Messages in this thread | | | Date | Thu, 26 Jul 2012 16:49:21 +0530 | Subject | Re: [RFC PATCH] drivers: regulator: add error checks for exported APIs | From | "ABRAHAM, KISHON VIJAY" <> |
| |
Hi,
On Thu, Jul 26, 2012 at 4:40 PM, Mark Brown <broonie@opensource.wolfsonmicro.com> wrote: > On Thu, Jul 26, 2012 at 04:00:03PM +0530, Kishon Vijay Abraham I wrote: > >> Added error checks for exported APIs in regulator framework so that >> incorrect use of these APIs does not result in a crash. > > The commit (especially the subject) should be more descriptive, it's not > that we're missing error checks it's that you're handling garbage passed > in from the user a bit more gracefully. > > I'm not sure how useful this is as-is - if you're not checking your > errors you're not checking your errors so it's probably just pushing the > robustness failures around, especially since... > >> + if (IS_ERR_OR_NULL(regulator)) >> + return -ENODEV; > > ...this is too quiet, this should at the very least be generating a loud > warning so people can see there's a problem (probably with WARN_ON so we > can see what called it). Can you respin with something noisier please?
Sure. I'll respin a new version. > > Also NULL is a valid regulator (though we don't use it right now). I see a lot of dereferencing *regulator*. How is it supposed to work when *regulator* is NULL?
Thanks Kishon
| |