[PATCH v1 2/4] arm64/signal: Include TPIDR2 in the signal context

[Szabolcs Nagy]

The 10/31/2022 15:34, Mark Brown wrote:
> On Mon, Oct 31, 2022 at 03:05:06PM +0000, Szabolcs Nagy wrote:
> 
> > this looks good from userspace point of view.
> 
> > originally the magic numbers made some sense, e.g.
> > esr used 0x45535201 which is ascii encoded "ESR\01",
> > the new ones seem random. probably does not matter.
> 
> Ah!  That looks like it was just the one I added for ZA that doesn't
> follow the convention, plus this and the ZT one that are on the list.
> The fact that there wasn't anything in the code mentioning this meant
> that when I was adding new values I had no idea that the numbers were
> anything other than random.  Nobody said anything in the many, many
> months the SME patches were on the list.  
> 
> We could start trying to follow that again and make this something like
> 0x54504902 (TPI\02 for TPIDR2)?  We should definitely at least add a
> comment about this if we're hoping to have people follow the convention.

i think this is not super important, but may help
when somebody is debugging a signal frame and does
not have to look up the magic value if it's self
documenting.

> 
> > Reviewed-by: Szabolcs Nagy <szabolcs.nagy at arm.com>
> 
> Thanks, if I just change the constant I'll preserve the tag if that's
> OK?

that's fine by me.