| Message ID | 172398527264.293426.2050093948411376857.stgit@devnote2 (mailing list archive) |
|---|---|
| Headers | show |
| Series | tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph | expand |
On Sun, Aug 18, 2024 at 5:58 PM Masami Hiramatsu <mhiramat@kernel.org> wrote: > > Hi, > > Where can I get the test programs? I would like to check what the programs > actually expected. > Hi Masami, This is part of BPF selftests (so it's under tools/testing/selftests/bpf). Specifically, this test is defined in - progs/kprobe_multi.c for BPF parts - prog_tests/kprobe_multi_testmod_test.c for user-space parts. It seems to be failing on arm64 (but not on x86-64, so something architecture-specific). > On Sun, 18 Aug 2024 13:51:30 +0000 (UTC) > bot+bpf-ci@kernel.org wrote: > > > Dear patch submitter, > > > > CI has tested the following submission: > > Status: FAILURE > > Name: [v13,00/20] tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph > > Patchwork: https://patchwork.kernel.org/project/netdevbpf/list/?series=880630&state=* > > Matrix: https://github.com/kernel-patches/bpf/actions/runs/10440799833 > > > > Failed jobs: > > test_progs-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439106 > > test_progs_no_alu32-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439234 > > test_progs-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911405063 > > test_progs_no_alu32-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911404959 > > veristat-x86_64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911401263 > > > > First test_progs failure (test_progs-aarch64-gcc): > > #126 kprobe_multi_testmod_test > > serial_test_kprobe_multi_testmod_test:PASS:load_kallsyms_local 0 nsec > > #126/1 kprobe_multi_testmod_test/testmod_attach_api_syms > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > #126/2 kprobe_multi_testmod_test/testmod_attach_api_addrs > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > > > > > Please note: this email is coming from an unmonitored mailbox. If you have > > questions or feedback, please reach out to the Meta Kernel CI team at > > kernel-ci@meta.com. > > > -- > Masami Hiramatsu (Google) <mhiramat@kernel.org>
+bpf On Mon, Aug 19, 2024 at 6:17 PM Masami Hiramatsu <mhiramat@kernel.org> wrote: > > On Mon, 19 Aug 2024 15:56:39 +0000 > Daniel Xu <dlxu@meta.com> wrote: > > > [sorry for outlook top post] > > > > Hi Masami, > > > > test_progs is checked into kernel tree. There should be source files in selftests > > for the test names. For example, for fill_link_info/kprobe_multi_invalid_ubuff > > failure: > > > > $ find . -name "*fill_link_info*" > > ./tools/testing/selftests/bpf/prog_tests/fill_link_info.c > > ./tools/testing/selftests/bpf/progs/test_fill_link_info.c > > > > veristat I'm less famiiar with. My understanding is that it checks for verifier > > regressions. Skimming your patchset, it's not obvious to me why verifier > > would regress. If you have issues debugging that, we can poke Andrii for > > help. > > Thanks for the information! Hmm, maybe kprobe_multi_testmod_check might check the > register which is not supported on the ftrace_regs. But I also don't have any idea This test is getting IP of the function using bpf_get_func_ip() helper. If that somehow started returning wrong value on arm64/s390x, then the test will basically not find expected addresses > about veristat. Is that also checks all pt_regs? Andrii, do you have any idea? I wouldn't worry about veristat, your changes shouldn't regress BPF verifier logic, so it's probably just an artifact of our BPF CI setup. The above test regression seems much more worrying. > > Thank you, > > > > > Thanks, > > Daniel > > > > > > ________________________________ > > From: Masami Hiramatsu <mhiramat@kernel.org> > > Sent: Sunday, August 18, 2024 5:58 PM > > To: bot+bpf-ci@kernel.org <bot+bpf-ci@kernel.org> > > Cc: kernel-ci <kernel-ci@meta.com>; andrii@kernel.org <andrii@kernel.org>; daniel@iogearbox.net <daniel@iogearbox.net>; martin.lau@linux.dev <martin.lau@linux.dev> > > Subject: Re: [PATCH v13 00/20] tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph > > > > Hi, > > > > Where can I get the test programs? I would like to check what the programs > > actually expected. > > > > On Sun, 18 Aug 2024 13:51:30 +0000 (UTC) > > bot+bpf-ci@kernel.org wrote: > > > > > Dear patch submitter, > > > > > > CI has tested the following submission: > > > Status: FAILURE > > > Name: [v13,00/20] tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph > > > Patchwork: https://patchwork.kernel.org/project/netdevbpf/list/?series=880630&state=* > > > Matrix: https://github.com/kernel-patches/bpf/actions/runs/10440799833 > > > > > > Failed jobs: > > > test_progs-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439106 > > > test_progs_no_alu32-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439234 > > > test_progs-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911405063 > > > test_progs_no_alu32-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911404959 > > > veristat-x86_64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911401263 > > > > > > First test_progs failure (test_progs-aarch64-gcc): > > > #126 kprobe_multi_testmod_test > > > serial_test_kprobe_multi_testmod_test:PASS:load_kallsyms_local 0 nsec > > > #126/1 kprobe_multi_testmod_test/testmod_attach_api_syms > > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > > #126/2 kprobe_multi_testmod_test/testmod_attach_api_addrs > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > > > > > > > > Please note: this email is coming from an unmonitored mailbox. If you have > > > questions or feedback, please reach out to the Meta Kernel CI team at > > > kernel-ci@meta.com. > > > > > > -- > > Masami Hiramatsu (Google) <mhiramat@kernel.org> > > > -- > Masami Hiramatsu (Google) <mhiramat@kernel.org>
Let me gently ping :) Let me check it should be rebased on tracing/for-next again. But basically this should be the same... On Sun, 18 Aug 2024 21:47:53 +0900 "Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote: > Hi, > > Here is the 13th version of the series to re-implement the fprobe on > function-graph tracer. The previous version is; > > https://lore.kernel.org/all/172000134410.63468.13742222887213469474.stgit@devnote2/ > > This version is based on v6.11-rc3. > In this version, I added a bugfix as [1/20], which should go to urgent > branch, and dropped the performance improvement patch which was introduced > in v12 because I found that does not work with new kernel. > > Overview > -------- > This series rewrites the fprobe on this function-graph. > The purposes of this change are; > > 1) Remove dependency of the rethook from fprobe so that we can reduce > the return hook code and shadow stack. > > 2) Make 'ftrace_regs' the common trace interface for the function > boundary. > > 1) Currently we have 2(or 3) different function return hook codes, > the function-graph tracer and rethook (and legacy kretprobe). > But since this is redundant and needs double maintenance cost, > I would like to unify those. From the user's viewpoint, function- > graph tracer is very useful to grasp the execution path. For this > purpose, it is hard to use the rethook in the function-graph > tracer, but the opposite is possible. (Strictly speaking, kretprobe > can not use it because it requires 'pt_regs' for historical reasons.) > > 2) Now the fprobe provides the 'pt_regs' for its handler, but that is > wrong for the function entry and exit. Moreover, depending on the > architecture, there is no way to accurately reproduce 'pt_regs' > outside of interrupt or exception handlers. This means fprobe should > not use 'pt_regs' because it does not use such exceptions. > (Conversely, kprobe should use 'pt_regs' because it is an abstract > interface of the software breakpoint exception.) > > This series changes fprobe to use function-graph tracer for tracing > function entry and exit, instead of mixture of ftrace and rethook. > Unlike the rethook which is a per-task list of system-wide allocated > nodes, the function graph's ret_stack is a per-task shadow stack. > Thus it does not need to set 'nr_maxactive' (which is the number of > pre-allocated nodes). > Also the handlers will get the 'ftrace_regs' instead of 'pt_regs'. > Since eBPF mulit_kprobe/multi_kretprobe events still use 'pt_regs' as > their register interface, this changes it to convert 'ftrace_regs' to > 'pt_regs'. Of course this conversion makes an incomplete 'pt_regs', > so users must access only registers for function parameters or > return value. > > Design > ------ > Instead of using ftrace's function entry hook directly, the new fprobe > is built on top of the function-graph's entry and return callbacks > with 'ftrace_regs'. > > Since the fprobe requires access to 'ftrace_regs', the architecture > must support CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS and > CONFIG_HAVE_FTRACE_GRAPH_FUNC, which enables to call function-graph > entry callback with 'ftrace_regs', and also > CONFIG_HAVE_FUNCTION_GRAPH_FREGS, which passes the ftrace_regs to > return_to_handler. > > All fprobes share a single function-graph ops (means shares a common > ftrace filter) similar to the kprobe-on-ftrace. This needs another > layer to find corresponding fprobe in the common function-graph > callbacks, but has much better scalability, since the number of > registered function-graph ops is limited. > > In the entry callback, the fprobe runs its entry_handler and saves the > address of 'fprobe' on the function-graph's shadow stack as data. The > return callback decodes the data to get the 'fprobe' address, and runs > the exit_handler. > > The fprobe introduces two hash-tables, one is for entry callback which > searches fprobes related to the given function address passed by entry > callback. The other is for a return callback which checks if the given > 'fprobe' data structure pointer is still valid. Note that it is > possible to unregister fprobe before the return callback runs. Thus > the address validation must be done before using it in the return > callback. > > Download > -------- > This series can be applied against the ftrace/for-next branch in > linux-trace tree. > > This series can also be found below branch. > > https://git.kernel.org/pub/scm/linux/kernel/git/mhiramat/linux.git/log/?h=topic/fprobe-on-fgraph > > Thank you, > > --- > > Masami Hiramatsu (Google) (20): > tracing: fgraph: Fix to add new fgraph_ops to array after ftrace_startup_subops() > tracing: Add a comment about ftrace_regs definition > tracing: Rename ftrace_regs_return_value to ftrace_regs_get_return_value > function_graph: Pass ftrace_regs to entryfunc > function_graph: Replace fgraph_ret_regs with ftrace_regs > function_graph: Pass ftrace_regs to retfunc > fprobe: Use ftrace_regs in fprobe entry handler > fprobe: Use ftrace_regs in fprobe exit handler > tracing: Add ftrace_partial_regs() for converting ftrace_regs to pt_regs > tracing: Add ftrace_fill_perf_regs() for perf event > tracing/fprobe: Enable fprobe events with CONFIG_DYNAMIC_FTRACE_WITH_ARGS > bpf: Enable kprobe_multi feature if CONFIG_FPROBE is enabled > ftrace: Add CONFIG_HAVE_FTRACE_GRAPH_FUNC > fprobe: Rewrite fprobe on function-graph tracer > tracing: Fix function timing profiler to initialize hashtable > tracing/fprobe: Remove nr_maxactive from fprobe > selftests: ftrace: Remove obsolate maxactive syntax check > selftests/ftrace: Add a test case for repeating register/unregister fprobe > Documentation: probes: Update fprobe on function-graph tracer > fgraph: Skip recording calltime/rettime if it is not nneeded > > > Documentation/trace/fprobe.rst | 42 + > arch/arm64/Kconfig | 2 > arch/arm64/include/asm/ftrace.h | 47 + > arch/arm64/kernel/asm-offsets.c | 12 > arch/arm64/kernel/entry-ftrace.S | 32 + > arch/arm64/kernel/ftrace.c | 20 + > arch/loongarch/Kconfig | 4 > arch/loongarch/include/asm/ftrace.h | 32 - > arch/loongarch/kernel/asm-offsets.c | 12 > arch/loongarch/kernel/ftrace_dyn.c | 10 > arch/loongarch/kernel/mcount.S | 17 - > arch/loongarch/kernel/mcount_dyn.S | 14 > arch/powerpc/Kconfig | 1 > arch/powerpc/include/asm/ftrace.h | 15 > arch/powerpc/kernel/trace/ftrace.c | 2 > arch/powerpc/kernel/trace/ftrace_64_pg.c | 10 > arch/riscv/Kconfig | 3 > arch/riscv/include/asm/ftrace.h | 26 - > arch/riscv/kernel/ftrace.c | 17 + > arch/riscv/kernel/mcount.S | 24 - > arch/s390/Kconfig | 3 > arch/s390/include/asm/ftrace.h | 39 + > arch/s390/kernel/asm-offsets.c | 6 > arch/s390/kernel/mcount.S | 9 > arch/x86/Kconfig | 4 > arch/x86/include/asm/ftrace.h | 37 + > arch/x86/kernel/ftrace.c | 50 +- > arch/x86/kernel/ftrace_32.S | 15 > arch/x86/kernel/ftrace_64.S | 17 - > include/linux/fprobe.h | 57 +- > include/linux/ftrace.h | 136 ++++ > kernel/trace/Kconfig | 23 + > kernel/trace/bpf_trace.c | 19 - > kernel/trace/fgraph.c | 127 +++- > kernel/trace/fprobe.c | 638 ++++++++++++++------ > kernel/trace/ftrace.c | 10 > kernel/trace/trace.h | 6 > kernel/trace/trace_fprobe.c | 147 ++--- > kernel/trace/trace_functions_graph.c | 10 > kernel/trace/trace_irqsoff.c | 6 > kernel/trace/trace_probe_tmpl.h | 2 > kernel/trace/trace_sched_wakeup.c | 6 > kernel/trace/trace_selftest.c | 11 > lib/test_fprobe.c | 51 -- > samples/fprobe/fprobe_example.c | 4 > .../test.d/dynevent/add_remove_fprobe_repeat.tc | 19 + > .../ftrace/test.d/dynevent/fprobe_syntax_errors.tc | 4 > 47 files changed, 1168 insertions(+), 630 deletions(-) > create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe_repeat.tc > > -- > Masami Hiramatsu (Google) <mhiramat@kernel.org> >
