Re: new SMP laptop (Core 2 Duo / 2.20Ghz ) ... kernel panics can't find MP table

[Andrew Atrens <atrens@xxxxxxxxxx>]

Heh,

I created a fake mp table (one specifying that 'defaults' be used for the apic_io, etc addresses as
 I noted from dmesg that Linux seems to use these same 'defaults') and now I get as far as
calibrating clocks completed before the box spontaneously reboots. :)  The screen posts immediately
so I guess I'll need a serial console to go any further .. :(

static mpfps_t  mpfps;
static struct MPFPS fake_mpfps = {
        .signature = "_MP_", /* the table "Intel" signature */
        .pap = 0,            /* not looked at when mpfb1 is non-zero */
        .length = 0,         /* ditto */
        .spec_rev = 0,       /* not used */
        .checksum = 0,       /* not used */
        .mpfb1 = 5,          /* ISA + PCI + builtin APIC */
        .mpfb2 = 0,          /* tried 0x80, no noticeable difference */
        .mpfb3 = 0,          /* can't see where 3, 4, 5 are used */
        .mpfb4 = 0,
        .mpfb5 = 0,
};
. ..
  if ( x < 0 )
      mpfps = &fake_mpfps;
  else
      mpfps = (mpfps_t)x;


I hope I'm wrong but this is beginning to look like a lot of work ;)

Any help/hope anyone can provide would be greatly appreciated!!

--Andrew

Andrew Atrens wrote:
> The title says it all.
> 
> I've done my best to scan around memory to find the table signature - _MP_
> but I'm beginning to think that the table doesn't exist.
> 
> Linux takes the novel approach of booting in UP mode, and then later switching to MP using
> info gleaned from ACPI MADT.
> 
> --Andrew