On Fri, Feb 13, 2004 at 10:28:44PM -0500, Mark M. Hoffman wrote: > * Greg KH <greg at kroah.com> [2004-02-13 15:40:31 -0800]: > > On Tue, Feb 10, 2004 at 08:26:51PM +0100, Jean Delvare wrote: > > > Greg, is it correct to have such a high number of sysfs files for a > > > driver? > > > > That's fine to do. You might want to start using attribute lists to > > make it easier to register all of those files :) > > True enough, but that's not the part that bothers me. > > It'd be nice if the driver could provide a parameter to the > sysfs show() and store() routines... i.e. > > static ssize_t show_temp(struct device *dev, char *buf, int param); > > DEVICE_ATTR(temp1_input, S_IRUGO, show_temp, NULL, 1) > DEVICE_ATTR(temp2_input, S_IRUGO, show_temp, NULL, 2) > DEVICE_ATTR(temp3_input, S_IRUGO, show_temp, NULL, 3) > > ... and so on. Then we could end the blizzard of tiny helper > functions and the macro abuse. Try 'nm --size-sort w83781d.o' > some time for an illustration. Well we can always emulate something like the above with some macros :) In looking at these drivers they can easily be shrunk today with just a bit of rewriting of the functions and macros. I sure wasn't thinking about driver size when writing the first few ones, and it looks like everyone has copied my "style" in the rest of them. > But, I can't see a way to do this without forcing a massive > rewrite of all sysfs-enabled drivers - oh well. Yeah, that's not going to happen, sorry. thanks, greg k-h
