one page too many (worse, it unmapped our stack page... Ouch!)
Also, use loadcrp() and such instead of calling pmap_activate
in various places so we have more control over TLB flush.
Now pmap_activate() is called only by locore (and that call is
done only if PMAP_DEBUG is defined). Other minor clean-up...
Define pmap_resident_count(), pmap_wired_count()
to call pmap_resident_pages(), pmap_wired_pages()
respectively, so we can keep those statistics on
lower-level structures instead of here.
to "struct trapframe" (the minimal frame) because in most cases, all we
care about is the minimal, invariant part. Further, if the declaration
says we have an arg of type "frame" (big union), then KGDB will try to
print the whole bloody thing out whether you want to see it or not.
trap.c:trap_kdebug() so it can be C code instead of assembly.
Make sure the initial frame pointer is zero so KGDB will know when to
stop trying to follow frame pointers during backtrace.
to "struct trapframe" (the minimal frame) because in most cases, all we
care about is the minimal, invariant part. Further, if the declaration
says we have an arg of type "frame" (big union), then KGDB will try to
print the whole bloody thing out whether you want to see it or not.
Also add the function trap_kdebug() which takes care of calling one of:
kgdb_trap (for KGDB), or kdb_trap, (for DDB) or the Sun PROM monitor.
This logic used to be in locore, but it is better as C code.
to "struct trapframe" (the minimal frame) because in most cases, all we
care about is the minimal, invariant part. Further, if the declaration
says we have an arg of type "frame" (big union), then KGDB will try to
print the whole bloody thing out whether you want to see it or not.
gwr