mirror of
https://github.com/torvalds/linux.git
synced 2026-03-09 19:26:35 +01:00
23e0a63c6d
Currently, running "perf script flamegraph -a -F 99 sleep 1" should produce flamegraph.html containing the flamegraph. Howevever, it gives a segmentation fault. This is caused because the flamegraph.py script is supposed to take as input the output of "perf record", which should be in stdin. This would require passing "-i -" to flamegraph.py. However, the "flamegraph-report" script causes "perf script" command to take the "-i -" option instead of flamegraph.py, which causes no problem for "perf script", but causes a seg fault since flamegraph.py has no input file. To fix this I added the "-i -" option directly to the flamegraph-report script to ensure flamegraph.py gets input from stdin. Signed-off-by: Anubhav Shelat <ashelat@redhat.com> Tested-by: Michael Petlan <mpetlan@redhat.com> Link: https://lore.kernel.org/r/20250131145704.3164542-2-ashelat@redhat.com Signed-off-by: Namhyung Kim <namhyung@kernel.org> |
||
|---|---|---|
| .. | ||
| bin | ||
| Perf-Trace-Util | ||
| arm-cs-trace-disasm.py | ||
| check-perf-trace.py | ||
| compaction-times.py | ||
| event_analyzing_sample.py | ||
| export-to-postgresql.py | ||
| export-to-sqlite.py | ||
| exported-sql-viewer.py | ||
| failed-syscalls-by-pid.py | ||
| flamegraph.py | ||
| futex-contention.py | ||
| gecko.py | ||
| intel-pt-events.py | ||
| libxed.py | ||
| mem-phys-addr.py | ||
| net_dropmonitor.py | ||
| netdev-times.py | ||
| parallel-perf.py | ||
| powerpc-hcalls.py | ||
| sched-migration.py | ||
| sctop.py | ||
| stackcollapse.py | ||
| stat-cpi.py | ||
| syscall-counts-by-pid.py | ||
| syscall-counts.py | ||
| task-analyzer.py | ||