It may prove helpful to find out what process was involved at the time of the trap. Let's find out what was running. To do this, we look at the cpu structures (there is only one on this system, so this is quick and easy), get the running thread , offset into the associated process ( proc ) structure, and pull the command line. $<cpus cpus: cpus: id flags thread idle_t 0 1b fc45c600 f03c1ec0 cpus+0x10: lwp callo fpu f05d09e0 0 f05e8d18 cpus+0x1c: next prev next on prev on f017caa0 f017caa0 0 0 cpus+0x2c: lock npri queue limit actmap 0 110 fc00a000 fc00a528 fc03b460 cpus+0x3c: maxrunpri max unb pri nrunnable -1 -1 0 cpus+0x4c: runrun kprnrn chsnlevel dispthread thread lock 0 0 0 fc45c600 0 cpus+0x58: intr_stack on_intr intr_thread intr_actv f03dffa0 0 f03dcec0 0 cpus+0x68: base_spl 0 fc45c600+a0/X 0xfc45c6a0: fc44f800 fc44f800+260/s tmp_mdevmap+0x5c88: /apps/dist/share/framemaker,v4.0/bin/sunxm.s5.sparc/fm_flb From this, we see that a FrameMaker program was running at the time. This is surprising and a bit humorous to the authors, as when we crashed the system on purpose, we firmly believed it crashed due to an ls / command we had executed. But that's not what the evidence shows! Let's find out what the fm_flb program was trying to accomplish when it caused the system to crash. |