[Skiboot] [PATCH v2 0/7] IMA Instrumentation Support
Hemant Kumar
hemant at linux.vnet.ibm.com
Fri Nov 18 13:10:53 AEDT 2016
Patchset adds support for In Memory Accumulation instrumentation (IMA)
services in OPAL for Power9. The entire IMA infrastructure consists of
two kinds of Performance Monitoring Units (PMUs) : nest ima pmus (chip
level) and core ima pmus (core level).
Nest IMA PMUs are off core but on chip. And these can be accessed via
in-band scoms. Programming these counters and accumulating the counter
data to memory is done via microcode running in one of the OCC Engines.
This patchset is to add nest IMA instrumentation support in the OPAL
side.
"IMA_CATALOG" partition in PNOR contains multiple device tree binaries
(DTB) in a compressed form with PVR tag. So, when loading the IMA_CATALOG
partition, OPAL passes the system PVR as a "subid" to the load_resource
API. If a catalog dtb found for a given pvr, it is decompressed and
linked to the main device tree.
Commit which adds the "IMA_CATALOG" partition to PNOR is :
https://github.com/open-power/pnor/commit/c940142c6dc64dd176096dc648f433c889919e84
Each event node in the device tree contains "event-name" and "offset".
Some of the PMUs may contain properties such as "scale" and "unit" which
reflects the fact that all the events inside this PMU will have the
same "scale" and "unit" values.
https://github.com/open-power/ima-catalog/commit/99b73ee691fb5273f502e879e07603c510db5f7a
talks about the DTS file for power8. For power9, it has a similar
format and adds more units for nest, core and thread level PMUs.
An excerpt from the dtb showing "mcs" pmu node and two of its event nodes:
/dts-v1/;
/ {
name = "";
compatible = "ibm,opal-in-memory-counters";
#address-cells = <0x1>;
#size-cells = <0x1>;
ima-nest-offset = <0x320000>;
ima-nest-size = <0x30000>;
version-id = "";
mcs0 {
compatible = "ibm,ima-counters-nest";
ranges;
#address-cells = <0x1>;
#size-cells = <0x1>;
unit = "MiB" ;
scale = "1.2207e-4" ;
event at 118 {
event-name = "PM_MCS_UP_128B_DATA_XFER_MC0" ;
reg = <0x118 0x8>;
desc = "Total Read Bandwidth seen on both MCS of MC0" ;
};
[SNIP]
So, from the OPAL side, we need to :
- Find out the current processor's PVR.
- Fetch the "IMA_CATALOG" partition.
- Fetch the correct subpartition based on the current processor's PVR.
- Decompress the blob taken from the subpartition.
- Expand the (now uncompressed) device tree binary and attach it to the
system's device tree, so that, it can now be discovered by the
kernel.
- Look at the IMA availability vector which denotes which of the nest
PMUs are available and remove the unavailable PMU nodes from the device tree.
Note that :
- The Catalog team is working on upstreaming the DTS files.
- The commit which adds the IMA_CATALOG partition to PNOR is mentioned
above.
- Since OPAL lacks a xz decompression library, an xz decompression
library has been reused from the hostboot repo (link has been
mentioned in patch 3/7).
- This patchset is for base enablement for IMA and hence, only contains
the nest IMA support.
- The last patch in the series is to add "chip-id" to reserved homer
region node in the device tree. This will give us the homer region's
associated chip in the kernel (which will be needed to fetch the
counter values from the required chip).
This Patchset does a couple of things :
1) At the time of boot, it detects the IMA_CATALOG resource. Based on
the current processor's PVR value, it fetches the appropriate
subpartition. The blob in this subpartition is then uncompressed and the
flattened device tree is obtained. This dtb is then expanded and then
linked to the system's device tree under
"/proc/device-tree/ima-counters". The node "ima-counters" is a new node
created in this patchset. The kernel can then discover this node based
on its compatibility field.
2) It implements an opal call to control a microcode running in one of the
OCC engines (responsible for nest IMA data collection) from kernel to
start/stop Nest PMU counter data collection.
This patchset is based on the initial work for Nest Instrumentation done
by Madhavan Srinivasan, which can be found here :
(https://lists.ozlabs.org/pipermail/skiboot/2016-March/002999.html).
TODOs:
- Add support for Core IMA.
Changes since v1 have been mentioned in the individual patches.
Hemant Kumar (6):
skiboot: Nest IMA macro definitions
skiboot: Find the IMA catalog
skiboot: Add a library for xz
skiboot: Fetch the IMA dtb
skiboot: Add opal call to enable/disable Nest IMA microcode
skiboot: Add documentation for nest IMA opal call
Vasant Hegde (1):
skiboot: Add chip id to HOMER reserved region
Makefile.main | 5 +-
core/flash.c | 1 +
core/init.c | 7 +
doc/opal-api/opal-nest-ima-counters.rst | 49 ++
hw/Makefile.inc | 2 +-
hw/homer.c | 30 +
hw/ima.c | 260 +++++++
include/ima.h | 114 +++
include/nest_ima.h | 130 ++++
include/opal-api.h | 3 +-
include/platform.h | 1 +
include/skiboot.h | 1 +
libxz/Makefile.inc | 7 +
libxz/xz.h | 312 ++++++++
libxz/xz_config.h | 133 ++++
libxz/xz_crc32.c | 67 ++
libxz/xz_dec_lzma2.c | 1183 +++++++++++++++++++++++++++++++
libxz/xz_dec_stream.c | 855 ++++++++++++++++++++++
libxz/xz_lzma2.h | 212 ++++++
libxz/xz_private.h | 164 +++++
libxz/xz_stream.h | 70 ++
21 files changed, 3602 insertions(+), 4 deletions(-)
create mode 100644 doc/opal-api/opal-nest-ima-counters.rst
create mode 100644 hw/ima.c
create mode 100644 include/ima.h
create mode 100644 include/nest_ima.h
create mode 100644 libxz/Makefile.inc
create mode 100644 libxz/xz.h
create mode 100644 libxz/xz_config.h
create mode 100644 libxz/xz_crc32.c
create mode 100644 libxz/xz_dec_lzma2.c
create mode 100644 libxz/xz_dec_stream.c
create mode 100644 libxz/xz_lzma2.h
create mode 100644 libxz/xz_private.h
create mode 100644 libxz/xz_stream.h
--
2.7.4
More information about the Skiboot
mailing list