On Mon, 19 Aug 2024 19:48:23 -0700 Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > +bpf > > On Mon, Aug 19, 2024 at 6:17 PM Masami Hiramatsu <mhiramat@kernel.org> wrote: > > > > On Mon, 19 Aug 2024 15:56:39 +0000 > > Daniel Xu <dlxu@meta.com> wrote: > > > > > [sorry for outlook top post] > > > > > > Hi Masami, > > > > > > test_progs is checked into kernel tree. There should be source files in selftests > > > for the test names. For example, for fill_link_info/kprobe_multi_invalid_ubuff > > > failure: > > > > > > $ find . -name "*fill_link_info*" > > > ./tools/testing/selftests/bpf/prog_tests/fill_link_info.c > > > ./tools/testing/selftests/bpf/progs/test_fill_link_info.c > > > > > > veristat I'm less famiiar with. My understanding is that it checks for verifier > > > regressions. Skimming your patchset, it's not obvious to me why verifier > > > would regress. If you have issues debugging that, we can poke Andrii for > > > help. > > > > Thanks for the information! Hmm, maybe kprobe_multi_testmod_check might check the > > register which is not supported on the ftrace_regs. But I also don't have any idea > > This test is getting IP of the function using bpf_get_func_ip() > helper. If that somehow started returning wrong value on arm64/s390x, > then the test will basically not find expected addresses Ah, bpf_get_func_ip() may get PC register in pt_regs. ftrace_regs doesn't have it. Hmm, but that information should be injected by bpf side because ftrace_regs -> pt_regs converted in trace_bpf, and the handler should know the IP. > > > about veristat. Is that also checks all pt_regs? Andrii, do you have any idea? > > I wouldn't worry about veristat, your changes shouldn't regress BPF > verifier logic, so it's probably just an artifact of our BPF CI setup. > The above test regression seems much more worrying. OK, trying to fix that. Maybe [07/20] caused this issue. Thank you, > > > > > > Thank you, > > > > > > > > Thanks, > > > Daniel > > > > > > > > > ________________________________ > > > From: Masami Hiramatsu <mhiramat@kernel.org> > > > Sent: Sunday, August 18, 2024 5:58 PM > > > To: bot+bpf-ci@kernel.org <bot+bpf-ci@kernel.org> > > > Cc: kernel-ci <kernel-ci@meta.com>; andrii@kernel.org <andrii@kernel.org>; daniel@iogearbox.net <daniel@iogearbox.net>; martin.lau@linux.dev <martin.lau@linux.dev> > > > Subject: Re: [PATCH v13 00/20] tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph > > > > > > Hi, > > > > > > Where can I get the test programs? I would like to check what the programs > > > actually expected. > > > > > > On Sun, 18 Aug 2024 13:51:30 +0000 (UTC) > > > bot+bpf-ci@kernel.org wrote: > > > > > > > Dear patch submitter, > > > > > > > > CI has tested the following submission: > > > > Status: FAILURE > > > > Name: [v13,00/20] tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph > > > > Patchwork: https://patchwork.kernel.org/project/netdevbpf/list/?series=880630&state=* > > > > Matrix: https://github.com/kernel-patches/bpf/actions/runs/10440799833 > > > > > > > > Failed jobs: > > > > test_progs-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439106 > > > > test_progs_no_alu32-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439234 > > > > test_progs-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911405063 > > > > test_progs_no_alu32-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911404959 > > > > veristat-x86_64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911401263 > > > > > > > > First test_progs failure (test_progs-aarch64-gcc): > > > > #126 kprobe_multi_testmod_test > > > > serial_test_kprobe_multi_testmod_test:PASS:load_kallsyms_local 0 nsec > > > > #126/1 kprobe_multi_testmod_test/testmod_attach_api_syms > > > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > > > #126/2 kprobe_multi_testmod_test/testmod_attach_api_addrs > > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > > > > > > > > > > > Please note: this email is coming from an unmonitored mailbox. If you have > > > > questions or feedback, please reach out to the Meta Kernel CI team at > > > > kernel-ci@meta.com. > > > > > > > > > -- > > > Masami Hiramatsu (Google) <mhiramat@kernel.org> > > > > > > -- > > Masami Hiramatsu (Google) <mhiramat@kernel.org>
On Sun, Aug 18, 2024 at 09:47:53PM +0900, Masami Hiramatsu (Google) wrote: > Hi, > > Here is the 13th version of the series to re-implement the fprobe on > function-graph tracer. The previous version is; > > https://lore.kernel.org/all/172000134410.63468.13742222887213469474.stgit@devnote2/ > > This version is based on v6.11-rc3. > In this version, I added a bugfix as [1/20], which should go to urgent > branch, and dropped the performance improvement patch which was introduced > in v12 because I found that does not work with new kernel. > > Overview > -------- > This series rewrites the fprobe on this function-graph. > The purposes of this change are; > > 1) Remove dependency of the rethook from fprobe so that we can reduce > the return hook code and shadow stack. > > 2) Make 'ftrace_regs' the common trace interface for the function > boundary. > > 1) Currently we have 2(or 3) different function return hook codes, > the function-graph tracer and rethook (and legacy kretprobe). > But since this is redundant and needs double maintenance cost, > I would like to unify those. From the user's viewpoint, function- > graph tracer is very useful to grasp the execution path. For this > purpose, it is hard to use the rethook in the function-graph > tracer, but the opposite is possible. (Strictly speaking, kretprobe > can not use it because it requires 'pt_regs' for historical reasons.) > > 2) Now the fprobe provides the 'pt_regs' for its handler, but that is > wrong for the function entry and exit. Moreover, depending on the > architecture, there is no way to accurately reproduce 'pt_regs' > outside of interrupt or exception handlers. This means fprobe should > not use 'pt_regs' because it does not use such exceptions. > (Conversely, kprobe should use 'pt_regs' because it is an abstract > interface of the software breakpoint exception.) > > This series changes fprobe to use function-graph tracer for tracing > function entry and exit, instead of mixture of ftrace and rethook. > Unlike the rethook which is a per-task list of system-wide allocated > nodes, the function graph's ret_stack is a per-task shadow stack. > Thus it does not need to set 'nr_maxactive' (which is the number of > pre-allocated nodes). > Also the handlers will get the 'ftrace_regs' instead of 'pt_regs'. > Since eBPF mulit_kprobe/multi_kretprobe events still use 'pt_regs' as > their register interface, this changes it to convert 'ftrace_regs' to > 'pt_regs'. Of course this conversion makes an incomplete 'pt_regs', > so users must access only registers for function parameters or > return value. > > Design > ------ > Instead of using ftrace's function entry hook directly, the new fprobe > is built on top of the function-graph's entry and return callbacks > with 'ftrace_regs'. > > Since the fprobe requires access to 'ftrace_regs', the architecture > must support CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS and > CONFIG_HAVE_FTRACE_GRAPH_FUNC, which enables to call function-graph > entry callback with 'ftrace_regs', and also > CONFIG_HAVE_FUNCTION_GRAPH_FREGS, which passes the ftrace_regs to > return_to_handler. > > All fprobes share a single function-graph ops (means shares a common > ftrace filter) similar to the kprobe-on-ftrace. This needs another > layer to find corresponding fprobe in the common function-graph > callbacks, but has much better scalability, since the number of > registered function-graph ops is limited. > > In the entry callback, the fprobe runs its entry_handler and saves the > address of 'fprobe' on the function-graph's shadow stack as data. The > return callback decodes the data to get the 'fprobe' address, and runs > the exit_handler. > > The fprobe introduces two hash-tables, one is for entry callback which > searches fprobes related to the given function address passed by entry > callback. The other is for a return callback which checks if the given > 'fprobe' data structure pointer is still valid. Note that it is > possible to unregister fprobe before the return callback runs. Thus > the address validation must be done before using it in the return > callback. > > Download > -------- > This series can be applied against the ftrace/for-next branch in > linux-trace tree. > > This series can also be found below branch. > > https://git.kernel.org/pub/scm/linux/kernel/git/mhiramat/linux.git/log/?h=topic/fprobe-on-fgraph > > Thank you, > > --- hi, I ran the kprobe_multi bench and I'm seeing lower throughput numbers with new fprobe implementation base (bpf-next/master) root@amd:/home/jolsa/bpf-next/tools/testing/selftests/bpf# ./bench -w2 -d5 -a trig-kprobe-multi Setting up benchmark 'trig-kprobe-multi'... Benchmark 'trig-kprobe-multi' started. Iter 0 ( 70.969us): hits 8.530M/s ( 8.530M/prod), drops 0.000M/s, total operations 8.530M/s Iter 1 (-24.985us): hits 8.541M/s ( 8.541M/prod), drops 0.000M/s, total operations 8.541M/s Iter 2 ( -2.976us): hits 8.537M/s ( 8.537M/prod), drops 0.000M/s, total operations 8.537M/s Iter 3 ( -1.502us): hits 8.540M/s ( 8.540M/prod), drops 0.000M/s, total operations 8.540M/s Iter 4 ( 0.392us): hits 8.540M/s ( 8.540M/prod), drops 0.000M/s, total operations 8.540M/s Iter 5 ( 1.308us): hits 8.545M/s ( 8.545M/prod), drops 0.000M/s, total operations 8.545M/s Iter 6 ( 2.783us): hits 8.530M/s ( 8.530M/prod), drops 0.000M/s, total operations 8.530M/s Iter 7 ( -3.775us): hits 8.535M/s ( 8.535M/prod), drops 0.000M/s, total operations 8.535M/s Summary: hits 8.538 ± 0.005M/s ( 8.538M/prod), drops 0.000 ± 0.000M/s, total operations 8.538 ± 0.005M/s your patchset on top of bpf-next/master root@amd:/home/jolsa/bpf-next/tools/testing/selftests/bpf# ./bench -w2 -d5 -a trig-kprobe-multi Setting up benchmark 'trig-kprobe-multi'... Benchmark 'trig-kprobe-multi' started. Iter 0 ( 55.712us): hits 7.224M/s ( 7.224M/prod), drops 0.000M/s, total operations 7.224M/s Iter 1 ( 15.226us): hits 7.222M/s ( 7.222M/prod), drops 0.000M/s, total operations 7.222M/s Iter 2 ( -1.353us): hits 7.221M/s ( 7.221M/prod), drops 0.000M/s, total operations 7.221M/s Iter 3 (-13.029us): hits 7.223M/s ( 7.223M/prod), drops 0.000M/s, total operations 7.223M/s Iter 4 ( 9.111us): hits 7.222M/s ( 7.222M/prod), drops 0.000M/s, total operations 7.222M/s Iter 5 ( -0.106us): hits 7.222M/s ( 7.222M/prod), drops 0.000M/s, total operations 7.222M/s Iter 6 ( 0.734us): hits 7.221M/s ( 7.221M/prod), drops 0.000M/s, total operations 7.221M/s Iter 7 (-10.233us): hits 7.220M/s ( 7.220M/prod), drops 0.000M/s, total operations 7.220M/s Summary: hits 7.221 ± 0.001M/s ( 7.221M/prod), drops 0.000 ± 0.000M/s, total operations 7.221 ± 0.001M/s I did perf profile (attached), but nothing stands out on the first look, I'll try to check on that later this week jirka --- # To display the perf.data header info, please use --header/--header-only options. # # # Total Lost Samples: 0 # # Samples: 42K of event 'cycles:k' # Event count (approx.): 44580427937 # # Overhead Command Shared Object Symbol # ........ ............... .................................................... ........................................................ # 17.47% bench [kernel.vmlinux] [k] find_kallsyms_symbol | ---find_kallsyms_symbol module_address_lookup kallsyms_lookup_buildid kallsyms_lookup print_rec t_show seq_read_iter seq_read vfs_read ksys_read __x64_sys_read x64_sys_call do_syscall_64 entry_SYSCALL_64 read 0 11.71% bench [kernel.vmlinux] [k] kprobe_multi_link_prog_run | ---kprobe_multi_link_prog_run | --11.56%--kprobe_multi_link_handler fprobe_entry function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 7.65% bench [kernel.vmlinux] [k] function_graph_enter_regs | ---function_graph_enter_regs | --7.65%--ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 7.55% bench [kernel.vmlinux] [k] kprobe_multi_link_handler | ---kprobe_multi_link_handler | |--4.35%--function_graph_enter_regs | ftrace_graph_func | 0xffffffffc29850ab | bpf_get_numa_node_id | bpf_prog_d9703036495d54b0_trigger_driver | __bpf_prog_test_run_raw_tp | bpf_prog_test_run_raw_tp | __sys_bpf | __x64_sys_bpf | x64_sys_call | do_syscall_64 | entry_SYSCALL_64 | syscall | bpf_prog_test_run_opts | trigger_producer_batch | 0x7a2a73a94ac3 | --3.20%--fprobe_entry function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 7.50% bench [kernel.vmlinux] [k] fprobe_entry | ---fprobe_entry | --7.50%--function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 7.08% bench bpf_prog_7790468e40a289ea_bench_trigger_kprobe_multi [k] bpf_prog_7790468e40a289ea_bench_trigger_kprobe_multi | ---bpf_prog_7790468e40a289ea_bench_trigger_kprobe_multi | |--4.10%--kprobe_multi_link_handler | fprobe_entry | function_graph_enter_regs | ftrace_graph_func | 0xffffffffc29850ab | bpf_get_numa_node_id | bpf_prog_d9703036495d54b0_trigger_driver | __bpf_prog_test_run_raw_tp | bpf_prog_test_run_raw_tp | __sys_bpf | __x64_sys_bpf | x64_sys_call | do_syscall_64 | entry_SYSCALL_64 | syscall | bpf_prog_test_run_opts | trigger_producer_batch | 0x7a2a73a94ac3 | --2.97%--kprobe_multi_link_prog_run kprobe_multi_link_handler fprobe_entry function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 6.11% bench [kernel.vmlinux] [k] migrate_enable | ---migrate_enable | |--3.37%--kprobe_multi_link_handler | fprobe_entry | function_graph_enter_regs | ftrace_graph_func | 0xffffffffc29850ab | bpf_get_numa_node_id | bpf_prog_d9703036495d54b0_trigger_driver | __bpf_prog_test_run_raw_tp | bpf_prog_test_run_raw_tp | __sys_bpf | __x64_sys_bpf | x64_sys_call | do_syscall_64 | entry_SYSCALL_64 | syscall | bpf_prog_test_run_opts | trigger_producer_batch | 0x7a2a73a94ac3 | --2.73%--kprobe_multi_link_prog_run kprobe_multi_link_handler fprobe_entry function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 5.90% bench [kernel.vmlinux] [k] ftrace_graph_func | ---ftrace_graph_func | --5.75%--0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 4.03% bench [kernel.vmlinux] [k] __rcu_read_lock | ---__rcu_read_lock | --4.00%--kprobe_multi_link_handler fprobe_entry function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 3.61% bench [kernel.vmlinux] [k] bpf_get_numa_node_id | ---bpf_get_numa_node_id __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 2.99% bench [nf_conntrack] [k] 0x000000000002c0b3 | ---0xffffffffc29850b3 bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 2.69% bench bpf_prog_d9703036495d54b0_trigger_driver [k] bpf_prog_d9703036495d54b0_trigger_driver | ---bpf_prog_d9703036495d54b0_trigger_driver | --2.68%--__bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 2.04% bench [kernel.vmlinux] [k] t_start | ---t_start seq_read_iter seq_read vfs_read ksys_read __x64_sys_read x64_sys_call do_syscall_64 entry_SYSCALL_64 read 0 1.90% bench [kernel.vmlinux] [k] srso_untrain_ret | ---srso_untrain_ret | |--0.68%--fprobe_entry | function_graph_enter_regs | ftrace_graph_func | 0xffffffffc29850ab | bpf_get_numa_node_id | bpf_prog_d9703036495d54b0_trigger_driver | __bpf_prog_test_run_raw_tp | bpf_prog_test_run_raw_tp | __sys_bpf | __x64_sys_bpf | x64_sys_call | do_syscall_64 | entry_SYSCALL_64 | syscall | bpf_prog_test_run_opts | trigger_producer_batch | 0x7a2a73a94ac3 | --0.54%--seq_read_iter seq_read vfs_read ksys_read __x64_sys_read x64_sys_call do_syscall_64 entry_SYSCALL_64 read 0 1.14% bench [kernel.vmlinux] [k] __rcu_read_unlock | ---__rcu_read_unlock | --1.12%--kprobe_multi_link_handler fprobe_entry function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 1.02% bench [kernel.vmlinux] [k] migrate_disable | ---migrate_disable kprobe_multi_link_handler fprobe_entry function_graph_enter_regs ftrace_graph_func 0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 0.99% bench [kernel.vmlinux] [k] t_func_next.isra.0 | ---t_func_next.isra.0 | --0.83%--t_start seq_read_iter seq_read vfs_read ksys_read __x64_sys_read x64_sys_call do_syscall_64 entry_SYSCALL_64 read 0 0.78% bench [nf_conntrack] [k] 0x000000000002c0d4 | ---0xffffffffc29850d4 __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 0.75% bench [nf_conntrack] [k] 0x000000000002c0ab | ---0xffffffffc29850ab bpf_get_numa_node_id bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 0.68% bench [nf_conntrack] [k] 0x000000000002c00d | ---0xffffffffc298500d bpf_prog_d9703036495d54b0_trigger_driver __bpf_prog_test_run_raw_tp bpf_prog_test_run_raw_tp __sys_bpf __x64_sys_bpf x64_sys_call do_syscall_64 entry_SYSCALL_64 syscall bpf_prog_test_run_opts trigger_producer_batch 0x7a2a73a94ac3 0.52% bench [kernel.vmlinux] [k] __x86_indirect_thunk_array | ---__x86_indirect_thunk_array 0.42% bench [nf_conntrack] [k] 0x000000000002c005 0.38% bench [kernel.vmlinux] [k] __sys_bpf 0.37% bench [kernel.vmlinux] [k] srso_return_thunk 0.24% bench [nf_conntrack] [k] 0x000000000002c001 0.23% bench [kernel.vmlinux] [k] read_hpet 0.21% bench [kernel.vmlinux] [k] entry_SYSCALL_64 0.19% bench [kernel.vmlinux] [k] bpf_prog_test_run_raw_tp 0.17% bench [nf_conntrack] [k] 0x000000000002c000 0.15% bench [kernel.vmlinux] [k] _copy_from_user 0.15% bench [kernel.vmlinux] [k] syscall_exit_to_user_mode 0.14% bench [nf_conntrack] [k] 0x000000000002c027 0.14% bench [nf_conntrack] [k] 0x000000000002c0ca 0.13% swapper [kernel.vmlinux] [k] read_hpet 0.12% bench [kernel.vmlinux] [k] _copy_to_user 0.12% bench [nf_conntrack] [k] 0x000000000002c086 0.11% bench [nf_conntrack] [k] 0x000000000002c0c0 0.11% bench [kernel.vmlinux] [k] __fdget 0.11% bench bench [.] bpf_prog_test_run_opts 0.11% bench [kernel.vmlinux] [k] __bpf_prog_test_run_raw_tp 0.10% bench [nf_conntrack] [k] 0x000000000002c059 0.09% bench [kernel.vmlinux] [k] get_symbol_offset 0.09% bench [kernel.vmlinux] [k] do_syscall_64 0.08% bench [kernel.vmlinux] [k] __bpf_prog_get 0.08% bench [kernel.vmlinux] [k] memchr_inv 0.07% bench [kernel.vmlinux] [k] syscall_return_via_sysret 0.06% bench [kernel.vmlinux] [k] x64_sys_call 0.06% bench [kernel.vmlinux] [k] rep_movs_alternative 0.06% bench [kernel.vmlinux] [k] ftrace_replace_code 0.04% bench [kernel.vmlinux] [k] __x64_sys_bpf 0.04% bench [kernel.vmlinux] [k] get_symbol_pos 0.04% bench bench [.] sys_bpf 0.04% bench [kernel.vmlinux] [k] vsnprintf 0.04% bench [kernel.vmlinux] [k] number 0.04% bench bench [.] ptr_to_u64 0.04% bench [kernel.vmlinux] [k] kfree 0.03% bench [kernel.vmlinux] [k] seq_read_iter 0.03% bench [kernel.vmlinux] [k] security_bpf 0.03% bench libc.so.6 [.] syscall 0.03% bench [kernel.vmlinux] [k] kallsyms_expand_symbol.constprop.0 0.03% kworker/2:2-eve [kernel.vmlinux] [k] memcpy_orig 0.03% swapper [amdgpu] [k] 0x0000000000002f75 0.03% bench bench [.] trigger_producer_batch 0.03% bench [kernel.vmlinux] [k] __bpf_prog_put 0.02% bench [kernel.vmlinux] [k] seq_printf 0.02% bench [kernel.vmlinux] [k] smp_call_function_many_cond 0.02% kworker/2:2-eve [amdgpu] [k] 0x0000000000002f75 0.02% bench [kernel.vmlinux] [k] __check_object_size 0.02% bench [kernel.vmlinux] [k] ftrace_test_record 0.02% bench [kernel.vmlinux] [k] seq_write 0.02% bench [kernel.vmlinux] [k] ftrace_lookup_ip 0.02% bench [kernel.vmlinux] [k] t_show 0.02% bench [kernel.vmlinux] [k] print_rec 0.02% swapper [kernel.vmlinux] [k] io_idle 0.01% bench [amdgpu] [k] 0x0000000000002f75 0.01% kworker/2:2-eve [kernel.vmlinux] [k] memcmp 0.01% bench [kernel.vmlinux] [k] format_decode 0.01% bench [kernel.vmlinux] [k] memcpy_orig 0.01% bench [kernel.vmlinux] [k] __ftrace_hash_rec_update.part.0 0.01% bench [kernel.vmlinux] [k] mod_find 0.01% swapper [kernel.vmlinux] [k] srso_untrain_ret 0.01% bench [kernel.vmlinux] [k] kallsyms_lookup_buildid 0.01% bench [kernel.vmlinux] [k] __seq_puts 0.01% bench [kernel.vmlinux] [k] native_read_msr 0.01% bench [kernel.vmlinux] [k] check_stack_object 0.01% bench [kernel.vmlinux] [k] fput 0.01% bench bench [.] libbpf_err_errno 0.01% bench [kernel.vmlinux] [k] fpregs_assert_state_consistent 0.01% bench [kernel.vmlinux] [k] ftrace_check_record 0.01% bench [kernel.vmlinux] [k] t_next 0.01% bench [kernel.vmlinux] [k] strlen 0.01% swapper [kernel.vmlinux] [k] native_sched_clock 0.01% bench [kernel.vmlinux] [k] __sysvec_apic_timer_interrupt 0.01% bench libc.so.6 [.] 0x00000000001a0f10 0.01% bench [kernel.vmlinux] [k] native_write_msr 0.01% bench [kernel.vmlinux] [k] __update_load_avg_se 0.01% bench [kernel.vmlinux] [k] amd_pmu_addr_offset 0.01% bench [kernel.vmlinux] [k] module_address_lookup 0.01% bench [kernel.vmlinux] [k] ftrace_rec_iter_next 0.01% bench [kernel.vmlinux] [k] string 0.01% bench [kernel.vmlinux] [k] asm_exc_page_fault 0.01% swapper [kernel.vmlinux] [k] __get_next_timer_interrupt 0.01% bench [kernel.vmlinux] [k] clear_page_rep 0.01% kworker/2:2-eve [kernel.vmlinux] [k] delay_halt_mwaitx 0.01% bench [kernel.vmlinux] [k] __handle_mm_fault 0.01% swapper [kernel.vmlinux] [k] ktime_get 0.01% kworker/2:2-eve [kernel.vmlinux] [k] clear_page_rep 0.01% swapper [kernel.vmlinux] [k] menu_select 0.00% perf [kernel.vmlinux] [k] clear_page_rep 0.00% bench [kernel.vmlinux] [k] ftrace_rec_iter_record 0.00% bench [kernel.vmlinux] [k] account_process_tick 0.00% bench [kernel.vmlinux] [k] __update_load_avg_cfs_rq 0.00% bench [kernel.vmlinux] [k] update_curr 0.00% bench [kernel.vmlinux] [k] do_sync_core 0.00% bench [kernel.vmlinux] [k] kallsyms_lookup 0.00% bench [kernel.vmlinux] [k] memset_orig 0.00% kworker/6:1-eve [kernel.vmlinux] [k] memcpy_orig 0.00% perf [kernel.vmlinux] [k] rep_movs_alternative 0.00% sshd [kernel.vmlinux] [k] p4d_offset 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000002f75 0.00% bench [kernel.vmlinux] [k] __mod_memcg_lruvec_state 0.00% swapper [kernel.vmlinux] [k] _raw_spin_unlock 0.00% swapper [kernel.vmlinux] [k] cpuidle_enter_state 0.00% swapper [kernel.vmlinux] [k] rb_erase 0.00% containerd [kernel.vmlinux] [k] read_hpet 0.00% kworker/6:1-eve [kernel.vmlinux] [k] memcmp 0.00% swapper [kernel.vmlinux] [k] irq_entries_start 0.00% swapper [kernel.vmlinux] [k] do_idle 0.00% kworker/2:2-eve [kernel.vmlinux] [k] read_hpet 0.00% swapper [kernel.vmlinux] [k] update_rq_clock 0.00% swapper [kernel.vmlinux] [k] psi_group_change 0.00% swapper [kernel.vmlinux] [k] sched_clock_cpu 0.00% swapper [kernel.vmlinux] [k] do_sync_core 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ec11 0.00% swapper [kernel.vmlinux] [k] _raw_spin_lock_irqsave 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_call_function_single 0.00% swapper [kernel.vmlinux] [k] update_sd_lb_stats.constprop.0 0.00% perf [kernel.vmlinux] [k] __filemap_get_folio 0.00% swapper [kernel.vmlinux] [k] irqentry_enter 0.00% bench [kernel.vmlinux] [k] amd_pmu_test_overflow_topbit 0.00% bench [amdgpu] [k] 0x0000000000124f4c 0.00% bench [kernel.vmlinux] [k] update_rq_clock 0.00% bench [kernel.vmlinux] [k] perf_event_task_tick 0.00% bench [amdgpu] [k] 0x0000000000126f3a 0.00% bench [kernel.vmlinux] [k] sched_balance_update_blocked_averages 0.00% bench [kernel.vmlinux] [k] hrtimer_active 0.00% bench [kernel.vmlinux] [k] _copy_to_iter 0.00% bench [kernel.vmlinux] [k] _raw_spin_lock_irqsave 0.00% bench [kernel.vmlinux] [k] calc_global_load_tick 0.00% bench [kernel.vmlinux] [k] ftrace_shutdown.part.0 0.00% bench libc.so.6 [.] 0x00000000001a0fb5 0.00% bench [kernel.vmlinux] [k] task_tick_mm_cid 0.00% bench libc.so.6 [.] 0x00000000001a0fba 0.00% bench [kernel.vmlinux] [k] sched_clock_cpu 0.00% bench [kernel.vmlinux] [k] update_cfs_group 0.00% bench [kernel.vmlinux] [k] irq_exit_rcu 0.00% bench [kernel.vmlinux] [k] native_sched_clock 0.00% bench [kernel.vmlinux] [k] sched_tick 0.00% bench [kernel.vmlinux] [k] sized_strscpy 0.00% bench [kernel.vmlinux] [k] task_tick_fair 0.00% bench [nf_conntrack] [k] 0x000000000002c036 0.00% bench [kernel.vmlinux] [k] perf_adjust_freq_unthr_context 0.00% bench [kernel.vmlinux] [k] idle_cpu 0.00% bench [kernel.vmlinux] [k] bpf_check_uarg_tail_zero 0.00% swapper [kernel.vmlinux] [k] __hrtimer_next_event_base 0.00% swapper [kernel.vmlinux] [k] native_write_msr 0.00% bench [kernel.vmlinux] [k] __module_address.part.0 0.00% bench [kernel.vmlinux] [k] lru_gen_add_folio 0.00% bench [kernel.vmlinux] [k] set_pte_range 0.00% bench [kernel.vmlinux] [k] pfn_pte 0.00% swapper [kernel.vmlinux] [k] error_entry 0.00% swapper [kernel.vmlinux] [k] idle_cpu 0.00% bench [kernel.vmlinux] [k] unmap_page_range 0.00% bench [kernel.vmlinux] [k] __count_memcg_events 0.00% bench [kernel.vmlinux] [k] __irq_exit_rcu 0.00% perf [kernel.vmlinux] [k] __perf_event_enable 0.00% bench [kernel.vmlinux] [k] seq_putc 0.00% perf [kernel.vmlinux] [k] srso_untrain_ret 0.00% swapper [kernel.vmlinux] [k] __irq_exit_rcu 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __kmalloc_cache_noprof 0.00% bench libc.so.6 [.] 