From: Chris Metcalf <cmetcalf@tilera.com>
To: Roland McGrath <roland@hack.frob.com>
Cc: Maxim Kuvyrkov <maxim@codesourcery.com>,
Andrew Haley <aph@redhat.com>, David Miller <davem@davemloft.net>,
<joseph@codesourcery.com>, <rdsandiford@googlemail.com>,
<libc-ports@sourceware.org>, <libc-alpha@sourceware.org>
Subject: Re: [PATCH] Unify pthread_spin_[try]lock implementations.
Date: Wed, 25 Jul 2012 19:04:00 -0000 [thread overview]
Message-ID: <50104312.70205@tilera.com> (raw)
In-Reply-To: <20120725181300.DD1812C0B5@topped-with-meat.com>
On 7/25/2012 2:13 PM, Roland McGrath wrote:
> Here I think the reasonable thing to do is:
>
> /* A machine-specific version can define SPIN_LOCK_READS_BETWEEN_CMPXCHG
> to the number of plain reads that it's optimal to spin on between uses
> of atomic_compare_and_exchange_val_acq. If spinning forever is optimal
> then use -1. If no plain reads here would ever be optimal, use 0. */
> #ifndef SPIN_LOCK_READS_BETWEEN_CMPXCHG
> # warning machine-dependent file should define SPIN_LOCK_READS_BETWEEN_CMPXCHG
> # define SPIN_LOCK_READS_BETWEEN_CMPXCHG 1000
> #endif
>
> Then ARM et al can do:
>
> /* Machine-dependent rationale about the selection of this value. */
> #define SPIN_LOCK_READS_BETWEEN_CMPXCHG 1000
> #include <nptl/pthread_spin_lock.c>
>
> while Teil will use -1.
The tile architecture is unlikely to use this generic version no matter
what; see http://sourceware.org/ml/libc-ports/2012-07/msg00030.html for the
details, but the primary point is that in a mesh-based architecture it's a
bad idea to ever end up in a situation where all the cores can be spinning
issues loads or cmpxchg as fast as they can, so some kind of backoff is
necessary.
--
Chris Metcalf, Tilera Corp.
http://www.tilera.com
next prev parent reply other threads:[~2012-07-25 19:04 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-06-14 4:36 [PATCH 2/3, MIPS] Rewrite MIPS' pthread_spin_[try]lock using __atomic_* builtins Maxim Kuvyrkov
2012-06-28 23:10 ` Joseph S. Myers
2012-07-11 5:58 ` [PATCH] Unify pthread_spin_[try]lock implementations Maxim Kuvyrkov
2012-07-11 8:15 ` Roland McGrath
2012-07-11 8:25 ` David Miller
2012-07-11 8:44 ` Andrew Haley
2012-07-11 8:54 ` Maxim Kuvyrkov
2012-07-11 9:02 ` Andrew Haley
2012-07-11 9:05 ` Maxim Kuvyrkov
2012-07-11 11:22 ` Roland McGrath
2012-07-11 14:52 ` Chris Metcalf
2012-07-11 22:16 ` Maxim Kuvyrkov
2012-07-25 18:13 ` Roland McGrath
2012-07-25 19:04 ` Chris Metcalf [this message]
2012-07-25 20:22 ` Roland McGrath
2012-07-25 20:29 ` Chris Metcalf
2012-07-25 20:43 ` Roland McGrath
2012-08-15 3:17 ` Maxim Kuvyrkov
2012-08-15 16:27 ` Roland McGrath
2012-08-15 16:39 ` Maxim Kuvyrkov
2012-08-15 16:44 ` Roland McGrath
2012-08-15 16:53 ` Maxim Kuvyrkov
2012-08-15 16:56 ` Roland McGrath
2012-08-15 17:05 ` Maxim Kuvyrkov
2012-08-15 17:05 ` Jeff Law
2012-08-15 17:11 ` Roland McGrath
2012-08-15 17:24 ` Jeff Law
2012-08-15 16:40 ` Joseph S. Myers
2012-08-15 16:43 ` Carlos O'Donell
2012-08-15 22:00 ` Andreas Schwab
2012-08-16 10:21 ` Torvald Riegel
2012-08-16 12:40 ` Chris Metcalf
2012-08-22 23:16 ` Maxim Kuvyrkov
2012-07-11 8:26 ` Maxim Kuvyrkov
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=50104312.70205@tilera.com \
--to=cmetcalf@tilera.com \
--cc=aph@redhat.com \
--cc=davem@davemloft.net \
--cc=joseph@codesourcery.com \
--cc=libc-alpha@sourceware.org \
--cc=libc-ports@sourceware.org \
--cc=maxim@codesourcery.com \
--cc=rdsandiford@googlemail.com \
--cc=roland@hack.frob.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).