Rainer Dorsch wrote: > to insert the second floppy, but the laptop was basically dead at this point. > mvrn from #debian-boot helped out: I booted with BOOT_DEBUG=3 and inserted > the second boot floppy in the debugshell after hardware detection. Then it > worked out very smooth. > > > If I do not change the floppy in the debugshell, the problem becomes obvious: > > During the first > > mount /dev/floppy/o -o ro -tvfat floppy > > I get a > > Unable to handle kernel NULL ponter dereference at virtual address 000000ff > [..] > > Then after the 5 sleep seconds the installation hangs at the second call of > > mount /dev/floppy/o -o ro -tvfat floppy > > > Though it is somehow cool, if the user only needs the change the floppy > without doing anything, I recommand to ask the user to change floppy and hit > the enter button, this would be a very save workaround for this issue. Thank you for your report, it's very helpful and explains some similar reports I got about kernel hangs when loading the second floppy. I wish the kernel were a bit more robust in this area. :-( I really don't understand why mounting the first floppy, which is there and has a valid filesystem on it, should crash the kernel. I'll think about adding a prompt, it is of course doable (even for USB keyboards I think), but kinda annoying. -- see shy jo
Attachment:
signature.asc
Description: Digital signature