On Wed, Apr 06, 2005 at 11:55:28AM -0500, Micah Anderson wrote: > > Here is a report of the hacking done last night (or day, depending on > your timezone) to port the vserver 1.9.5.12 patches to the Debian > kernel 2.6.8. Bertl is truely an amazing kernel hacker! > > Debian's Sarge, which will freeze "any day now"(tm), will be shipped > with the 2.6.8 kernel as the primarly supported kernel. It contains > backports of critical issues (including security fixes and major > crashes), is the 2.6 kernel that is part of the Debian Installer that > will be a part of the Sarge release, and will be the primary supported > kernel for Sarge. This means that most Debian users will use the > Debian 2.6.8 kernel, however there will be a 2.6.11 kernel available > as a fall-back in the case that 2.6.8 has a major problem that cannot > be backported. > > It is for these reasons that there was work done to get the vserver > 1.9.5 patch sets to work with this version of the Debian kernel. > Having a working version of this patch against the Debian kernel will > make it easier to get it to work for future Debian kernel revisions, > and will be included in the debian package kernel-patch-vserver. > > Plan: > . Get a vanilla/mainline 2.6.11.6-vs1.9.5.12 kernel up and > running on one machine > > . Find a way to easily check the debian 2.6.8-15-vs1.9.5.12 > kernel, compare them and check for differences > > . Port the split vserver 1.9.5.12 patches, by using a diff > against the vanilla 2.6.8 source and the debian 2.6.8-15 > source (this diff has the md5sum: 823fee2ebc230fdddff90824662fae62) > > . Compile, reboot, test test test. > > Progress: > . Got the 2.6.11-vs1.9.5.12 kernel up and running > . Installed the vserver utils 0.30.205 + 2 patches > . Converted legacy vserver configurations to new format > . Bertl finished adapting each split patch, reassembled and > was able to build successfully > . Filed a couple Debian kernel-source-2.6.8 bug reports > . First compile did not boot in qemu > . After disabling power-management it boots fine (apparantly > ACPI and PM got selected in kernel config) > . 02:08 < Bertl> Linux-VServer Test [V0.11] (C) 2003-2005 H.Poetzl > 02:08 < Bertl> chcontext is working. > 02:08 < Bertl> chbind is working. > 02:08 < Bertl> Linux 2.6.8-vs1.9.5 i686/0.30.204/0.30.204 [Ea] > 02:08 < Bertl> VCI: 0001:0025 273 03110076 > 02:08 < Bertl> --- > 02:08 < Bertl> [001]# succeeded. > 02:08 < Bertl> [011]# succeeded. > 02:08 < Bertl> [031]# succeeded. > 02:08 < Bertl> [101]# succeeded. > 02:08 < Bertl> [102]# succeeded. > 02:08 < Bertl> [201]# succeeded. > 02:08 < Bertl> [202]# succeeded. > > . Bertl tried with a different split (different from default) > and it was broken somehow > > . http://vserver.13thfloor.at/Experimental/patch-2.6.8-15-vs1.9.5-1.diff > was the first one > > . The latest version is > > http://vserver.13thfloor.at/Experimental/patch-2.6.8-15-vs1.9.5-1.1.diff
okay, update here: for some reason (I really do not want to discuss it ;) the first patch was done against a vanilla 2.6.8 kernel instead of the supposed to be 2.6.8-15 debian kernel ... okay, okay I admit it, I forgot to patch the all-in-one debian patch ;) so now we have a basically untested 2.6.8-vs1.9.5 kernel and because debian 2.6.8-15 isn't that far away from that I was able to update it to 2.6.8-15 which is now available as: http://vserver.13thfloor.at/Experimental/patch-2.6.8-15-vs1.9.5-2.diff the varhz and split patches were dropped, because the kernel is just too old to support them properly ... we'll see if FOR-1.9.6 stuff will fit to that later, now the basic testing has to be done (debian folks please start testing the stuff ... and report issues back) also I would like to get a statement from the debian folks who will in future maintain the 'debian' kernel version of linux-vserver ... because I do the 'initial' port, but I will not have time to keep it up to date ... HTH, Herbert > Notes/TODO: > . Need to test NFS > > . Need to compile, boot and test several times with small > changes, loooking at various features (from vserver paper and > changelogs) to know what to test > > . Need to check ext2/ext3 xattr with disk limit accounting > > . The ext3 xattr/dquot stuff is broken in Debian, Bertl > submitted a patch a few months ago that was included in > mainline (http://lkml.org/lkml/2005/1/26/174), this gets > quota wrong (filed debian bug: #303426) > > . Debian is missing > http://www.ussg.iu.edu/hypermail/linux/kernel/0501.3/1026.html which > is a spinlock_assert issue will crash your kernel when > compiled with up/spinlock_debug (filed debian bug: #303424) > > . Debian has atomic_inc_return for all archs, except x86 (but > not really a 'bug' because kernel does not need inc_and_return()) > > . Had to add some rcu stuff (what was this Bertl?) > > . Need to check task_state (in proc) regarding initpid virtualization > > . Virtual process uptime probably needs fixing > > . FIXADDR mapping is missing (what is this Bertl?) > > . Need to test cross-compile for other arches > (http://vserver.13thfloor.at/Stuff/Cross/howto.info) > > . Need to compile x25 in, as this is one of the typical bug > sources in recent vserver kernels > > > Additional Info: > > . The Debian kernel has a lot of modules configured by > default, needed to find a way to recompile the vserver > debian kernel with minimal recompile times: > > using the official debian image packages, you can nuke > the extra configs, and drop a lot of modules > out of the remaining configuration: > rm config/{386,k7,k7-smp,686-smp} > cp config/686 linux-2.6.11/.config, cd linux-2.6.11 > && make menuconfig; cp linix-2.6.11/.config > config/686 > also, use ccache when compiling, this will drastically increase > compile times > _______________________________________________ > Vserver mailing list > [email protected] > http://list.linux-vserver.org/mailman/listinfo/vserver _______________________________________________ Vserver mailing list [email protected] http://list.linux-vserver.org/mailman/listinfo/vserver
