On 10.08.2010 09:16, Nikolai Kondrashov wrote: > On 08/10/2010 03:03 AM, Robert Schedel wrote: >> Would not it make sense to declare the returned buffer from report_fixup >> const? With the desired new definition (no buffer copy/return private >> reference) you should never want anyone to change the returned buffer >> content afterwards. Of course, const would propagate a bit into hid-core.c >> (hid_parse_report, fetch_item). > Sorry for the late response, I managed to overlook your message somehow. My email was late. > Yes, I thought about it too. However, as this is my first patch, I'm still > unsure about const policy in the kernel (some people don't like it) and I > wanted to make the patch footprint as little as possible to minimize the > errors and thus rejection possibility. It is not only const-cleanness. Those static descriptor arrays in your new driver (approx. 200 bytes, I guess) are currently also non-const, due to this. They should probably also have a chance to be "rodata". > I could make an additional patch, I think, since the original one is already > applied. Jiri should decide what he prefers as I am not aware of const guidelines either -- only wanted to point it out. Regards, Robert Schedel -- 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
