Messages in this thread | | | Date | Thu, 19 Feb 2015 11:33:44 +0100 | From | Ingo Molnar <> | Subject | Re: [PATCH] x86/intel/quark: fix simple_return.cocci warnings |
| |
* Bryan O'Donoghue <pure.logic@nexus-software.ie> wrote:
> On 19/02/15 10:25, Ingo Molnar wrote: > > >>>- ret = iosf_mbi_read(QRK_MBI_UNIT_MM, QRK_MBI_MM_READ, > >>>+ return iosf_mbi_read(QRK_MBI_UNIT_MM, QRK_MBI_MM_READ, > >>> reg++, &imr->wmask); > >>>- if (ret) > >>>- return ret; > >>>- > >>>- return 0; > >>> } > >>> > >>> /** > >>> > >> > >>This flow was a change asked for and supplied in review > >>feedback for Andy Shevchenko so NAK to this patch. > > > >But this pattern: > > > > if (ret) > > return ret; > > > > return 0; > > > >makes very little sense. Why is it done? > > > >Thanks, > > > > Ingo > > Feedback at review was that it's more consistent with the > code that comes before.
But that feedback makes very little sense. In C we don't ever want to write:
if (ret) return ret;
return 0;
Because we can return the fine value straight away:
return ret;
regardless of what comes before.
Thanks,
Ingo
| |