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

Re: initrd, lvm, and devfs

[Brian May]
> Whatever happened to the idea of even numbered kernels being
> "stable"?

You didn't get the memo?  That's an obsolete standard - the 2.6.x line
of development has been much more aggressive than past stable series,
as far as allowed tree changes, and last July or so (I think it was),
Andrew Morton announced that this policy would be purposely continued
as it seemed to be working well.

They've forfeited the responsibility to produce absolutely stable
kernels in favor of more aggressive development.  The kernel people now
say that the distribution trees are where final stabilization needs to
happen.  Which is not to say the kernel.org kernels are meant to be as
disruptive as, say, 2.3.7, but they are no longer attempting to
asymptotically approach 'perfectly stable', like 2.2.25.

Andrew Morton left open the possibility of opening up a 2.7 kernel tree
some time when some proposed tree changes were going to be *too*
disruptive (like real-SMP and dcache were in 2.1.x, I guess), but this
won't necessarily happen for awhile.

> I guess this means sarge won't work "out-of-the-box" with 2.6.11 and
> LVM unless you compile your own kernel (one that doesn't require an
> initrd image), or fix this initrd image.

If a 2.6.11 kernel makes it into sarge, then yes, someone'll have to
fix up initrd-tools for sarge as well.  Since the Debian kernel team
also maintains initrd-tools, I don't expect this issue to go unnoticed.


Attachment: signature.asc
Description: Digital signature

Reply to: