[openbeosnetteam] Re: Netstack status...

I'm still seeing an occasional trip to kernel debugger land when we get the
INT13 error.

Apart from that it seems to be nice and stable...

david
----- Original Message -----
From: "Cain O'Sullivan" <cos@xxxxxxxxxxxx>
To: <openbeosnetteam@xxxxxxxxxxxxx>
Sent: Wednesday, March 20, 2002 3:43 PM
Subject: [openbeosnetteam] Re: Netstack status...


> Good stuff guys!!!
>
> The Network Stack (as of yesterday) runs properly on my BeOS without
> crashing...
>
> I promise I'll get stuck into some coding very soon... Off on holliday on
> the weekend for a few days, and will read up about networking stack issues
I
> am not so certain upon!!!
>
>
> Regards,
>
> Cain
>
>
>
>
>
> ----- Original Message -----
> From: David Reid <dreid@xxxxxxxxxxxx>
> To: OpenBeOS Network Team <openbeosnetteam@xxxxxxxxxxxxx>
> Sent: Wednesday, 20 March 2002 10:20
> Subject: [openbeosnetteam] Netstack status...
>
>
> > Well, finally the CVS tree is updated correctly and is building well on
my
> > machine. I have a few more minor tweaks to come (mainly removing debug
> code)
> > and then it'll be time to look at the next round of additions.
> >
> > However, the code is running and doesn't seem to crash (famous last
words
> no
> > doubt)!
> >
> > My thanks to Marcus for his efforts at some very unfriendly hour of the
> > morning this morning helping me get the area_malloc stuff in and so
> allowing
> > us to track the source of the problem!  I really must try to get to bed
> > before 4am one night this week!
> >
> > So, my next question is how many people have actually tried building
this?
> > Come on now, don't be shy.
> >
> > Why don't we move to a kernel only codebase? Well, I'd like to, but to
be
> > honest the userland version is much better for debugging and as long as
we
> > keep it in line with the other one allows quicker bug fixes than using
the
> > kernel only version. At some point soon (once we have more functionality
> in
> > the tree) I think we should swap to being kernel be default and userland
> by
> > specifying the build. That will be just a little way away yet.
> >
> > Also I have a question - who wants to do some coding?! No, don't worry,
> it's
> > not something that only folks who can spout RFC821 need get involved in.
I
> > think (and feel free to object here) that we should have a function
> > net_log_error(). This should by default log errors into syslog and
perhaps
> > by build time switches dump them to serial debugging or just the command
> > line if we're userland.
> >
> > The prototype should go into include/net_misc.h and the code into
> > net_server/misc.c. I won't get involved too much by even suggesting what
> it
> > should/shouldn't do. Perhaps people feel I get too involved? Don't know.
> > Anyway, it's just a straight be coding project!
> >
> > So, any volunteers for this?
> >
> > david
> >
> >
> >
> >
>
>
>

Other related posts: