From: Antonio V. <wi...@co...> - 2003-07-18 16:01:09
|
On Sun, Jul 13, 2003 at 10:16:19PM -0400, David Coulson wrote: > > > -------- Original Message -------- > Subject: [lugc-talk] RedHat 9 UML Host Kernel Packages > Date: Sun, 13 Jul 2003 22:15:12 -0400 (EDT) > From: William R. Lorenz <wr...@ex...> > Reply-To: ta...@cl... > To: ta...@cl... > > With the Calc II final behind me now, I wanted to post a quick link to > some RedHat 9 host UML kernel packages that I built. These are made to > use for the host one which you will be running UML instances and include > the skas3 and fairsched patches, most definately required for good hosts. > > ftp://wrl.umlcoop.net/pub/kernel.fairsched+skas3/ > > These kernel RPM packages (SRPM & RPM) include the following changes from > the most recent stock RedHat-provided kernel for RH9 distributions: > > ( all are included in the changelog w/ specific patch numbers) > > - added support for fairsched & skas3 host patch (User Mode Linux) > - added defconfig definitions for fairsched & skas3 host patch > - modified kernel config definitions in SOURCES tree (kernel-2.4*.config) > - removed O(1) scheduler & threading (based on v2.5) backport Some months ago, I started work on a new fairsched patch for 2.5 kernels with o(1) complexity ( goto http://wind.cocodriloo.com/~wind/fairsched/ ). I was stalled, but perhaps someone else, with better kernel skills, could take a look a try to make it work. Mind you, this may be important in order to get a good kernel for the host side. > - removed LDT allocation to allow > 1200 threaded proc backport > - removed VM tuning patch for CONFIG_HIGHMEM64G special DMA stuff > - touched up ptrace()/kmod patch to work with UML patches > > I hope to also have a RedHat 9 pristine minimal image available via FTP > soon, and I'll also work on building a tutorial for those interested in > getting a UML system running in under 5 minutes and with networking. > > This is the under-5 approach to the host kernel, and it also helps to have > the kernel installed with MD5 prints and using manageable packages. :) > > More info coming soon! > Greets, Antonio -- 1. Dado un programa, siempre tiene al menos un fallo. 2. Dadas varias lineas de codigo, siempre se pueden acortar a menos lineas. 3. Por induccion, todos los programas se pueden reducir a una linea que no funciona. |