Re: [PATCH for-8.0] target/arm: Fix generated code for cpreg reads when HSTR is active

[Richard Henderson <richard.henderson@linaro.org>]

On 3/28/23 11:27, Peter Maydell wrote:
> On Tue, 28 Mar 2023 at 18:27, Richard Henderson
> <richard.henderson@linaro.org> wrote:
>>
>> On 3/28/23 09:28, Peter Maydell wrote:
>>> +            /*
>>> +             * gen_exception_insn() will set is_jmp to DISAS_NORETURN,
>>> +             * but since we're conditionally branching over it, we want
>>> +             * to retain the existing value.
>>> +             */
>>> +            old_is_jmp = s->base.is_jmp;
>>>                gen_exception_insn(s, 0, EXCP_UDEF, syndrome);
>>> +            s->base.is_jmp = old_is_jmp;
>>
>> A third solution is to simply set is_jmp = DISAS_NEXT here.
> 
> I wasn't confident enough that the previous is_jmp had
> to be DISAS_NEXT to do that -- there are a lot of
> different values and it's not clear to me which are ones you
> might find lying around in is_jmp at the start of an insn.

At the very start of an insn, you will *only* find DISAS_NEXT.
Anything else will have exited the TB for the previous insn.


r~