[Top][All Lists]

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

Re: grub-probe detects ext4 wronly as ext2

From: Javier Martín
Subject: Re: grub-probe detects ext4 wronly as ext2
Date: Fri, 04 Jul 2008 16:33:50 +0200

El vie, 04-07-2008 a las 16:09 +0200, Robert Millan escribió:
> On Fri, Jul 04, 2008 at 02:00:34PM +0200, Javier Martín wrote:
> > In the nearly eight years from ext2 release to the merge of ext3 into
> > the kernel, some incompatible features were introduced in ext2 proper,
> > like filetype and compression. Even now, with ext4 in development and
> > merged into the kernel less than five years after ext3, incompatible
> > features are being backported to ext2 like the lazy inode initializing I
> > mentioned earlier. Those can hit us HARD.
> So you're saying we should make our decisions based on the assumption that
> Linux developers could be jeopardizing our efforts by introducing incompatible
> features by surprise, without any coordination with us, and that distributors
> are going to pass those features down to stable releases, knowing that they
> break GRUB, and again without any coordination with us?
I'm not saying that they will _consciously_ try to break us, just that
new flags can be introduced "just like that" and that, while ext4 is
definitely making quite a fuss, the real danger is in flags like
EXT2_FEATURE_INCOMPAT_META_BG, the lazy inode writing flag I mentioned
to Bean, since a "best-effort" reader that ignores it can (with a very
high probability) read old metadata, since if the same partition is
re-formatted with ext2 the inodes will be on the same places. Thus,
files from the old FS could "resurrect" (with their data in unknown
state) and other nasty things.

These kind of flags don't make a lot of noise in developer circles
outside LKML, they are introduced without long times in "development"
state and then our driver becomes "broken" in a way that will be very
difficult to bugtrace. That's why I said that the bootloader should
adopt a conservative stance WRT incompatible feature flags too, but the
user override should be present in order to give informed users a choice
without having to rebuild GRUB from source.

Attachment: signature.asc
Description: Esta parte del mensaje está firmada digitalmente

reply via email to

[Prev in Thread] Current Thread [Next in Thread]