panic during cpu_attach() on core i7

[Tobias Nygren <tnn%NetBSD.org@localhost>]

Looks like we explode in rdmsr() on recent intel processors.
The value of the %cx register is 0xcd which corresponds to MSR_FSB_FREQ.
I'll try a kernel with speedstep disabled and see what happens ...

NetBSD 5.99.15 (GENERIC.g33m) #0: Sun Aug 30 17:45:07 CEST 2009
        
tnn%g33m.int.nygren.pp.se@localhost:/x/obj/sys/arch/amd64/compile/GENERIC.g33m
total memory = 6135 MB
avail memory = 5931 MB
mainbus0 (root)
cpu0 at mainbus0: Intel 686-class, 2807MHz, id 0x106e5
fatal protection fault in supervisor mode
trap type 4 code 0 rip ffffffff8022dd56 cs 8 rflags 10246 cr2  0 cpl 8
rsp ffffffff80edcc28 kernel: protection fault trap, code=0
Stopped in pid 0.1 (system) at  netbsd:rdmsr+0x6:       rdmsr
db{0}> bt
rdmsr() at netbsd:rdmsr+0x6
est_init_once() at netbsd:est_init_once+0x3dc
_run_once() at netbsd:_run_once+0x62
cpu_identify() at netbsd:cpu_identify+0x163
cpu_attach() at netbsd:cpu_attach+0x357
config_attach_loc() at netbsd:config_attach_loc+0x15d
mainbus_attach() at netbsd:mainbus_attach+0x15a
config_attach_loc() at netbsd:config_attach_loc+0x15d
cpu_configure() at netbsd:cpu_configure+0x26
main() at netbsd:main+0x1b3
db{0}>