Subject: Re: fsync performance hit on 1.6.1
To: Christos Zoulas <christos@zoulas.com>
From: Greg A. Woods <woods@weird.com>
List: tech-kern
Date: 07/08/2003 21:47:58
[ On Tuesday, July 8, 2003 at 22:29:55 (GMT), Christos Zoulas wrote: ]
> Subject: Re: fsync performance hit on 1.6.1
>
> In article <m19ZzWe-000B44C@proven.weird.com>,
> Greg A. Woods <tech-kern@NetBSD.ORG> wrote:
> >
> >The only part of the whole SysV IPC API that's really rather inelegant
> >is the semaphores part (and of course that's due to the minor flaw that
> >requires two separate system calls to create and then initialize a
> >semaphore). IIUC this problem is fixed with POSIX semaphores, but of
> >course NetBSD doesn't yet have any implementation of POSIX semaphores.
>
> The above statement is incorrect.
If there is a sem_open() implementation for an existing NetBSD release
then I'd like to know where to find it.
Or do you mean that POSIX semaphores do need two system calls to create
and initialize a semaphore (and thus require the same care to avoid a
race condition when two processes or threads try to create it at the
same time). The synopsis for sem_open() suggests otherwise:
"initialize and open a named semaphore".
--
Greg A. Woods
+1 416 218-0098; <g.a.woods@ieee.org>; <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>