Messages in this thread | | | Subject | Re: [PATCH v3 0/9] lib/string: introduce match_string() helper | From | Andy Shevchenko <> | Date | Fri, 08 Jan 2016 16:13:18 +0200 |
| |
On Fri, 2016-01-08 at 13:19 +0000, Al Viro wrote: > On Fri, Jan 08, 2016 at 03:09:09PM +0200, Andy Shevchenko wrote: > > There are users of a simple string matching in the array. Let's do > > a common > > helper for that. > > What's the reason for making it return -ENODATA when no match is > found?
What else can be suitable?
> That one of the callers wants to return that as error in such case? > At least one other is returning -EINVAL in the same situation...
Linus Acked this, maybe he missed that one.
Linus, do we still need to return -EINVAL in pinmux?
In general our error reporting sucks, you know. So, any return value will be not ideal and self-explanatory (see json approach in perf). But I prefer return some return code instead of opaque -1, for example. This at least helps some users.
-- Andy Shevchenko <andriy.shevchenko@linux.intel.com> Intel Finland Oy
| |