Carl-Daniel Hailfinger wrote: > On 28.01.2008 21:55, Marc Jones wrote: >> >> Carl-Daniel Hailfinger wrote: >>> On 25.01.2008 19:17, ron minnich wrote: >>>> On Jan 25, 2008 10:15 AM, Jordan Crouse <[EMAIL PROTECTED]> wrote: >>>> >>>> >>>>> Hmm - probably the right move, but the changes in buildrom are not >>>>> trivial for v2. We would need to figure out what the ROM size is >>>>> (by greping >>>>> Config.lb, probably), compress the VSA, calculate the difference and >>>>> pad. Its not impossible, but there is lots that can go wrong if we >>>>> are not careful. >>>>> >>>> We need a VSA.v3 and leave the old one as before. VSA.v3 is just the >>>> uncompressed VSA. >>>> >>> Fully agreed. Maybe pick another name (VSA.uncompressed or >>> VSA.corebootv3.uncompressed), but let's simply add another download >>> option, this time suited for v3. IIRC the VSA for v3 may also use some >>> v3 infrastructure and be incompatible to v2 by definition. >>> >> I disagree. We will have an uncompressed solution that will work for >> both v2 and v3. VSA does simple hardware emulation so it should work >> for both. We are working out the details for buildrom now. I will also >> write down how to do it by hand on the wiki. > > I was under the impression that VSA in v3 was using x86emu for BIOS INT > services and using separate INT routines in v2. That's what led me to > the conclusion about different VSAs for v2 and v3.
VSA having int15 calls was bad. It made sense for a normal BIOS but not for coreboot. VSA can just as easily calculate what it was asking for. I am correcting that for this next release (real soon now). v2 patches also coming. Marc -- Marc Jones Senior Firmware Engineer (970) 226-9684 Office mailto:[EMAIL PROTECTED] http://www.amd.com/embeddedprocessors -- coreboot mailing list [email protected] http://www.coreboot.org/mailman/listinfo/coreboot