0x00000000001a0bd7 0.00% perf [kernel.vmlinux] [k] _raw_spin_lock 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __bio_split_to_limits 0.00% bench [kernel.vmlinux] [k] do_anonymous_page 0.00% perf [kernel.vmlinux] [k] __es_insert_extent 0.00% bench [kernel.vmlinux] [k] __mem_cgroup_charge 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __wait_for_common 0.00% bench [kernel.vmlinux] [k] __folio_throttle_swaprate 0.00% swapper [kernel.vmlinux] [k] __switch_to_asm 0.00% swapper [kernel.vmlinux] [k] add_interrupt_randomness 0.00% kworker/2:2-eve [kernel.vmlinux] [k] srso_untrain_ret 0.00% kworker/2:2-eve [kernel.vmlinux] [k] memset_orig 0.00% kworker/u65:2-e [kernel.vmlinux] [k] percpu_counter_add_batch 0.00% swapper [kernel.vmlinux] [k] tick_irq_enter 0.00% swapper [kernel.vmlinux] [k] tick_check_broadcast_expired 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000045d28e 0.00% perf [kernel.vmlinux] [k] __lruvec_stat_mod_folio 0.00% bench [kernel.vmlinux] [k] kernfs_fop_read_iter 0.00% bench [kernel.vmlinux] [k] __mod_zone_page_state 0.00% kworker/2:2-eve [kernel.vmlinux] [k] delay_halt 0.00% perf [kernel.vmlinux] [k] __mod_memcg_lruvec_state 0.00% perf [kernel.vmlinux] [k] xas_load 0.00% perf [kernel.vmlinux] [k] workingset_update_node 0.00% perf [kernel.vmlinux] [k] _raw_spin_unlock 0.00% swapper [kernel.vmlinux] [k] __handle_irq_event_percpu 0.00% kworker/2:2-eve [kernel.vmlinux] [k] post_alloc_hook 0.00% bench [kernel.vmlinux] [k] refill_obj_stock 0.00% perf [kernel.vmlinux] [k] __alloc_pages_noprof 0.00% swapper [kernel.vmlinux] [k] cpuidle_enter 0.00% swapper [kernel.vmlinux] [k] sched_balance_domains 0.00% perf [kernel.vmlinux] [k] __handle_mm_fault 0.00% swapper [kernel.vmlinux] [k] ct_kernel_enter.constprop.0 0.00% bench [kernel.vmlinux] [k] mab_mas_cp 0.00% swapper [amdgpu] [k] 0x00000000004619a7 0.00% swapper [kernel.vmlinux] [k] tick_nohz_irq_exit 0.00% perf [kernel.vmlinux] [k] __memcg_slab_post_alloc_hook 0.00% bench [kernel.vmlinux] [k] __alloc_pages_noprof 0.00% swapper [kernel.vmlinux] [k] irq_exit_rcu 0.00% bench [kernel.vmlinux] [k] memcmp 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_call_function 0.00% perf [kernel.vmlinux] [k] perf_poll 0.00% swapper [kernel.vmlinux] [k] enqueue_task_fair 0.00% bench [kernel.vmlinux] [k] __rmqueue_pcplist 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __iommu_map 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000992cd2 0.00% kworker/6:1-eve [amdgpu] [k] 0x00000000004b2ade 0.00% swapper [kernel.vmlinux] [k] wakeup_preempt 0.00% migration/11 [kernel.vmlinux] [k] enqueue_task 0.00% swapper [kernel.vmlinux] [k] hrtimer_update_next_event 0.00% swapper [kernel.vmlinux] [k] native_apic_mem_eoi 0.00% bench [kernel.vmlinux] [k] is_sync_callback_calling_insn 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000992cb3 0.00% systemd-network libc.so.6 [.] clock_gettime 0.00% kworker/u65:4-e [kernel.vmlinux] [k] filemap_get_folios_tag 0.00% kworker/u65:4-e [kernel.vmlinux] [k] fscrypt_mergeable_bio_bh 0.00% swapper [kernel.vmlinux] [k] hrtimer_start_range_ns 0.00% multipathd [kernel.vmlinux] [k] cpuacct_charge 0.00% swapper [kernel.vmlinux] [k] hrtimer_get_next_event 0.00% kworker/12:1-ev [amdgpu] [k] 0x00000000004b09fc 0.00% kworker/12:1-ev [kernel.vmlinux] [k] memcpy_orig 0.00% kworker/u65:4-e [kernel.vmlinux] [k] wbt_track 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __lruvec_stat_mod_folio 0.00% kworker/12:1-ev [amdgpu] [k] 0x0000000000002f75 0.00% bench [kernel.vmlinux] [k] do_jit 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __folio_start_writeback 0.00% bench [kernel.vmlinux] [k] uncharge_folio 0.00% perf [kernel.vmlinux] [k] do_fault 0.00% perf [kernel.vmlinux] [k] __kmalloc_noprof 0.00% containerd [kernel.vmlinux] [k] blkcg_maybe_throttle_current 0.00% kworker/12:1-ev [amdgpu] [k] 0x0000000000523901 0.00% kworker/12:1-ev [amdgpu] [k] 0x0000000000538414 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000465f98 0.00% irqbalance [kernel.vmlinux] [k] format_decode 0.00% swapper [kernel.vmlinux] [k] handle_edge_irq 0.00% kworker/12:1-ev [kernel.vmlinux] [k] srso_untrain_ret 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __find_get_block 0.00% systemd-journal [kernel.vmlinux] [k] __put_user_nocheck_4 0.00% systemd-journal libsystemd-shared-249.so [.] 0x00000000001c9c3e 0.00% swapper [kernel.vmlinux] [k] local_touch_nmi 0.00% bench [kernel.vmlinux] [k] __mem_cgroup_uncharge_folios 0.00% swapper [kernel.vmlinux] [k] _raw_spin_lock 0.00% swapper [kernel.vmlinux] [k] select_task_rq_fair 0.00% swapper [kernel.vmlinux] [k] need_update 0.00% bench [kernel.vmlinux] [k] rcu_segcblist_enqueue 0.00% swapper [kernel.vmlinux] [k] tick_nohz_get_sleep_length 0.00% containerd [kernel.vmlinux] [k] _raw_spin_unlock 0.00% containerd [kernel.vmlinux] [k] ep_autoremove_wake_function 0.00% kworker/u66:1-e [kernel.vmlinux] [k] __percpu_counter_sum 0.00% swapper [kernel.vmlinux] [k] ktime_get_mono_fast_ns 0.00% swapper [kernel.vmlinux] [k] sysvec_apic_timer_interrupt 0.00% swapper [kernel.vmlinux] [k] _raw_spin_unlock_irqrestore 0.00% migration/10 [kernel.vmlinux] [k] dequeue_entity 0.00% swapper [kernel.vmlinux] [k] handle_irq_event 0.00% perf [kernel.vmlinux] [k] srso_return_thunk 0.00% perf [kernel.vmlinux] [k] xas_find_conflict 0.00% perf-exec [kernel.vmlinux] [k] mas_next_slot 0.00% swapper [kernel.vmlinux] [k] fetch_next_timer_interrupt.constprop.0 0.00% bench [kernel.vmlinux] [k] get_page_from_freelist 0.00% swapper [amdgpu] [k] 0x0000000000125079 0.00% kworker/u65:4-e [kernel.vmlinux] [k] sched_balance_newidle 0.00% migration/6 [kernel.vmlinux] [k] raw_spin_rq_unlock 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000444a05 0.00% swapper [kernel.vmlinux] [k] fetch_pte 0.00% perf [kernel.vmlinux] [k] __pte_offset_map 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000534985 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000003ac5 0.00% bench [kernel.vmlinux] [k] perf_iterate_sb 0.00% swapper [kernel.vmlinux] [k] get_cpu_device 0.00% perf [kernel.vmlinux] [k] xas_start 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000098fea5 0.00% perf [kernel.vmlinux] [k] mark_buffer_dirty 0.00% containerd containerd [.] runtime.cleantimers 0.00% bench [kernel.vmlinux] [k] __cond_resched 0.00% kworker/2:2-eve [drm_kms_helper] [k] 0x00000000000008eb 0.00% kworker/2:2-eve [raid6_pq] [k] 0x000000000000c24b 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ea79 0.00% perf [kernel.vmlinux] [k] cpu_util 0.00% swapper [kernel.vmlinux] [k] asm_common_interrupt 0.00% perf [kernel.vmlinux] [k] __filemap_add_folio 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000007d1eab 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051a45f 0.00% perf [kernel.vmlinux] [k] ext4_es_lookup_extent 0.00% swapper [kernel.vmlinux] [k] __flush_smp_call_function_queue 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000519af6 0.00% swapper [kernel.vmlinux] [k] refresh_cpu_vm_stats 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000477759 0.00% perf-exec [kernel.vmlinux] [k] kmem_cache_free 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000539303 0.00% kworker/2:2-eve [drm] [k] 0x0000000000020c1c 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_apic_timer_interrupt 0.00% bench [kernel.vmlinux] [k] alloc_pages_mpol_noprof 0.00% bench [kernel.vmlinux] [k] filemap_map_pages 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000477767 0.00% swapper [kernel.vmlinux] [k] update_load_avg 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000045d3d6 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051d905 0.00% kworker/2:2-eve [drm_kms_helper] [k] 0x0000000000004ae8 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __mod_memcg_lruvec_state 0.00% kworker/u65:2-e [kernel.vmlinux] [k] bio_associate_blkg 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004916dd 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000465ecb 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004b780a 0.00% swapper [kernel.vmlinux] [k] sysvec_call_function 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051e764 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ed72 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051c7bf 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ebe0 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000523fba 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000832660 0.00% swapper [kernel.vmlinux] [k] acpi_idle_do_entry 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097ddfc 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000045d334 0.00% kworker/u65:2-e [kernel.vmlinux] [k] soft_cursor 0.00% perf [kernel.vmlinux] [k] sched_balance_rq 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __fprop_add_percpu 0.00% swapper [kernel.vmlinux] [k] rb_next 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000977ece 0.00% perf [kernel.vmlinux] [k] __folio_mark_dirty 0.00% perf [kernel.vmlinux] [k] filemap_alloc_folio_noprof 0.00% perf [kernel.vmlinux] [k] get_mem_cgroup_from_mm 0.00% perf [kernel.vmlinux] [k] ext4_da_reserve_space 0.00% perf [kernel.vmlinux] [k] read_hpet 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051cd3e 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000532f0f 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __mod_node_page_state 0.00% systemd-journal [kernel.vmlinux] [k] mod_objcg_state 0.00% jbd2/dm-0-8 [kernel.vmlinux] [k] dm_submit_bio 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000517b6d 0.00% kcompactd0 [kernel.vmlinux] [k] idle_cpu 0.00% kworker/2:2-eve [kernel.vmlinux] [k] free_tail_page_prepare 0.00% swapper [amdgpu] [k] 0x00000000002f7f65 0.00% swapper [kernel.vmlinux] [k] tick_program_event 0.00% containerd [kernel.vmlinux] [k] psi_group_change 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000053776f 0.00% swapper [kernel.vmlinux] [k] srso_return_thunk 0.00% perf [kernel.vmlinux] [k] ext4_claim_free_clusters 0.00% perf [kernel.vmlinux] [k] generic_perform_write 0.00% perf [kernel.vmlinux] [k] mem_cgroup_commit_charge 0.00% perf [kernel.vmlinux] [k] file_modified 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051e779 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000053974c 0.00% perf [kernel.vmlinux] [k] ext4_da_write_begin 0.00% perf [kernel.vmlinux] [k] filemap_add_folio 0.00% perf libc.so.6 [.] write 0.00% perf [kernel.vmlinux] [k] __es_remove_extent 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000052244a 0.00% swapper [kernel.vmlinux] [k] ct_nmi_exit 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000034bee5 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __rcu_read_lock 0.00% perf [kernel.vmlinux] [k] __dquot_alloc_space 0.00% swapper [kernel.vmlinux] [k] native_read_msr 0.00% perf [kernel.vmlinux] [k] folio_unlock 0.00% perf [kernel.vmlinux] [k] radix_tree_node_ctor 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000006c5b82 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000527259 0.00% perf [kernel.vmlinux] [k] __mod_node_page_state 0.00% perf [kernel.vmlinux] [k] __radix_tree_lookup 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __rcu_read_unlock 0.00% perf [kernel.vmlinux] [k] ext4_get_reserved_space 0.00% swapper [kernel.vmlinux] [k] hrtimer_cancel 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004ba8ce 0.00% swapper [amdgpu] [k] 0x0000000000126f3a 0.00% perf perf [.] perf_cpu_map__idx 0.00% swapper [kernel.vmlinux] [k] sched_balance_update_blocked_averages 0.00% systemd-journal [kernel.vmlinux] [k] do_timerfd_settime 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097ddc5 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097e5a4 0.00% perf [kernel.vmlinux] [k] folio_batch_move_lru 0.00% swapper [amdgpu] [k] 0x000000000046f8a5 0.00% perf [kernel.vmlinux] [k] xas_nomem 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051d24c 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097df7c 0.00% tuned [kernel.vmlinux] [k] psi_task_switch 0.00% swapper [kernel.vmlinux] [k] __dequeue_entity 0.00% swapper [amdgpu] [k] 0x00000000006e7c5d 0.00% swapper [kernel.vmlinux] [k] rcu_preempt_deferred_qs 0.00% kworker/7:1-eve [kernel.vmlinux] [k] __free_one_page 0.00% swapper [kernel.vmlinux] [k] memchr_inv 0.00% perf [kernel.vmlinux] [k] psi_task_switch 0.00% swapper [kernel.vmlinux] [k] _nohz_idle_balance.isra.0 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000009a6fbd 0.00% swapper [kernel.vmlinux] [k] acpi_idle_enter 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000456a9f 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004ac642 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000477888 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000529a52 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004b36d2 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000525b33 0.00% swapper [kernel.vmlinux] [k] tmigr_active_up 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000452444 0.00% swapper [amdgpu] [k] 0x000000000046f865 0.00% swapper [kernel.vmlinux] [k] next_timer_interrupt 0.00% kworker/6:1-eve [kernel.vmlinux] [k] __free_pages_ok 0.00% perf [kernel.vmlinux] [k] exc_page_fault 0.00% swapper [kernel.vmlinux] [k] __schedule 0.00% irqbalance [kernel.vmlinux] [k] __kmalloc_cache_noprof 0.00% swapper [kernel.vmlinux] [k] cpuidle_not_available 0.00% containerd [kernel.vmlinux] [k] pick_next_task_fair 0.00% kworker/6:1-eve [amdgpu] [k] 0x00000000009a70b7 0.00% swapper [kernel.vmlinux] [k] tick_nohz_next_event 0.00% swapper [kernel.vmlinux] [k] tick_check_oneshot_broadcast_this_cpu 0.00% swapper [kernel.vmlinux] [k] irq_chip_ack_parent 0.00% irqbalance [kernel.vmlinux] [k] show_stat 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004b8295 0.00% swapper [amdgpu] [k] 0x000000000045d385 0.00% swapper [kernel.vmlinux] [k] tick_nohz_idle_stop_tick 0.00% containerd [kernel.vmlinux] [k] __schedule 0.00% swapper [kernel.vmlinux] [k] switch_mm_irqs_off 0.00% swapper [amdgpu] [k] 0x0000000000461a21 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000051d7c2 0.00% swapper [kernel.vmlinux] [k] try_to_wake_up 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051e7cf 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ed4e 0.00% perf [kernel.vmlinux] [k] perf_mmap_to_page 0.00% migration/7 [kernel.vmlinux] [k] update_sd_lb_stats.constprop.0 0.00% irqbalance [kernel.vmlinux] [k] __rcu_read_lock 0.00% migration/3 [kernel.vmlinux] [k] __update_load_avg_cfs_rq 0.00% swapper [br_netfilter] [k] 0x0000000000000775 0.00% bench [kernel.vmlinux] [k] remove_vma 0.00% swapper [kernel.vmlinux] [k] sched_clock_noinstr 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000529169 0.00% swapper [amdgpu] [k] 0x000000000000df07 0.00% swapper [kernel.vmlinux] [k] __switch_to 0.00% swapper [kernel.vmlinux] [k] place_entity 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000097dd9a 0.00% swapper [amdgpu] [k] 0x00000000006e7c5a 0.00% irqbalance [kernel.vmlinux] [k] __rcu_read_unlock 0.00% migration/6 [kernel.vmlinux] [k] enqueue_entity 0.00% swapper [kernel.vmlinux] [k] __update_load_avg_cfs_rq 0.00% perf [kernel.vmlinux] [k] event_function 0.00% perf [kernel.vmlinux] [k] do_user_addr_fault 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064eaee 0.00% migration/4 [kernel.vmlinux] [k] __update_load_avg_cfs_rq 0.00% swapper [amdgpu] [k] 0x00000000006e79ff 0.00% containerd [kernel.vmlinux] [k] schedule_hrtimeout_range_clock 0.00% irqbalance [kernel.vmlinux] [k] security_file_open 0.00% perf [kernel.vmlinux] [k] _find_next_and_bit 0.00% multipathd [kernel.vmlinux] [k] __rseq_handle_notify_resume 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_r12 0.00% swapper [kernel.vmlinux] [k] rcu_note_context_switch 0.00% bench [kernel.vmlinux] [k] free_unref_folios 0.00% migration/0 [kernel.vmlinux] [k] migration_cpu_stop 0.00% migration/1 [kernel.vmlinux] [k] rcu_note_context_switch 0.00% kworker/u66:1-e [kernel.vmlinux] [k] n_tty_receive_buf2 0.00% swapper [kernel.vmlinux] [k] __common_interrupt 0.00% swapper [drm] [k] 0x000000000003d658 0.00% migration/5 [kernel.vmlinux] [k] psi_group_change 0.00% swapper [kernel.vmlinux] [k] clockevents_program_event 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_array 0.00% perf [kernel.vmlinux] [k] set_pte_range 0.00% perf [kernel.vmlinux] [k] __schedule 0.00% swapper [kernel.vmlinux] [k] __update_load_avg_se 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_reschedule_ipi 0.00% swapper [kernel.vmlinux] [k] local_clock_noinstr 0.00% swapper [kernel.vmlinux] [k] tick_nohz_stop_idle 0.00% swapper [amdgpu] [k] 0x000000000044a039 0.00% kworker/4:1-mm_ [kernel.vmlinux] [k] native_queued_spin_lock_slowpath 0.00% rcu_preempt [kernel.vmlinux] [k] __update_idle_core 0.00% rcu_preempt [kernel.vmlinux] [k] pick_next_task_fair 0.00% rcu_preempt [kernel.vmlinux] [k] prepare_to_swait_event 0.00% rcu_preempt [kernel.vmlinux] [k] sched_balance_rq 0.00% rcu_preempt [kernel.vmlinux] [k] update_rq_clock 0.00% rs:main Q:Reg [kernel.vmlinux] [k] ext4_inode_csum 0.00% swapper [kernel.vmlinux] [k] tmigr_inactive_up 0.00% swapper [kernel.vmlinux] [k] hrtimer_forward 0.00% swapper [kernel.vmlinux] [k] cpuidle_reflect 0.00% perf [kernel.vmlinux] [k] percpu_counter_add_batch 0.00% swapper [amdgpu] [k] 0x000000000031e745 0.00% swapper [kernel.vmlinux] [k] sched_idle_set_state 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000045675c 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000051cbb1 0.00% swapper [amdgpu] [k] 0x000000000002dce5 0.00% systemd-journal [kernel.vmlinux] [k] __ext4_ext_check 0.00% swapper [kernel.vmlinux] [k] sched_core_idle_cpu 0.00% swapper [drm] [k] 0x000000000003de52 0.00% swapper [kernel.vmlinux] [k] __msecs_to_jiffies 0.00% swapper [amdgpu] [k] 0x000000000046f9c4 0.00% swapper [kernel.vmlinux] [k] call_cpuidle 0.00% swapper [kernel.vmlinux] [k] cpuidle_governor_latency_req 0.00% kworker/7:1-eve [kernel.vmlinux] [k] __switch_to_asm 0.00% dockerd [kernel.vmlinux] [k] apparmor_file_permission 0.00% swapper [amdgpu] [k] 0x0000000000126f37 0.00% swapper [amdgpu] [k] 0x00000000003218b6 0.00% swapper [amdgpu] [k] 0x00000000006e8d65 0.00% swapper [amdgpu] [k] 0x000000000096f62d 0.00% systemd-journal [kernel.vmlinux] [k] __es_tree_search.isra.0 0.00% swapper [kernel.vmlinux] [k] update_group_capacity 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000005225c3 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000990c4f 0.00% swapper [kernel.vmlinux] [k] timer_base_try_to_set_idle 0.00% swapper [kernel.vmlinux] [k] ct_idle_exit 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_rbx 0.00% swapper [kernel.vmlinux] [k] fast_mix 0.00% swapper [kernel.vmlinux] [k] sched_clock_idle_sleep_event 0.00% swapper [kernel.vmlinux] [k] touch_softlockup_watchdog_sched 0.00% containerd [kernel.vmlinux] [k] psi_task_switch 0.00% swapper [amdgpu] [k] 0x0000000000123820 0.00% swapper [kernel.vmlinux] [k] sched_clock 0.00% swapper [amdgpu] [k] 0x00000000002f46a5 0.00% swapper [amdgpu] [k] 0x0000000000449ff5 0.00% containerd [kernel.vmlinux] [k] _copy_to_user 0.00% swapper [kernel.vmlinux] [k] enqueue_task 0.00% swapper [amdgpu] [k] 0x000000000096f625 0.00% swapper [amdgpu] [k] 0x00000000006e7c66 0.00% swapper [kernel.vmlinux] [k] hrtimer_interrupt 0.00% swapper [kernel.vmlinux] [k] irq_enter_rcu 0.00% perf [kernel.vmlinux] [k] _raw_spin_lock_irqsave 0.00% perf [kernel.vmlinux] [k] poll_freewait 0.00% rcu_preempt [kernel.vmlinux] [k] __rcu_read_lock 0.00% swapper [kernel.vmlinux] [k] wake_q_add 0.00% kworker/7:1H-kb [kernel.vmlinux] [k] __switch_to_asm 0.00% swapper [amdgpu] [k] 0x00000000002f8058 0.00% swapper [kernel.vmlinux] [k] sched_clock_tick 0.00% swapper [kernel.vmlinux] [k] slab_update_freelist.constprop.0.isra.0 0.00% swapper [kernel.vmlinux] [k] tick_nohz_get_next_hrtimer 0.00% swapper [kernel.vmlinux] [k] ct_kernel_enter_state 0.00% swapper [kernel.vmlinux] [k] hrtimer_next_event_without 0.00% swapper [kernel.vmlinux] [k] resched_curr 0.00% swapper [amdgpu] [k] 0x0000000000003a55 0.00% swapper [amdgpu] [k] 0x000000000045d3df 0.00% swapper [amdgpu] [k] 0x000000000098aeb9 0.00% containerd [kernel.vmlinux] [k] __switch_to 0.00% swapper [kernel.vmlinux] [k] rcu_sched_clock_irq 0.00% kworker/9:1-eve [kernel.vmlinux] [k] sched_clock_cpu 0.00% swapper [amdgpu] [k] 0x0000000000124ef5 0.00% containerd [kernel.vmlinux] [k] update_curr 0.00% swapper [drm] [k] 0x0000000000035f4e 0.00% swapper [kernel.vmlinux] [k] tmigr_update_events 0.00% swapper [kernel.vmlinux] [k] profile_pc 0.00% swapper [nvme] [k] 0x0000000000004315 0.00% swapper [kernel.vmlinux] [k] raw_spin_rq_unlock 0.00% containerd [kernel.vmlinux] [k] posix_get_monotonic_timespec 0.00% swapper [amdgpu] [k] 0x0000000000002f4b 0.00% swapper [amdgpu] [k] 0x000000000034bf25 0.00% perf [kernel.vmlinux] [k] get_page_from_freelist 0.00% rcu_preempt [kernel.vmlinux] [k] __mod_timer 0.00% perf [kernel.vmlinux] [k] mutex_unlock 0.00% swapper [amdgpu] [k] 0x0000000000460f85 0.00% swapper [amdgpu] [k] 0x0000000000977ec5 0.00% swapper [amdgpu] [k] 0x0000000000977ed6 0.00% swapper [kernel.vmlinux] [k] sched_clock_idle_wakeup_event 0.00% swapper [amdgpu] [k] 0x0000000000124016 0.00% swapper [kernel.vmlinux] [k] pick_next_task_fair 0.00% swapper [raid6_pq] [k] 0x000000000001acf5 0.00% swapper [amdgpu] [k] 0x0000000000977ec6 0.00% systemd-journal [kernel.vmlinux] [k] security_inode_setattr 0.00% swapper [kernel.vmlinux] [k] __sysvec_apic_timer_interrupt 0.00% swapper [amdgpu] [k] 0x0000000000461985 0.00% swapper [amdgpu] [k] 0x0000000000123825 0.00% swapper [amdgpu] [k] 0x0000000000126ee9 0.00% swapper [amdgpu] [k] 0x000000000096f6cd 0.00% perf [kernel.vmlinux] [k] do_syscall_64 0.00% systemd-journal [kernel.vmlinux] [k] unmap_mapping_range 0.00% systemd-journal libc.so.6 [.] 0x0000000000090a64 0.00% containerd [kernel.vmlinux] [k] dequeue_entity 0.00% systemd-journal [kernel.vmlinux] [k] dax_layout_busy_page 0.00% swapper [kernel.vmlinux] [k] apic_ack_irq 0.00% swapper [drm] [k] 0x000000000003d83b 0.00% containerd [kernel.vmlinux] [k] __futex_queue 0.00% swapper [kernel.vmlinux] [k] complete_all 0.00% swapper [amdgpu] [k] 0x000000000098b0d6 0.00% swapper [kernel.vmlinux] [k] can_stop_idle_tick 0.00% swapper [amdgpu] [k] 0x000000000000dee5 0.00% swapper [amdgpu] [k] 0x000000000044ecef 0.00% swapper [amdgpu] [k] 0x000000000000dee6 0.00% migration/14 [kernel.vmlinux] [k] update_curr_common 0.00% rs:main Q:Reg [kernel.vmlinux] [k] ext4_get_group_desc 0.00% swapper [kernel.vmlinux] [k] tick_do_update_jiffies64 0.00% swapper [kernel.vmlinux] [k] tmigr_cpu_new_timer 0.00% kworker/7:1-eve [kernel.vmlinux] [k] free_pcppages_bulk 0.00% dockerd [kernel.vmlinux] [k] srso_untrain_ret 0.00% dockerd [kernel.vmlinux] [k] update_curr 0.00% swapper [amdgpu] [k] 0x0000000000124f0c 0.00% swapper [drm] [k] 0x000000000003f455 0.00% swapper [amdgpu] [k] 0x00000000003218b5 0.00% swapper [kernel.vmlinux] [k] ct_irq_enter 0.00% swapper [kernel.vmlinux] [k] ct_irq_exit 0.00% swapper [kernel.vmlinux] [k] blk_stat_add 0.00% swapper [amdgpu] [k] 0x00000000006e8dd0 0.00% swapper [amdgpu] [k] 0x0000000000124f4c 0.00% swapper [amdgpu] [k] 0x000000000098aef6 0.00% swapper [drm] [k] 0x000000000003f468 0.00% swapper [kernel.vmlinux] [k] nohz_balance_enter_idle 0.00% swapper [kernel.vmlinux] [k] __wake_up_common 0.00% swapper [amdgpu] [k] 0x00000000001240ea 0.00% swapper [kernel.vmlinux] [k] __wake_up 0.00% swapper [amdgpu] [k] 0x00000000002f7f81 0.00% swapper [amdgpu] [k] 0x000000000046f97f 0.00% swapper [drm] [k] 0x000000000003f45c 0.00% swapper [kernel.vmlinux] [k] __mod_memcg_lruvec_state 0.00% swapper [kernel.vmlinux] [k] mempool_free_slab 0.00% swapper [amdgpu] [k] 0x0000000000977ef0 0.00% swapper [kernel.vmlinux] [k] x86_pmu_disable_all 0.00% swapper [amdgpu] [k] 0x00000000002f7f71 0.00% kworker/3:1H-kb [kernel.vmlinux] [k] read_hpet 0.00% swapper [kernel.vmlinux] [k] psi_task_change 0.00% swapper [kernel.vmlinux] [k] perf_adjust_freq_unthr_context 0.00% swapper [kernel.vmlinux] [k] account_idle_ticks 0.00% swapper [amdgpu] [k] 0x000000000045d365 0.00% swapper [kernel.vmlinux] [k] ttwu_do_activate 0.00% swapper [raid6_pq] [k] 0x000000000001acc5 0.00% swapper [kernel.vmlinux] [k] enqueue_entity 0.00% wpa_supplicant [kernel.vmlinux] [k] read_hpet 0.00% swapper [amdgpu] [k] 0x0000000000012205 0.00% swapper [drm] [k] 0x000000000003d649 0.00% containerd [kernel.vmlinux] [k] __update_load_avg_se 0.00% rs:main Q:Reg [kernel.vmlinux] [k] futex_wake 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_r13 0.00% swapper [kernel.vmlinux] [k] sched_balance_softirq 0.00% swapper [kernel.vmlinux] [k] timerqueue_del 0.00% swapper [kernel.vmlinux] [k] cpu_util # # (Tip: Limit to show entries above 5% only: perf report --percent-limit 5) #
