00:13 <somlo> daveshah: I like the idea of *not* having a pr_info -- it all depends on whether the log output generated by BUG_ON() is enough to make it clear to the driver developer that they called litex_[get|set]_reg with the wrong register size

00:14 <somlo> I'll run some tests where I screw this up on purpose to see what the BUG_ON log output looks like

14:20 <somlo> shorne: on second thought, I can always update the unreleased drivers in https://github.com/litex-hub/linux/tree/litex-rebase to *not* use [_]litex_[get|set]_reg() and replace all call sites with the appropriate litex_[read|write][X]() -- and not have to deal with BUG() at all. There's really no situation in which I can see an accessor determine the width of its access at *run time*

14:20 <somlo> * s/determine/need to determine/