profclock overrides?

[Warner Losh]

On Mar 7, 2013, at 7:35 PM, Tim Kientzle wrote:

> As part of my poking and prodding towards a GENERIC kernel, I'm
> trying to see what happens when I actually combine the sources for
> different ARM systems together.
> 
> Here's one problem that seems to appear with a lot of the ARM systems:
> 
> ixp425_timer.o: In function `cpu_startprofclock':
> /usr/home/tim/projects/FreeBSD/svn/head/sys/arm/xscale/ixp425/ixp425_timer.c:263: multiple definition of `cpu_startprofclock'
> kern_clocksource.o:/usr/home/tim/projects/FreeBSD/svn/head/sys/kern/kern_clocksource.c:734: first defined here
> ixp425_timer.o: In function `cpu_stopprofclock':
> /usr/home/tim/projects/FreeBSD/svn/head/sys/arm/xscale/ixp425/ixp425_timer.c:268: multiple definition of `cpu_stopprofclock'
> kern_clocksource.o:/usr/home/tim/projects/FreeBSD/svn/head/sys/kern/kern_clocksource.c:754: first defined here
> 
> It looks like a bunch of systems have overridden these two functions
> in order to disable the profclock.
> 
> Is this an intrinsic board difference (that should be included in
> board/SoC-specific startup vectors), or is there a better way to
> do this?

I think we should add this to a platform-sw function table that each platform sets....

Warner