Re: om-2.6 patch for kcomd (pre-alpha)

"Matt Dew" <[email protected]>
Newsgroups gmane.linux.cluster.openmosix.devel
Message-ID <[email protected]>
Hi Tab,
 
On Tue, 2006-08-01 at 21:53 +0000, Vincent Hanquez  wrote:
> On Fri, Jul 28, 2006 at 12:22:23PM -0600, Matt Dew wrote:
> >    In the continuing efforts of a non-programmer to contribute to
the
> > project, here's my second installation of kcomd patches.
> > 
> >   Here's another patch for kcomd at
> > http://www.osource.org/openmosix/om-26.php.
> > 
> > Vanilla kernel patch:
> >
http://www.osource.org/openmosix/patches/om-2.6-full-20060727.patch.bz2
> > 
> > Patch against git:
> >
http://www.osource.org/openmosix/patches/om-2.6-git-20060727.patch.bz2
> 
> Matt, I tried to do something with the patch (twice), but eventually
> gave up, because I have no idea what is useful from what is not.

Right now the code is in flux as I try to get to a working state. I
found out I was having to rewrite some older code as I added things
because there were bugs or bad design decisions I made earlier. 
I feel that submitting smaller patches now is wasted effort as the code
continues to change.

I left commented code in as a signpost for what not do do as I continue
forward.  Once things are working, I plan to pull the comments out and
clean things up, but for now that commented code is helping me not
repeat past mistakes.

Currently, I'm working on getting syscalls and signals to work,
hopefully by mid-next week.  Once that's done, I'll start the clean up
and submitting smaller patches for you to review.  Until then, my
feelings are that it's not worth the effort to submit smaller patches.
Plus I don't know what the diffs between 2.6.16 and 2.6.17 are, but I
won't patch against 2.6.17 until this is working.  


 A question I do have is:

Are signals the best way to tell kcomd a packet is waiting to be sent,
and processes they have packets to be processed?  It doesn't seem to be
very fast.


> 
> Please remove all different commented versions of the same code.
> Do not comment code that is not useful anymore, just delete it.
> Try to split the patch in some small feature:
> - patch 1: move function to ..., patch 2: add X functions that does Y
>   patch 3: remove dead code in Z, patch 4: convert A to new B ...
> 
> That way I can understand what is going on and apply/correct
> everything very quickly.
> 
> I'll try to give the big patch a shot if I don't see anything coming,
> but that would be really helpful to have small patches ...
> 
> Cheers,


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.