On Thu, 2024-06-06 at 17:12 +0300, Andy Shevchenko wrote: > On Thu, Jun 06, 2024 at 02:27:03PM +0200, Nuno Sá wrote: > > On Thu, 2024-06-06 at 13:17 +0300, Andy Shevchenko wrote: > > > On Thu, Jun 06, 2024 at 09:22:38AM +0200, Nuno Sa wrote: > > ... > > > > > + return dev_err_ptr_probe(&st->spi->dev, - > > > > EINVAL, > > > > > > You can make all these lines shorter by using > > > > > > struct device *dev = &st->spi->dev; // or analogue > > > > > > at the top of the function. > > > > > > > Well, I had that in v2 (making the whole driver coherent with the local > > struct > > device helper but you kind of "complained" for a precursor patch (on a > > devm_kzalloc() call). So basically I deferred that change for a follow up > > patch. > > Hmm... I don't remember the story behind this, but probably it's good to have > this done one (precursor) or the other way (follow up). Just check how many > changes will be done, whichever diff is shorter, choose that one. > Well that has not much to do with the current series. I would prefer to have a follow up when we're done with the current changes. Right now I would really prefer to focus on the new dev_err_* APIs and see if anything else is needed for this to be acceptable. - Nuno Sá
