[PATCH 0/6] arm64: Disabling SVE/SME from the command-line

Mark Brown broonie at kernel.org
Mon Jun 27 09:49:44 PDT 2022


On Mon, Jun 27, 2022 at 04:30:36PM +0100, Marc Zyngier wrote:
> Mark Brown <broonie at kernel.org> wrote:

> > Oh dear, I've not seen any reports of this.  I'm guessing it's TF-A just
> > not enabling the requisite traps, coupled with people using prebuilt
> > kernels that turned off SVE when doing validation?

> Most probably something along these lines. The case I'm aware of
> involves booting a vanilla kernel on a system with a half baked
> firmware. I have little sympathy for these systems, but this looks
> like a reasonable ask to be able to disable a feature from the command
> line in order to pinpoint the issue.

It does, given the number of features that we've got in the pipeline
that require higher EL enablement I was actually wondering if we might
be better off going further and making this more of a standard idiom for
how we add any feature that requires setup at EL2/3 to be functional so
we don't need to go add overrides separately and most likely get them
backported or whatever.  There's the AMU which needs CPTR_EL[23].TAM for
example.

> Will someone ship a system with this baked into the command line? I
> don't believe it for a second (/me ducks).

I simply can't understand what you're saying here, must be something
inconceivable.
-------------- 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/20220627/64508423/attachment.sig>


More information about the linux-arm-kernel mailing list