Re: [PATCH v7 13/14] arm/cpu: switch to a generated cpu-sysregs.h.inc

[Cornelia Huck <cohuck@redhat.com>]

On Thu, Jun 12 2025, Peter Maydell <peter.maydell@linaro.org> wrote:

> On Thu, 15 May 2025 at 16:40, Cornelia Huck <cohuck@redhat.com> wrote:
>>
>> Generated against Linux 6.14-rc1.
>>
>> Reviewed-by: Sebastian Ott <sebott@redhat.com>
>> Reviewed-by: Eric Auger <eric.auger@redhat.com>
>> Signed-off-by: Cornelia Huck <cohuck@redhat.com>
>> ---
>
> Does it make sense to include all these extra registers?
> Some of them are not simple ID registers at all,
> for example ID_CCSIDR_EL1 is an array of registers, not
> a single value (in QEMU, stored in cpu->ccsidr[]),
> and CSSELR_EL1 is not an ID register at all, it's a
> normal read/write sysreg.

Oops, I think the generator needs to check for op1=={0,1,3} instead of
including op1==2...

[Another issue that came up in recent discussions for the cpu model
series is that recent kernels have made {MIDR,REVIDR,AIDR}_EL1 writable,
but even though they are in the range the kernel reports writable
registers in, the source sysreg file does not include them. I've changed
the script to add them manually so far, but not sure if that is the
best solution here.]