cvs commit: src/sys/ia64/ia64 machdep.c

[M. Warner Losh]

In message: <420322CA.3090201 at root.org>
            Nate Lawson <nate at root.org> writes:
: Marcel Moolenaar wrote:
: > 
: > On Feb 3, 2005, at 11:01 PM, Nate Lawson wrote:
: > 
: >> Marcel Moolenaar wrote:
: >>
: >>> marcel      2005-02-04 06:38:58 UTC
: >>>   FreeBSD src repository
: >>>   Modified files:
: >>>     sys/ia64/ia64        machdep.c   Log:
: >>>   Include sys/bus.h before sys/cpu.h. The latter needs device_t.
: >>>     Revision  Changes    Path
: >>>   1.195     +1 -0      src/sys/ia64/ia64/machdep.c
: >>
: >>
: >> Thanks, I've made similar changes to the other arches too.
: > 
: > 
: > No worries. Sorry I didn't had the time to test before committing.
: > 
: > BTW, I still have a link error:
: > 
: > pluto2# make
: > linking kernel.debug
: > acpi_cpu.o(.text+0x2302): In function `acpi_cpu_startup':
: > ../../../dev/acpica/acpi_cpu.c:757: undefined reference to `cpu_devclass'
: > acpi_cpu.o(.data+0x1a8): In function `acpi_cpu_probe':
: > ./bus_if.h:107: undefined reference to `cpu_devclass'
: > *** Error code 1
: > 
: > Stop in /q/6.x/src/sys/ia64/compile/PLUTO2.
: > 
: > Can it be that you missed a file?
: 
: No, it's something I need to fix one way or the other and had on my todo 
: list.  I'm not sure if my approach is correct -- should different 
: drivers for the same hardware share the same devclass?  I have 
: legacy::cpu and acpi::cpu sharing the same devclass value since neither 
: can be attached at the same time.  I thought in the past there might be 
: some advantage to this but can't remember the rationale.

In general, we've said that all devices who have the same name must
share the same devclass, or you wind up with two devices that could be
called 'cpu0' which can either be harmless, or lead to big problems
depending on what that device is...

Warner