Re: [parisc-linux-cvs] pjlahaie, interrupted with code 28....

Re: [parisc-linux-cvs] pjlahaie, interrupted with code 28....

[Helge Deller]

Am Wed, 21 Jul 1999 schrieb Paul Lahaie:
> Update of /home/cvs/parisc/linux/drivers/gecko
> In directory puffin.external.hp.com:/tmp/cvs-serv735
> 
> Modified Files:
> 	dino.c 
> Log Message:
> Fix dino probing code so that systems without Dinos do not crash while trying
> to read the configuration registers.

Hi Paul !

Those changes seem to initialize the interrupt-routine, so that I get the whole
time  "Interrupted with code 28, regs xxxxx"...... (I don´t have a dino in my
hp-machine!)

Could you/someone maybe take a look into it ?
It does not happen, when you don´t include DINO-support into the kernel.
Please correct me, if I´m totally wrong here.....

PS: This also happened here with my machine, before I applied/committed my
latest changes...

Helge.

Re: [parisc-linux-cvs] pjlahaie, interrupted with code 28....

[Helge Deller]

Hi Paul, 
Hi Philipp !

Sorry for my last mail, but I found some more background-info for my
last mail:

I think it´s something wrong in the interrupt-routine (but I´m really not sure
about that !!), since when I don´t use DINO/PCI-support the kernel freezes too.
Info:
I added a stopkernel_with_rolling_leds()-call (endless-loop in setup.c)
directly behind the "POSIX"-text in main.c. This means, that the kernel should
loop forever in this routine. But the kernel freezes after some time,
at other boot-ups I get endless messages "Interrupted with code 9" after some time 
behind the output of "POSIX...". Could it be, that the stack
overflows after some time (maybe because of interrupts in the background, which
does not restore the stack correctly (Philipp??)) ?

Maybe also my kernel-origin-change in the Makefile to 0x10000 (we used
0x12000 before) could be the reason ? I don´t know yet...

Comments ?

Helge.