Re: [PATCH v2 1/2] arm64/sme: Restore SMCR on exit from suspend

[Mark Brown]

On Fri, Feb 09, 2024 at 04:46:48PM +0000, Will Deacon wrote:

> Looking at the other places where we touch SMCR_EL1, it looks like we
> always use a read-modify-write sequence. However, doesn't that mean we
> inherit a bunch of unknown bits on cold boot? I'm basically wondering
> whether we should be initialising these registers to a well-known value
> earlier in the CPU init path, a bit like we do for the EL2 variants.

That'd be safer from a future proofing point of view, yes - we've got a
similar pattern with ZCR as well from the looks of it.  I'll send a
separate series.

Attachment: signature.asc
Description: PGP signature