On Wed, Mar 20, 2002 at 12:27:24AM +0100, Andreas W?st wrote: > Hello Chris > > To make it straight: it finally worked (well, not getting the 'Initialize..' > to work, but to get debugging output out of the system)!! > > > Only that I said I thought it had been fixed. Let me 'splain a little > > more about that. I was seeing dbootstrap restart at the point during > > kernel install when it mounted a partition, in December I think. There > > were several bugs at the time about it, IIRC. > > > > Anyway, I made a bunch of changes in the code that mounts and unmounts > > partitions, and cleaned that problem up. I think my initial assumption, > > that you were seeing the same problem, was wrong. > > > > I just had an inspiration. Someone put in debugging code to help when > > we get segfault problems. It won't show the registers for your arch, > > but it will at least give a backtrace which may prove helpful to > > someone around here. > > > > The way you activate it, is add debug to your boot arguments. It's > > supposed to give you a backtrace on stderr. Can you try that with > > the troublesome kernel? > > I now invoked the whole thing with > > bootstrap --apus -v -k apus/vmlinuz-2.4.17-apus -r apus/images-1.44/root.bin > root=/dev/ram video=pm2fb:mode:1024x768-75 debug > > It started up normally, well hola, I could see some more output now on > vc1 (which is quite interesting when compared to the dmesg output, as it > doesn't match.. the dmesg AmigaOS side output I've posted in my last mail > seemed to be a bit screwed up. Could it be that logging to mem doesn't worked > correctly?), then I went on in the installation until the now 1000 times > mentioned menu point, and tataa, it presented me with: > > *** Segmentation fault > Backtrace: > /sbin/dbootstrap [0x1000ac60] > [0x7fffe958] > /sbin/dbootstrap [0x10015520] > /sbin/dbootstrap [0x1001047c] > /sbin/dbootstrap [0x1000f378] > Debian GNU/Lin /lib/libc.so.6 (__libc_start_main+0x170) [0xfecb640] > *** Press RETURN to continue > > Well, I don't know if the output is represented correctly this way, as it was > quite screwed up on the display (the beginning of the lines were shifted > against each others). The "Debian GNU/Lin" part in the message seems to come > from the screen displayed before the debug output appeared. > > I hope I have copied all the numbers correctly. > > In parallel, on vc3 I could read (right before the segfault): > > Mar 19 ... (none) user.info dbootstrap [60]: formatting swap partition > /dev/hda2 > > then, the segfault output appeared, and, after pressing return, it > went onwards with: > > Mar 19 ... (none) user.info init: Process 'sbin/udbootstrap' (pid 57) > exited. Scheduling it for restart. > Mar 19 ... (none) user.info init: Starting pid 75, console /dev/console: > '/sbin/udbootstrap' > Mar 19 ... (none) user.info dbootstrap[78]: dbootstrap starting; built > 2002.03.15-16:02+0000 for powerpc > > Of course there has been some more output on vc3, but those lines I copied > here are right around the 'Initialize..' (and no other line was between the > lines listed here!). > > Well, now I hope you can do something with the runes printed here!! ;)
Cool, thanks Andi. Matt decided to take a look at the code, and I think he found the problem. He marked the bug pending. So, after the next build it should be better. -- *------v--------- Installing Debian GNU/Linux 3.0 --------v------* | <http://www.debian.org/releases/woody/installmanual> | | debian-imac (potato): <http://debian-imac.sourceforge.net> | | Chris Tillman [EMAIL PROTECTED] | | May the Source be with you | *----------------------------------------------------------------* -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
