> After doing this at the debugger, it hung up again
> instead of returning to the prompt. At this point
> I'm thinking perhaps I should file a bug report?
A very good idea.
I'm not sure how important the agpmaster
driver is, maybe the system can boot
without it? You can try to boot with option
"-B agpmaster-disabled=true", maybe that
will allow the system to boot?
But I suspect that access to pci configuration
space for pci devices is somehow broken
on this system, and I guess that a lot of
device drivers will be affected by this...
As a further test to you can repeat the same
kmdb boot with a breakpoint in agpmaster_cap_find:
::bp agpmaster`agpmaster_cap_find
::cont
::step out
Again, I suspect the ::step out will hang. This
would confirm that the boot process is hanging
in agpmaster_cap_find.
Or use "::step over" a few times - a shortcut for
::step over is to use the '[' key; repeat a few
times until you stepped over the call to
pci_config_get16 and look at the value returned,
using the command " <rax=X ". The returned
value in the RAX register should be different from
0xffff, but if you always get 0xffff this would explain
the endless loop in agpmaster_cap_find.
> I checked out your links too. It looks like that guy
> got Solaris working, though at that time he would
> have been using an older version, which could explain
> why the 2008.11 ISO booted.
Yep, that was two years ago. There probably have
been lots of changes to opensolaris since then that
could explain that a current kernel is hanging.
Or he had a different model of the ibm e326
hardware or a different hardware setup?
Or a different version of the system's BIOS?
--
This message posted from opensolaris.org
_______________________________________________
opensolaris-help mailing list
[email protected]
