[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-aio
Subject:    Re: open posix test results
From:       Joel Becker <jlbec () evilplan ! org>
Date:       2005-01-05 15:44:06
Message-ID: 20050105154406.GY17261 () parcelfarce ! linux ! theplanet ! co ! uk
[Download RAW message or body]

On Wed, Jan 05, 2005 at 09:50:42AM +0100, Laurent Vivier wrote:
> Le mar 04/01/2005 à 20:08, Joel Becker a écrit :
> > 	This is correct.  aio_read(2) is, effectively, defined to be
> > read(2) except with submission and completion semantics for
> > asynchronousness.  So, if read(2) would behave a certain way with a
> > given file descriptor, so should aio_read(2).  Ergo, if O_DIRECT places
> > a restriction on read(2), it places the same restriction on aio_read(2).
> > That conforms to POSIX (and SUSv2) just fine.
> 
> OK, but if you don't open with O_DIRECT all is done synchronously, and I
> don't think a synchronous aio_read conforms to POSIX...
> I agree with you for all other things.

	That's a current bug/limitation of the 2.6 AIO code.  It's
waiting on fixing.  No one is claiming that conforms to POSIX.
Actually, POSIX might just be OK with a hack like that, but we certainly
aren't.

Joel


-- 

Life's Little Instruction Book #20

	"Be forgiving of yourself and others."

			http://www.jlbec.org/
			jlbec@evilplan.org
--
To unsubscribe, send a message with 'unsubscribe linux-aio' in
the body to majordomo@kvack.org.  For more info on Linux AIO,
see: http://www.kvack.org/aio/
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
[prev in list] [next in list] [prev in thread] [next in thread]