[openbeosnetteam] Daily kernel progress report - 19 mar 2002

Well, after a pleasant and unexpected evening, a bit of coding and things
are starting to settle down a little.

The server is still not as stable as I'd like with segfaults every so often
that I just don't know what's casuing them! Example...

13: General protection fault occured at eip = 60429db3
Fault (13) not handled
kernel debugger: Welcome to Kernel Debugging Land...
 eax 018f7fa0  ebp fc02f958  cs 0008 | area 05cfccf8  (core_seg0)
 ebx 30039324  esp fc02f948  ss 0010 | addr 60420000  size 0000d000
 ecx 00000001  edi 000005dc  ds 0010 |
 edx ffffffff  esi 00002ee0  es 0010 | Thread: ec9xx0_rx_thread
 eip 60429db3 flag 00010286  fs 0000 | Team:  kernel_team
trap 0000000d  err 00000000  gs 0000 | Stack Trace follows:

00000000  60429db3  _end+baadb
fc02f958  60427d0b  _end+b8a33
fc02f984  6042aa17  _end+bb73f

I've also seen a lot of segfaults where malloc seems to play a part, but
they've been in different teams and so no thread has emerged.

On a brighter note, the userland side is working well and the kernel side,
when it runs, doe seem to be working almost as per prior to it's move, which
to be honest is as much as I could have hoped for in 2 weeks! In fact in a
lot of ways we're further forward as we have the start of a user app in the
tree now as well - and yes, it does work. It may not do much, but it works.

Where next? Well, there are some obvious bits to fix (ioctl for ethernet)
and then I think we should try to see if me can't make it more crash proof
and stable. The main problem is, I'm not sure how!

Thinking about it, the code isn't quite where it was as we don't have sendto
or recvfrom yet, so maybe I'll tackle those as well.

Suggestions on how to debug and improve stability will be greatly
appreciated.

Now, has anyone else tried running this beast yet?

david

Other related posts: