It is possible for spurious interrupt(s) to be posted to an AP cpu
authorMatthew Dillon <dillon@dragonflybsd.org>
Mon, 30 Apr 2007 17:41:15 +0000 (17:41 +0000)
committerMatthew Dillon <dillon@dragonflybsd.org>
Mon, 30 Apr 2007 17:41:15 +0000 (17:41 +0000)
commitd19f6edf1b5b0ff3e740de59f480375e3ca20442
tree58ac721a8e96bf153b01310fd0e480683e1b43d4
parent0b692e79417510633878c41a7fee73c8a6eede77
It is possible for spurious interrupt(s) to be posted to an AP cpu
while the cpu is being bootstrapped.  Explicitly enable interrupts
while still in a critical section near the end of ap_init to allow
the cpu to issue the interrupt vector(s), then clear the software
interrupt pending fields in the globaldata structure.

This stops the single report to the console of a spurious interrupt
occuring during AP booting.  This problem is unrelated to the multitude
of spurious interrupts that can be reported during device initialization.
Those will probably still occur.
sys/platform/pc32/i386/mp_machdep.c
sys/platform/pc32/i386/swtch.s