Re: [PATCH 7/8] arm64: cpufeature: Relax checks for AArch32 support at EL[0-2]

[Suzuki K Poulose]

On 04/15/2020 01:29 PM, Will Deacon wrote:
> On Wed, Apr 15, 2020 at 12:37:31PM +0100, Suzuki K Poulose wrote:
> > On 04/15/2020 11:58 AM, Will Deacon wrote:
> > > On Wed, Apr 15, 2020 at 11:50:58AM +0100, Suzuki K Poulose wrote:
> > > > On 04/14/2020 10:31 PM, Will Deacon wrote:
> > > > > We don't need to be quite as strict about mismatched AArch32 support,
> > > > > which is good because the friendly hardware folks have been busy
> > > > > mismatching this to their hearts' content.
> > > > >
> > > > >      * We don't care about EL2 or EL3 (there are silly comments concerning
> > > > >        the latter, so remove those)
> > > > >
> > > > >      * EL1 support is gated by the ARM64_HAS_32BIT_EL1 capability and handled
> > > > >        gracefully when a mismatch occurs
> > > > >
> > > > >      * EL1 support is gated by the ARM64_HAS_32BIT_EL0 capability and handled
> > > >
> > > > s/EL1/EL0
> > > >
> > > > >        gracefully when a mismatch occurs
> > > > >
> > > > > Relax the AArch32 checks to FTR_NONSTRICT.
> > > >
> > > > Agreed. We should do something similar for the features exposed by the
> > > > ELF_HWCAP, of course in a separate series.
> > >
> > > Hmm, I didn't think we needed to touch the HWCAPs, as they're derived from
> > > the sanitised feature register values. What am I missing?
> >
> > sorry, that was cryptic. I was suggesting to relax the ftr fields to
> > NONSTRICT for the fields covered by ELF HWCAPs (and other CPU hwcaps).
>
> Ah, gotcha. Given that the HWCAPs usually describe EL0 features, I say we
> can punt this down the road until people give us hardware with mismatched
> AArch32 at EL0.

Btw, this is not just mismatched AArch32, but mismatched AArch64 HWCAPs
too, which I believe exists. Anyways as you said, we can delay this
until we get the reports :-)

Suzuki