In message <HAnsJx.FI1@tac.nyc.ny.us>, Christos Zoulas writes:
>I've seen this with 1.4 on linux [native-threads]. I think it is because
>we are not queuing real time signals. Try killing your process with SIGRT0
>or SIGRT1 [the numbers of the signals that the solaris emulation expects
>for these] and see if you can get it going again. I believe RT0 is used
>to suspend, and RT1 is used to resume, but I may have gotten them backwards.

Solaris emulation?  I assume you mean Linux emulation.

Anyway, I have experimented with this, and it doesn't seem to help.  The
problem appears to mostly show up when more than one program is running in a
VM.  *sigh*.

-s