Message-ID: <3AB68A9D.346346FC@uow.edu.au>
Date: Mon, 19 Mar 2001 22:39:25 +0000
From: Andrew Morton <andrewm@uow.edu.au>
MIME-Version: 1.0
To: Rajagopal Ananthanarayanan <ananth@sgi.com>
CC: kaio@oss.sgi.com
Subject: Re: spotted in the kaio patch
References: <3AB54B71.80DB2CC9@uow.edu.au> <3AB64E89.7374E767@sgi.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-kaio@oss.sgi.com
Precedence: bulk

Rajagopal Ananthanarayanan wrote:
> 
> ...
> 
> >
> > In kaio_set_slave_name(), do this:
> >
> >         current->mm->arg_start = current->mm->arg_end = 0;
> >
> > and the slaves will look right in `ps'.
> 
> Ah, so that's the trick, thanks. I had always wanted to fix this,
> but never got around to it. Interestingly top reports the correct
> names even without this fix.

mm..  I assume `ps' sees an empty argv and then falls back to
tsk->comm and puts "[]" around it.

It seems that `top' uses ->comm[] all the time, unless you give
it the 'c' option.

> Just curious: did you find these bugs by running the patch or
> by code examination?

>From reading the code.  It's *very* readable code.  Thank you!
It makes me want to read the raw I/O patch :)

Question, please:

- I note that Ben LaHaise's embryonic aio patch
  saves a copy_from_user by pinning the user pages down
  (map_user_kiobuf).  I assume they're marked COW while I/O
  is in progress.  Have you considered doing this?  Do
  you expect there would be benefit (despite Linus'
  predictions)?

Thanks.