On 2002.07.23 David Walser wrote: > >--- "J.A. Magallon" <[EMAIL PROTECTED]> wrote: >> What would be nice is that Mandrake shipped a full >> -aa kernel, instead of >> just taking the vm part (as I understand from >> changelogs). >> This way you get XFS, UML, Tux, for free and many >> performance candy designed for >> big (entreprise, smp) boxes that does not hurt at >> all on workstations. >> That would be an 'advantage' over RedHat, that ships >> -ac kernels (heavily >> patched, of course). >> >> You can take a look at performance comparisons at >> >> http://home.earthlink.net/~rwhron/kernel/bigbox.html > >If they're gonna ship an -aa kernel, why not ship a >-jam kernel while they're at it, that page says it's >based on -aa, and the guy who makes the -jam kernel is >on this list. >
je, je... >Hey, I just noticed who I was replying to :o) You >make the -jam kernel. Do you think it would be appropriate? > Currently, it just contains some small bug fixes and the bproc stuff. I think it is too experimental for a distro, or contains things that just not fit on a distro kernel: - enhancements for PII and up, that have no use on a general i586 kernel - Andre Hedirck IDE updates, that work fine but are now a bit incompatible with -aa (Promise...) - bttv and sensors, that are also tracked by Juan (btw, 2.6.4 is _out_) - the bproc stuf... (If someday I have the time, perhaps I will put a kernel-bproc-xxxx-mdk+libbproc out there. Problem is maintenance...) Of course, if anything in -jam is useful for a mdk kernel, there it is. Some people are just trying to know which of -jam patches made that low latecy effect. -aa is a secure approach and much more tested. -- J.A. Magallon \ Software is like sex: It's better when it's free mailto:[EMAIL PROTECTED] \ -- Linus Torvalds, FSF T-shirt Linux werewolf 2.4.19-rc3-jam1, Mandrake Linux 9.0 (Cooker) for i586 gcc (GCC) 3.1.1 (Mandrake Linux 8.3 3.1.1-0.10mdk)
