Re: [PATCH 08/12] printk: Replace strncmp with str_has_prefix

[Joe Perches]

On Thu, 2019-08-01 at 23:23 +0800, Chuhong Yuan wrote:
> Joe Perches <joe@xxxxxxxxxxx> ä2019å7æ30æåä äå8:16åéï
> > On Mon, 2019-07-29 at 23:15 +0800, Chuhong Yuan wrote:
> > > strncmp(str, const, len) is error-prone.
> > > We had better use newly introduced
> > > str_has_prefix() instead of it.
> > []
> > > diff --git a/kernel/printk/braille.c b/kernel/printk/braille.c
> > []
> > > @@ -11,10 +11,10 @@
> > > 
> > >  int _braille_console_setup(char **str, char **brl_options)
> > >  {
> > > -     if (!strncmp(*str, "brl,", 4)) {
> > > +     if (str_has_prefix(*str, "brl,")) {
> > >               *brl_options = "";
> > >               *str += 4;
> > > -     } else if (!strncmp(*str, "brl=", 4)) {
> > > +     } else if (str_has_prefix(*str, "brl=")) {
> > >               *brl_options = *str + 4;
> > 
> > Better to get rid of the += 4 uses too by storing the result
> > of str_has_prefix and using that as the addend.
> > 
> > Perhaps
> >         size_t len;
> > 
> >         if ((len = str_has_prefix(*str, "brl,"))) {
> >                 *brl_options = "";
> >                 *str += len;
> >         } else if ((len = str_has_prefix(*str, "brl="))) {
> >                 etc...
> > 
> 
> I find that checkpatch.pl forbids assignment in if condition.
> So this seems to be infeasible...

checkpatch is a stupid script and doesn't forbid
anything.  It's just a suggestion guide.

Ignore checkpatch when you know better.