Big memory per process troubles

[Amos Shapir]

The configuration: VAX 785 running 4.3-beta with 2 Eagles (uda), 2 rm03,
1 rp07 deuna ethernet, various dz's and dmf's, and 16 Mb memory.
Deviation: each process has 30 Mb virtual mem, by #defining the max swap
segment to 2Mb.

The phenomenon: crash, usually reserved-operand panic when executing
'ldpctx' in Resume, when the system's load is high.

Suspicion: 'resume' is attempted before a big swap-in is completed
(note that the ldpctx gets its args from the kernel stack, which is at the very
end of the process' virtual memory, and if that's not back in from the swap,
the crash would happen just as it appears to be).

Question: has anyone there seen/heard of anything like it? is this a known
bug on 4.3? has anyone tried very big virtual mem on 4.2/3 ?
	Thanks
-- 
	Amos Shapir

National Semiconductor (Israel)
6 Maskit st. P.O.B. 3007, Herzlia 46104, Israel
(01-972) 52-522261  amos%nsta at nsc
34.48'E 32.10'N