On 19/01/2023 16:58, Arnaldo Carvalho de Melo wrote:
Em Thu, Jan 19, 2023 at 03:42:59PM +0000, James Clark escreveu:
Changes since v3:
- Scale time estimates by INSTR_PER_NS, rather than assuming 1 instruction = 1ns
- Add a new commit that fixes some issues around timestamps going backwards
- Use nanoseconds inside cs-etm-decoder.c, rather than storing the raw time values and converting when a sample is synthesized. This simplifies some of the code like estimating the first timestamp.
I would check this myself, but since Suzuki had some review comments and you may consider a v5, please check this:
Cover: ./v4_20230119_james_clark_perf_cs_etm_basic_support_for_virtual_kernel_timestamps.cover Link: https://lore.kernel.org/r/20230119154308.3815108-1-james.clark@arm.com git checkout -b v4_20230119_james_clark_arm_com 69b41ac87e4a664de78a395ff97166f0b2943210 git am ./v4_20230119_james_clark_perf_cs_etm_basic_support_for_virtual_kernel_timestamps.mbx [acme@quaco perf]$ git am ./v4_20230119_james_clark_perf_cs_etm_basic_support_for_virtual_kernel_timestamps.mbx Applying: perf: Remove duplication around EVENT_SOURCE_DEVICE_PATH error: patch failed: tools/perf/util/pmu.c:1993 error: tools/perf/util/pmu.c: patch does not apply error: patch failed: tools/perf/util/pmu.h:259 error: tools/perf/util/pmu.h: patch does not apply Patch failed at 0001 perf: Remove duplication around EVENT_SOURCE_DEVICE_PATH hint: Use 'git am --show-current-patch=diff' to see the failed patch When you have resolved this problem, run "git am --continue". If you prefer to skip this patch, run "git am --skip" instead. To restore the original branch and stop patching, run "git am --abort". [acme@quaco perf]$
Probably a conflict with one of these:
[acme@quaco perf]$ git log --oneline -5 tools/perf/util/pmu.c acef233b7ca749fd perf pmu: Add #slots literal support for arm64 336b92da1aa4228a perf tool: Move pmus list variable to a new file 49bd97c28b7e7f01 perf tools: Use dedicated non-atomic clear/set bit helpers e5c6109f4813246a perf list: Reorganize to use callbacks to allow honouring command line options eb2d4514a5971444 perf pmu: Restructure print_pmu_events() to avoid memory allocations [acme@quaco perf]$
Update to my perf/core branch or even tmp.perf/core which is newer but still needs the full set of container build tests.
Oops I must have just missed the latest perf/core update. I fixed the conflict and sent a v5, and decided to not make any changes for Suzuki's comments.
Thanks James