ETM is a hardware feature available on arm64 devices. It collects the instruction stream running on each cpu. ARM uses ETM as an alternative for LBR (last branch record) on x86. Simpleperf supports collecting ETM data, and converting it to input files for AutoFDO, which can then be used for PGO (profile-guided optimization) during compilation.
On ARMv8, ETM is considered as an external debug interface (unless ARMv8.4 Self-hosted Trace extension is impelemented). So it needs to be enabled explicitly in the bootloader, and isn‘t available on user devices. For Pixel devices, it’s available on EVT and DVT devices on Pixel 4, Pixel 4a (5G) and Pixel 5. To test if it's available on other devices, you can follow commands in this doc and see if you can record any ETM data.
Below are examples collecting ETM data for AutoFDO. It has two steps: first recording ETM data, second converting ETM data to AutoFDO input files.
Record ETM data:
# preparation: we need to be root to record ETM data $ adb root $ adb shell redfin:/ \# cd data/local/tmp redfin:/data/local/tmp \# # Do a system wide collection, it writes output to perf.data. # If only want ETM data for kernel, use `-e cs-etm:k`. # If only want ETM data for userspace, use `-e cs-etm:u`. redfin:/data/local/tmp \# simpleperf record -e cs-etm --duration 3 -a # To reduce file size and time converting to AutoFDO input files, we recommend converting ETM data # into an intermediate branch-list format. redfin:/data/local/tmp \# simpleperf inject --output branch-list -o branch_list.data
Converting ETM data to AutoFDO input files needs to read binaries. So for userspace libraries, they can be converted on device. For kernel, it needs to be converted on host, with vmlinux and kernel modules available.
Convert ETM data for userspace libraries:
# Injecting ETM data on device. It writes output to perf_inject.data. # perf_inject.data is a text file, containing branch counts for each library. redfin:/data/local/tmp \# simpleperf inject -i branch_list.data
Convert ETM data for kernel:
# pull ETM data to host. host $ adb pull /data/local/tmp/branch_list.data # download vmlinux and kernel modules to <binary_dir> # host simpleperf is in <aosp-top>/system/extras/simpleperf/scripts/bin/linux/x86_64/simpleperf, # or you can build simpleperf by `mmma system/extras/simpleperf`. host $ simpleperf inject --symdir <binary_dir> -i branch_list.data
The generated perf_inject.data may contain branch info for multiple binaries. But AutoFDO only accepts one at a time. So we need to split perf_inject.data. The format of perf_inject.data is below:
executed range with count info for binary1 branch with count info for binary1 // name for binary1 executed range with count info for binary2 branch with count info for binary2 // name for binary2 ...
We need to split perf_inject.data, and make sure one file only contains info for one binary.
Then we can use AutoFDO to create profile. AutoFDO only works for binaries having an executable segment as its first loadable segment. But binaries built in Android may not follow this rule. Simpleperf inject command knows how to work around this problem. But there is a check in AutoFDO forcing binaries to start with an executable segment. We need to disable the check in AutoFDO, by commenting out L127-L136 in https://github.com/google/autofdo/commit/188db2834ce74762ed17108ca344916994640708#diff-2d132ecbb5e4f13e0da65419f6d1759dd27d6b696786dd7096c0c34d499b1710R127-R136. Then we can use create_llvm_prof in AutoFDO to create profiles used by clang.
# perf_inject_binary1.data is split from perf_inject.data, and only contains branch info for binary1. host $ autofdo/create_llvm_prof -profile perf_inject_binary1.data -profiler text -binary path_of_binary1 -out a.prof -format binary # perf_inject_kernel.data is split from perf_inject.data, and only contains branch info for [kernel.kallsyms]. host $ autofdo/create_llvm_prof -profile perf_inject_kernel.data -profiler text -binary vmlinux -out a.prof -format binary
Then we can use a.prof for PGO during compilation, via -fprofile-sample-use=a.prof. Here are more details.
etm_test_loop.cpp is an example to show the complete process. The source code is in etm_test_loop.cpp. The build script is in Android.bp. It builds an executable called etm_test_loop, which runs on device.
Step 1: Build etm_test_loop binary.
(host) <AOSP>$ . build/envsetup.sh (host) <AOSP>$ lunch aosp_arm64-trunk_staging-userdebug (host) <AOSP>$ make etm_test_loop
Step 2: Run etm_test_loop on device, and collect ETM data for its running.
(host) <AOSP>$ adb push out/target/product/generic_arm64/system/bin/etm_test_loop /data/local/tmp (host) <AOSP>$ adb root (host) <AOSP>$ adb shell (device) / # cd /data/local/tmp (device) /data/local/tmp # chmod a+x etm_test_loop (device) /data/local/tmp # simpleperf record -e cs-etm:u ./etm_test_loop simpleperf I cmd_record.cpp:729] Recorded for 0.0370068 seconds. Start post processing. simpleperf I cmd_record.cpp:799] Aux data traced: 1689136 (device) /data/local/tmp # simpleperf inject -i perf.data --output branch-list -o branch_list.data simpleperf W dso.cpp:557] failed to read min virtual address of [vdso]: File not found (device) /data/local/tmp # exit (host) <AOSP>$ adb pull /data/local/tmp/branch_list.data
Step 3: Convert ETM data to AutoFDO data.
