Re: [Qemu-devel] [RFC v2 01/11] tcg: move tb_invalidated_flag to CPUState

[Sergey Fedorov]

On 05/04/16 18:44, Paolo Bonzini wrote:
>
> On 05/04/2016 17:32, Alex Bennée wrote:
>> +    cpu->tb_invalidated_flag = false;
>>      tb = tb_gen_code(cpu, orig_tb->pc, orig_tb->cs_base, orig_tb->flags,
>>                       max_cycles | CF_NOCACHE
>>                           | (ignore_icount ? CF_IGNORE_ICOUNT : 0));
>> -    tb->orig_tb = tcg_ctx.tb_ctx.tb_invalidated_flag ? NULL : orig_tb;
>> +    tb->orig_tb = cpu->tb_invalidated_flag ? NULL : orig_tb;
>>      cpu->current_tb = tb;
> Based on discussion, it's safest to save the old value at the beginning
> of the hunk, and "OR" it into cpu->tb_invalidated_flag here.
>
>>      /* if no translated code available, then translate it now */
>> +    cpu->tb_invalidated_flag = false;
> Please remove this...
>
>>      tb = tb_gen_code(cpu, pc, cs_base, flags, 0);
>>  
>>  #ifdef CONFIG_USER_ONLY
>> @@ -509,12 +509,11 @@ int cpu_exec(CPUState *cpu)
>>                  tb = tb_find_fast(cpu);
>>                  /* Note: we do it here to avoid a gcc bug on Mac OS X when
>>                     doing it in tb_find_slow */
>> -                if (tcg_ctx.tb_ctx.tb_invalidated_flag) {
>> +                if (cpu->tb_invalidated_flag) {
>>                      /* as some TB could have been invalidated because
>> -                       of memory exceptions while generating the code, we
>> +                       of a tb_flush while generating the code, we
>>                         must recompute the hash index here */
>>                      next_tb = 0;
>> -                    tcg_ctx.tb_ctx.tb_invalidated_flag = 0;
> ... and leave the zeroing here.
>

Yes, I'm going to do something like this in a separate patch series or
just append it back to the "TCG misc clean-up" patch series when
re-spinning it.

Kind regards,
Sergey