On Tue, 29 Mar 2011, Randy Dunlap wrote: > In 2.6.39-rc1 on x86_64, I get 92 total section mismatch warnings. > usb/gadget accounts for 75 of those. > > Does usb/gadget code have any special considerations for hotplug-ability > that is different from the rest of the kernel source tree? As far as I know, there are no special requirements. The mismatches probably arose for historical reasons. People working on gagdget drivers tend to be oriented toward embedded systems, where minimizing memory consumption is a prime concern. Therefore they try hard to make sure that code and data can be unloaded when no longer needed. But when the first gadget drivers were being written, the rules for doing this correctly weren't clearly understood and mistakes were made. Later drivers copied those mistakes, and nobody cared about the mismatch warnings badly enough to fix them. In theory, fixing them should be a straightforward, almost mechanical job. Provided one can track down the exact causes, of course. There used to be a Kbuild option for diagnosing section mismatches, but I can't find it now -- and I don't remember how much detail it provided. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
