comp-mail-mush@srhqla.uucp (05/19/89)
From: island!maui!argv (Dan Heller)
I am forwarding this to mush-users and comp.mail.mush because there seem
to be enough mushtool users who can acknowledge this or give other info
that they might think is causing the problem. *I* believe it's a bad cpu
board -- not a memory board.
----------
On May 17, 9:16pm, Leonard Jacobs wrote:
The latest version of mush, 6.5, resolved the problem I had been
experiencing running mushtool on a Sun 3/160 under SunOS3.5, i.e.,
the long delays after updating a folder from mushtool.
With 6.5 this problem seems to be gone and the updates are almost
instantaneous.
However, a new twist began not long after making mush 6.5. Although it
performed flawlessly for several weeks, I began to get parity errors
and crashes more & more frequently as I was running the new mush.
I thought it may have been a bad memory board so I was able to do a
trade in with Sun for a new 4Mb board.
As soon as it was installed, same problem. Check out
/usr/adm/messages:
--------------------------------------------------------------------
May 17 16:08
Window mutex lock broken after process 1364148 went away
Window mush:
trap address 0x8, pid 145, pc = f01f6e2, sr = 2500, stkfmt b, context 1
Bus Error Reg 80<INVALID>
data fault address 0 faultc 0 faultb 0 dfault 1 rw 1 size 1 fcode 5
KERNEL MODE
page map 0
D0-D7 126 9 c 0 0 0 25 78
A0-A7 0 ef6c ffff87e4 0 f067cc0 f07403b f067ca0 f067c84
Begin traceback...fp = f067ca0, sp = f067c84
Called from f01f534, fp=f067cb4, args=f074032 f067cc0 0 f067ce4
Called from f061cdc, fp=f067ce4, args=f074032 0 0 0
Called from f05d9fa, fp=f067d00, args=f081c88 6 0 f081b34
Called from f064c52, fp=f067d2c, args=f081b34 6 0 f064af4
Called from f019ada, fp=f067d50, args=0 0 0 3
Called from f0199a8, fp=f067d74, args=150000 0 2 efffd84
Called from f004468, fp=efff25c, args=ef6c 150000 0 e0000001
End traceback...
panic: Bus error
syncing disks... 23 23 22 19 15 10 6 2 2 2 2 2 2 2 2 2 2 2 2 2 done
dumping to dev 701, offset 17104
dump succeeded
Rebooting Unix...
Sun UNIX 4.2 Release 3.5 (EWJ_SUN) #12: Wed Nov 23 22:43:00 EST 1988
-----------------------------------------------------------------------
Now this is completely replicable and troubling. Not only is it
amazing that a user program can crash the system, but that having run
mush 6.[0-4] without hardly a hitch, this new version is unusable.
I have not modified the source files except config.h. I am compiling the
makefile.sun as delivered. The problem was not one that ocurred in
the beginning of using 6.5, but began increasing in frequency the
more mail I received.
Oh well, a posting to sun-spots and a letter to you (I thought posting
something to comp.mail.mush may be too limited since practically
everyone is using curses mode.) might resolve the problem.
Thanks for contributing this program and if this problem is of
interest, I would appreciate hearing from you.
Thanks.
--
Leonard Jacobs
ewj!lj@ursa-major.spdcc.com OR
{ima,harvard,rayssd,linus}!spdcc!ewj!lj
--dan