RE: pyxis_machine_check ---- strange message!

[Loic Prylli <Loic.Prylli@ens-lyon.fr>]

dannyhp@pop.web.net writes:
 > I think the address is constant.  Here is a larger example, that includes a vi
 > memory violation:

Could you also tells us in which function is the address 0xfffffc00003159c8
(from System.map) and  disassemble some code around  (for instance
with "echo disas 0xfffffc00003159a8 0xfffffc00003159e8 | gdb vmlinux".

By the way what is your arch? And you use MILO, right?

I may be wrong (I amd not very familiar with machine checks) but I
interpret the data as follows 

 > Mar 21 19:09:43 arrhenius kernel:  +170 ffffffff80000080 10

PYXIS_ERR (see milo/palcode/lx164/impure.h to know which data
correspond to which register) tells there was a PCI Master Abort

 > Mar 21 19:09:43 arrhenius kernel:  +1a0 205000e 800804
 > Mar 21 19:09:43 arrhenius kernel:  +1b0 800804 0

PCI_ERR0 and PCI_ERR2 are telling the error was on reading PCI address
800804 (which looks strange).

More information on the location of the error in the kernel
could probably help.


Loic