Re: [PATCH -v3 0/5] x86, cacheinfo, amd: L3 Cache Index Disable fixes
From: Ingo Molnar
Date: Sat Jan 23 2010 - 04:01:53 EST
* Borislav Petkov <petkovbb@xxxxxxxxxxxxxx> wrote:
> On Sat, Jan 23, 2010 at 07:59:53AM +0100, Ingo Molnar wrote:
> > * H. Peter Anvin <hpa@xxxxxxxxx> wrote:
> > > On 01/22/2010 09:40 AM, Borislav Petkov wrote:
> > > >>>
> > > >>> Those patches are also good -stable candidates.
> > > >>
> > > >> Hmmm... I'm not sure I see a strong justification for a late -rc push
> > > >> into Linus/stable push for for these... I think you would have to
> > > >> explicitly make the case if you want them to be considered as such.
> > > >
> > > > Well, on the one hand, they fix real bugs in the L3 cache index disable
> > > > code and since they're bugfixes, they are eligible late -rc candidates.
> > >
> > > Bugfixes are *early* -rc candidates. Regression fixes are *late* -rc
> > > candidates, at least that seems to be the policy Linus currently implements.
> > > -stable seems to use slightly less strict criteria (the whole point is that
> > > -final needs to be a stabilization point, backported fixes/drivers can then
> > > come onto a stable base) which is why you seem some patches which are
> > > "straight to .1".
> > Yes.
> Ok, thanks for the clarification - my only trouble was that the current
> code is b0rked as is and those fixes are needed. However, backporting
> them at a later point seems much more riskfree and I will do so later.
Well, if there's a crasher in there, then a minimal fix to address just that
is preferred for .33 - and that can be tagged for -stable immediately.
Anything more complex (these handful of patches) should go via the usual route
of .34-rc1 and then -stable if it's problem-free.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/