mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Jeremy Fitzhardinge <jeremy@goop.org>
To: Jan Beulich <JBeulich@novell.com>
Cc: Ingo Molnar <mingo@elte.hu>,
	Jun Nakajima <jun.nakajima@intel.com>,
	Xiaohui Xin <xiaohui.xin@intel.com>, Xin Li <xin.li@intel.com>,
	Xen-devel <xen-devel@lists.xensource.com>,
	Nick Piggin <npiggin@suse.de>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	"H. Peter Anvin" <hpa@zytor.com>
Subject: Re: [Xen-devel] Performance overhead of paravirt_ops on			 nativeidentified
Date: Wed, 20 May 2009 15:42:08 -0700	[thread overview]
Message-ID: <4A148740.1070405@goop.org> (raw)
In-Reply-To: <4A11280402000078000014E2@vpn.id2.novell.com>

Jan Beulich wrote:
>>>> Jeremy Fitzhardinge <jeremy@goop.org> 15.05.09 20:50 >>>
>>>>         
>> Jan Beulich wrote:
>>     
>>> A patch for the pv-ops kernel would require some time. What I can give you
>>> right away - just for reference - are the sources we currently use in our kernel:
>>> attached.
>>>       
>> Hm, I see.  Putting a call out to a pv-ops function in the ticket lock 
>> slow path looks pretty straightforward.  The need for an extra lock on 
>> the contended unlock side is a bit unfortunate; have you measured to see 
>> what hit that has?  Seems to me like you could avoid the problem by 
>> using per-cpu storage rather than stack storage (though you'd need to 
>> copy the per-cpu data to stack when handling a nested spinlock).
>>     
>
> Not sure how you'd imagine this to work: The unlock code has to look at all
> cpus' data in either case, so an inner lock would still be required imo.
>   

Well, rather than an explicit lock I was thinking of an optimistic 
scheme like the pv clock update algorithm.

"struct spinning" would have a version counter it could update using the 
even=stable/odd=unstable algorithm.  The lock side would simply save the 
current per-cpu "struct spinning" state onto its own stack (assuming you 
can even have nested spinning), and then update the per-cpu copy with 
the current lock.  The kick side can check the version counter to make 
sure it gets a consistent snapshot of the target cpu's current lock state.

I think that only the locking side requires locked instructions, and the 
kick side can be all unlocked.

>> What's the thinking behind the xen_spin_adjust() stuff?
>>     
>
> That's the placeholder for implementing interrupt re-enabling in the irq-save
> lock path. The requirement is that if a nested lock attempt hits the same
> lock on the same cpu that it failed to get acquired on earlier (but got a ticket
> already), tickets for the given (lock, cpu) pair need to be circularly shifted
> around so that the innermost requestor gets the earliest ticket. This is what
> that function's job will become if I ever get to implement this.
>   

Sounds fiddly.

>>> static __always_inline void __ticket_spin_lock(raw_spinlock_t *lock) { 
>>> unsigned int token, count; bool free; __ticket_spin_lock_preamble; if 
>>> (unlikely(!free)) token = xen_spin_adjust(lock, token); do { count = 1 
>>> << 10; __ticket_spin_lock_body; } while (unlikely(!count) && 
>>> !xen_spin_wait(lock, token)); } 
>>>       
>> How does this work?  Doesn't it always go into the slowpath loop even if 
>> the preamble got the lock with no contention?
>>     
>
> It indeed always enters the slowpath loop, but only for a single pass through
> part of its body (the first compare in the body macro will make it exit the loop
> right away: 'token' is not only the ticket here, but the full lock->slock
> contents). But yes, I think you're right, one could avoid entering the body
> altogether by moving the containing loop into the if(!free) body. The logic
> went through a number of re-writes, so I must have overlooked that
> opportunity on the last round of adjustments.
>   

I was thinking of something like this: (completely untested)

void __ticket_spin_lock(raw_spinlock_t *lock)
{
	unsigned short inc = 0x100;
	unsigned short token;

	do {
		unsigned count = 1 << 10;
		asm volatile (
			"	lock xaddw %w0, %1\n"
			"1:	cmpb %h0, %b0\n"
			"	je 2f\n"
			"	rep ; nop\n"
			"	movb %1, %b0\n"
			/* don't need lfence here, because loads are in-order */
			"	sub $1,%2\n"
			"	jnz 1b\n"
			"2:"
			: "+Q" (inc), "+m" (lock->slock), "+r" (count)
			:
			: "memory", "cc");

		if (likely(count != 0))
			break;

		token = inc;
		inc = 0;
	} while (unlikely(!xen_spin_wait(lock, token)));
}


where xen_spin_wait() would actually be a pvops call, and perhaps the 
asm could be pulled out into an inline to deal with the 8/16 bit ticket 
difference.

    J

  reply	other threads:[~2009-05-20 22:42 UTC|newest]

Thread overview: 79+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-05-14  0:16 Performance overhead of paravirt_ops on native identified Jeremy Fitzhardinge
2009-05-14  1:10 ` H. Peter Anvin
2009-05-14  8:25   ` Peter Zijlstra
2009-05-14 14:05     ` H. Peter Anvin
2009-05-14 17:36   ` Jeremy Fitzhardinge
2009-05-14 17:50     ` H. Peter Anvin
2009-05-14  8:05 ` [Xen-devel] Performance overhead of paravirt_ops on nativeidentified Jan Beulich
2009-05-14  8:33   ` Peter Zijlstra
2009-05-14 17:45   ` Jeremy Fitzhardinge
2009-05-15  8:10     ` Jan Beulich
2009-05-15 18:50       ` Jeremy Fitzhardinge
2009-05-18  7:19         ` Jan Beulich
2009-05-20 22:42           ` Jeremy Fitzhardinge [this message]
2009-05-15 18:18 ` [tip:x86/urgent] x86: Fix performance regression caused by paravirt_ops on native kernels tip-bot for Jeremy Fitzhardinge
2009-05-21 22:42 ` Performance overhead of paravirt_ops on native identified Chuck Ebbert
2009-05-21 22:48   ` Jeremy Fitzhardinge
2009-05-21 23:10     ` H. Peter Anvin
2009-05-22  1:26     ` Xin, Xiaohui
2009-05-22  3:39       ` H. Peter Anvin
2009-05-22  4:27       ` Jeremy Fitzhardinge
2009-05-22  5:59         ` Xin, Xiaohui
2009-05-22 16:33           ` H. Peter Anvin
2009-05-22 22:44             ` Jeremy Fitzhardinge
2009-05-22 22:47               ` H. Peter Anvin
2009-05-25  9:15 ` [benchmark] 1% performance overhead of paravirt_ops on native kernels Ingo Molnar
2009-05-26 18:42   ` Jeremy Fitzhardinge
2009-05-28  6:17     ` Nick Piggin
2009-05-28 20:57       ` Jeremy Fitzhardinge
2009-05-30 10:23       ` Ingo Molnar
2009-06-02 14:18         ` Chris Mason
2009-06-02 14:49           ` Ulrich Drepper
2009-06-02 15:03             ` Chris Mason
2009-06-02 15:22               ` Ulrich Drepper
2009-06-02 16:20                 ` Chris Mason
2009-06-02 18:13                   ` Pekka Enberg
2009-06-02 18:06               ` Pekka Enberg
2009-06-02 18:27                 ` Chris Mason
2009-06-03  6:33             ` Jeremy Fitzhardinge
2009-06-02 19:14           ` Thomas Gleixner
2009-06-02 19:51             ` Chris Mason
2009-06-03 12:38         ` Rusty Russell
2009-06-03 16:09           ` Linus Torvalds
     [not found]             ` <200906041554.37102.rusty@rustcorp.com.au>
2009-06-04 15:02               ` Linus Torvalds
2009-06-04 21:52                 ` Dave McCracken
2009-06-05  7:31                   ` Gerd Hoffmann
2009-06-05 14:31                     ` Rusty Russell
2009-06-06 18:54                   ` Anders K. Pedersen
2009-06-05  4:46                 ` Rusty Russell
2009-06-05 14:54                   ` Linus Torvalds
2009-06-07  0:53                     ` Rusty Russell
2009-06-08 14:53                       ` Linus Torvalds
2009-06-09  9:39                 ` Nick Piggin
2009-06-09 11:17                   ` Ingo Molnar
2009-06-09 12:10                     ` Nick Piggin
2009-06-09 12:25                       ` Ingo Molnar
2009-06-09 12:42                         ` Nick Piggin
2009-06-09 12:56                         ` Avi Kivity
2009-06-09 15:18                         ` Linus Torvalds
2009-06-09 23:33                         ` Paul Mackerras
2009-06-10  1:26                           ` Ingo Molnar
2009-06-09 15:07                       ` Linus Torvalds
2009-06-09 15:09                     ` H. Peter Anvin
2009-06-09 18:06                       ` Linus Torvalds
2009-06-09 18:07                         ` Linus Torvalds
2009-06-09 22:48                           ` Matthew Garrett
2009-06-09 22:54                             ` H. Peter Anvin
2009-06-09 14:54                   ` Linus Torvalds
2009-06-09 14:57                     ` Ingo Molnar
2009-06-09 15:55                       ` Avi Kivity
2009-06-09 15:38                     ` Nick Piggin
2009-06-09 16:00                       ` Linus Torvalds
2009-06-09 16:21                         ` Nick Piggin
2009-06-09 16:26                           ` Linus Torvalds
2009-06-09 16:45                             ` Nick Piggin
2009-06-09 17:08                               ` Linus Torvalds
2009-06-10  5:53                                 ` Nick Piggin
2009-06-17  9:40                                   ` Pavel Machek
2009-06-17  9:56                                     ` Nick Piggin
2009-06-10  6:29                             ` 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=4A148740.1070405@goop.org \
    --to=jeremy@goop.org \
    --cc=JBeulich@novell.com \
    --cc=hpa@zytor.com \
    --cc=jun.nakajima@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=npiggin@suse.de \
    --cc=xen-devel@lists.xensource.com \
    --cc=xiaohui.xin@intel.com \
    --cc=xin.li@intel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome