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

Re: [etch] driver updates for 2.6.18?



On Thu, Aug 23, 2007 at 07:53:49PM +0200, maximilian attems wrote:
> xen in etch is way beyond broken.
> i don't even start looking when the bug report mentions xen.
> the bts has *way* too much xen bugs.
> i doubt it is worth the effort to isolate them to some specific
> changesets.
> 
> please take the bts fact in consideration for the etch update xen decision.

Though I certainly empathize, accepting such a large patch is in
conflict with SRM policy.

My interpretation of this policy is that breaking existing users is
the worst possible outcome. If we accept this patch and enable 100 new
people to run Xen, but at the same time introduce some corner-case
regression that breaks 10 previously-stable users, then we've
failed. Our primary goal in a stable release is to keep existing happy
users happy. Enabling new users and fixing bugs for existing users are
good things, but secondary.

I want users to have no reason to think that rebooting their computer
after a point release is going to cause them any downtime - after all,
they really don't have the option of staying on their old/working bits
if they want to accept future security updates.

I think the only option to get such a large patch into etch at this
point would be w/ etch 1/2, though I've no clue if upstream Xen is at
all caught up w/ recent 2.6.

-- 
dann frazier



Reply to: