[PATCH v4 1/2] arm64: mte: make the per-task SCTLR_EL1 field usable elsewhere

[Peter Collingbourne]

On Mon, Nov 30, 2020 at 4:51 AM Catalin Marinas <catalin.marinas at arm.com> wrote:
>
> Hi Peter,
>
> On Fri, Nov 27, 2020 at 07:13:43PM +0000, Catalin Marinas wrote:
> > On Thu, 19 Nov 2020 19:29:46 -0800, Peter Collingbourne wrote:
> > > In an upcoming change we are going to introduce per-task SCTLR_EL1
> > > bits for PAC. Move the existing per-task SCTLR_EL1 field out of the
> > > MTE-specific code so that we will be able to use it from both the
> > > PAC and MTE code paths and make the task switching code more efficient.
> >
> > Applied to arm64 (for-next/pac-keys-ctrl), thanks!
> >
> > [1/2] arm64: mte: make the per-task SCTLR_EL1 field usable elsewhere
> >       https://git.kernel.org/arm64/c/e710c29e0177
> > [2/2] arm64: Introduce prctl(PR_PAC_{SET,GET}_ENABLED_KEYS)
> >       https://git.kernel.org/arm64/c/284c64e9eba3
> >
> > If Dave has any other comments next week, we can apply them on top.
>
> I'm dropping these patches from the arm64 tree as they conflict
> (and not in a trivial way) with the MTE patches queued in the mm tree.
>
> So we either wait until 5.12 or we try to merge them via Andrew on top
> of the mm tree. Given the interaction with MTE, I'm slightly inclined
> towards the former.

Okay, I don't mind waiting until 5.12.

In the process of resolving the conflict I happened to notice a couple
of minor bugs (one in your original MTE patch series and one in this
patch). I sent a patch to fix the former and v5 of this series to
resolve the conflicts and fix the latter.

I also took the opportunity to do some perf measurements with the real
instruction sequences now that I have access to capable hardware, and
I included the results in the commit message.

Peter