[PATCH] Staging: unisys: virtpci: fixed a brace coding style issue
Julia Lawall
julia.lawall at lip6.fr
Wed Jul 1 09:35:46 UTC 2015
On Wed, 1 Jul 2015, Sohny Thomas wrote:
> >>>> i = virtpci_device_del(NULL /*no parent bus */, VIRTHBA_TYPE,
> >>>> &scsi.wwnn, NULL);
> >>>> - if (i) {
> >>>> + if (i)
> >>>> return 1;
> >>>> - }
> >>>> - return 0;
> >>>> + else
> >>>> + return 0;
> >>> No, now this will introduce a new checkpatch warning that "else is not
> >>> required after return". why did you introduce this "else"?
> >> I did this so that the code is more readable and understandable, I checked and
> >> checkpatch didn't call this out , so its clean.
> >>
> >> Otherwise the above code looks like this
> >>
> >> if(i)
> >> return 1;
> >> return 0;
> >
> > That looks fine.
> >
> > I haven't looked at the code in detail. Is it normal that the return
> > values seem to be 0 1 and -1? Which values represent success and which
> > represent an error? It is nicer to have the errors under if and success
> > as a direct return at the end.
> Here in this driver directory, return 1 means SUCCESS and return 0 means FAILURE
What is -1?
> So you mean my code change is fine?
I think it would be best to have the success case that is not under an if.
So if (!i)
return 0;
return 1;
I guess some day the driver would need more normal error codes?
julia
More information about the devel
mailing list