The ARMv8.2 architecture introduces the optional Statistical Profiling
Extension (SPE).
SPE can be used to profile a population of operations in the CPU pipeline
after instruction decode. These are either architected instructions (i.e.
a dynamic instruction trace) or CPU-specific uops and the choice is fixed
statically in the hardware and advertised to userspace via caps/. Sampling
is controlled using a sampling interval, similar to a regular PMU counter,
but also with an optional random perturbation to avoid falling into patterns
where you continuously profile the same instruction in a hot loop.
After each operation is decoded, the interval counter is decremented. When
it hits zero, an operation is chosen for profiling and tracked within the
pipeline until it retires. Along the way, information such as TLB lookups,
cache misses, time spent to issue etc is captured in the form of a sample.
The sample is then filtered according to certain criteria (e.g. load
latency) that can be specified in the event config (described under
format/) and, if the sample satisfies the filter, it is written out to
memory as a record, otherwise it is discarded. Only one operation can
be sampled at a time.
The in-memory buffer is linear and virtually addressed, raising an
interrupt when it fills up. The PMU driver handles these interrupts to
give the appearance of a ring buffer, as expected by the AUX code.
The in-memory trace-like format is self-describing (though not parseable
in reverse) and written as a series of records, with each record
corresponding to a sample and consisting of a sequence of packets. These
packets are defined by the architecture, although some have CPU-specific
fields for recording information specific to the microarchitecture.
As a simple example, a record generated for a branch instruction may
consist of the following packets:
0 (Address) : Virtual PC of the branch instruction
1 (Type) : Conditional direct branch
2 (Counter) : Number of cycles taken from Dispatch to Issue
3 (Address) : Virtual branch target + condition flags
4 (Counter) : Number of cycles taken from Dispatch to Complete
5 (Events) : Mispredicted as not-taken
6 (END) : End of record
It is also possible to toggle properties such as timestamp packets in
each record.
This patch adds support for SPE in the form of a new perf driver.
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
All PMU drivers are going to depend on PERF_EVENTS, so let's make this
dependency common and simplify the individual Kconfig entries.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This patch adds framework code to handle parsing PMU data out of the
MADT, sanity checking this, and managing the association of CPUs (and
their interrupts) with appropriate logical PMUs.
For the time being, we expect that only one PMU driver (PMUv3) will make
use of this, and we simply pass in a single probe function.
This is based on an earlier patch from Jeremy Linton.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This adds a new dynamic PMU to the Perf Events framework to program
and control the L3 cache PMUs in some Qualcomm Technologies SOCs.
The driver supports a distributed cache architecture where the overall
cache for a socket is comprised of multiple slices each with its own PMU.
Access to each individual PMU is provided even though all CPUs share all
the slices. User space needs to aggregate to individual counts to provide
a global picture.
The driver exports formatting and event information to sysfs so it can
be used by the perf user space tools with the syntaxes:
perf stat -a -e l3cache_0_0/read-miss/
perf stat -a -e l3cache_0_0/event=0x21/
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Agustin Vega-Frias <agustinv@codeaurora.org>
[will: fixed sparse issues]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Adds perf events support for L2 cache PMU.
The L2 cache PMU driver is named 'l2cache_0' and can be used
with perf events to profile L2 events such as cache hits
and misses on Qualcomm Technologies processors.
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Neil Leeder <nleeder@codeaurora.org>
[will: minimise nesting in l2_cache_associate_cpu_with_cluster]
[will: use kstrtoul for unsigned long, remove redunant .owner setting]
Signed-off-by: Will Deacon <will.deacon@arm.com>
This patch adds a driver for the SoC-wide (AKA uncore) PMU hardware
found in APM X-Gene SoCs.
Signed-off-by: Tai Nguyen <ttnguyen@apm.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Now that the arm_pmu framework has been factored out to drivers/perf we
can make use of it for arm64, gaining support for heterogeneous PMUs
and unifying the two codebases before they diverge further.
The as yet unused PMU name for PMUv3 is changed to armv8_pmuv3, matching
the style previously applied to the 32-bit PMUs.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
To enable sharing of the arm_pmu code with arm64, this patch factors it
out to drivers/perf/. A new drivers/perf directory is added for
performance monitor drivers to live under.
MAINTAINERS is updated accordingly. Files added previously without a
corresponsing MAINTAINERS update (perf_regs.c, perf_callchain.c, and
perf_event.h) are also added.
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
[will: augmented Kconfig help slightly]
Signed-off-by: Will Deacon <will.deacon@arm.com>
