Messages in this thread | | | Date | Mon, 22 Oct 2012 16:45:38 -0700 | From | Andrew Morton <> | Subject | Re: [PATCH 2/5] ACPI: Reorder IPMI driver before any other ACPI drivers |
| |
On Tue, 16 Oct 2012 15:53:37 -0500 minyard@acm.org wrote:
> From: Matthew Garrett <mjg@redhat.com> > > Drivers may make calls that require the ACPI IPMI driver to have been > initialised already, so make sure that it appears earlier in the build > order. > > ... > > index 47199e2..82422fe 100644 > --- a/drivers/acpi/Makefile > +++ b/drivers/acpi/Makefile > @@ -47,6 +47,10 @@ acpi-y += video_detect.o > endif > > # These are (potentially) separate modules > + > +# IPMI may be used by other drivers, so it has to initialise before them > +obj-$(CONFIG_ACPI_IPMI) += acpi_ipmi.o > + > obj-$(CONFIG_ACPI_AC) += ac.o > obj-$(CONFIG_ACPI_BUTTON) += button.o > obj-$(CONFIG_ACPI_FAN) += fan.o > @@ -70,6 +74,5 @@ processor-y += processor_idle.o processor_thermal.o > processor-$(CONFIG_CPU_FREQ) += processor_perflib.o > > obj-$(CONFIG_ACPI_PROCESSOR_AGGREGATOR) += acpi_pad.o > -obj-$(CONFIG_ACPI_IPMI) += acpi_ipmi.o > > obj-$(CONFIG_ACPI_APEI) += apei/
Relying upon link ordering is the old-fashioned way of doing things, and I have vague memories that it only works by luck - that there's no hard-and-fast rule that the linker has to obey what we think we asked it to do.
The usual way of doing this sort of thing is to use the initcall priority levels - core_initcall(), postcore_initcall(), etc. Can that be done here?
| |