On Wed, 11 Sep 2024 09:24:34 +0900 Masami Hiramatsu (Google) <mhiramat@kernel.org> wrote: > Let me gently ping :) > > Let me check it should be rebased on tracing/for-next again. But basically > this should be the same... I found a build error on riscv. Let me fix that. Thanks, > > On Sun, 18 Aug 2024 21:47:53 +0900 > "Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote: > > > Hi, > > > > Here is the 13th version of the series to re-implement the fprobe on > > function-graph tracer. The previous version is; > > > > https://lore.kernel.org/all/172000134410.63468.13742222887213469474.stgit@devnote2/ > > > > This version is based on v6.11-rc3. > > In this version, I added a bugfix as [1/20], which should go to urgent > > branch, and dropped the performance improvement patch which was introduced > > in v12 because I found that does not work with new kernel. > > > > Overview > > -------- > > This series rewrites the fprobe on this function-graph. > > The purposes of this change are; > > > > 1) Remove dependency of the rethook from fprobe so that we can reduce > > the return hook code and shadow stack. > > > > 2) Make 'ftrace_regs' the common trace interface for the function > > boundary. > > > > 1) Currently we have 2(or 3) different function return hook codes, > > the function-graph tracer and rethook (and legacy kretprobe). > > But since this is redundant and needs double maintenance cost, > > I would like to unify those. From the user's viewpoint, function- > > graph tracer is very useful to grasp the execution path. For this > > purpose, it is hard to use the rethook in the function-graph > > tracer, but the opposite is possible. (Strictly speaking, kretprobe > > can not use it because it requires 'pt_regs' for historical reasons.) > > > > 2) Now the fprobe provides the 'pt_regs' for its handler, but that is > > wrong for the function entry and exit. Moreover, depending on the > > architecture, there is no way to accurately reproduce 'pt_regs' > > outside of interrupt or exception handlers. This means fprobe should > > not use 'pt_regs' because it does not use such exceptions. > > (Conversely, kprobe should use 'pt_regs' because it is an abstract > > interface of the software breakpoint exception.) > > > > This series changes fprobe to use function-graph tracer for tracing > > function entry and exit, instead of mixture of ftrace and rethook. > > Unlike the rethook which is a per-task list of system-wide allocated > > nodes, the function graph's ret_stack is a per-task shadow stack. > > Thus it does not need to set 'nr_maxactive' (which is the number of > > pre-allocated nodes). > > Also the handlers will get the 'ftrace_regs' instead of 'pt_regs'. > > Since eBPF mulit_kprobe/multi_kretprobe events still use 'pt_regs' as > > their register interface, this changes it to convert 'ftrace_regs' to > > 'pt_regs'. Of course this conversion makes an incomplete 'pt_regs', > > so users must access only registers for function parameters or > > return value. > > > > Design > > ------ > > Instead of using ftrace's function entry hook directly, the new fprobe > > is built on top of the function-graph's entry and return callbacks > > with 'ftrace_regs'. > > > > Since the fprobe requires access to 'ftrace_regs', the architecture > > must support CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS and > > CONFIG_HAVE_FTRACE_GRAPH_FUNC, which enables to call function-graph > > entry callback with 'ftrace_regs', and also > > CONFIG_HAVE_FUNCTION_GRAPH_FREGS, which passes the ftrace_regs to > > return_to_handler. > > > > All fprobes share a single function-graph ops (means shares a common > > ftrace filter) similar to the kprobe-on-ftrace. This needs another > > layer to find corresponding fprobe in the common function-graph > > callbacks, but has much better scalability, since the number of > > registered function-graph ops is limited. > > > > In the entry callback, the fprobe runs its entry_handler and saves the > > address of 'fprobe' on the function-graph's shadow stack as data. The > > return callback decodes the data to get the 'fprobe' address, and runs > > the exit_handler. > > > > The fprobe introduces two hash-tables, one is for entry callback which > > searches fprobes related to the given function address passed by entry > > callback. The other is for a return callback which checks if the given > > 'fprobe' data structure pointer is still valid. Note that it is > > possible to unregister fprobe before the return callback runs. Thus > > the address validation must be done before using it in the return > > callback. > > > > Download > > -------- > > This series can be applied against the ftrace/for-next branch in > > linux-trace tree. > > > > This series can also be found below branch. > > > > https://git.kernel.org/pub/scm/linux/kernel/git/mhiramat/linux.git/log/?h=topic/fprobe-on-fgraph > > > > Thank you, > > > > --- > > > > Masami Hiramatsu (Google) (20): > > tracing: fgraph: Fix to add new fgraph_ops to array after ftrace_startup_subops() > > tracing: Add a comment about ftrace_regs definition > > tracing: Rename ftrace_regs_return_value to ftrace_regs_get_return_value > > function_graph: Pass ftrace_regs to entryfunc > > function_graph: Replace fgraph_ret_regs with ftrace_regs > > function_graph: Pass ftrace_regs to retfunc > > fprobe: Use ftrace_regs in fprobe entry handler > > fprobe: Use ftrace_regs in fprobe exit handler > > tracing: Add ftrace_partial_regs() for converting ftrace_regs to pt_regs > > tracing: Add ftrace_fill_perf_regs() for perf event > > tracing/fprobe: Enable fprobe events with CONFIG_DYNAMIC_FTRACE_WITH_ARGS > > bpf: Enable kprobe_multi feature if CONFIG_FPROBE is enabled > > ftrace: Add CONFIG_HAVE_FTRACE_GRAPH_FUNC > > fprobe: Rewrite fprobe on function-graph tracer > > tracing: Fix function timing profiler to initialize hashtable > > tracing/fprobe: Remove nr_maxactive from fprobe > > selftests: ftrace: Remove obsolate maxactive syntax check > > selftests/ftrace: Add a test case for repeating register/unregister fprobe > > Documentation: probes: Update fprobe on function-graph tracer > > fgraph: Skip recording calltime/rettime if it is not nneeded > > > > > > Documentation/trace/fprobe.rst | 42 + > > arch/arm64/Kconfig | 2 > > arch/arm64/include/asm/ftrace.h | 47 + > > arch/arm64/kernel/asm-offsets.c | 12 > > arch/arm64/kernel/entry-ftrace.S | 32 + > > arch/arm64/kernel/ftrace.c | 20 + > > arch/loongarch/Kconfig | 4 > > arch/loongarch/include/asm/ftrace.h | 32 - > > arch/loongarch/kernel/asm-offsets.c | 12 > > arch/loongarch/kernel/ftrace_dyn.c | 10 > > arch/loongarch/kernel/mcount.S | 17 - > > arch/loongarch/kernel/mcount_dyn.S | 14 > > arch/powerpc/Kconfig | 1 > > arch/powerpc/include/asm/ftrace.h | 15 > > arch/powerpc/kernel/trace/ftrace.c | 2 > > arch/powerpc/kernel/trace/ftrace_64_pg.c | 10 > > arch/riscv/Kconfig | 3 > > arch/riscv/include/asm/ftrace.h | 26 - > > arch/riscv/kernel/ftrace.c | 17 + > > arch/riscv/kernel/mcount.S | 24 - > > arch/s390/Kconfig | 3 > > arch/s390/include/asm/ftrace.h | 39 + > > arch/s390/kernel/asm-offsets.c | 6 > > arch/s390/kernel/mcount.S | 9 > > arch/x86/Kconfig | 4 > > arch/x86/include/asm/ftrace.h | 37 + > > arch/x86/kernel/ftrace.c | 50 +- > > arch/x86/kernel/ftrace_32.S | 15 > > arch/x86/kernel/ftrace_64.S | 17 - > > include/linux/fprobe.h | 57 +- > > include/linux/ftrace.h | 136 ++++ > > kernel/trace/Kconfig | 23 + > > kernel/trace/bpf_trace.c | 19 - > > kernel/trace/fgraph.c | 127 +++- > > kernel/trace/fprobe.c | 638 ++++++++++++++------ > > kernel/trace/ftrace.c | 10 > > kernel/trace/trace.h | 6 > > kernel/trace/trace_fprobe.c | 147 ++--- > > kernel/trace/trace_functions_graph.c | 10 > > kernel/trace/trace_irqsoff.c | 6 > > kernel/trace/trace_probe_tmpl.h | 2 > > kernel/trace/trace_sched_wakeup.c | 6 > > kernel/trace/trace_selftest.c | 11 > > lib/test_fprobe.c | 51 -- > > samples/fprobe/fprobe_example.c | 4 > > .../test.d/dynevent/add_remove_fprobe_repeat.tc | 19 + > > .../ftrace/test.d/dynevent/fprobe_syntax_errors.tc | 4 > > 47 files changed, 1168 insertions(+), 630 deletions(-) > > create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe_repeat.tc > > > > -- > > Masami Hiramatsu (Google) <mhiramat@kernel.org> > > > > > -- > Masami Hiramatsu (Google) <mhiramat@kernel.org>
On Wed, 11 Sep 2024 20:27:27 +0200 Jiri Olsa <olsajiri@gmail.com> wrote: > On Sun, Aug 18, 2024 at 09:47:53PM +0900, Masami Hiramatsu (Google) wrote: > > Hi, > > > > Here is the 13th version of the series to re-implement the fprobe on > > function-graph tracer. The previous version is; > > > > https://lore.kernel.org/all/172000134410.63468.13742222887213469474.stgit@devnote2/ > > > > This version is based on v6.11-rc3. > > In this version, I added a bugfix as [1/20], which should go to urgent > > branch, and dropped the performance improvement patch which was introduced > > in v12 because I found that does not work with new kernel. > > > > Overview > > -------- > > This series rewrites the fprobe on this function-graph. > > The purposes of this change are; > > > > 1) Remove dependency of the rethook from fprobe so that we can reduce > > the return hook code and shadow stack. > > > > 2) Make 'ftrace_regs' the common trace interface for the function > > boundary. > > > > 1) Currently we have 2(or 3) different function return hook codes, > > the function-graph tracer and rethook (and legacy kretprobe). > > But since this is redundant and needs double maintenance cost, > > I would like to unify those. From the user's viewpoint, function- > > graph tracer is very useful to grasp the execution path. For this > > purpose, it is hard to use the rethook in the function-graph > > tracer, but the opposite is possible. (Strictly speaking, kretprobe > > can not use it because it requires 'pt_regs' for historical reasons.) > > > > 2) Now the fprobe provides the 'pt_regs' for its handler, but that is > > wrong for the function entry and exit. Moreover, depending on the > > architecture, there is no way to accurately reproduce 'pt_regs' > > outside of interrupt or exception handlers. This means fprobe should > > not use 'pt_regs' because it does not use such exceptions. > > (Conversely, kprobe should use 'pt_regs' because it is an abstract > > interface of the software breakpoint exception.) > > > > This series changes fprobe to use function-graph tracer for tracing > > function entry and exit, instead of mixture of ftrace and rethook. > > Unlike the rethook which is a per-task list of system-wide allocated > > nodes, the function graph's ret_stack is a per-task shadow stack. > > Thus it does not need to set 'nr_maxactive' (which is the number of > > pre-allocated nodes). > > Also the handlers will get the 'ftrace_regs' instead of 'pt_regs'. > > Since eBPF mulit_kprobe/multi_kretprobe events still use 'pt_regs' as > > their register interface, this changes it to convert 'ftrace_regs' to > > 'pt_regs'. Of course this conversion makes an incomplete 'pt_regs', > > so users must access only registers for function parameters or > > return value. > > > > Design > > ------ > > Instead of using ftrace's function entry hook directly, the new fprobe > > is built on top of the function-graph's entry and return callbacks > > with 'ftrace_regs'. > > > > Since the fprobe requires access to 'ftrace_regs', the architecture > > must support CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS and > > CONFIG_HAVE_FTRACE_GRAPH_FUNC, which enables to call function-graph > > entry callback with 'ftrace_regs', and also > > CONFIG_HAVE_FUNCTION_GRAPH_FREGS, which passes the ftrace_regs to > > return_to_handler. > > > > All fprobes share a single function-graph ops (means shares a common > > ftrace filter) similar to the kprobe-on-ftrace. This needs another > > layer to find corresponding fprobe in the common function-graph > > callbacks, but has much better scalability, since the number of > > registered function-graph ops is limited. > > > > In the entry callback, the fprobe runs its entry_handler and saves the > > address of 'fprobe' on the function-graph's shadow stack as data. The > > return callback decodes the data to get the 'fprobe' address, and runs > > the exit_handler. > > > > The fprobe introduces two hash-tables, one is for entry callback which > > searches fprobes related to the given function address passed by entry > > callback. The other is for a return callback which checks if the given > > 'fprobe' data structure pointer is still valid. Note that it is > > possible to unregister fprobe before the return callback runs. Thus > > the address validation must be done before using it in the return > > callback. > > > > Download > > -------- > > This series can be applied against the ftrace/for-next branch in > > linux-trace tree. > > > > This series can also be found below branch. > > > > https://git.kernel.org/pub/scm/linux/kernel/git/mhiramat/linux.git/log/?h=topic/fprobe-on-fgraph > > > > Thank you, > > > > --- > > hi, > I ran the kprobe_multi bench and I'm seeing lower throughput numbers > with new fprobe implementation > > base (bpf-next/master) > > root@amd:/home/jolsa/bpf-next/tools/testing/selftests/bpf# ./bench -w2 -d5 -a trig-kprobe-multi > Setting up benchmark 'trig-kprobe-multi'... > Benchmark 'trig-kprobe-multi' started. > Iter 0 ( 70.969us): hits 8.530M/s ( 8.530M/prod), drops 0.000M/s, total operations 8.530M/s > Iter 1 (-24.985us): hits 8.541M/s ( 8.541M/prod), drops 0.000M/s, total operations 8.541M/s > Iter 2 ( -2.976us): hits 8.537M/s ( 8.537M/prod), drops 0.000M/s, total operations 8.537M/s > Iter 3 ( -1.502us): hits 8.540M/s ( 8.540M/prod), drops 0.000M/s, total operations 8.540M/s > Iter 4 ( 0.392us): hits 8.540M/s ( 8.540M/prod), drops 0.000M/s, total operations 8.540M/s > Iter 5 ( 1.308us): hits 8.545M/s ( 8.545M/prod), drops 0.000M/s, total operations 8.545M/s > Iter 6 ( 2.783us): hits 8.530M/s ( 8.530M/prod), drops 0.000M/s, total operations 8.530M/s > Iter 7 ( -3.775us): hits 8.535M/s ( 8.535M/prod), drops 0.000M/s, total operations 8.535M/s > Summary: hits 8.538 ± 0.005M/s ( 8.538M/prod), drops 0.000 ± 0.000M/s, total operations 8.538 ± 0.005M/s > > your patchset on top of bpf-next/master > > root@amd:/home/jolsa/bpf-next/tools/testing/selftests/bpf# ./bench -w2 -d5 -a trig-kprobe-multi > Setting up benchmark 'trig-kprobe-multi'... > Benchmark 'trig-kprobe-multi' started. > Iter 0 ( 55.712us): hits 7.224M/s ( 7.224M/prod), drops 0.000M/s, total operations 7.224M/s > Iter 1 ( 15.226us): hits 7.222M/s ( 7.222M/prod), drops 0.000M/s, total operations 7.222M/s > Iter 2 ( -1.353us): hits 7.221M/s ( 7.221M/prod), drops 0.000M/s, total operations 7.221M/s > Iter 3 (-13.029us): hits 7.223M/s ( 7.223M/prod), drops 0.000M/s, total operations 7.223M/s > Iter 4 ( 9.111us): hits 7.222M/s ( 7.222M/prod), drops 0.000M/s, total operations 7.222M/s > Iter 5 ( -0.106us): hits 7.222M/s ( 7.222M/prod), drops 0.000M/s, total operations 7.222M/s > Iter 6 ( 0.734us): hits 7.221M/s ( 7.221M/prod), drops 0.000M/s, total operations 7.221M/s > Iter 7 (-10.233us): hits 7.220M/s ( 7.220M/prod), drops 0.000M/s, total operations 7.220M/s > Summary: hits 7.221 ± 0.001M/s ( 7.221M/prod), drops 0.000 ± 0.000M/s, total operations 7.221 ± 0.001M/s > > > I did perf profile (attached), but nothing stands out on the first look, Hm, we may need to compare both before and after, and need annotate in the code to find the bottleneck. Anyway, performance optimization can be done afterwards, so I'm not so worried it :) > I'll try to check on that later this week Thank you for your help! Thanks, > > jirka > > > --- > # To display the perf.data header info, please use --header/--header-only options. > # > # > # Total Lost Samples: 0 > # > # Samples: 42K of event 'cycles:k' > # Event count (approx.): 44580427937 > # > # Overhead Command Shared Object Symbol > # ........ ............... .................................................... ........................................................ > # > 17.47% bench [kernel.vmlinux] [k] find_kallsyms_symbol > | > ---find_kallsyms_symbol > module_address_lookup > kallsyms_lookup_buildid > kallsyms_lookup > print_rec > t_show > seq_read_iter > seq_read > vfs_read > ksys_read > __x64_sys_read > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > read > 0 > > 11.71% bench [kernel.vmlinux] [k] kprobe_multi_link_prog_run > | > ---kprobe_multi_link_prog_run > | > --11.56%--kprobe_multi_link_handler > fprobe_entry > function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 7.65% bench [kernel.vmlinux] [k] function_graph_enter_regs > | > ---function_graph_enter_regs > | > --7.65%--ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 7.55% bench [kernel.vmlinux] [k] kprobe_multi_link_handler > | > ---kprobe_multi_link_handler > | > |--4.35%--function_graph_enter_regs > | ftrace_graph_func > | 0xffffffffc29850ab > | bpf_get_numa_node_id > | bpf_prog_d9703036495d54b0_trigger_driver > | __bpf_prog_test_run_raw_tp > | bpf_prog_test_run_raw_tp > | __sys_bpf > | __x64_sys_bpf > | x64_sys_call > | do_syscall_64 > | entry_SYSCALL_64 > | syscall > | bpf_prog_test_run_opts > | trigger_producer_batch > | 0x7a2a73a94ac3 > | > --3.20%--fprobe_entry > function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 7.50% bench [kernel.vmlinux] [k] fprobe_entry > | > ---fprobe_entry > | > --7.50%--function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 7.08% bench bpf_prog_7790468e40a289ea_bench_trigger_kprobe_multi [k] bpf_prog_7790468e40a289ea_bench_trigger_kprobe_multi > | > ---bpf_prog_7790468e40a289ea_bench_trigger_kprobe_multi > | > |--4.10%--kprobe_multi_link_handler > | fprobe_entry > | function_graph_enter_regs > | ftrace_graph_func > | 0xffffffffc29850ab > | bpf_get_numa_node_id > | bpf_prog_d9703036495d54b0_trigger_driver > | __bpf_prog_test_run_raw_tp > | bpf_prog_test_run_raw_tp > | __sys_bpf > | __x64_sys_bpf > | x64_sys_call > | do_syscall_64 > | entry_SYSCALL_64 > | syscall > | bpf_prog_test_run_opts > | trigger_producer_batch > | 0x7a2a73a94ac3 > | > --2.97%--kprobe_multi_link_prog_run > kprobe_multi_link_handler > fprobe_entry > function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 6.11% bench [kernel.vmlinux] [k] migrate_enable > | > ---migrate_enable > | > |--3.37%--kprobe_multi_link_handler > | fprobe_entry > | function_graph_enter_regs > | ftrace_graph_func > | 0xffffffffc29850ab > | bpf_get_numa_node_id > | bpf_prog_d9703036495d54b0_trigger_driver > | __bpf_prog_test_run_raw_tp > | bpf_prog_test_run_raw_tp > | __sys_bpf > | __x64_sys_bpf > | x64_sys_call > | do_syscall_64 > | entry_SYSCALL_64 > | syscall > | bpf_prog_test_run_opts > | trigger_producer_batch > | 0x7a2a73a94ac3 > | > --2.73%--kprobe_multi_link_prog_run > kprobe_multi_link_handler > fprobe_entry > function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 5.90% bench [kernel.vmlinux] [k] ftrace_graph_func > | > ---ftrace_graph_func > | > --5.75%--0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 4.03% bench [kernel.vmlinux] [k] __rcu_read_lock > | > ---__rcu_read_lock > | > --4.00%--kprobe_multi_link_handler > fprobe_entry > function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 3.61% bench [kernel.vmlinux] [k] bpf_get_numa_node_id > | > ---bpf_get_numa_node_id > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 2.99% bench [nf_conntrack] [k] 0x000000000002c0b3 > | > ---0xffffffffc29850b3 > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 2.69% bench bpf_prog_d9703036495d54b0_trigger_driver [k] bpf_prog_d9703036495d54b0_trigger_driver > | > ---bpf_prog_d9703036495d54b0_trigger_driver > | > --2.68%--__bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 2.04% bench [kernel.vmlinux] [k] t_start > | > ---t_start > seq_read_iter > seq_read > vfs_read > ksys_read > __x64_sys_read > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > read > 0 > > 1.90% bench [kernel.vmlinux] [k] srso_untrain_ret > | > ---srso_untrain_ret > | > |--0.68%--fprobe_entry > | function_graph_enter_regs > | ftrace_graph_func > | 0xffffffffc29850ab > | bpf_get_numa_node_id > | bpf_prog_d9703036495d54b0_trigger_driver > | __bpf_prog_test_run_raw_tp > | bpf_prog_test_run_raw_tp > | __sys_bpf > | __x64_sys_bpf > | x64_sys_call > | do_syscall_64 > | entry_SYSCALL_64 > | syscall > | bpf_prog_test_run_opts > | trigger_producer_batch > | 0x7a2a73a94ac3 > | > --0.54%--seq_read_iter > seq_read > vfs_read > ksys_read > __x64_sys_read > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > read > 0 > > 1.14% bench [kernel.vmlinux] [k] __rcu_read_unlock > | > ---__rcu_read_unlock > | > --1.12%--kprobe_multi_link_handler > fprobe_entry > function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 1.02% bench [kernel.vmlinux] [k] migrate_disable > | > ---migrate_disable > kprobe_multi_link_handler > fprobe_entry > function_graph_enter_regs > ftrace_graph_func > 0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 0.99% bench [kernel.vmlinux] [k] t_func_next.isra.0 > | > ---t_func_next.isra.0 > | > --0.83%--t_start > seq_read_iter > seq_read > vfs_read > ksys_read > __x64_sys_read > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > read > 0 > > 0.78% bench [nf_conntrack] [k] 0x000000000002c0d4 > | > ---0xffffffffc29850d4 > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 0.75% bench [nf_conntrack] [k] 0x000000000002c0ab > | > ---0xffffffffc29850ab > bpf_get_numa_node_id > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 0.68% bench [nf_conntrack] [k] 0x000000000002c00d > | > ---0xffffffffc298500d > bpf_prog_d9703036495d54b0_trigger_driver > __bpf_prog_test_run_raw_tp > bpf_prog_test_run_raw_tp > __sys_bpf > __x64_sys_bpf > x64_sys_call > do_syscall_64 > entry_SYSCALL_64 > syscall > bpf_prog_test_run_opts > trigger_producer_batch > 0x7a2a73a94ac3 > > 0.52% bench [kernel.vmlinux] [k] __x86_indirect_thunk_array > | > ---__x86_indirect_thunk_array > > 0.42% bench [nf_conntrack] [k] 0x000000000002c005 > 0.38% bench [kernel.vmlinux] [k] __sys_bpf > 0.37% bench [kernel.vmlinux] [k] srso_return_thunk > 0.24% bench [nf_conntrack] [k] 0x000000000002c001 > 0.23% bench [kernel.vmlinux] [k] read_hpet > 0.21% bench [kernel.vmlinux] [k] entry_SYSCALL_64 > 0.19% bench [kernel.vmlinux] [k] bpf_prog_test_run_raw_tp > 0.17% bench [nf_conntrack] [k] 0x000000000002c000 > 0.15% bench [kernel.vmlinux] [k] _copy_from_user > 0.15% bench [kernel.vmlinux] [k] syscall_exit_to_user_mode > 0.14% bench [nf_conntrack] [k] 0x000000000002c027 > 0.14% bench [nf_conntrack] [k] 0x000000000002c0ca > 0.13% swapper [kernel.vmlinux] [k] read_hpet > 0.12% bench [kernel.vmlinux] [k] _copy_to_user > 0.12% bench [nf_conntrack] [k] 0x000000000002c086 > 0.11% bench [nf_conntrack] [k] 0x000000000002c0c0 > 0.11% bench [kernel.vmlinux] [k] __fdget > 0.11% bench bench [.] bpf_prog_test_run_opts > 0.11% bench [kernel.vmlinux] [k] __bpf_prog_test_run_raw_tp > 0.10% bench [nf_conntrack] [k] 0x000000000002c059 > 0.09% bench [kernel.vmlinux] [k] get_symbol_offset > 0.09% bench [kernel.vmlinux] [k] do_syscall_64 > 0.08% bench [kernel.vmlinux] [k] __bpf_prog_get > 0.08% bench [kernel.vmlinux] [k] memchr_inv > 0.07% bench [kernel.vmlinux] [k] syscall_return_via_sysret > 0.06% bench [kernel.vmlinux] [k] x64_sys_call > 0.06% bench [kernel.vmlinux] [k] rep_movs_alternative > 0.06% bench [kernel.vmlinux] [k] ftrace_replace_code > 0.04% bench [kernel.vmlinux] [k] __x64_sys_bpf > 0.04% bench [kernel.vmlinux] [k] get_symbol_pos > 0.04% bench bench [.] sys_bpf > 0.04% bench [kernel.vmlinux] [k] vsnprintf > 0.04% bench [kernel.vmlinux] [k] number > 0.04% bench bench [.] ptr_to_u64 > 0.04% bench [kernel.vmlinux] [k] kfree > 0.03% bench [kernel.vmlinux] [k] seq_read_iter > 0.03% bench [kernel.vmlinux] [k] security_bpf > 0.03% bench libc.so.6 [.] syscall > 0.03% bench [kernel.vmlinux] [k] kallsyms_expand_symbol.constprop.0 > 0.03% kworker/2:2-eve [kernel.vmlinux] [k] memcpy_orig > 0.03% swapper [amdgpu] [k] 0x0000000000002f75 > 0.03% bench bench [.] trigger_producer_batch > 0.03% bench [kernel.vmlinux] [k] __bpf_prog_put > 0.02% bench [kernel.vmlinux] [k] seq_printf > 0.02% bench [kernel.vmlinux] [k] smp_call_function_many_cond > 0.02% kworker/2:2-eve [amdgpu] [k] 0x0000000000002f75 > 0.02% bench [kernel.vmlinux] [k] __check_object_size > 0.02% bench [kernel.vmlinux] [k] ftrace_test_record > 0.02% bench [kernel.vmlinux] [k] seq_write > 0.02% bench [kernel.vmlinux] [k] ftrace_lookup_ip > 0.02% bench [kernel.vmlinux] [k] t_show > 0.02% bench [kernel.vmlinux] [k] print_rec > 0.02% swapper [kernel.vmlinux] [k] io_idle > 0.01% bench [amdgpu] [k] 0x0000000000002f75 > 0.01% kworker/2:2-eve [kernel.vmlinux] [k] memcmp > 0.01% bench [kernel.vmlinux] [k] format_decode > 0.01% bench [kernel.vmlinux] [k] memcpy_orig > 0.01% bench [kernel.vmlinux] [k] __ftrace_hash_rec_update.part.0 > 0.01% bench [kernel.vmlinux] [k] mod_find > 0.01% swapper [kernel.vmlinux] [k] srso_untrain_ret > 0.01% bench [kernel.vmlinux] [k] kallsyms_lookup_buildid > 0.01% bench [kernel.vmlinux] [k] __seq_puts > 0.01% bench [kernel.vmlinux] [k] native_read_msr > 0.01% bench [kernel.vmlinux] [k] check_stack_object > 0.01% bench [kernel.vmlinux] [k] fput > 0.01% bench bench [.] libbpf_err_errno > 0.01% bench [kernel.vmlinux] [k] fpregs_assert_state_consistent > 0.01% bench [kernel.vmlinux] [k] ftrace_check_record > 0.01% bench [kernel.vmlinux] [k] t_next > 0.01% bench [kernel.vmlinux] [k] strlen > 0.01% swapper [kernel.vmlinux] [k] native_sched_clock > 0.01% bench [kernel.vmlinux] [k] __sysvec_apic_timer_interrupt > 0.01% bench libc.so.6 [.] 0x00000000001a0f10 > 0.01% bench [kernel.vmlinux] [k] native_write_msr > 0.01% bench [kernel.vmlinux] [k] __update_load_avg_se > 0.01% bench [kernel.vmlinux] [k] amd_pmu_addr_offset > 0.01% bench [kernel.vmlinux] [k] module_address_lookup > 0.01% bench [kernel.vmlinux] [k] ftrace_rec_iter_next > 0.01% bench [kernel.vmlinux] [k] string > 0.01% bench [kernel.vmlinux] [k] asm_exc_page_fault > 0.01% swapper [kernel.vmlinux] [k] __get_next_timer_interrupt > 0.01% bench [kernel.vmlinux] [k] clear_page_rep > 0.01% kworker/2:2-eve [kernel.vmlinux] [k] delay_halt_mwaitx > 0.01% bench [kernel.vmlinux] [k] __handle_mm_fault > 0.01% swapper [kernel.vmlinux] [k] ktime_get > 0.01% kworker/2:2-eve [kernel.vmlinux] [k] clear_page_rep > 0.01% swapper [kernel.vmlinux] [k] menu_select > 0.00% perf [kernel.vmlinux] [k] clear_page_rep > 0.00% bench [kernel.vmlinux] [k] ftrace_rec_iter_record > 0.00% bench [kernel.vmlinux] [k] account_process_tick > 0.00% bench [kernel.vmlinux] [k] __update_load_avg_cfs_rq > 0.00% bench [kernel.vmlinux] [k] update_curr > 0.00% bench [kernel.vmlinux] [k] do_sync_core > 0.00% bench [kernel.vmlinux] [k] kallsyms_lookup > 0.00% bench [kernel.vmlinux] [k] memset_orig > 0.00% kworker/6:1-eve [kernel.vmlinux] [k] memcpy_orig > 0.00% perf [kernel.vmlinux] [k] rep_movs_alternative > 0.00% sshd [kernel.vmlinux] [k] p4d_offset > 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000002f75 > 0.00% bench [kernel.vmlinux] [k] __mod_memcg_lruvec_state > 0.00% swapper [kernel.vmlinux] [k] _raw_spin_unlock > 0.00% swapper [kernel.vmlinux] [k] cpuidle_enter_state > 0.00% swapper [kernel.vmlinux] [k] rb_erase > 0.00% containerd [kernel.vmlinux] [k] read_hpet > 0.00% kworker/6:1-eve [kernel.vmlinux] [k] memcmp > 0.00% swapper [kernel.vmlinux] [k] irq_entries_start > 0.00% swapper [kernel.vmlinux] [k] do_idle > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] read_hpet > 0.00% swapper [kernel.vmlinux] [k] update_rq_clock > 0.00% swapper [kernel.vmlinux] [k] psi_group_change > 0.00% swapper [kernel.vmlinux] [k] sched_clock_cpu > 0.00% swapper [kernel.vmlinux] [k] do_sync_core > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ec11 > 0.00% swapper [kernel.vmlinux] [k] _raw_spin_lock_irqsave > 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_call_function_single > 0.00% swapper [kernel.vmlinux] [k] update_sd_lb_stats.constprop.0 > 0.00% perf [kernel.vmlinux] [k] __filemap_get_folio > 0.00% swapper [kernel.vmlinux] [k] irqentry_enter > 0.00% bench [kernel.vmlinux] [k] amd_pmu_test_overflow_topbit > 0.00% bench [amdgpu] [k] 0x0000000000124f4c > 0.00% bench [kernel.vmlinux] [k] update_rq_clock > 0.00% bench [kernel.vmlinux] [k] perf_event_task_tick > 0.00% bench [amdgpu] [k] 0x0000000000126f3a > 0.00% bench [kernel.vmlinux] [k] sched_balance_update_blocked_averages > 0.00% bench [kernel.vmlinux] [k] hrtimer_active > 0.00% bench [kernel.vmlinux] [k] _copy_to_iter > 0.00% bench [kernel.vmlinux] [k] _raw_spin_lock_irqsave > 0.00% bench [kernel.vmlinux] [k] calc_global_load_tick > 0.00% bench [kernel.vmlinux] [k] ftrace_shutdown.part.0 > 0.00% bench libc.so.6 [.] 