All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jamie Lokier <jamie@shareable.org>
To: Ingo Molnar <mingo@elte.hu>
Cc: Ulrich Drepper <drepper@redhat.com>,
	Arjan van de Ven <arjan@infradead.org>,
	akpm@linux-foundation.org, hugh@veritas.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org, briangrant@google.com,
	cgd@google.com, mbligh@google.com,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	"H. Peter Anvin" <hpa@zytor.com>
Subject: Re: pthread_create() slow for many threads; also time to revisit 64b context switch optimization?
Date: Fri, 15 Aug 2008 16:54:57 +0100	[thread overview]
Message-ID: <20080815155457.GA5210@shareable.org> (raw)
In-Reply-To: <20080813160218.GB18037@elte.hu>

Ingo Molnar wrote:
> As unimplemented flags just get ignored by the kernel, if this flag goes 
> into v2.6.27 as-is and is ignored by the kernel (i.e. we just use a 
> plain old 64-bit [47-bit] allocation), then you could do the glibc 
> change straight away, correct? So then if people complain we can fix it 
> in the kernel purely.
> 
> how about this then?

> +#define MAP_64BIT_STACK 0x20000         /* give out 32bit addresses on old CPUs */

I think the flag makes sense but it's name is confusing - 64BIT for a
flag which means "maybe request 32-bit stack"!  Suggest:

+#define MAP_STACK       0x20000         /* 31bit or 64bit address for stack, */
+                                        /* whichever is faster on this CPU */

Also, is this _only_ useful for thread stacks, or are there other
memory allocations where 31-bitness affects execution speed on old P4s?

-- Jamie

WARNING: multiple messages have this Message-ID (diff)
From: Jamie Lokier <jamie@shareable.org>
To: Ingo Molnar <mingo@elte.hu>
Cc: Ulrich Drepper <drepper@redhat.com>,
	Arjan van de Ven <arjan@infradead.org>,
	akpm@linux-foundation.org, hugh@veritas.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org, briangrant@google.com,
	cgd@google.com, mbligh@google.com,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	"H. Peter Anvin" <hpa@zytor.com>
Subject: Re: pthread_create() slow for many threads; also time to revisit 64b context switch optimization?
Date: Fri, 15 Aug 2008 16:54:57 +0100	[thread overview]
Message-ID: <20080815155457.GA5210@shareable.org> (raw)
In-Reply-To: <20080813160218.GB18037@elte.hu>

Ingo Molnar wrote:
> As unimplemented flags just get ignored by the kernel, if this flag goes 
> into v2.6.27 as-is and is ignored by the kernel (i.e. we just use a 
> plain old 64-bit [47-bit] allocation), then you could do the glibc 
> change straight away, correct? So then if people complain we can fix it 
> in the kernel purely.
> 
> how about this then?

> +#define MAP_64BIT_STACK 0x20000         /* give out 32bit addresses on old CPUs */

I think the flag makes sense but it's name is confusing - 64BIT for a
flag which means "maybe request 32-bit stack"!  Suggest:

+#define MAP_STACK       0x20000         /* 31bit or 64bit address for stack, */
+                                        /* whichever is faster on this CPU */

Also, is this _only_ useful for thread stacks, or are there other
memory allocations where 31-bitness affects execution speed on old P4s?

-- Jamie

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2008-08-15 15:55 UTC|newest]

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <af8810200808121736q76640cc1kb814385072fe9b29@mail.gmail.com>
2008-08-13  0:45 ` pthread_create() slow for many threads; also time to revisit 64b context switch optimization? Pardo
2008-08-13  0:45   ` Pardo
2008-08-13 10:44   ` Ingo Molnar
2008-08-13 10:44     ` Ingo Molnar
2008-08-13 13:35     ` Arjan van de Ven
2008-08-13 13:35       ` Arjan van de Ven
2008-08-13 14:21       ` Ulrich Drepper
2008-08-13 14:21         ` Ulrich Drepper
2008-08-13 14:25         ` Ingo Molnar
2008-08-13 14:25           ` Ingo Molnar
2008-08-13 14:36           ` Ulrich Drepper
2008-08-13 14:36             ` Ulrich Drepper
2008-08-13 15:10             ` Ingo Molnar
2008-08-13 15:10               ` Ingo Molnar
2008-08-13 15:21               ` Ulrich Drepper
2008-08-13 15:21                 ` Ulrich Drepper
2008-08-13 15:40                 ` Ingo Molnar
2008-08-13 15:40                   ` Ingo Molnar
2008-08-13 15:55                   ` Ulrich Drepper
2008-08-13 15:55                     ` Ulrich Drepper
2008-08-13 16:02                     ` Ingo Molnar
2008-08-13 16:02                       ` Ingo Molnar
2008-08-15 15:54                       ` Jamie Lokier [this message]
2008-08-15 15:54                         ` Jamie Lokier
2008-08-15 16:03                         ` Ingo Molnar
2008-08-15 16:03                           ` Ingo Molnar
2008-08-15 17:13                         ` Ulrich Drepper
2008-08-15 17:13                           ` Ulrich Drepper
2008-08-15 17:19                           ` Ingo Molnar
2008-08-15 17:19                             ` Ingo Molnar
2008-08-15 17:23                             ` Ulrich Drepper
2008-08-15 17:23                               ` Ulrich Drepper
2008-08-15 19:00                               ` Ingo Molnar
2008-08-15 19:00                                 ` Ingo Molnar
2008-08-13 17:09                     ` Linus Torvalds
2008-08-13 17:09                       ` Linus Torvalds
2008-08-13 18:04                       ` Ulrich Drepper
2008-08-13 18:04                         ` Ulrich Drepper
2008-08-13 18:16                         ` Arjan van de Ven
2008-08-13 18:16                           ` Arjan van de Ven
2008-08-13 18:22                           ` Ulrich Drepper
2008-08-13 18:22                             ` Ulrich Drepper
2008-08-13 16:05                 ` H. Peter Anvin
2008-08-13 16:05                   ` H. Peter Anvin
2008-08-13 20:42               ` Andi Kleen
2008-08-13 20:42                 ` Andi Kleen
2008-08-13 20:56                 ` Andrew Morton
2008-08-13 20:56                   ` Andrew Morton
2008-08-13 21:46                   ` Andi Kleen
2008-08-13 21:46                     ` Andi Kleen
2008-08-15 12:43                 ` Ingo Molnar
2008-08-15 12:43                   ` Ingo Molnar
2008-08-15 13:33                   ` Andi Kleen
2008-08-15 13:33                     ` Andi Kleen
2008-08-21 16:35 Leon Bottou

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=20080815155457.GA5210@shareable.org \
    --to=jamie@shareable.org \
    --cc=akpm@linux-foundation.org \
    --cc=arjan@infradead.org \
    --cc=briangrant@google.com \
    --cc=cgd@google.com \
    --cc=drepper@redhat.com \
    --cc=hpa@zytor.com \
    --cc=hugh@veritas.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mbligh@google.com \
    --cc=mingo@elte.hu \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.org \
    /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.