Re: [PATCH v14 1/7] x86: lib: Separate instruction decoder MMIO type from MMIO trace

[Jason A. Donenfeld]

On Wed, Jan 4, 2023 at 9:26 PM Ingo Molnar <mingo@xxxxxxxxxx> wrote:
>
>
> * Ingo Molnar <mingo@xxxxxxxxxx> wrote:
>
> >
> > * Jason A. Donenfeld <Jason@xxxxxxxxx> wrote:
> >
> > > On Tue, Jan 3, 2023 at 6:29 PM Borislav Petkov <bp@xxxxxxxxx> wrote:
> > > >
> > > > On Tue, Jan 03, 2023 at 06:00:46PM +0100, Ingo Molnar wrote:
> > > > > > I guess you missed the conversation with Borislav yesterday about that.
> > > > > > He mentioned that I'd just take it through random.git when this whole
> > > > > > series goes in.
> > > > >
> > > > > Please base your tree off on tip:x86/asm then (or pull it in) - it only
> > > >
> > > > My idea was a lot simpler: avoid the tree inter-dependency by us acking this
> > > > patch so that it can go through the random.git tree.
> > >
> > > Indeed I would prefer this.
> > >
> > > Or... just put this in 6.2 because it's trivial anyway? Heck, even mark
> > > it as stable@ so make future backporting easier. Then it'll meet tip's
> > > urgent criteria.
> >
> > Yeah - that's sensible too, it does fix a header namespace bug - I've put
> > it into tip:x86/urgent.
>
> This namespace clash fix is now upstream as of 512dee0c00ad & later kernels.

Thanks. I've rebased my vdso branch on that now. I guess at this point
it probably doesn't matter that much since everyone hates my use of
the instruction decoder anyway, so I'll see what else I can do for
v15, but still, it'll at least make it easier to experiment.

Jason