Re: [PATCH 4/9] perf/x86/intel: Support hardware TopDown metrics

[Peter Zijlstra]

On Wed, May 29, 2019 at 10:42:10AM -0400, Liang, Kan wrote:
> On 5/29/2019 3:54 AM, Peter Zijlstra wrote:

> > cd09c0c40a97 ("perf events: Enable raw event support for Intel unhalted_reference_cycles event")
> > 
> > We used the fake event=0x00, umask=0x03 for CPU_CLK_UNHALTED.REF_TSC,
> > because that was not available as a generic event, *until now* it seems.
> > I see ICL actually has it as a generic event, which means we need to fix
> > up the constraint mask for that differently.
> > 
> 
> There is no change for REF_TSC on ICL.

Well, if I look at the SDM for May'19 (latest afaict), Volume 3, Chapter
19.3 'Performance Monitoring Events for Future Intel (C) Core(tm)
Processors' the table lists:

 Event Num.	Umask Value	Event Mask Mnemonic

 00H		03H		CPU_CLK_UNHALTED.REF_TSC

as a generic event, without constraints, unlike any of the preceding
uarchs, where that event was not available except through FIXED2.

That is most certainly a change.

> > But note that for all previous uarchs this event did not in fact exist.
> > 
> > It appears the TOPDOWN.SLOTS thing, which is available in in FIXED3 is
> > event=0x00, umask=0x04, is indeed a generic event too.
> 
> The SLOTS do have a generic event, TOPDOWN.SLOTS_P, event=0xA4, umask=0x1.
> 
> I think we need a fix as below for ICL, so the SLOT event can be extended to
> generic event.
> -	FIXED_EVENT_CONSTRAINT(0x0400, 3),	/* SLOTS */
> +	FIXED_EVENT_CONSTRAINT(0x01a4, 3),	/* TOPDOWN.SLOTS */

Then WTH is that 00H, 04H event listed in the table? Note the distinct
lack of 'Fixed Counter' or any other contraints in the 'Comments'
column.