On Monday 28 February 2005 21:54, Maarten wrote: [replying to myself, with a solution]
I solved this problem (after a lot of fruitless searching with google). What happened is that the guest kernel ran in SKAS mode, yet the _host_ kernel did not support that mode, since it wasn't patched for that, but it must be. Unless I'm blind, the Gentoo UML howto at http://www.gentoo.org/doc/en/uml.xml doesn't mention this fact anywhere, so maybe this fact needs some emphasis: When configuring the UML kernel you CANNOT enable SKAS ("Separate Kernel Address Space support") if you didn't also patch your main (host-) kernel for that. The kernel config help is misleading IMO; it states in so many words that it is "safe to say Yes", but from what I experienced it is definitely NOT safe to enable that option if you didn't patch your host kernel. This patch is, AFAIK, not included in the gentoo-sources, but it applies cleanly (http://www.user-mode-linux.org/~blaisorblade/) Just hoping this might help other people... Maarten > When I booted linux, it hung(*) because it couldn't find any modules, so I > went back and built and installed modules target, mounted the rootfs as > loopdevice and copied all of the fresh /lib/modules/x.y.z to /mnt/loop/lib/ > modules. Nothing is mentioned about modules in the howto so I just had to > second-guess it... > (*) I assumed it hung because on the host system I saw many many shells > spawned by the usermode system > > So now the modules are there, but the real problem, the endless spawning of > shells, still persists. Like so: > > [host system] ps ax > <snip usual processes> > 4890 pts/0 T+ 0:00 linux [/sbin/init] > 4896 pts/0 T+ 0:00 linux [/bin/bash] > 4904 pts/0 T+ 0:00 linux [/bin/bash] > 4914 pts/0 T+ 0:00 linux [/bin/bash] > 4926 pts/0 T+ 0:00 linux [/bin/bash] > 4932 pts/0 T+ 0:00 linux [/bin/bash] > > <snip couple of thousand lines...> > > 25661 pts/0 T+ 0:00 linux [/bin/bash] > 25665 pts/0 T+ 0:00 linux [/bin/bash] > 25669 pts/0 T+ 0:00 linux [/sbin/init] > 25671 pts/0 T+ 0:00 linux [/sbin/init] > 25673 pts/0 S+ 0:00 linux [/sbin/agetty] > 25676 pts/0 T+ 0:00 linux [/sbin/init] > 25678 pts/0 S+ 0:00 linux [/sbin/agetty] > > Meanwhile, this is what the -unresponsive at this point- uml bootscreen > shows: > > <snip> > * Bringing lo up... [ ok ] > * Initializing random number generator... [ ok ] > INIT: Entering runlevel: 3 > * Starting metalog... [ ok ] > * Mounting network filesystems... [ ok ] > * Starting vixie-cron... [ ok ] > * Starting local... [ ok ] > > And there it hangs, or seems to. > > Are all those shells normal ? I kind of doubt it. I've made sure to log > into the host system while using the -X flag to ssh, so I assume the xterms > should be able to open, but I have no way to check for that as the host > system has no X installed (nor do I want to install it). > > Am I nearly there, or is there something really wrong here ? > > TIA, > Maarten -- bash-2.05b$ emerge ncy -- gentoo-user@gentoo.org mailing list
