On Thu, Apr 14, 2016 at 04:08:43PM +0200, Enric Balletbo i Serra wrote: > Hi Emil, > > On 14/04/16 16:06, Emil Velikov wrote: > > Hi Enric, > > > > On 14 April 2016 at 14:42, Enric Balletbo i Serra > > <enric.balletbo@xxxxxxxxxxxxx> wrote: > > > The patch was implemented first without OR'ing error codes. The reason why I > > > changed this is because I received the comments that checking the error on > > > every regmap_* didn't help the readability of the driver and is likely to > > > not fail if the first call doesn't fail. > > > > > > For example, originally the code was like this: > > > http://pastebin.com/rPgyji8k > > > but I changed to this > > > http://pastebin.com/rPgyji8k > > > > > Both links are the same ;-) But I believe we all get what you meant. > > > > Ooops, sorry. This is the other link > > http://pastebin.com/e2KpGxHy I like the explicit checks better. Like I said, OR'ing the error codes makes you loose all context and likely give you many similar failures if for some reason regmap accesses always fail. Also you're code most likely is required to run in entirety for an operation to be successful. If the first part of the operation doesn't succeed it is natural to assume that the whole operation will fail, so there's no need (and a waste of time) to attempt any subsequent steps. Thierry
Attachment:
signature.asc
Description: PGP signature
_______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
