[LinuxBIOS] [PATCH] AMD Barcelona(family 10) support 3 of 4

Rudolf Marek r.marek at assembler.cz
Fri Dec 14 00:21:15 CET 2007


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Marc,

Just a quick note due to total -ENOTIME.

+unsigned long acpi_fill_mcfg(unsigned long current)
+{
+	/* Just a dummy */
+	return current;

You don't need this, the acpi_fill_mcfg is a weak symbol. if it does not exists
everything compiles fine.

Oh just one thing. I wanted to get the S3 working with K8 revE. I have written
some experimental implementation, but memory after resume is unusable. I'm just
triggering the bits DCL_ESR | DCL_SRS to exit self-refresh but no luck. I'm
working-out all known revE errata. Is there a way how to get access to some more
advance documentation? Is really sufficient just to trigger those to bits
instead of InitRAM? Do I need to deal with the DisDqsHys even when from resume
(of course I tried both). The PME registers should respond to S3 sb generated
cycle. Anyway this can wait and it is totally lowprio, but still I'm curios ;)

Thanks,
Rudolf
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHYb5r3J9wPJqZRNURAic8AKCEj7H2+VLR9bgRLQjAHT2xsi2QpQCdF3/Z
lRPnycGrpcKhTAXuOFnhE1Q=
=Ic13
-----END PGP SIGNATURE-----




More information about the coreboot mailing list