On Wed, 2010-12-08 at 01:12 +0900, Namhyung Kim wrote: > 2010-12-07 (í), 09:57 -0600, James Bottomley: > > Um so this is definitely used in the MODULE_DEVICE_TABLE just below. > > > > I deduce you have some type of config which causes that to be a nop > > (say !CONFIG_MODULE)? In which case, surely the correct fix is in > > module.h? > > > > Hi, > > MODULE_DEVICE_TABLE expands to nop when !MODULE (not !CONFIG_MODULE) so > allyesconfig build always emits the warnings. Possible patch for this > would be wraping it up in #ifdefs but when I sent a same patch to Greg > KH he didn't want to add more #ifdefs in .c files. I didn't mean changing any c files. I meant fixing the module include to make the table used even in the !MODULE case. That way we only alter one file instead of a significant proportion of the drivers. James --- diff --git a/include/linux/module.h b/include/linux/module.h index 7575bbb..d49aa9c 100644 --- a/include/linux/module.h +++ b/include/linux/module.h @@ -90,7 +90,10 @@ extern const struct gtype##_id __mod_##gtype##_table \ extern struct module __this_module; #define THIS_MODULE (&__this_module) #else /* !MODULE */ -#define MODULE_GENERIC_TABLE(gtype,name) +/* Flag name as used to avoid warnings about the unused device id table */ +#define MODULE_GENERIC_TABLE(gtype,name) \ +static const struct gtype##_id __mod_##gtype##_table \ + __attribute__ ((unused, alias(__stringify(name)))) #define THIS_MODULE ((struct module *)0) #endif -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html