Re: [RFC PATCH v2 8/9] sparc/sun4m: Use one cpu_reset() function for main and secondary CPUs

[Thiago Jung Bauermann <bauerman@linux.ibm.com>]

Philippe Mathieu-Daudé <philmd@redhat.com> writes:

> On 7/22/20 5:50 AM, Thiago Jung Bauermann wrote:
>> If we rely on cpu_common_reset() setting CPUState::halted according to the
>> start-powered-off property, both reset functions become equivalent and we
>> can use only one.
>>
>> Signed-off-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
>> ---
>>  hw/sparc/sun4m.c | 21 ++++-----------------
>>  1 file changed, 4 insertions(+), 17 deletions(-)
>>
>> NB: I was only able to test that this patch builds. I wasn't able to
>> run it.
>>
>> diff --git a/hw/sparc/sun4m.c b/hw/sparc/sun4m.c
>> index 7b3042a801..deb5e9f027 100644
>> --- a/hw/sparc/sun4m.c
>> +++ b/hw/sparc/sun4m.c
>> @@ -218,16 +218,7 @@ static void dummy_cpu_set_irq(void *opaque, int irq, int level)
>>  {
>>  }
>>
>> -static void main_cpu_reset(void *opaque)
>> -{
>> -    SPARCCPU *cpu = opaque;
>> -    CPUState *cs = CPU(cpu);
>> -
>> -    cpu_reset(cs);
>> -    cs->halted = 0;
>> -}
>> -
>> -static void secondary_cpu_reset(void *opaque)
>> +static void sun4m_cpu_reset(void *opaque)
>>  {
>>      SPARCCPU *cpu = opaque;
>>      CPUState *cs = CPU(cpu);
>> @@ -818,7 +809,6 @@ static const TypeInfo ram_info = {
>>  static void cpu_devinit(const char *cpu_type, unsigned int id,
>>                          uint64_t prom_addr, qemu_irq **cpu_irqs)
>>  {
>> -    CPUState *cs;
>>      SPARCCPU *cpu;
>>      CPUSPARCState *env;
>>
>> @@ -826,12 +816,9 @@ static void cpu_devinit(const char *cpu_type, unsigned int id,
>>      env = &cpu->env;
>>
>>      cpu_sparc_set_id(env, id);
>> -    if (id == 0) {
>> -        qemu_register_reset(main_cpu_reset, cpu);
>
> IMO it is easier to review this patch in 2, first drop main_cpu_reset
> as it is pointless (we rely on cpu_common_reset), then set the
> "start-powered-off" property and drop secondary_cpu_reset().

That's a good idea. I made those patches for v3.

>> -    } else {
>> -        qemu_register_reset(secondary_cpu_reset, cpu);
>> -        cs = CPU(cpu);
>> -        object_property_set_bool(OBJECT(cs), "start-powered-off", true,
>> +    qemu_register_reset(sun4m_cpu_reset, cpu);
>
> Why do you still keep it?

I didn't know that not registering a reset function would cause
cpu_reset() to be caused.

>> +    if (id != 0) {
>> +        object_property_set_bool(OBJECT(cpu), "start-powered-off", true,
>>                                   &error_abort);
>
> At this point the CPU is realized, so this is correct.

Great. Thanks for confirming!

> I'd use directly:
>
>        object_property_set_bool(OBJECT(cpu), "start-powered-off", !!id,
>                                 &error_abort);

I used a slight variation of your suggestion, with `id != 0` instead of
`!!id` because I think it makes the code easier to read.

--
Thiago Jung Bauermann
IBM Linux Technology Center