Re: [PATCH] intc/i8259: avoid (false positive) gcc warning

[Christian Borntraeger <borntraeger@de.ibm.com>]

On 18.03.21 17:03, Paolo Bonzini wrote:
> On 18/03/21 16:47, Christian Borntraeger wrote:
>> some copiler versions are smart enough to detect a potentially
>> uninitialized variable, but are not smart enough to detect that this
>> cannot happen due to the code flow:
>>
>> ../hw/intc/i8259.c: In function ‘pic_read_irq’:
>> ../hw/intc/i8259.c:203:13: error: ‘irq2’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
>>     203 |         irq = irq2 + 8;
>>         |         ~~~~^~~~~~~~~~
>>
>> Let us initialize irq2 to -1 to avoid this warning as the most simple
>> solution.
> 
> What about:

This also works, but if you want to go down that path then it would be good if you
could do this patch as I do not have the testing infrastructure to do proper x86
changes.
> 
> diff --git a/hw/intc/i8259.c b/hw/intc/i8259.c
> index 344fd04db1..bf28c179de 100644
> --- a/hw/intc/i8259.c
> +++ b/hw/intc/i8259.c
> @@ -189,20 +189,18 @@ int pic_read_irq(DeviceState *d)
>                   irq2 = 7;
>               }
>               intno = slave_pic->irq_base + irq2;
> +            irq = irq2 + 8;
> +            pic_intack(s, 2);
>           } else {
>               intno = s->irq_base + irq;
> +            pic_intack(s, irq);
>           }
> -        pic_intack(s, irq);
>       } else {
>           /* spurious IRQ on host controller */
>           irq = 7;
>           intno = s->irq_base + irq;
>       }
> 
> -    if (irq == 2) {
> -        irq = irq2 + 8;
> -    }
> -
>   #ifdef DEBUG_IRQ_LATENCY
>       printf("IRQ%d latency=%0.3fus\n",
>              irq,
> 
> 
> ?
> 
> Paolo
>