On Sat, 15 Jul 2000, Russell King - ARM Linux Admin wrote:
> > - When did binutils and gcc start supporting ARM-based formats?
>
> See my site.
All you say is "At the start of 1995, GCC supported the ARM processors,
but binutils was lacking. [...] Binutils didn't have particularly
brilliant support for ARM - it didn't link the bl instruction properly."
I was really looking for something a bit more specific, like who
contributed this work to gcc and binutils, and for what purpose
originally. I find it a bit unlikely that there would have been some
substantial amount of ARM support in these tools but that it was
heavily broken -- that doesn't make much sense!
> It was a mini-C library written by me on the A5000 to get things just
> "smash", the fileutils, etc going.
I was more after where the ARM support in libc4 came from, or equally
where NetBSD got its ARM support in its C library. (I'm not sure about
the code topology sharing between NetBSD 1.0 and Linux 1.1's C libraries.)
> Minix patches were available, and of course RISCiX. NetBSD/arm32 was
> in very early days (no console at least iirc). I started the kernel
> in early `94.
Right; are there any details on these Minix patches?
And NetBSD/arm32 was, it seemed to me, out there and running X and
friends a year or few before Linux/ARM -- am I imagining this?
> > (Yes, I know there's some story on Russell's site, but it seems apocryphal.)
>
> Well, thanks a lot. I've had to put up with a lot of arsey comments
> over the past 6 years, and I certainly don't need people like you
> trying to be offensive and down right rude to me, or trying to make
> stuff I put on my site to be useful or interesting to be termed
> "a work of fiction".
>
> Why do you doubt it? If you'd been at UKUUG Linux99, then you'd have
> heard the full story, which starts off more or less with that.
>
> Oh, and if you want to check up on the info on that page, you could
> talk to either Martin Ebourne of !Zap fame, or Tim Chown, the
> Electronics and Computer Science CS admin at Southampton University
> (92 to at least 96). Martin, however, has the most detailed
> independent knowledge of the progress.
>
> Oh, and please, don't write an apocryphal story about it, lest you
> piss me off even more. After all, there is only ONE person who can
> really review such a story, and say "its accurate" isn't there?
I didn't term that story `a work of fiction'. I said it /seemed/
apocryphal, meaning `of questionable authenticity', where I take
`authenticity' to mean `undisputed credibility'.
Given that story is only your account, having been ratified by no one
else to my knowledge, and also that parts of it seem a little strange
(other kernel hackers have cast doubt upon the ability of any version
of Norcroft C to compile /any/ version of the kernel), it seemed fair
to say that its credibility is not unquestionable; anyone making an
accurate account of history to which many people must agree surely must
make this judgement of this source!
I'm not about to write an apocryphal story: that's why I'm being careful
about blindly accepting the facts you present on your site. Obviously
I have no wish to `piss you off' without reason, but I certainly won't
stop at presenting what I believe the truth to be /purely/ because it
may irritate you.
I think your statement "there is only ONE person who can really
review such a story" is indicative of a minor attitude problem on your
part; you've won your kudos by being the first to contribute ARM
code to the Linux kernel -- don't now lose it by being a bad team
player. (Yes, that's harsh, but it's my sincere opinion.)
c.
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
++ Please use [EMAIL PROTECTED] for ++
++ kernel-related discussions. ++
