On Thu, Jul 20, 2023 at 02:42:49PM +0200, Oliver Neukum wrote: > On 20.07.23 14:01, Johan Hovold wrote: > > On Thu, Jul 20, 2023 at 01:44:06PM +0200, Oliver Neukum wrote: > > >> + * These tables are NOT in order of device id by intention > >> + * > >> + * Keep them and add new entries sorted by > >> + * > >> + * 1. Alphabetical order of the vendor name > >> + * 2. Alphabetical order of the product name > >> + * > >> + */ > > > > No, this is not correct. The tables are sorted alphabetically based on > > the symbol names, but the entries in each table is sorted by VID/PID as > > usual. > > OK, I stand corrected. You do understand that this ordering is > extremely peculiar? I am sorry, but deducing that from the example, > which the current order is, will just not work. Nobody gets that > without a clear and present explanation. It's not peculiar. And if this wasn't implemented using macros, I'm pretty sure you would not think so. Take a look at io_edgeport.c, for example, where edgeport_2port_device comes before edgeport_4port_device, etc. > This just needs a comment in the text. I disagree, with the last few entries sorted this should be clear enough. Johan
