[OLPC-devel] Re: LinuxBIOS and integrated system debug status (Friday night).

[Ronald G Minnich]

Jim Gettys wrote:

> So, are we concluding that what I saw (that I could not mount the file
> system after a improper shutdown) is a USB problem and not a LinuxBIOS
> problem?  It sounds like we should from the information as I understand
> it.

it's more interesting than that. It's a 'what do we do' problem.

If the file system was not shut down cleanly, then the ext2 in the 
BIOS-based linux will not happily mount it. But, Grub will happily go 
after it and load files just fine. Grub is not picky. It knows it's 
always going to be read-only. So, the question for me is, should an ext2 
file system in a  linux embedded in a FLASH BIOS be as strict as a 
normal linux, or as free and easy as  grub?

If it is as strict as linux, then I think in the general case we may 
need fsck in busybox. If it is free and easy as grub, then we have to 
beat on the kernel ext2 code to tolerate mounting (readonly!) and using 
a non-clean partition.

if it is strict, then the BIOS-based linux can use the USB flash in the 
way any linux can: it can write to it. If it is loose, it can never 
mount the partition writeable, but it can boot an unclean partition.

There are no good answers here; just a decision we have to make.

ron