In message E188n1J-0002qw-00@hall.mail.mindspring.net you wrote:
Any comments about this patch? Should I merge it?
Looks OK to me. This should work for all Xscale.
Thanks, Kyle.
On a side note Wolfgang. Have you ever considered supporting the
if (machine_is_board_x()) {}
I thought about it once, but not really seriously.
I know you're busy and this kind of thing is low priority. But it does make code, IMHO, more readable and cleaner.
Well, maybe. If you check the #ifdef-mess we have right now, you will find that
1) many #ifdef's come clustered for seral boards, and I don;t see much difference between
#if defined(CONFIG_CCM) || \ defined(CONFIG_COGENT) || \ defined(CONFIG_CPCI405) || \ defined(CONFIG_EVB64260) || \ defined(CONFIG_HYMOD) || \ defined(CONFIG_LWMON) || \ defined(CONFIG_PCU_E) || \ defined(CONFIG_W7O) || \ defined(CONFIG_MISC_INIT_R) /* miscellaneous platform dependent initialisations */ misc_init_r (); #endif
and
if (machine_is_CCM() || \ machine_is_COGENT() || \ machine_is_CPCI405() || \ machine_is_EVB64260() || \ machine_is_HYMOD() || \ machine_is_LWMON() || \ machine_is_PCU_E() || \ machine_is_W7O() || \ machine_is_MISC_INIT_R() ) { /* miscellaneous platform dependent initialisations */ misc_init_r (); }
2) Many of these #ifdef's are used to conditionammy define variables, or initializers, where an if() cannot be used, so we would have to support BOTH forms, which is actually worse.
But as I said: I didn't reallys pend much thought on this, so maybe I overlooked something. Feel free to come up with a more detailed proposal / patch.
Best regards,
Wolfgang Denk