Message-ID: <03273D4BC7F4D511AC440008C7E6F8815421CF@nlntmail2b.ats.nld.alcatel.nl> From: "Witvliet, Hans" <Hans.Witvliet@Alcatel.nl> Date: Tue, 19 Feb 2002 09:27:48 +0100 Subject: FW: [suse-sparc] FW: 2.4.x compile on SPARC?
So it seems that all the "official hackers" were to busy for the last 6
months... alas
Means we still can not use ip-tables ;-(
Perhaps now that SUN and linux are getting a closer relationship, thing
might (!) improve.
Hans.
-----Original Message-----
From: rolf@diesing.net [SMTP:rolf@diesing.net]
Sent: Tuesday, February 19, 2002 3:18 AM
To: Witvliet, Hans
Subject: RE: [suse-sparc] sun-32
Hello Hans,
Thorsten Kukuk says, it (2.4.x) is NOT working for sun4m, only
sun4u.
I don't know, why.
Imho it is possible to compile it but it will NOT work correctly.
This is something i realy know!
I found out, a newer Kernel from Kernel.org (2.4.18preX) will work
on Sun4m
but NOT so fast like the good old 2.2.20 on SMP-Engines.
This is something i tryed.
The 2.4.x Kernels seem to have trouble with flushing the Caches
partialy
in SMP-Mode. ( in my Case one!!! MByte per 2 RT626-CPU's )
In this Way Taskswitching gets very very slow and with high
CPU-latency.
(Because they flush the Caches over the Memory-Bus multiple Times
per
Seconds)
On SingleCPU-Sparcs without CPU-Cache the 2.4.18 Kernel maybe a good
Thing
but
in all other Situations use the stable and well _running_ 2.2.20
Kernel.
This is what i think about.
-----Original Message-----
From: Joshua Uziel [SMTP:uzi@uzix.org] <mailto:[SMTP:uzi@uzix.org]>
Sent: Tuesday, May 29, 2001 11:31 AM
To: Witvliet, Hans
Cc: 'suse-sparc@suse.com'; 'jim@ic.uva.nl'; 'pkrul@ultralinux.nl'
Subject: Re: [suse-sparc] FW: 2.4.x compile on SPARC?
* Witvliet, Hans <Hans.Witvliet@Alcatel.nl <mailto:Hans.Witvliet@Alcatel.nl>
> [010528 23:22]:
> I guess if the statement came from "guru" David Miller himself, it is quit
> official, as he 's the one who maintained sparc for a very long time.....
> Hope a new maintainer can be found real soon (Jakub Jelinek, Jim Mintha,
> Pieter Krul......?????) before serious bitrot will occur.
> BTW, who else could afford to spent that much time, to me it looks like
> almost a verious serious commitment.
> Not just considering the needed experience, but also in other requirements
> like hardware and much, very much time.
So, this seems to be something on the many SPARC/Linux mailing lists...
as it goes, not very many people are working on the SPARC/Linux kernel.
Dave Miller is still actively maintaining sparc64... that hasn't
changed, but he doesn't want to do sparc32 anymore. He says he'd rather
do one port really well than two of them half-assed.
So where does this leave sparc32? Anton Blanchard has been it's
maintainer for a while now... he's just taken a new job at IBM hacking
on Linux/PPC64... needless to say he's got a lot of cool hardware to
muck with there. He still seems to have intent on maintaining sun4m
support, but doesn't much care to do sun4c... and I'll probably be
takin' over sun4c in the little free-time that I have while backin' up
Anton on the sun4m's (since I have quite a few of 'em to test with).
So yeah, it's not that people are givin' on up it... just that they're
all very busy with other stuff and will work on it when there's some
free cycles to do it.
-- To unsubscribe, e-mail: suse-sparc-unsubscribe@suse.com For additional commands, e-mail: suse-sparc-help@suse.com
This archive was generated by hypermail 2.1.0 : Tue Feb 19 2002 - 00:25:46 PST