On Fri, 29 Oct 2021, Philippe Mathieu-Daudé wrote:
> On 10/28/21 21:27, BALATON Zoltan wrote:
>> Use g_new0 instead of g_malloc0 and avoid some unneeded temporary
>> variable assignments.
>>
>> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
>> ---
>>  hw/intc/sh_intc.c | 13 +++----------
>>  1 file changed, 3 insertions(+), 10 deletions(-)
>>
>> diff --git a/hw/intc/sh_intc.c b/hw/intc/sh_intc.c
>> index eb58707e83..ed0a5f87cc 100644
>> --- a/hw/intc/sh_intc.c
>> +++ b/hw/intc/sh_intc.c
>> @@ -400,21 +400,14 @@ int sh_intc_init(MemoryRegion *sysmem,
>>      /* Allocate 4 MemoryRegions per register (2 actions * 2 aliases) */
>>      desc->iomem_aliases = g_new0(MemoryRegion,
>>                                   (nr_mask_regs + nr_prio_regs) * 4);
>> -
>> -    j = 0;
>> -    i = sizeof(struct intc_source) * nr_sources;
>> -    desc->sources = g_malloc0(i);
>> -
>> +    desc->sources = g_new0(struct intc_source, nr_sources);
>
> g_new() is enough, since all get initialized in the next line.

Only their parent fields get init not the whole struct so I think g_new0 
is still needed.

>>      for (i = 0; i < desc->nr_sources; i++) {
>
> Even clearer as:
>
>       for (i = 0; i < nr_sources; i++) {

This may be a small improvement but not too much, desc->sources is 
assigned a few lines before. I consider this change but not sure about the 
g_new0.

Regards,
BALATON Zoltan

>> -        struct intc_source *source = &desc->sources[i];
>> -
>> -        source->parent = desc;
>> +        desc->sources[i].parent = desc;
>>      }
>> -
>>      desc->irqs = qemu_allocate_irqs(sh_intc_set_irq, desc, nr_sources);
>>      memory_region_init_io(&desc->iomem, NULL, &sh_intc_ops, desc, "intc",
>>                            0x100000000ULL);
>> -
>> +    j = 0;
>>      if (desc->mask_regs) {
>>          for (i = 0; i < desc->nr_mask_regs; i++) {
>>              struct intc_mask_reg *mr = &desc->mask_regs[i];
>>
>
>