Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
On Wed, Oct 23, 2002 at 11:35:36AM +0200, christian mock wrote: > Did the stability of SMP sun4m boxen improve for anyone going from > 2.4.18 to 2.4.19? Shall I invest the multiple CPU hours to recompile > for SMP? Well, I tried 2.4.20-pre11 from osinvestor.com, and here it sits, going spin_lock() CPU#X stuck at ... for both CPUs, over and over -- same problem as with earlier 2.4SMP kernels. ciao, cm. -- ** christian mock in vienna, austria -- http://www.tahina.priv.at/~cm/ ** VIBE!AT http://www.vibe.at/ ** wir sind nicht zum spass hier. Ich kenne auch ein Klo, wo "Austria Email" draufsteht. Das ist wahrscheinlich eine Art Rohrpost. -- Robert Bihlmeyer in at.sonstiges
Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
* christian mock <[EMAIL PROTECTED]> [021023 02:41]: > Did the stability of SMP sun4m boxen improve for anyone going from > 2.4.18 to 2.4.19? Shall I invest the multiple CPU hours to recompile > for SMP? Absolutely... there were a _ton_ of fixes that went into the 2.4.19-pre series for 2.4.x ...
Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
My opinion is that Ben knows a lot more about this stuff than I do, plus he's the one that would end up having to support it. I personally think of 2.4 on sparc32 as experimental. sun4d is completely broken, sun4c breaks at whim, and sun4m has a bunch of smaller problems. The only reason I put up my debs is because they're what I use personally, that I don't guarantee anything by putting them up, and that the userbase is so much smaller than Debian's that I don't need to worry about getting 20 e-mails a day about swapon being broken on SMP. Regards, Rob Radez
Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
On Wed, Oct 23, 2002 at 09:05:22AM -0400, Ben Collins wrote: > On Wed, Oct 23, 2002 at 10:35:21AM +0200, Thomas 'Mike' Michlmayr wrote: > > On Tue, Oct 22, 2002 at 21:52:17 -0400, > > Ben Collins <[EMAIL PROTECTED]> wrote: > > [...] > > > Because kernel 2.4.x is not stable on sun4cdm yet. You've noted that > > > already :) > > > > actually, 2.4.19 was behaving quite well on my SS20 (even in SMP mode) > > and LX. i'm running the dual-SS20 as my two-headed desktop, and it works > > better with 2.4 than with 2.2 (there were issues with the mouse that the > > kernel-upgrade solved). and the 2.4.20-pre11 image from rob seems to have > > fixed the last outstanding annoyance (problems with swap on SMP kernels). > > as far as i can tell from my desktop and firewall, 2.4.current on sun4cdm > > is no worse than on i386. > > > > could we please start to reevaluate the '2.4 is bad for sun4cdm' mantra > > and actually give 2.4 a chance? > > When you have to answer to a few thousand Linux-sparc users for your > decisions, then you can make that choice. > > Answer me this, is the swapon segfault on boot fixed? If not, then I'll > leave 2.4.x out for sparc32 for now. I don't need 20 emails a day to > this list asking "Is this swapon segfault bad?". > I suggest you talk with Rob and see what his opinions are. Cheers, /ChJ
Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
On Wed, Oct 23, 2002 at 10:35:21AM +0200, Thomas 'Mike' Michlmayr wrote: > On Tue, Oct 22, 2002 at 21:52:17 -0400, > Ben Collins <[EMAIL PROTECTED]> wrote: > [...] > > Because kernel 2.4.x is not stable on sun4cdm yet. You've noted that > > already :) > > actually, 2.4.19 was behaving quite well on my SS20 (even in SMP mode) > and LX. i'm running the dual-SS20 as my two-headed desktop, and it works > better with 2.4 than with 2.2 (there were issues with the mouse that the > kernel-upgrade solved). and the 2.4.20-pre11 image from rob seems to have > fixed the last outstanding annoyance (problems with swap on SMP kernels). > as far as i can tell from my desktop and firewall, 2.4.current on sun4cdm > is no worse than on i386. > > could we please start to reevaluate the '2.4 is bad for sun4cdm' mantra > and actually give 2.4 a chance? When you have to answer to a few thousand Linux-sparc users for your decisions, then you can make that choice. Answer me this, is the swapon segfault on boot fixed? If not, then I'll leave 2.4.x out for sparc32 for now. I don't need 20 emails a day to this list asking "Is this swapon segfault bad?". -- Debian - http://www.debian.org/ Linux 1394 - http://www.linux1394.org/ Subversion - http://subversion.tigris.org/ Deqo - http://www.deqo.com/
Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
On Wed, Oct 23, 2002 at 11:35:36AM +0200, christian mock wrote: > On Wed, Oct 23, 2002 at 10:35:21AM +0200, Thomas 'Mike' Michlmayr wrote: > > > actually, 2.4.19 was behaving quite well on my SS20 (even in SMP mode) > > and LX. i'm running the dual-SS20 as my two-headed desktop, and it works > > IBTD; I've compiled 2.4.19 as non-SMP to finally get my 2 CPU SS10 > stable; it's running since 29 days now on one CPU, with 2.4.18SMP it > was getting stuck about every other day. > > Did the stability of SMP sun4m boxen improve for anyone going from > 2.4.18 to 2.4.19? Shall I invest the multiple CPU hours to recompile > for SMP? > I'd suggest you download rob's iso and try them out, I've done it for some months and they're just fine. Both the 2.4.19_2 ones and the 2.4.20-pre11 ones. Cheers, /ChJ
Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
On Wed, Oct 23, 2002 at 10:35:21AM +0200, Thomas 'Mike' Michlmayr wrote: > actually, 2.4.19 was behaving quite well on my SS20 (even in SMP mode) > and LX. i'm running the dual-SS20 as my two-headed desktop, and it works IBTD; I've compiled 2.4.19 as non-SMP to finally get my 2 CPU SS10 stable; it's running since 29 days now on one CPU, with 2.4.18SMP it was getting stuck about every other day. Did the stability of SMP sun4m boxen improve for anyone going from 2.4.18 to 2.4.19? Shall I invest the multiple CPU hours to recompile for SMP? ciao, cm. -- ** christian mock in vienna, austria -- http://www.tahina.priv.at/~cm/ Vielleicht ist es aber auch nur einfach Zeit, dem Wolfi wieder den root-Account wegzunehmen. Er ist noch zu klein dafuer. Ferdinand Goldmann in at.gesellschaft.politik
Re: official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
On Wed, Oct 23, 2002 at 10:35:21AM +0200, Thomas 'Mike' Michlmayr wrote: > On Tue, Oct 22, 2002 at 21:52:17 -0400, > Ben Collins <[EMAIL PROTECTED]> wrote: > [...] > > Because kernel 2.4.x is not stable on sun4cdm yet. You've noted that > > already :) > > actually, 2.4.19 was behaving quite well on my SS20 (even in SMP mode) > and LX. i'm running the dual-SS20 as my two-headed desktop, and it works > better with 2.4 than with 2.2 (there were issues with the mouse that the > kernel-upgrade solved). and the 2.4.20-pre11 image from rob seems to have > fixed the last outstanding annoyance (problems with swap on SMP kernels). > as far as i can tell from my desktop and firewall, 2.4.current on sun4cdm > is no worse than on i386. > > could we please start to reevaluate the '2.4 is bad for sun4cdm' mantra > and actually give 2.4 a chance? I can support this, I have been running 2.4.19 sun4m SMP (SS20, dual) and 2.4.20-preXX with various results, the 2.4.20-pre11smp that I use off of Rob's site works fine as far as I can tell. Sure, it has bugs, what piece of source code does not... Cheers, /ChJ
official 2.4 debs for sun4cdm (was: Re: kernel 2.4.x on sun4c)
On Tue, Oct 22, 2002 at 21:52:17 -0400, Ben Collins <[EMAIL PROTECTED]> wrote: [...] > Because kernel 2.4.x is not stable on sun4cdm yet. You've noted that > already :) actually, 2.4.19 was behaving quite well on my SS20 (even in SMP mode) and LX. i'm running the dual-SS20 as my two-headed desktop, and it works better with 2.4 than with 2.2 (there were issues with the mouse that the kernel-upgrade solved). and the 2.4.20-pre11 image from rob seems to have fixed the last outstanding annoyance (problems with swap on SMP kernels). as far as i can tell from my desktop and firewall, 2.4.current on sun4cdm is no worse than on i386. could we please start to reevaluate the '2.4 is bad for sun4cdm' mantra and actually give 2.4 a chance? -- Thomas 'Mike' Michlmayr | ignorami: n: The BOFH art of folding problem <[EMAIL PROTECTED]> | lusers into representational shapes. pgpNqh161XHxv.pgp Description: PGP signature