# Build simpleperf tool on host. (host) <AOSP>$ make simpleperf_ndk (host) <AOSP>$ simpleperf_ndk64 inject -i branch_list.data -o perf_inject_etm_test_loop.data --symdir out/target/product/generic_arm64/symbols/system/bin simpleperf W cmd_inject.cpp:505] failed to build instr ranges for binary [vdso]: File not found (host) <AOSP>$ cat perf_inject_etm_test_loop.data 13 1000-1010:1 1014-1050:1 ... 112c->0:1 // /data/local/tmp/etm_test_loop (host) <AOSP>$ create_llvm_prof -profile perf_inject_etm_test_loop.data -profiler text -binary out/target/product/generic_arm64/symbols/system/bin/etm_test_loop -out etm_test_loop.afdo -format binary (host) <AOSP>$ ls -lh etm_test_loop.afdo rw-r--r-- 1 user group 241 Aug 29 16:04 etm_test_loop.afdo
Step 4: Use AutoFDO data to build optimized binary.
(host) <AOSP>$ mkdir toolchain/pgo-profiles/sampling/ (host) <AOSP>$ cp etm_test_loop.afdo toolchain/pgo-profiles/sampling/ (host) <AOSP>$ vi toolchain/pgo-profiles/sampling/Android.bp # edit Android.bp to add a fdo_profile module # soong_namespace {} # # fdo_profile { # name: "etm_test_loop_afdo", # profile: ["etm_test_loop.afdo"], # }
soong_namespace is added to support fdo_profile modules with the same name
In a product config mk file, update PRODUCT_AFDO_PROFILES with
PRODUCT_AFDO_PROFILES += etm_test_loop://toolchain/pgo-profiles/sampling:etm_test_loop_afdo
(host) <AOSP>$ vi system/extras/simpleperf/runtest/Android.bp # edit Android.bp to enable afdo for etm_test_loop. # cc_binary { # name: "etm_test_loop", # srcs: ["etm_test_loop.cpp"], # afdo: true, # } (host) <AOSP>$ make etm_test_loop
If comparing the disassembly of out/target/product/generic_arm64/symbols/system/bin/etm_test_loop before and after optimizing with AutoFDO data, we can see different preferences when branching.
Android also has a daemon collecting ETM data periodically. It only runs on userdebug and eng devices. The source code is in https://android.googlesource.com/platform/system/extras/+/main/profcollectd/.
To let simpleperf use ETM function, we need to enable Coresight driver in the kernel, which lives in <linux_kernel>/drivers/hwtracing/coresight.
The Coresight driver can be enabled by below kernel configs:
CONFIG_CORESIGHT=y CONFIG_CORESIGHT_LINK_AND_SINK_TMC=y CONFIG_CORESIGHT_SOURCE_ETM4X=y
On Kernel 5.10+, we recommend building Coresight driver as kernel modules. Because it works with GKI kernel.
CONFIG_CORESIGHT=m CONFIG_CORESIGHT_LINK_AND_SINK_TMC=m CONFIG_CORESIGHT_SOURCE_ETM4X=m
Android common kernel 5.10+ should have all the Coresight patches needed to collect ETM data. Android common kernel 5.4 misses two patches. But by adding patches in https://android-review.googlesource.com/q/topic:test_etm_on_hikey960_5.4, we can collect ETM data on hikey960 with 5.4 kernel. For Android common kernel 4.14 and 4.19, we have backported all necessary Coresight patches.
Besides Coresight driver, we also need to add Coresight devices in device tree. An example is in https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/arm/juno-base.dtsi. There should be a path flowing ETM data from ETM device through funnels, ETF and replicators, all the way to ETR, which writes ETM data to system memory.
One optional flag in ETM device tree is “arm,coresight-loses-context-with-cpu”. It saves ETM registers when a CPU enters low power state. It may be needed to avoid “coresight_disclaim_device_unlocked” warning when doing system wide collection.
One optional flag in ETR device tree is “arm,scatter-gather”. Simpleperf requests 4M system memory for ETR to store ETM data. Without IOMMU, the memory needs to be contiguous. If the kernel can't fulfill the request, simpleperf will report out of memory error. Fortunately, we can use “arm,scatter-gather” flag to let ETR run in scatter gather mode, which uses non-contiguous memory.
Each CPU has an ETM device, which has a unique trace_id assigned from the kernel. The formula is: trace_id = 0x10 + cpu * 2, as in https://github.com/torvalds/linux/blob/master/include/linux/coresight-pmu.h#L37. If the formula is modified by local patches, then simpleperf inject command can't parse ETM data properly and is likely to give empty output.
Unless ARMv8.4 Self-hosted Trace extension is implemented, ETM is considered as an external debug interface. It may be disabled by fuse (like JTAG). So we need to check if ETM is disabled, and if bootloader provides a way to reenable it.
We can tell if ETM is disable by checking its TRCAUTHSTATUS register, which is exposed in sysfs, like /sys/bus/coresight/devices/coresight-etm0/mgmt/trcauthstatus. To reenable ETM, we need to enable non-Secure non-invasive debug on ARM CPU. The method depends on chip vendors(SOCs).