Re: SMP Mtrr Probs with 2.2.1

Kurt Garloff (K.Garloff@ping.de)
Mon, 8 Feb 1999 16:58:13 +0100


--mYCpIKhGyMATD0i+
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable

On Mon, Feb 08, 1999 at 02:29:15PM +0100, Jan Blunck wrote:
> Im running an Dual Celeron Gigabyte 686 BXDS with linux-2.2.1-ac3.
>=20
> It's running stable but on bootup the Kernel alerts about wrong
> Mtrr-settings:
>=20
> mtrr: your CPUs had inconsistent variable MTRR settings
> mtrr: probably your BIOS does not setup all CPUs
>=20
> But the BIOS detects both CPU's. WinNT and Solaris 7 are also detecting
> both CPU's.
>=20
> cat /proc/mtrr:
>=20
> reg00: base=3D0x00000000 ( 0MB), size=3D 64MB: write-back, count=3D1
> reg01: base=3D0xe8000000 (3712MB), size=3D 8MB: write-combining, count=
=3D1
>=20
> Is this because off the Celeron CPU's ? I couldn't test it with orig.
> Pentium 2 CPU=20
> because I have non.
> Any Idea's ???

Everything is OK. Your BIOS probably sets up write-back for your main memory
and write-combining for your framebuffer (VGA) onyl on hte first (boot) CPU.
Linux detects it and sets the second CPU the same way and tells you about i=
t.
Everything fine with Linux.

I don't know what NT or Solaris do, but they don't tell you ...
Note that inconsistent settings in this case don't affect system stability,
only performance ...

BTW: A lot of BIOSes only set the framebuffer wrt-comb. on the first CPU,
AFAIK.

--=20
Kurt Garloff <kurt@garloff.de> [Dortmund, FRG]
Plasma physics, high perf. computing [Linux-ix86,-axp, DUX]
PGP key: see mailheader [Linux SCSI driver: DC390]

--mYCpIKhGyMATD0i+
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3in

iQCVAwUBNr8JlRaQN/7O/JIVAQERzAP+KYvOHyMgABBaXvGir3ctptlc4zZy330P
P6uNzfUNOjqIykEVcrAxaLV4tEeTk9guNu66lFnFtTDT1LQFFuoS0+rxmKF0ljfJ
30chQfW8KKt8U66c5y921CKFkbo9nUHVXGAEw7gdZHR5FygTcCJLwCTzciqs0Gcg
vqKtGOjwgmc=
=18WF
-----END PGP SIGNATURE-----

--mYCpIKhGyMATD0i+--

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/