Hi Geert, Am 11.04.2018 um 18:51 schrieb Geert Uytterhoeven: > Hi Michael, > > On Tue, Apr 10, 2018 at 11:50 PM, Michael Schmitz <schmitzmic@xxxxxxxxx> wrote: >>>> Short of a complete rewrite of the Zorro driver support code to be >>>> closer to what PCI does, I don' see what can be done about the use of >>>> Zorro IDs. I don't think such a rewrite is planned in the near future, >>>> Geert? >>> >>> I think what Christoph means is the use of the define >>> ZORRO_PROD_PHASE5_BLIZZARD_1230_II_FASTLANE_Z3_CYBERSCSI_CYBERSTORM060 >>> versus hardcoded numbers, or ZORRO_ID(PHASE5, 0x0B, 0). >>> >>> We have a long list of ZORRO_PROD_* definitions in >>> include/uapi/linux/zorro_ids.h because of historical reasons. The list >>> isn't really changing (no new IDs in git history) due to almost no new >>> Zorro boards being made, unlike for PCI, where keeping an in-kernel list >>> is a lot of work, and not desirable. >> >> now I see, thanks. >> >> I could change the device table to use ZORRO_ID(PHASE5, ...) style IDs >> instead of the longish defines if you're OK with that. > > I don't have a preference. If you think it makes the driver easier to read, > go for it. It does get rid of a few 'line over 80 characters' warnings from checkpatch, and with comments in the device table describing what the IDs mean, it isn't any less readable. Let's see ... > Note that we can't get rid of the longish defines anyway, as they're in a > uapi header file, so you're free to keep on using it. I don't mind these defines remaining in the header file at all - the only place they are now used is in the device table, and hardcoded IDs are as good as the defines there, I think. Have a look at the result in v8 - I can easily revert to the version with the defines if that's the consensus, it's all in git after all. Cheers, Michael
