Port-vax archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Which NetBSD on a VS3176?



Holm Tiffe skrev 2014-05-20 13:25:
Anders Magnusson wrote:

Holm Tiffe skrev 2014-05-20 12:01:
/tmp//EdDk.a00665a: 29 lines, 1031 characters.
vs3176# newfs /dev/rsd1a
/dev/rsd1a: 200.0MB (409600 sectors) block size 8192, fragment size 1024
         using 5 cylinder groups of 40.00MB, 5120 blks, 9920 inodes.
[1]   Segmentation fault (core dumped) newfs /dev/rsd1a
vs3176# newfs /dev/rsd1f
/dev/rsd1f: 200.0MB (409600 sectors) block size 8192, fragment size 1024
         using 5 cylinder groups of 40.00MB, 5120 blks, 9920 inodes.
[1]   Segmentation fault (core dumped) newfs /dev/rsd1f
vs3176# newfs /dev/rsd1e
/dev/rsd1e: 3121.7MB (6393144 sectors) block size 16384, fragment size 2048
         using 16 cylinder groups of 195.11MB, 12487 blks, 24704 inodes.
panic: rmpage: pm 0x8088d760 br 0x80fc7560
cpu0: Begin traceback...
panic: rmpage: pm 0x8088d760 br 0x80fc7560
Stack traceback :
          Process is executing in user space.
cpu0: End traceback...

Hm, I would think that the segfault is due to something else than a
missing instruction.
CVAX has all usual instructions (no need to emulate anything).

The rmpage panic is more likely to be the root cause; it's a fail in the
page removing routine.
It may be because of something in the real CVAX do not behave as simh,
especially in the
cache or TLB.  It's a request for removal of a mapping where the mapping
itself cannot be found.

-- Ragge
Hmm, this is a little bit too complicated for my horizont.
I do know almost nothing about VM Systems..
It can be quite difficult to find. Hm, the only special thing in newfs I can think of should be mmap.
Can I do anything to further debug this?
Should I do make the machine accessable from the net?

I have a CVAX (mv3900) myself, but have no idea of when it can be possible for me to debug it.

About the problem itself; the code path there hasn't changed for a long time, and the bug do not reveal itself on other targets. Therefore I'm thinking of if it can be a late change in gcc that optimize something that causes this, which it did not do in older gcc.
But this is only speculations.

--Ragge






Home | Main Index | Thread Index | Old Index