[PATCH v4] arm64: errata: Work around AmpereOne's erratum AC04_CPU_23

[Mark Brown]

On Fri, May 23, 2025 at 04:00:06PM +0100, Marc Zyngier wrote:
> On Fri, 23 May 2025 15:15:53 +0100,
> Mark Brown <broonie at kernel.org> wrote:
> > On Tue, May 13, 2025 at 11:45:14AM -0700, D Scott Phillips wrote:

> > > On AmpereOne AC04, updates to HCR_EL2 can rarely corrupt simultaneous
> > > translations for data addresses initiated by load/store instructions.

> > This patch, which is in -next as fed55f49fad181be9dfb93c0, breaks the
> > build of at least the vDSO selftests:

> > $ make -C tools/testing/selftests ARCH=arm64 LLVM=1 TARGETS=vDSO

> >   CC       vdso_test_chacha
> > In file included from vgetrandom-chacha.S:9:
> > In file included from ./../../../../arch/arm64/kernel/vdso/vgetrandom-chacha.S:5:
> > In file included from /home/broonie/git/bisect/tools/testing/selftests/../../../arch/arm64/include/asm/assembler.h:21:
> > In file included from /home/broonie/git/bisect/tools/testing/selftests/../../../arch/arm64/include/asm/cpufeature.h:13:
> > /home/broonie/git/bisect/tools/testing/selftests/../../../arch/arm64/include/asm/sysreg.h:1097:5: error: function-like macro 'IS_ENABLED' is not defined
> >  1097 | #if IS_ENABLED(CONFIG_AMPERE_ERRATUM_AC04_CPU_23)
> >       |     ^
> > 1 error generated.

> This:

...

> should solve it.

Yes, that works.

Tested-by: Mark Brown <broonie at kernel.org>

> But it also outlines that the vdso is getting built using stuff that
> is not meant for userspace code.

Indeed, adding Jason for that.  It's the same for all architectures.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20250523/c28d5f09/attachment.sig>