Re: Dropping sparc32 for lenny
On Wed, 2007-04-25 at 09:01 +0200, BERTRAND Joël wrote:
> I don't aggree. Some new hardware are created today with Leon and some
> other sparc v8 processors (in fpga or other technology). If sparc32 port
> is dropped, no one usable system will be able to use these systems.
> 1/ Sparc32 (with sun4m) support was dropped with solaris 10 and solaris
> 9 is not stable with more than two procs;
> 2/ NetBSD 3.1 is not stable when it runs on SMP configuration;
> 3/ NetBSD 4.x is not usable (on SS20 with CG14);
> 4/ OpenBSD cannot run on any SMP station;
> 5/ Linux sparc32 is not dropped, some bugs remain in sparc32 SMP kernel,
> but is better than all other systems on sparc32.
> I use several SS20 (with raid, CG14, SMP, X, windowmaker...) with only
> one trouble: a bug in read_pipe()... And I have worked on another
> project that uses Leon. Berkeley University work today on a calculator
> that shall use several Leon processors... Sparc32 is not died even Sun
> does not sell Sparc32 anymore.
> Thus dropped sparc32 arch will be a big strategic error for Debian.
I don't dispute the facts you've given, and personally, I'd like to see
Debian ported as widely as possible* but ... if I may play devil's
advocate for a moment:
1. There are still sparc32 systems in production and in use; but what
proportion of them are 1. publically available (as in COTS rather than
Verliog like Leon or the UltraSPARC T1), 2. run Linux, 3. run Debian?
If they care about having a Debian/sparc32 port then where are they?
And why aren't they offering some support in keeping it running? For
developers with experience of custom designed embeded systems, helping
maintain kernel code shouldn't be too difficult.
2. This unlikely to make a major difference until a year after lenny is
released, when support will be fully stopped. Given most users seem to
be using old Sun hardware; what proportion of that will still be in use
in 2/3 years time?
I guess it comes down to a question of how much work it is to maintain
verses how much the Debian projects gains by the extra user base / being
the only option of free OS for these platforms.
I note that according to popcon : http://popcon.debian.org/ sparc is
one of the larger ports. I guess with access to the raw data, it should
be possible to extrapolate / guestimate the number of sparc32 systems in
*. but as I can't directly contribute to making this happen I don't
think anyone should pay attention to what I think on the matter. I'd
like it if they did but I understand completely if they say "you want
it; you go do it".