On 4/18/21 9:08 PM, Richard Henderson wrote:
> On 4/18/21 9:31 AM, Philippe Mathieu-Daudé wrote:
>> +static inline hwaddr do_translate_address(CPUMIPSState *env,
>> + target_ulong address,
>> + MMUAccessType access_type,
>> + uintptr_t retaddr)
>> +{
>> + hwaddr paddr;
>> + CPUState *cs = env_cpu(env);
>> +
>> + paddr = cpu_mips_translate_address(env, address, access_type);
>> +
>> + if (paddr == -1LL) {
>> + cpu_loop_exit_restore(cs, retaddr);
>> + } else {
>> + return paddr;
>> + }
>> +}
>
> Code motion, so,
> Reviewed-by: Richard Henderson <[email protected]>
>
> but I think that it's a mistake for cpu_mips_translate_address to split
> the raise_mmu_exception from the cpu_loop_exit_restore. If you make the
> changes to env to indicate the exception, you *must* then go back to the
> main loop.
>
> There seems to be exactly one caller, this one, so it should be trivial
> to change, which the lets do_translate_address vanish entirely.
I'm taking note of this comment and will try to address it later.
