Kernel support of socket async I/O

Kernel support of socket async I/O

[marco]

Hello Benjamin and all of the guys on the lists,
I'm pretty much interested in socket async I/O for a project at work. I
read the document at 

http://lse.sourceforge.net/io/aionotes.txt

and I saw that patches are available at

http://kanga.kvack.org/~blah/aio/

I also searched linux-kernel archives for some status information, but
couldn't gain much info (other than a couple of discussion threads back in
late 1999).
What we need is a standard aio/thread-pool-in-sigwaitinfo architecture and
we wouldn't like to use select/poll.

I would also like to contribute on this project if it could somehow help 
(if my boss lets me, of course ...)

I promise to narrow the distribution as long as I know who to talk to!
Thanks,
Marco.



---------------------------------------------------

      ~
     . .
     /V\     Computers are like air conditioners.
    // \\   They stop working when you open Windows.
   /(   )\
    ^`~'^

Re: Kernel support of socket async I/O

[Benjamin LaHaise]

On Mon, Feb 25, 2002 at 05:05:41PM +0100, marco wrote:
> Hello Benjamin and all of the guys on the lists,
> I'm pretty much interested in socket async I/O for a project at work. I
> read the document at 
...
> I also searched linux-kernel archives for some status information, but
> couldn't gain much info (other than a couple of discussion threads back in
> late 1999).
> What we need is a standard aio/thread-pool-in-sigwaitinfo architecture and
> we wouldn't like to use select/poll.
...

Recent development activity was mostly geared at raw block device and 
filesystem aio (as that's what the testers are using).  I did post patches 
for network aio, but dropped them due to time constraints.  Now that the 
most recent flurry of fixes to the aio core is complete (it's looking 
very good now), I need to get the core bits updated for 2.5 and see how 
much can be merged.  As for help, the network bits need polish and testing 
to get to the near-release quality of the rest of the code.

		-ben

Re: Kernel support of socket async I/O

[marco]

Where can I find the latest network aio stuff? Just to start having a look
at it ....

marco


On Mon, 25 Feb 2002, Benjamin LaHaise wrote:

> On Mon, Feb 25, 2002 at 05:05:41PM +0100, marco wrote:
> > Hello Benjamin and all of the guys on the lists,
> > I'm pretty much interested in socket async I/O for a project at work. I
> > read the document at 
> ...
> > I also searched linux-kernel archives for some status information, but
> > couldn't gain much info (other than a couple of discussion threads back in
> > late 1999).
> > What we need is a standard aio/thread-pool-in-sigwaitinfo architecture and
> > we wouldn't like to use select/poll.
> ...
> 
> Recent development activity was mostly geared at raw block device and 
> filesystem aio (as that's what the testers are using).  I did post patches 
> for network aio, but dropped them due to time constraints.  Now that the 
> most recent flurry of fixes to the aio core is complete (it's looking 
> very good now), I need to get the core bits updated for 2.5 and see how 
> much can be merged.  As for help, the network bits need polish and testing 
> to get to the near-release quality of the rest of the code.
> 
> 		-ben
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

---------------------------------------------------

      ~
     . .
     /V\     Computers are like air conditioners.
    // \\   They stop working when you open Windows.
   /(   )\
    ^`~'^