SCHED_DEBUG?
If they suspect performance problems and want to analyze them?
The vast majority of users do not and usually cannot
On Wed, 14 Jan 2009 22:14:58 +0100
Ingo Molnar mi...@elte.hu wrote:
* Andrew Morton a...@linux-foundation.org wrote:
On Wed, 14 Jan 2009 21:51:22 +0100
Ingo Molnar mi...@elte.hu wrote:
* Andrew Morton a...@linux-foundation.org wrote:
Do people enable
* Ingo Molnar mi...@elte.hu wrote:
You just disproved your own case :(
how so? 80% is not enough? I also checked Fedora and it has
SCHED_DEBUG=y in its kernel rpms.
Ubuntu has CONFIG_SCHED_DEBUG=y as well in their kernels.
note that there's also a performance issue here: we generally
On Wed, Jan 14, 2009 at 22:41, Ingo Molnar mi...@elte.hu wrote:
* Ingo Molnar mi...@elte.hu wrote:
You just disproved your own case :(
how so? 80% is not enough? I also checked Fedora and it has
SCHED_DEBUG=y in its kernel rpms.
Ubuntu has CONFIG_SCHED_DEBUG=y as well in their kernels.
* Kay Sievers kay.siev...@vrfy.org wrote:
On Wed, Jan 14, 2009 at 22:41, Ingo Molnar mi...@elte.hu wrote:
* Ingo Molnar mi...@elte.hu wrote:
You just disproved your own case :(
how so? 80% is not enough? I also checked Fedora and it has
SCHED_DEBUG=y in its kernel rpms.
* Andrew Morton a...@linux-foundation.org wrote:
I also checked Fedora and it has SCHED_DEBUG=y
in its kernel rpms.
If all distros set SCHED_DEBUG=y then fine.
95% of the distros and significant majority of the lkml traffic.
And no, we dont generally dont provide knobs for essential
On Wed, Jan 14, 2009 at 06:22:36PM +0100, Peter Zijlstra wrote:
On Wed, 2009-01-14 at 18:18 +0100, Nick Piggin wrote:
@@ -173,21 +237,21 @@ __mutex_lock_common(struct mutex *lock,
spin_unlock_mutex(lock-wait_lock, flags);
On Wed, Jan 14, 2009 at 07:23:12PM +0200, Avi Kivity wrote:
Nick Piggin wrote:
(no they're not, Nick's ticket locks still spin on a shared cacheline
IIRC -- the MCS locks mentioned could fix this)
It reminds me. I wrote a basic variation of MCS spinlocks a while back. And
converted
On Wed, Jan 14, 2009 at 01:35:29PM -0800, Andrew Morton wrote:
You're taking a whizzy new feature which drastically changes a critical
core kernel feature and jamming it into mainline with a vestigial
amount of testing coverage without giving sufficient care and thought
to the practical
On Tue, 2009-01-13 at 15:43 +0100, Eric Sesterhenn wrote:
* Chris Mason (chris.ma...@oracle.com) wrote:
On Tue, 2009-01-13 at 15:21 +0100, Eric Sesterhenn wrote:
Hi,
when mounting an intentionally corrupted btrfs filesystem i get the
following warning and bug message. The image can
On Tue, 2009-01-13 at 18:18 +0530, Goldwyn Rodrigues wrote:
Add man/mkfs.btrfs.8.in
Kept the name with the name in, so that further processing such as
BUILD_DATE BUILD_VERSION etc. could be included later.
All man pages included in the man directory to avoid file cluttering.
Thanks for
There functions are only called by 'static int __init init_btrfs_fs(void)',
so also mark them as '__init'.
Signed-off-by: Qinghuang Feng qhfeng.ker...@gmail.com
---
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index eee060f..7e03ec8 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@
On Thu, 2009-01-15 at 01:46 +0100, Nick Piggin wrote:
Hmm, well this is rather a slow path, I would say. I'd prefer not to
modify schedule in this way (if we just get scheduled back on after
being switched away, the subsequent call to schedule is going to be
cache hot and not do too much
On Thu, Jan 15, 2009 at 08:44:03AM +0100, Peter Zijlstra wrote:
On Thu, 2009-01-15 at 01:46 +0100, Nick Piggin wrote:
Hmm, well this is rather a slow path, I would say. I'd prefer not to
modify schedule in this way (if we just get scheduled back on after
being switched away, the
34 matches
Mail list logo