@ erklaerbaer post #15

you can always rebuild the kernel as Slackware does not break with vanilla kernels from kernel.org

i used to be scared to rebuild a kernel as i used to break things, and i was determined to learn this procedure so i used an extra partition with a clean slackware install in it and kept the stock kernel in it in case of failure, and after going thru make menuconfig a few times and looking at Pat's config for the bare.i and what the kernel developers have in make defconfig i finally got the method down to building my own kernel quite well, trimming the unnecessary items out that my hardware does not use and what i don't use or need, and changing a few things to what i prefer, i find building a kernel mostly a mundane task but a necessary thing to do if the stock kernel does not suit you for long term use...

that was not my question

i was just interested, why -smp kernels are actually considered "better" for non-smp machines. "just as good" makes perfekt , but; hmm well i guess it doesn't matter.

EDIT: kernel ~> machines ; -since

Well, I didn't know this one off the top of my head, but after some consultation with Pat and PiterPunk, here's the verdict: :-)
1. Some new hardware has broken static IRQ routing, and the SMP kernel is build with local APIC enabled (which should take care of that, as I understand it).
2. An SMP-capable kernel tests if the machine is a uniproc at boot, and if so, some functions are replaced so that there's (supposedly) no performance penalty induced by running the SMP-enabled kernel.
3. The kernel sources and includes shipped with Slackware are for SMP kernels. There's a "fix" for that in /extra, but based on the above information, there's no good reason to use a uniproc kernel if the smp kernel will run.

Thanks for the explanation Robby:-) Will the next release of Slack have the 2.6x SMP capable kernel as the default choice when we do the installation? Or will we have a choice of using a non SMP enabled kernel.
Based on your explanation I will go with a SMP enabled kernel.

the SMP should not cause any performance loss to single processor machines...

i will use it for a while until a newer kernel is released then i will roll another one

Well, I really don't *know* what the next release of Slackware will do in this regard, but if the present behavior of a new -current installation is any indication, it will be something like this with a full installation (assuming the release ships with 2.6.21.3 - adjust as needed):
kernel-generic-2.6.21.3, kernel-huge-2.6.21.3 and kernel-modules-2.6.21.3 for uniprocessor systems as well as kernel-generic-smp-2-6.21.3-smp, kernel-huge-smp-2.6.21.3-smp and kernel-modules-smp-2.6.21.3-smp for SMP systems will all be installed. Since kernel-huge-smp is the last of the packages to be installed, the default /boot/vmlinuz symlink will point to it. This will be a safe choice for *most* people.
As noted in CHANGES_AND_HINTS / UPGRADE.TXT, the generic kernel is recommended for daily use, but of course, that's up to the user. The huge kernel is a good safe choice which will allow the user to boot into the system and create the initrd necessary for the generic kernel. This can, of course, be done prior to the first reboot after install, but it requires manually chrooting into the new installation and mounting /proc.

Thanks for the reply, Robby!

Any clue on what Xorg version will be used? 6.9 or 7.0?

Neither.

If the release happens today (which it won't, obviously), it will have an xorg version that's somewhere between 7.2 and upcoming 7.3 (which is slated to release in August, from what I understand). Pat has been pulling stable releases from the /individual directories on the xorg mirrors (which is essentially what happens on a major version (7.x) release from xorg.

Oh, because since 11.0 I think xorg was at 6.8 though there was already a 6.9 release. Now I can't remember what xorg version 11.0 uses, I'm so dumb

[edit]
Well according to ftp://ftp.slackware.com/pub/slackwar.../ANNOUNCE.11_0

So I just answered my own question.