On Fri, May 01, 2015 at 12:41:59PM -0500, Timur Tabi wrote:
On 05/01/2015 12:32 PM, Guenter Roeck wrote:
devm_ioremap_resource already prints a message. For this reason, elsewhere in the kernel the check for !res before calling devm_ioremap_resource is being removed, leaving the error handling to devm_ioremap_resource. I would suggest to do the same here.
I see what you're saying, but I leave the error handling to devm_ioremap_resource(), then no one will know whether it's because the control address or the refresh address is missing. The user will just see "invalid resource" and won't what resource is actually invalid.
You are saying that pretty much everyone in the kernel is, in your opinion, doing the Wrong Thing (tm) and you insist in doing it differently.
As maintainer, I have seen lots of patches which remove this very same error checking as unnecessary. If we accept your code, we can be all but sure to see such a patch at some point, probably right after your patch was accepted and shows up in linux-next. So besides arguing about something we should not have to argue about in the first place, you are trying to create even more maintainer work going forward.
Is that really neceessary ?
And then people wonder why maintainers sometimes get grumpy :-(.
Guenter