FreeLists / openbeosnetteam / [openbeosnetteam] Re: BUG: ethernet

[Waldemar Kornewald <wkornew@xxxxxxx>]

Philippe Houdoin wrote:
> > Thread: ethernet_input
> > Stack trace:
> > xxxx close
>
>
> The thread stack crawl is just that: "close"?
> Nothing below that could provide more call context???

Yep, sounds crazy, but it's true.

> What's the KDL error title? Memory fault?

I think so (forgot and don't want any more crash-tests ;).

> If it crash on close(), maybe something corrupt or never initialized the fd
> value...

What has happened to the net_stack? It worked fine half a year ago, but 
now my PPP stack is crashing and does not unload correctly, stop_stack 
is crashing after a second "ifconfig -a" and the ethernet module is 
crashing! It's really time for a clean netstack.
Why does the stack driver unload itself (I remember this causing crashes 
a long time ago)? The netstack does not crash when it's kept in memory.

> ethernet_input() doesn't do close() anything itself but does call higher
> protocols code that may do it. Meat for thought.
> ;-)

Is only my computer infected? Maybe someone put a spell on it when I 
left the project? :)

> That's a better defensive code, so it's welcome. Plus killing a thread is
> considered violent and bad design compared to semaphore deletion detection.

Also, etherq was never deleted (leak).

> > Is there a bug when killing a real-time kernel thread?
>
> Under R5/BONE? Dunno. Does changing the thread priority fix something?

Seems not. :)

I want to have a new netstack, Philippe. ;))

Bye,
Waldemar