The patched source tarball and SFS are here:
Comments:Posted on 4 Dec 2009, 19:03 by panzerpuppy
Why use the old kernel when 2.6.32 is out?
2.6.32 is out and it features an improved scheduler and memory management that makes the system more responsive on both old and new hardware. There are sound driver improvements as well.
Posted on 4 Dec 2009, 19:43 by BarryK
I patched and compiled 18.104.22.168 back when it was the latest, but I have only got around to uploading and announcing it now.
Posted on 5 Dec 2009, 2:01 by Feverfew
The Philosophy of Limiting Bloat
Hello Mr. Kauler. Your posts "Quirky idea #1 works", and "Smaller Xorg" got me remembering jest now about some stuff I stumbled upon sometime ago, and I just don't know if you have heard about some of it., FBUI (FrameBufferUI)for example. From what I've read IMHO you, and Mr. Smith (author of FBUI) have similarities in your philosophy's. So I thought. Hay this program might be helpful B).
Posted on 5 Dec 2009, 3:36 by Iguleder
So ... when do you plan to release Quirky? For what architectures?
I think you should shift from i486 to i686, i486 is dying slowly ... and a x86_64 flavor can be awesome. You can also drop all ISDN/old NIC stuff and save valuable space.
I built a Linux distro that connects to the internet through DHCP with a browser in about 3 MB. Just think of it, the possibilities are endless ... you can free up tenths of MBs this way.
Posted on 5 Dec 2009, 8:11 by Raffy
Perhaps in terms of Puppy's niche, it will be good to go in the direction of MIPS-compatible processor types like the Loongson
The gdium (laptop)is one mature product using the Loongson.
Posted on 5 Dec 2009, 22:52 by Michalis
BFS Scheduler patch?
What is your opinion about patching the kernel with Con Kolivas BFS Scheduler? Seems much more suitable to puppy than the general scheduler used in kernel.
More info in a linux magazine article and the actual [code=http://ck.kolivas.org/patches/bfs/]patch page[/code].
I also believe that a move of puppy/woof/quirky to i686 is a good choice since nowdays a computer with a cpu of pentiumII or amd-k6 is the bare minimum for everyday use. But I don't know whether i686 can influence the sizeof the packages.
Posted on 6 Dec 2009, 3:17 by Sit Heel Speak
on the BFS patch
I have recently tried the bfs scheduler patch. However, I could not get it to apply. The kernel scheduler code is lately undergoing changes with every new kernel subversion, so the bfs patch is kernel-subversion-specific. It might help if Mr. Kolivas would reveal which kernel subversion he wrote the patch for. Also, the bfs patch directly conflicts with Ingo Molnar's rt patch. So it looks like you can have bfs, or you can have rt, but not both.
Posted on 6 Dec 2009, 3:35 by Sit Heel Speak
gcc 4.4.1, heavily patched
I invite you to have a look at my gcc-4.4.1, extravagantly patched. Source code is at
Puppy forum page:
I admit it could use the attention of someone who knows how and what to strip.
I have been using it with unbroken success on both upup-450 and Pup-4.2-rt-smp. playdayz compiled fglrx (the ATi Catalyst driver) with it on upup-432. I just finished compiling full Tex with it on ttuuxxx's upup-450. It seems invincible.