[openib-general] Re: Re: Userspace testing results (manykernels, many svn trees)

[Michael S. Tsirkin]

Quoting r. Roland Dreier <rdreier at cisco.com>:
> Subject: Re: [openib-general] Re: Re: Userspace testing results (manykernels, many svn trees)
> 
>     Michael> I just added some cycle calibration code to
>     Michael> get_cpu_mhz().  Check it out (you can just run
>     Michael> clock_test).
> 
> Seems to work well.  On the same JS20 system:
> 
> Warning: measured CPU frequency value14.317 differs from nominal 2194.62
> 1 sec = 1.00254e+06 usec
> 1 sec = 1.00203e+06 usec
> 
> The only minor issue is:
> 
>     get_clock.c: In function get_clock.c:97: warning: long long int format, long unsigned int arg (arg 4)
> 
>  - R.

What about 32 bit?
I see this in timex.h:

	asm volatile (
		"98:	mftb %0\n"
		"99:\n"
		".section __ftr_fixup,\"a\"\n"
		"	.long %1\n"
		"	.long 0\n"
		"	.long 98b\n"
		"	.long 99b\n"
		".previous"
		: "=r" (ret) : "i" (CPU_FTR_601));

Could some PPC gurus here chime up on what this does?
Cant I just "mftb %0, 268"?


-- 
MST