diff options
| author | Miguel Ojeda <ojeda@kernel.org> | 2026-08-16 15:32:33 +0200 |
|---|---|---|
| committer | Miguel Ojeda <ojeda@kernel.org> | 2026-08-23 21:27:17 +0200 |
| commit | 5febf432df1cfa5b25d99b54c32103fafbdd0eb9 (patch) | |
| tree | 93a1ed1d337bc53adf1e572c2586d347703b286e /tools/perf/python/twatch.py | |
| download | linux-stable-5febf432df1cfa5b25d99b54c32103fafbdd0eb9.tar.gz linux-stable-5febf432df1cfa5b25d99b54c32103fafbdd0eb9.zip | |
kbuild: rust: keep Rust objects out of Clang LTO with inline helpersgrafted
Under `CONFIG_LTO_CLANG` + `CONFIG_RUST_INLINE_HELPERS`, one may hit
`objtool` errors such as:
vmlinux.o: error: objtool: _R..._3Gsp4boot+0xd6a:
can't find jump dest instruction at .text._R..._3Gsp4boot+0x1dfd
The reason is that in such builds, the Clang invocation that compiles
the combined Rust plus helpers bitcode emits LLVM bitcode (again) --
the final code generation happens in the linker's LTO step, which the
`-mllvm` trap options passed to Clang do not reach.
This, in turn, means that unreachable traps are missing, and the
impossible paths do not merely fallthrough to the next symbol, but past
the end of their own section, since LTO builds place each function in
its own section.
Thus filter `CC_FLAGS_LTO` out of the Clang invocation, so that it always
emits machine code directly, with the traps in place.
Assisted-by: LLM
Cc: Gary Guo <gary@garyguo.net>
Cc: Boqun Feng <boqun@kernel.org>
Cc: Alice Ryhl <aliceryhl@google.com>
Cc: Matthew Maurer <mmaurer@google.com>
Cc: Josh Poimboeuf <jpoimboe@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: stable@vger.kernel.org
Fixes: 3a2486cc1da5 ("kbuild: rust: provide an option to inline C helpers into Rust")
Acked-by: Gary Guo <gary@garyguo.net>
Link: https://patch.msgid.link/20260816133233.197500-2-ojeda@kernel.org
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Diffstat (limited to 'tools/perf/python/twatch.py')
| -rwxr-xr-x | tools/perf/python/twatch.py | 61 |
1 files changed, 61 insertions, 0 deletions
diff --git a/tools/perf/python/twatch.py b/tools/perf/python/twatch.py new file mode 100755 index 000000000..04f3db29b --- /dev/null +++ b/tools/perf/python/twatch.py @@ -0,0 +1,61 @@ +#! /usr/bin/env python +# SPDX-License-Identifier: GPL-2.0-only +# -*- python -*- +# -*- coding: utf-8 -*- +# twatch - Experimental use of the perf python interface +# Copyright (C) 2011 Arnaldo Carvalho de Melo <acme@redhat.com> +# + +import perf + +def main(context_switch = 0, thread = -1): + cpus = perf.cpu_map() + threads = perf.thread_map(thread) + evsel = perf.evsel(type = perf.TYPE_SOFTWARE, + config = perf.COUNT_SW_DUMMY, + task = 1, comm = 1, mmap = 0, freq = 0, + wakeup_events = 1, watermark = 1, + sample_id_all = 1, context_switch = context_switch, + sample_type = perf.SAMPLE_PERIOD | perf.SAMPLE_TID | perf.SAMPLE_CPU) + + """What we want are just the PERF_RECORD_ lifetime events for threads, + using the default, PERF_TYPE_HARDWARE + PERF_COUNT_HW_CYCLES & freq=1 + (the default), makes perf reenable irq_vectors:local_timer_entry, when + disabling nohz, not good for some use cases where all we want is to get + threads comes and goes... So use (perf.TYPE_SOFTWARE, perf_COUNT_SW_DUMMY, + freq=0) instead.""" + + evsel.open(cpus = cpus, threads = threads); + evlist = perf.evlist(cpus, threads) + evlist.add(evsel) + evlist.mmap() + while True: + evlist.poll(timeout = -1) + for cpu in cpus: + event = evlist.read_on_cpu(cpu) + if not event: + continue + print("cpu: {0}, pid: {1}, tid: {2} {3}".format(event.sample_cpu, + event.sample_pid, + event.sample_tid, + event)) + +if __name__ == '__main__': + """ + To test the PERF_RECORD_SWITCH record, pick a pid and replace + in the following line. + + Example output: + +cpu: 3, pid: 31463, tid: 31593 { type: context_switch, next_prev_pid: 31463, next_prev_tid: 31593, switch_out: 1 } +cpu: 1, pid: 31463, tid: 31489 { type: context_switch, next_prev_pid: 31463, next_prev_tid: 31489, switch_out: 1 } +cpu: 2, pid: 31463, tid: 31496 { type: context_switch, next_prev_pid: 31463, next_prev_tid: 31496, switch_out: 1 } +cpu: 3, pid: 31463, tid: 31491 { type: context_switch, next_prev_pid: 31463, next_prev_tid: 31491, switch_out: 0 } + + It is possible as well to use event.misc & perf.PERF_RECORD_MISC_SWITCH_OUT + to figure out if this is a context switch in or out of the monitored threads. + + If bored, please add command line option parsing support for these options :-) + """ + # main(context_switch = 1, thread = 31463) + main() |
