On Fri, 10 Jun 2022 at 17:07, Peter Maydell <[email protected]> wrote:
>
> From: Richard Henderson <[email protected]>
>
> With ARMv8, this field is always RES0.
> With ARMv7, targeting EL2 and TA=0, it is always 0xA.

I was just looking at this change again because we still
have the loose end of syn_simd_access_trap() not being used,
and I realized that the claim in this commit message and the
comment isn't right. The "RES0 or fill in TA/copro fields"
test is not v8 vs v7, but "are we reporting this syndrome
to AArch64 in ESR_ELx or to AArch32 in HSR?".

I filed https://gitlab.com/qemu-project/qemu/-/issues/1153
to make a note of this since we might not get around to
fixing this for a while, given it's not very important.

thanks
-- PMM

Reply via email to