ad@NetBSD.org wrote:

> > so softintr(9) occurred (or was enabled) as soon as
> > a kernel jumped to lwp_trampoline() before lwp_startup()
> > was called.
> 
> Is this with or without your patch?

Without my patch.

With the patch (#ifdef'ed out PSL_LOWIPL), this doesn't happen.
---
Izumi Tsustui