Text 2842, 98 rader
Skriven 2011-03-04 20:21:40 av Jonathan de Boyne Pollard
Ärende: Re: AMD releases open source boot code
==============================================
Gecko/20110221 Thunderbird/3.1.8
UTC)
comp.os.os2.ecomstation:876 comp.os.os2.beta:32
From: Jonathan de Boyne Pollard <J.deBoynePollard-newsgroups@NTLWorld.COM>

>>> Can't we learn something from it though? There will be examples of 
>>> tapping into interfaces that our loader currently has no clue about.
>>>
>> Once again: It's not an operating system or an operating system 
>> loader. It's replacement firmware. There won't be any "tapping into 
>> interfaces" because it isn't a loader. It's what loaders (the 
>> so-called "payload") run on top of and use.
>>
> "The final mainboardinit fragment is mainboard-specific, in C, called 
> romstage.c. For non-cache-as-RAM targets, it is compiled with romcc. 
> It includes and uses other C-code fragments for: [...]"
>
> Really?  None of that is useful?
>

Correct.  None of it is useful.  What part of "is mainboard-specific" is 
unclear here?  No-one working on the OS/2 kernel loader or a clone who 
is in their right mind would make the loader specific to a particular 
mainboard.  Loaders are generic, that use the abstract interfaces 
provided *by* the firmware, without regard to the details of the 
particular mainboard at hand.

> Just because this stuff is designed to execute from ROM doesn't mean 
> we can't make use of it.
>

No-one said anything about it executing from ROM being the problem.  The 
fact that it's firmware, not an operating system loader, is what makes 
it useless for the purposes of an operating system loader.  It's 
firmware.  It's not an operating system loader.  It's code that does a 
completely different job, at a different layer of abstraction.  How many 
times does this point have to be made before it sinks in?

>> And what's this "our loader" business? I for one don't have any stake 
>> in the IBM OS/2 kernel loader. It isn't mine. It's IBM's, last that I 
>> looked.
>>
> The loader that we, who use OS/2, use.  If you prefer not to be 
> included in that group, then I won't complain if you don't read 
> yourself into my statement.
>

You clearly haven't thought that through.  Merely using a loader doesn't 
make your the owner of it.  And if you aren't the person who is capable 
of changing the loader, *even if* this replacement firmware were somehow 
relevant to the task of doing so, then you don't belong to the relevant 
"we", which in this case is IBM and the people who can modify OS2LDR, 
and the other groups that I alluded to.  In other words: You're not 
included in the relevant "we" any more than I am.  You're mistaken to 
think that you are.

>> I know of only three groups outwith IBM that have stakes in 
>> OS/2-clone kernel loaders. The OSFree people have their own loader, 
>> which is (we'retold) based upon the old FreeLDR loader from David C. 
>> Zimmerli (not to be confused with the ReactOS FreeLDR). They're the 
>> only people with an interest in "our loader", and they'll tell you as 
>> I do that firmware source code doesn't show how to *use* firmware APIs.
>>
> I'd be mining for data from under every rock if I were them, and 
> hopefully they are not as close-minded as you seem to assume.
>

This isn't about close-mindedness.  This is about having some degree of 
Clue as to what the task of writing an operating system loader actually 
entails, and thus what is and isn't helpful to that task.  What is 
helpful are things like the PCI BIOS specification, the PnP BIOS 
specification, the EFI specification, the Phoenix/IBM/Microsoft INT 13h 
extensions specification, the AMD and Intel CPUID specifications, the 
AMD and Intel software developers' manuals, and so forth.  What isn't 
helpful is the source code to one particular machine firmware.  One 
doesn't need to know how the firmware does its job internally.  One 
needs to know the interface contract that you and it work to.  Rare 
indeed is the implementation source code that is useful for knowing the 
interface contract.  Try reading the Bochs firmware source code and 
working out the PnP BIOS or PCI BIOS specifications from it, for 
example.  (Those in the know will be highly amused at that suggestion.)

You're getting all excited just because of a press announcement for a 
five-year-old project, announcing some "free" stuff, and your excitement 
is taking the fact that it's "free" to be far more important than the 
fact of whether it's even relevant.  I'm the person with Clue telling 
you that this sort of thing is about as relevant to the job as someone 
announcing that they are giving away free pictures of Jamie Lynn 
Spears.  It's an entertaing sideshow for those with an interest in it, 
but it isn't at all useful for the task at hand.

For a fourth time:  It's replacement firmware.  It's not an operating 
system loader, nor an operating system.  It's what loaders (the 
so-called "payload") run on top of and use.  The fact that something is 
"free" and recently (re-)announced doesn't automatically make it relevant.

--- Internet Rex 2.31
 * Origin: virginmedia.com (1:261/20.999)