Situation:

RiscPC 233 34MB running a -current 1.3G kernel with PMAP_STATUS_L1S set
to 256.

Small C program called loop which is:

void main(void)
{
  for(;;);
}

Small script which is:

#!/bin/sh
while :
do
  ./loop &
done

Run top on one ttyv and then run the script above on the other.

Load average goes barmy (unsurprisingly) and processes rack up to about
210 or so. Then the machines locks completely. You can't ping it,
etc. There's 44M swap free at this point.

I appreciate this is a futile gesture, but I'd rather it didn't die.

Stephen Borrill