Re: [PATCH 2/4] perf/x86/ibs: Use CPUID region helper

[Robert Richter]

On 22.04.24 19:20:55, Borislav Petkov wrote:
> On Thu, Apr 18, 2024 at 02:05:49PM +0200, Robert Richter wrote:
> > There is the X86_FEATURE_IBS bit (Fn8000_0001_ECX, bit 10) which is
> > available from the beginning of IBS (all Fam10h production releases,
> > revB onwards).
> > 
> > And right, IBS_CPUID_FEATURES (CPUID Fn8000_001B) was introduced with
> > revC. The capabilities of revB are set in IBS_CAPS_DEFAULT.
> > 
> > It doesn't look broken to me, simply the ibs caps field was introduced
> > later which can be determined checking the return code of
> > get_cpuid_region_leaf().
> 
> Right.
> 
> > My preference would be:
> > 
> > 	[...]
> > 	if (!get_cpuid_region_leaf(IBS_CPUID_FEATURES, CPUID_EAX, &caps))
> 
> Right, checking get_cpuid_region_leaf() retval should happen.
> 
> > 		return IBS_CAPS_DEFAULT;
> > 
> > 	if (caps & IBS_CAPS_AVAIL)
> > 		return caps;
> > 
> > 	return 0;
> > 	[...]
> > 
> > Not too complex?
> 
> I'm wondering should we even support those old things? All revBs should
> be probably dead by now. But ok, it's not like it is a lot of code or
> so...

Since we check get_cpuid_region_leaf()'s return code here we know if
the cpud leaf exists and return IBS_CAPS_DEFAULT if not. That would
not change the refB behaviour.

Though I think that case is rare or even not existing, I would just
keep the implementation like that and as it was for for years.

> 
> > This slightly modifies the functionality so that 0 is return if
> > !IBS_CAPS_AVAIL (instead of IBS_CAPS_DEFAULT).
> 
> If !IBS_CAPS_AVAIL, then this is revB. But then you want to return
> IBS_CAPS_DEFAULT there.

No, on a rebB get_cpuid_region_leaf() would be false, meaning the
cpuid leaf is missing, function returns with IBS_CAPS_DEFAULT then.

-Robert