[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

GNU/Hurd in foreign filesystems?



Hello,

one little annoying aspect in the current version of the Hurd (and gnumach)
is, that it requires a partition of its own with an ext2fs filesystem.

It would be nice to have the additional option to install the Hurd
filesystem inside an existing "foreign" filesystem of another OS.

There are two ways to do it:

1. Use a regular file on the foreign filesystem and mke2fs on it, then
   boot directly off this file-based filesystem. The big file would
   contain an (ext2fs) image of the hurd filesystem.

2. Use a subdirectory of the foreign filesystem and put all Hurd files
   there. One problem here is that most likely, that foreign filesystem
   won't have the extra space for metadata like translator bits etc...
   A workaround would be to add a file (either per directory or for all
   files) that would contain such metadata (somehow like Linux' umsdosfs).

The forein filesystem should not necessarily be ext2fs. Support for
msdosfs, ext2fs, ufs, ffs (and of course *BSD disklabel) and maybe also
ntfs should be included in the bootloaders (e.g. GRUB) and in the booted
gnumach kernel/modules for this to work.

There is of course a performance penalty associated with such an indirect
approach, but for developers and first-time users, it may be worth it.

It may be easier to start implementing a user-level filesystem server
inside the Hurd and have this filesystem server access a regular file
(on the Hurd filesystem as well as on misc. foreign filesystems) in the
first place. If and when this works, GRUB could be extended to support/load
this filesystem-server at boot time.

What do you think about this?

[don't ask me for code, it's just an idea...]

-Farid.

-- 
Farid Hajji -- Unix Systems and Network Administrator | Phone: +49-2131-67-555
Broicherdorfstr. 83, D-41564 Kaarst, Germany          | farid.hajji@ob.kamp.net
- - - - - - - - - - - - - - - - - - - - - - - - - - - + - - - - - - - - - - - -
Fermat: ...I've found a remarkable proof for this: Let x,y @#$!@$!2@ NO CARRIER



Reply to: