On Wednesday 24 January 2007 18:19, Luck, Tony wrote: > > Some of the IA64 code assumes that ACPI is optional, > > and some assumes that it is not. > > The situation now is that the simulator doesn't want ACPI, and everything > else needs it (even "allnoconfig" ends up with CONFIG_ACPI=y!). So the > solution that you suggest near the end of your mail of making CONFIG_ACPI=y > required, and somehow cleverly deal with the fallout in sim_defconfig would > be the neatest (if we were clever enough to do this without an explosion > of #ifdefs all over the place ... I don't know how complex this would get. > A brief forray into Kconfig-file-land showed multiple levels of defenses > preventing ACPI ever getting near CONFIG_HP_SIM ... I ripped out the first > two layers and still didn't get asked whether I wanted ACPI (and couldn't > immediately see where the third layer of defenses was located). > > > But right behind this build issue is the IA64 PCI code assuming ACPI as well... > > Which is probably in the same category ... PCI gets excluded by HP_SIM, but > for everyone else (again including allnoconfig) CONFIG_PCI=y CONFIG_PCI=y even for IA64 allnoconfig. > > How do you want to proceed? > > I assume that > > 1. configs that are theoretically possible, but would run on nothing, should be discouraged > > as they become maintainer make-work that benefits nobody. > Discouraged ... sure. So if there is an easy tweak to a Kconfig file > to stop someone from configuring something that won't work, that's > definitely a good thing. How much effort to expend on more complex > cases to stop people shooting themselves in the foot is a much greyer > area. Right now we let people configure in all sorts of devices that > have zero probabability of ever being installed in an ia64 box. I'm not really excited about the use of select in arch/ia64/Kconfig, but I think if we simply tweak that part for ACPI to make all of IA64 except the simulator depend on it, we are good. We'd do the same for PCI, just for consistency. This seems to make it through my evil config script that tries everything -- by making the two failing configs I sent you simply impossible. thanks, -Len > > 2. littering the code with #ifdef CONFIG_ACPI should be minimized. > Yup ... #fidef's are bad, #ifdef CONFIG_* doubly so (the Makefile syntax > of "objects-$(CONFIG_FOO) += foo.o" is pure magic). > > So no definitive answers ... just a pointer to what I think is > the right direction. > > -Tony > - To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
