Re: [PATCH 2/8] pnv/psi: Clean up local variable shadowing

[Markus Armbruster <armbru@redhat.com>]

Cédric Le Goater <clg@kaod.org> writes:

> On 9/19/23 08:57, Harsh Prateek Bora wrote:
>> On 9/18/23 20:28, Cédric Le Goater wrote:
>>> to fix :
>>>
>>>    ../hw/ppc/pnv_psi.c: In function ‘pnv_psi_p9_mmio_write’:
>>>    ../hw/ppc/pnv_psi.c:741:24: warning: declaration of ‘addr’ shadows a parameter [-Wshadow=compatible-local]
>>>      741 |                 hwaddr addr = val & ~(PSIHB9_ESB_CI_VALID | PSIHB10_ESB_CI_64K);
>>>          |                        ^~~~
>>>    ../hw/ppc/pnv_psi.c:702:56: note: shadowed declaration is here
>>>      702 | static void pnv_psi_p9_mmio_write(void *opaque, hwaddr addr,
>>>          |                                                 ~~~~~~~^~~~
>>>
>>> Signed-off-by: Cédric Le Goater <clg@kaod.org>
>>> ---
>>>   hw/ppc/pnv_psi.c | 5 +++--
>>>   1 file changed, 3 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/hw/ppc/pnv_psi.c b/hw/ppc/pnv_psi.c
>>> index daaa2f0575fd..26460d210deb 100644
>>> --- a/hw/ppc/pnv_psi.c
>>> +++ b/hw/ppc/pnv_psi.c
>>> @@ -738,8 +738,9 @@ static void pnv_psi_p9_mmio_write(void *opaque, hwaddr addr,
>>>               }
>>>           } else {
>>>               if (!(psi->regs[reg] & PSIHB9_ESB_CI_VALID)) {
>>> -                hwaddr addr = val & ~(PSIHB9_ESB_CI_VALID | PSIHB10_ESB_CI_64K);
>>> -                memory_region_add_subregion(sysmem, addr,
>>> +                hwaddr esb_addr =
>>
>> While at it, we may want to move the declaration to the beginning of the function. 
>
> I am more in favor of declaring the variables where they are needed.
> I think it is better pratice since it identifies a functional block
> which could be move in a external routine at some point if it becomes
> too complex.

I'm old-fashioned and prefer my declarations in one place, where I can
find them easily, except perhaps for for-loop counters.  I suspect when
a function is big enough so that moving declarations to inner blocks
improves it, factoring out these inner blocks would improve it more.

My other reason is the risk for accidental shadowing, but us enabling
-Wshadow=local will render that point moot.

Maintainers get to decide such matters of taste, and I'm not the
maintainer here :)

>> Anyways,
>> Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
>
>
> Thanks,
>
> C.
>
>
>> 
>>> +                    val & ~(PSIHB9_ESB_CI_VALID | PSIHB10_ESB_CI_64K);
>>> +                memory_region_add_subregion(sysmem, esb_addr,
>>>                                               &psi9->source.esb_mmio);
>>>               }
>>>           }