async io (was: Re: Good point of Linux over NT)

Bryn Paul Arnold Jones (bpaj@gytha.demon.co.uk)
Wed, 22 Jan 1997 20:07:26 +0000 (GMT)

This message is in MIME format. The first part should be readable text,
while the remaining parts are likely unreadable without MIME-aware tools.
Send mail to mime@docserver.cac.washington.edu for more info.

---1734412634-934127028-853960996=:1981
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
Content-ID: <Pine.LNX.3.95.970122200649.1981E@gytha.demon.co.uk>

On Wed, 22 Jan 1997, Roger Espel Llima wrote:

>
> Hmm, clone() is the basis for kernel-threads, which can be used to
> simulate async I/O (and sound like a more modern design), but does Linux
> have a direct equivalent of SunOS/Solaris's aioread() and aiowrite() ?
> That is, a way to queue a read() or a write() and have the call return
> immediately.
>
> According to Sun's manpages these interfaces look like they're
> Sun-specific, but some POSIX spec gets mentioned in the manpage for
> aio_read (notice the underscore), which isn't "yet implemented on
> Solaris".
>
> Roger

Hmm, looking at the man pages for aioread(3), and aiowrite(3), we could do
this exactly this way. As it stands, it's a userspace problem in the
libaries so this isn't the place (where is a better place tho ?).

Bryn

--
PGP key pass phrase forgotten,   \ Overload -- core meltdown sequence 
again :( and I don't care ;)      |            initiated.
                                 / This space is intentionally left   
                                |  blank, apart from this text ;-)
                                 \____________________________________


---1734412634-934127028-853960996=:1981--