Re: [PATCH 1/7] perf: introduce raw_type attribute to specify thetype of a raw sample

From: Robert Richter
Date: Thu May 20 2010 - 19:06:47 EST

On 20.05.10 02:51:43, Ingo Molnar wrote:

> Whether IBS fits that category is an open question -
> ideally it should have similar high level usage as the
> PEBS/LBR bits on Intel CPUs. (at least as far as tooling
> goes)


high level usage is from the perspective of perf tools. IBS delivers
high detailed diagnostic data that is also processed by userland tools
other than perf. If we would introduce another interface layer it
becomes much harder for such tools to use the pmu via this layer. The
raw data is of much more interest here for users. If we think of
perf_event as a subsystem controlling the pmu in a single instance, it
should provide access to set and get raw pmu registers. I see this as
a necessary feature.

Nevertheless I agree that key features should be exposed also with a
high level interface and if possible in an abstracted non-model
specific way.


Advanced Micro Devices, Inc.
Operating System Research Center
email: robert.richter@xxxxxxx

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at