On Tue, 15 Jan 2002, Erik Mouw wrote: > On Tue, Jan 15, 2002 at 01:54:29PM -0500, Nicolas Pitre wrote: > > On Tue, 15 Jan 2002, Erik Mouw wrote: > > > The idea that the drivers usually need some small mods was actually one > > > of the main reasons to define CONFIG_ISA for ARM. > > > > Bogus argument. CONFIG_ISA isn't strictly needed, and the presence of > > it only move the cluttering inside the drivers themselves. > > Lots of devices function on embedded stuff without even the slightest > hint of an ISA bus.
ISA isn't just a wiring standard. It's also lists of typical IO address to probe for the existence of some particular hardware. That stuff is of course part of the .c driver file. On the SA11x0 all those probes must of course be configured out and surrounding them with #ifdef CONFIG_ISA instead of #if !defined(CONFIG_SA1100_X) && !defined(CONFIG_FOO) && !defined(CONFIG_BAR) seems much more tasteful to me. > > Enabling CONFIG_ISA will just allow the false selection of many > > more drivers that shouldn't be available. This is worse than few scattered > > cluttering of Config.in files IMHO, until "Aunt Tilly" and "Penelope" make > > that clutter be handled gracefully. > > Except that cluttering up Config.in files wasn't accepted as a patch > and cluttering up C files was. By who? AFAIK both are being rejected atm. And you think I'll take that as a good answer anyway? ;-) > Note that the cluttering up of C files > is needed independent of the the configuration issue. See above. Nicolas _______________________________________________ http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm Please visit the above address for information on this list.
