[LinuxBIOS] Intel refactoring and microcode updates

joe at smittys.pointclark.net joe at smittys.pointclark.net
Sat Jan 12 16:13:20 CET 2008


Quoting Corey Osgood <corey.osgood at gmail.com>:

> joe at smittys.pointclark.net wrote:
>> Quoting Uwe Hermann <uwe at hermann-uwe.de>:
>>
>>
>>> On Fri, Jan 11, 2008 at 03:54:18PM +0100, Peter Stuge wrote:
>>>
>>>> On Fri, Jan 11, 2008 at 09:42:00AM -0500, Corey Osgood wrote:
>>>>
>>>>> Peter Stuge wrote:
>>>>>
>>>>>> I think we need to make it configurable.
>>>>>>
>>>>> I don't like that. With a factory bios, you expect the correct
>>>>> microcode update for your CPU to be present, no matter what CPU you
>>>>> put in a socket.
>>>>>
>>>> (Actually no, not always.)
>>>>
>>>>
>>>>
>>>>> We should be able to do the same.
>>>>>
>>>> I agree, but we should also try to be even more flexible. I think we
>>>> should allow inclusion of 0<=n<=all microcode updates. Definately an
>>>> advanced option, but still.
>>>>
>>> Yep, that's what I meant. It's fine if the default is "all microcode
>>> updates", but there should be an option for advanced users to only use
>>> the one(s) you really want or need in order to save time and space.
>>>
>>>
>>> Uwe.
>>>
>>>
>> That makes perfect sense.....I like the advanced option idea:-)
>>
>> Thanks - Joe
>
> This seems like it would get very messy, very quickly. The only way I
> can come up with to do it is #if's for every single ID, perhaps in some
> intel_custom folder, so it doesn't make a mess of the existing stuff.
>
> Anyways, I'll get on top of testing this weekend, I have slot 1 (440zx),
> socket 370 (i810), and socket 479 (i830) boards I can test with,
> assuming they all still work. If those work, I'll bet the farm that the
> rest of them do. Everything passes abuild, btw.
>
> -Corey
>
Yeh, even though it is a good idea it probibly would complicate things  
alot. For now it may be better to just commit a piece at a time for  
each processor tested. I'm really interested in how the socket 479  
(Pentium M) testing goes. Let us know:-)


Thanks - Joe




More information about the coreboot mailing list