Power 9 has In-Memory-Accumulation (IMA) infrastructure which contains various Performance Monitoring Units (PMUs) at Nest level (these are on-chip but off-core). These Nest PMU counters are handled by a Nest IMA microcode. This microcode runs in the OCC (On-Chip Controller) complex and its purpose is to program the nest counters, collect the counter data and move the counter data to memory.
The IMA infrastructure encapsulates nest (per-chip), core and thread level counters. While the nest IMA PMUs are handled by the nest IMA microcode, the core and thread level PMUs are handled by the Core-HPMC engine. This patchset enables the nest IMA PMUs and is based on the initial work done by Madhavan Srinivasan. "Nest Instrumentation Support" : https://lists.ozlabs.org/pipermail/linuxppc-dev/2015-August/132078.html v1 for this patchset can be found here : https://lwn.net/Articles/705475/ Nest events: Per-chip nest instrumentation provides various per-chip metrics such as memory, powerbus, Xlink and Alink bandwidth. PMU Events' Information: OPAL obtains the Nest PMU and event information from the IMA Catalog and passes on to the kernel via the device tree. The events' information contains : - Event name - Event Offset - Event description and, maybe : - Event scale - Event unit Some PMUs may have a common scale and unit values for all their supported events. For those cases, the scale and unit properties for those events must be inherited from the PMU. The event offset in the memory is where the counter data gets accumulated. The OPAL-side patches are posted upstream : https://lists.ozlabs.org/pipermail/skiboot/2016-November/005552.html The kernel discovers the IMA counters information in the device tree at the "ima-counters" device node which has a compatible field "ibm,opal-in-memory-counters". Parsing of the Events' information: To parse the IMA PMUs and events information, the kernel has to discover the "ima-counters" node and walk through the pmu and event nodes. Here is an excerpt of the dt showing the ima-counters and mcs node: /dts-v1/; [...] ima-counters { ima-nest-offset = <0x320000>; compatible = "ibm,opal-in-memory-counters"; ima-nest-size = <0x30000>; #address-cells = <0x1>; #size-cells = <0x1>; phandle = <0x10000238>; version-id = [00]; mcs0 { compatible = "ibm,ima-counters-chip"; ranges; #address-cells = <0x1>; #size-cells = <0x1>; phandle = <0x10000279>; scale = "1.2207e-4"; unit = "MiB"; event@528 { event-name = "PM_MCS_UP_128B_DATA_XFER_MC0" ; desc = "Total Read Bandwidth seen on both MCS of MC0"; phandle = <0x1000028c>; reg = <0x118 0x8>; }; [...] >From the device tree, the kernel parses the PMUs and their events' information. After parsing the nest IMA PMUs and their events, the PMUs and their attributes are registered in the kernel. Example Usage : # perf list [...] nest_mcs0/PM_MCS_DOWN_128B_DATA_XFER_MC0/ [Kernel PMU event] nest_mcs0/PM_MCS_DOWN_128B_DATA_XFER_MC0_LAST_SAMPLE/ [Kernel PMU event] [...] # perf stat -e "nest_mcs0/PM_MCS_DOWN_128B_DATA_XFER_MC0/" -a --per-socket TODOs: - Add support for Core IMA. - Add support for thread IMA. Comments/feedback/suggestions are welcome. Changelog: v1 -> v2 : - Account for the cases where a PMU can have a common scale and unit values for all its supported events (Patch 3/6). - Fixed a Build error (for maple_defconfig) by enabling ima_pmu.o only for CONFIG_PPC_POWERNV=y (Patch 4/6) - Read from the "event-name" property instead of "name" for an event node (Patch 3/6). Cc: Madhavan Srinivasan <ma...@linux.vnet.ibm.com> Cc: Michael Ellerman <m...@ellerman.id.au> Cc: Benjamin Herrenschmidt <b...@kernel.crashing.org> Cc: Paul Mackerras <pau...@samba.org> Cc: Anton Blanchard <an...@samba.org> Cc: Sukadev Bhattiprolu <suka...@linux.vnet.ibm.com> Cc: Michael Neuling <mi...@neuling.org> Cc: Stewart Smith <stew...@linux.vnet.ibm.com> Cc: Daniel Axtens <d...@axtens.net> Cc: Stephane Eranian <eran...@google.com> Signed-off-by: Hemant Kumar <hem...@linux.vnet.ibm.com> Hemant Kumar (6): powerpc/powernv: Data structure and macros definitions powerpc/powernv: Autoload IMA device driver module powerpc/powernv: Detect supported IMA units and its events powerpc/perf: Add event attribute and group to IMA pmus powerpc/perf: Generic ima pmu event functions powerpc/perf: IMA pmu cpumask and cpu hotplug support arch/powerpc/include/asm/ima-pmu.h | 75 ++++ arch/powerpc/include/asm/opal-api.h | 3 +- arch/powerpc/include/asm/opal.h | 2 + arch/powerpc/perf/Makefile | 6 +- arch/powerpc/perf/ima-pmu.c | 383 ++++++++++++++++++++ arch/powerpc/platforms/powernv/Makefile | 2 +- arch/powerpc/platforms/powernv/opal-ima.c | 480 +++++++++++++++++++++++++ arch/powerpc/platforms/powernv/opal-wrappers.S | 1 + arch/powerpc/platforms/powernv/opal.c | 13 + 9 files changed, 962 insertions(+), 3 deletions(-) create mode 100644 arch/powerpc/include/asm/ima-pmu.h create mode 100644 arch/powerpc/perf/ima-pmu.c create mode 100644 arch/powerpc/platforms/powernv/opal-ima.c -- 2.7.4