xen-devel
Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal)
> > Which OSes are you thinking of here? Linux currently supports (assuming
> > it works well enough for read access) the boot filesystems of everything
> > that runs on Xen, plus more.
>
> Presuming they all work well enough. Stuff like UFS is the sort of
> corner code that doesn't get much love, because so few people use it.
> Making boot rely on it not breaking seems worrying. To some degree, the
> same goes for FAT.
I think FAT'll be OK - it has to work reasonably well for people with digital
cameras, MP3 players, etc.
I agree UFS is a bit less likely to be well maintained, but I imagine if read
support works it's relatively unlikely to break... It's something that
probably would merit investigation before anyone tries to mandate a standard
Xen bootloader (if, indeed, that is a good idea in the first place).
> > It's a fair point that it increases effort for OS maintainers if they
> > have a grub driver for a filesystem that Linux doesn't already support -
> > I'm not clear what this could be though. Are you thinking about ZFS,
> > perhaps?
>
> Yes.
Really, with a suitable bootsequence we could probably have a much more
guest-based bootloader setup such that users could specify their own
bootloader using a "boot block", which could be quite a nice compromise.
The ZFS-FUSE stuff looks like it might help with this, since in userspace it's
able reuse the Solaris code without licensing worries. However, it's still
in the pipeline, without a guaranteed completion date (although it *ought* to
work by the SoC deadline, there's not a lot else to do other than cross
fingers here).
Until somebody produces a Xenified version of Grub it looks like the most
straightforward solution for most people is to use pygrub and friends, or try
kboot if they're feeling adventurous. Once a Grub/Xen version is available I
guess we can weight up the pros/cons of different approaches then.
I assume you're using Grub 1 - if it were based on Grub2 presumably the ZFS
driver would still need porting?
Cheers,
Mark
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
<Prev in Thread] |
Current Thread |
[Next in Thread>
|
- [Xen-devel] Re: Xen bootloader, (continued)
- [Xen-devel] Re: Xen bootloader, Ronald G Minnich
- [Xen-devel] Re: Xen bootloader, Ronald G Minnich
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), Michael Loehr
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), John Levon
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), Mark Williamson
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), John Levon
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), Mark Williamson
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), John Levon
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), Mark Williamson
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal), John Levon
- Re: Xen bootloader (was: Re: [Xen-devel] Xen Roadmap proposal),
Mark Williamson <=
- [Xen-devel] Re: Xen bootloader, Ronald G Minnich
- [Xen-devel] Re: Xen bootloader, Mark Williamson
- [Xen-devel] Re: Xen bootloader, Ronald G Minnich
- Re: [Xen-devel] Re: Xen bootloader, Richard Miller
- Re: [Xen-devel] Re: Xen bootloader, Mark Williamson
- Re: [Xen-devel] Re: Xen bootloader, Jacob Gorm Hansen
- [Xen-devel] Re: Xen bootloader, Gerd Hoffmann
Re: [Xen-devel] Xen Roadmap proposal, Harry Butterworth
|
|
|