[OpenBIOS] State of AIX

[Andreas Färber]

Am 29.01.2011 um 17:19 schrieb Blue Swirl:

> On Sat, Jan 29, 2011 at 1:37 PM, Andreas Färber <andreas.faerber at web.de 
> > wrote:
>> Am 12.01.2011 um 19:58 schrieb Blue Swirl:
>>
>>> On Tue, Jan 11, 2011 at 9:19 PM, Andreas Färber <andreas.faerber at web.de 
>>> >
>>> wrote:
>>>>
>>>> Am 07.01.2011 um 23:03 schrieb Grieve, Shane:
>>>>
>>>>> I have been following the AIX booting issue for a while now and  
>>>>> was
>>>>> wondering if there is any progress. I have access to AIX  
>>>>> hardware and I
>>>>> am
>>>>> happy to help test.
>>>>
>>>> The next open issue still is the prom_arg_t ppc64  
>>>> incompatibility. The
>>>> patches are in need of review, in particular I'm waiting for  
>>>> feedback
>>>> from
>>>> Blue whose suggestions I tried to pick up in v2. Those patches  
>>>> may then
>>>> need
>>>> rebasing since Mark incorporated PRI*32 additions into a sparc32  
>>>> commit
>>>> recently.
>>>
>>> Which patches? I thought I had commented on all of them.
>>
>> http://lists.openbios.org/pipermail/openbios/2010-December/ 
>> 006028.html
>> http://lists.openbios.org/pipermail/openbios/2010-December/ 
>> 006029.html
>>
>> The only comment was from Mark concerning obsoleted v1.
>> And since this touches a central part of OpenBIOS I don't just  
>> commit that
>> without acks.
>>
>> I'll try to look into upstreaming all the PRI* stuff this weekend,  
>> which are
>> semi-correct in SVN.
>
> The patches do not apply anymore, but they look OK to me.
>
> Output formatting for 64 bit types is a bit tricky with system types.
> Sometimes on 64 bit machines, long is used for int64_t and so %llx
> generates a warning. In other cases, long long is used and so %lx is
> not correct. I think the patches are always using long long, so they
> should be OK.

Blame Mark. ;) Seriously, everything else was automatically git-rebase- 
able, I've pushed a testable version:
http://repo.or.cz/w/openbios/afaerber.git/shortlog/refs/heads/ppc64-boot

Not sure what you're referring to above. I don't think I modified the  
type definitions, I went with what they were defined like. Most are  
long long as you say, but that's something I would like to be able to  
change in the future. Thus, I'm defining the cell types in term of  
POSIX integer types, using their PRI* macros to get away from the ll  
vs. l issue. Refactoring that into a separate patch now.

As for the client interface part, I take it that this still makes  
sense despite the ppc64 cell width discussion and will work towards  
committing.

Andreas