0x00000000001a0fb5 > 0.00% bench [kernel.vmlinux] [k] task_tick_mm_cid > 0.00% bench libc.so.6 [.] 0x00000000001a0fba > 0.00% bench [kernel.vmlinux] [k] sched_clock_cpu > 0.00% bench [kernel.vmlinux] [k] update_cfs_group > 0.00% bench [kernel.vmlinux] [k] irq_exit_rcu > 0.00% bench [kernel.vmlinux] [k] native_sched_clock > 0.00% bench [kernel.vmlinux] [k] sched_tick > 0.00% bench [kernel.vmlinux] [k] sized_strscpy > 0.00% bench [kernel.vmlinux] [k] task_tick_fair > 0.00% bench [nf_conntrack] [k] 0x000000000002c036 > 0.00% bench [kernel.vmlinux] [k] perf_adjust_freq_unthr_context > 0.00% bench [kernel.vmlinux] [k] idle_cpu > 0.00% bench [kernel.vmlinux] [k] bpf_check_uarg_tail_zero > 0.00% swapper [kernel.vmlinux] [k] __hrtimer_next_event_base > 0.00% swapper [kernel.vmlinux] [k] native_write_msr > 0.00% bench [kernel.vmlinux] [k] __module_address.part.0 > 0.00% bench [kernel.vmlinux] [k] lru_gen_add_folio > 0.00% bench [kernel.vmlinux] [k] set_pte_range > 0.00% bench [kernel.vmlinux] [k] pfn_pte > 0.00% swapper [kernel.vmlinux] [k] error_entry > 0.00% swapper [kernel.vmlinux] [k] idle_cpu > 0.00% bench [kernel.vmlinux] [k] unmap_page_range > 0.00% bench [kernel.vmlinux] [k] __count_memcg_events > 0.00% bench [kernel.vmlinux] [k] __irq_exit_rcu > 0.00% perf [kernel.vmlinux] [k] __perf_event_enable > 0.00% bench [kernel.vmlinux] [k] seq_putc > 0.00% perf [kernel.vmlinux] [k] srso_untrain_ret > 0.00% swapper [kernel.vmlinux] [k] __irq_exit_rcu > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __kmalloc_cache_noprof > 0.00% bench libc.so.6 [.] 0x00000000001a0bd7 > 0.00% perf [kernel.vmlinux] [k] _raw_spin_lock > 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __bio_split_to_limits > 0.00% bench [kernel.vmlinux] [k] do_anonymous_page > 0.00% perf [kernel.vmlinux] [k] __es_insert_extent > 0.00% bench [kernel.vmlinux] [k] __mem_cgroup_charge > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __wait_for_common > 0.00% bench [kernel.vmlinux] [k] __folio_throttle_swaprate > 0.00% swapper [kernel.vmlinux] [k] __switch_to_asm > 0.00% swapper [kernel.vmlinux] [k] add_interrupt_randomness > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] srso_untrain_ret > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] memset_orig > 0.00% kworker/u65:2-e [kernel.vmlinux] [k] percpu_counter_add_batch > 0.00% swapper [kernel.vmlinux] [k] tick_irq_enter > 0.00% swapper [kernel.vmlinux] [k] tick_check_broadcast_expired > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000045d28e > 0.00% perf [kernel.vmlinux] [k] __lruvec_stat_mod_folio > 0.00% bench [kernel.vmlinux] [k] kernfs_fop_read_iter > 0.00% bench [kernel.vmlinux] [k] __mod_zone_page_state > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] delay_halt > 0.00% perf [kernel.vmlinux] [k] __mod_memcg_lruvec_state > 0.00% perf [kernel.vmlinux] [k] xas_load > 0.00% perf [kernel.vmlinux] [k] workingset_update_node > 0.00% perf [kernel.vmlinux] [k] _raw_spin_unlock > 0.00% swapper [kernel.vmlinux] [k] __handle_irq_event_percpu > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] post_alloc_hook > 0.00% bench [kernel.vmlinux] [k] refill_obj_stock > 0.00% perf [kernel.vmlinux] [k] __alloc_pages_noprof > 0.00% swapper [kernel.vmlinux] [k] cpuidle_enter > 0.00% swapper [kernel.vmlinux] [k] sched_balance_domains > 0.00% perf [kernel.vmlinux] [k] __handle_mm_fault > 0.00% swapper [kernel.vmlinux] [k] ct_kernel_enter.constprop.0 > 0.00% bench [kernel.vmlinux] [k] mab_mas_cp > 0.00% swapper [amdgpu] [k] 0x00000000004619a7 > 0.00% swapper [kernel.vmlinux] [k] tick_nohz_irq_exit > 0.00% perf [kernel.vmlinux] [k] __memcg_slab_post_alloc_hook > 0.00% bench [kernel.vmlinux] [k] __alloc_pages_noprof > 0.00% swapper [kernel.vmlinux] [k] irq_exit_rcu > 0.00% bench [kernel.vmlinux] [k] memcmp > 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_call_function > 0.00% perf [kernel.vmlinux] [k] perf_poll > 0.00% swapper [kernel.vmlinux] [k] enqueue_task_fair > 0.00% bench [kernel.vmlinux] [k] __rmqueue_pcplist > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __iommu_map > 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000992cd2 > 0.00% kworker/6:1-eve [amdgpu] [k] 0x00000000004b2ade > 0.00% swapper [kernel.vmlinux] [k] wakeup_preempt > 0.00% migration/11 [kernel.vmlinux] [k] enqueue_task > 0.00% swapper [kernel.vmlinux] [k] hrtimer_update_next_event > 0.00% swapper [kernel.vmlinux] [k] native_apic_mem_eoi > 0.00% bench [kernel.vmlinux] [k] is_sync_callback_calling_insn > 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000992cb3 > 0.00% systemd-network libc.so.6 [.] clock_gettime > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] filemap_get_folios_tag > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] fscrypt_mergeable_bio_bh > 0.00% swapper [kernel.vmlinux] [k] hrtimer_start_range_ns > 0.00% multipathd [kernel.vmlinux] [k] cpuacct_charge > 0.00% swapper [kernel.vmlinux] [k] hrtimer_get_next_event > 0.00% kworker/12:1-ev [amdgpu] [k] 0x00000000004b09fc > 0.00% kworker/12:1-ev [kernel.vmlinux] [k] memcpy_orig > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] wbt_track > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __lruvec_stat_mod_folio > 0.00% kworker/12:1-ev [amdgpu] [k] 0x0000000000002f75 > 0.00% bench [kernel.vmlinux] [k] do_jit > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __folio_start_writeback > 0.00% bench [kernel.vmlinux] [k] uncharge_folio > 0.00% perf [kernel.vmlinux] [k] do_fault > 0.00% perf [kernel.vmlinux] [k] __kmalloc_noprof > 0.00% containerd [kernel.vmlinux] [k] blkcg_maybe_throttle_current > 0.00% kworker/12:1-ev [amdgpu] [k] 0x0000000000523901 > 0.00% kworker/12:1-ev [amdgpu] [k] 0x0000000000538414 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000465f98 > 0.00% irqbalance [kernel.vmlinux] [k] format_decode > 0.00% swapper [kernel.vmlinux] [k] handle_edge_irq > 0.00% kworker/12:1-ev [kernel.vmlinux] [k] srso_untrain_ret > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] __find_get_block > 0.00% systemd-journal [kernel.vmlinux] [k] __put_user_nocheck_4 > 0.00% systemd-journal libsystemd-shared-249.so [.] 0x00000000001c9c3e > 0.00% swapper [kernel.vmlinux] [k] local_touch_nmi > 0.00% bench [kernel.vmlinux] [k] __mem_cgroup_uncharge_folios > 0.00% swapper [kernel.vmlinux] [k] _raw_spin_lock > 0.00% swapper [kernel.vmlinux] [k] select_task_rq_fair > 0.00% swapper [kernel.vmlinux] [k] need_update > 0.00% bench [kernel.vmlinux] [k] rcu_segcblist_enqueue > 0.00% swapper [kernel.vmlinux] [k] tick_nohz_get_sleep_length > 0.00% containerd [kernel.vmlinux] [k] _raw_spin_unlock > 0.00% containerd [kernel.vmlinux] [k] ep_autoremove_wake_function > 0.00% kworker/u66:1-e [kernel.vmlinux] [k] __percpu_counter_sum > 0.00% swapper [kernel.vmlinux] [k] ktime_get_mono_fast_ns > 0.00% swapper [kernel.vmlinux] [k] sysvec_apic_timer_interrupt > 0.00% swapper [kernel.vmlinux] [k] _raw_spin_unlock_irqrestore > 0.00% migration/10 [kernel.vmlinux] [k] dequeue_entity > 0.00% swapper [kernel.vmlinux] [k] handle_irq_event > 0.00% perf [kernel.vmlinux] [k] srso_return_thunk > 0.00% perf [kernel.vmlinux] [k] xas_find_conflict > 0.00% perf-exec [kernel.vmlinux] [k] mas_next_slot > 0.00% swapper [kernel.vmlinux] [k] fetch_next_timer_interrupt.constprop.0 > 0.00% bench [kernel.vmlinux] [k] get_page_from_freelist > 0.00% swapper [amdgpu] [k] 0x0000000000125079 > 0.00% kworker/u65:4-e [kernel.vmlinux] [k] sched_balance_newidle > 0.00% migration/6 [kernel.vmlinux] [k] raw_spin_rq_unlock > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000444a05 > 0.00% swapper [kernel.vmlinux] [k] fetch_pte > 0.00% perf [kernel.vmlinux] [k] __pte_offset_map > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000534985 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000003ac5 > 0.00% bench [kernel.vmlinux] [k] perf_iterate_sb > 0.00% swapper [kernel.vmlinux] [k] get_cpu_device > 0.00% perf [kernel.vmlinux] [k] xas_start > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000098fea5 > 0.00% perf [kernel.vmlinux] [k] mark_buffer_dirty > 0.00% containerd containerd [.] runtime.cleantimers > 0.00% bench [kernel.vmlinux] [k] __cond_resched > 0.00% kworker/2:2-eve [drm_kms_helper] [k] 0x00000000000008eb > 0.00% kworker/2:2-eve [raid6_pq] [k] 0x000000000000c24b > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ea79 > 0.00% perf [kernel.vmlinux] [k] cpu_util > 0.00% swapper [kernel.vmlinux] [k] asm_common_interrupt > 0.00% perf [kernel.vmlinux] [k] __filemap_add_folio > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000007d1eab > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051a45f > 0.00% perf [kernel.vmlinux] [k] ext4_es_lookup_extent > 0.00% swapper [kernel.vmlinux] [k] __flush_smp_call_function_queue > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000519af6 > 0.00% swapper [kernel.vmlinux] [k] refresh_cpu_vm_stats > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000477759 > 0.00% perf-exec [kernel.vmlinux] [k] kmem_cache_free > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000539303 > 0.00% kworker/2:2-eve [drm] [k] 0x0000000000020c1c > 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_apic_timer_interrupt > 0.00% bench [kernel.vmlinux] [k] alloc_pages_mpol_noprof > 0.00% bench [kernel.vmlinux] [k] filemap_map_pages > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000477767 > 0.00% swapper [kernel.vmlinux] [k] update_load_avg > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000045d3d6 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051d905 > 0.00% kworker/2:2-eve [drm_kms_helper] [k] 0x0000000000004ae8 > 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __mod_memcg_lruvec_state > 0.00% kworker/u65:2-e [kernel.vmlinux] [k] bio_associate_blkg > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004916dd > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000465ecb > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004b780a > 0.00% swapper [kernel.vmlinux] [k] sysvec_call_function > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051e764 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ed72 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051c7bf > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ebe0 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000523fba > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000832660 > 0.00% swapper [kernel.vmlinux] [k] acpi_idle_do_entry > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097ddfc > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000045d334 > 0.00% kworker/u65:2-e [kernel.vmlinux] [k] soft_cursor > 0.00% perf [kernel.vmlinux] [k] sched_balance_rq > 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __fprop_add_percpu > 0.00% swapper [kernel.vmlinux] [k] rb_next > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000977ece > 0.00% perf [kernel.vmlinux] [k] __folio_mark_dirty > 0.00% perf [kernel.vmlinux] [k] filemap_alloc_folio_noprof > 0.00% perf [kernel.vmlinux] [k] get_mem_cgroup_from_mm > 0.00% perf [kernel.vmlinux] [k] ext4_da_reserve_space > 0.00% perf [kernel.vmlinux] [k] read_hpet > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051cd3e > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000532f0f > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __mod_node_page_state > 0.00% systemd-journal [kernel.vmlinux] [k] mod_objcg_state > 0.00% jbd2/dm-0-8 [kernel.vmlinux] [k] dm_submit_bio > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000517b6d > 0.00% kcompactd0 [kernel.vmlinux] [k] idle_cpu > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] free_tail_page_prepare > 0.00% swapper [amdgpu] [k] 0x00000000002f7f65 > 0.00% swapper [kernel.vmlinux] [k] tick_program_event > 0.00% containerd [kernel.vmlinux] [k] psi_group_change > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000053776f > 0.00% swapper [kernel.vmlinux] [k] srso_return_thunk > 0.00% perf [kernel.vmlinux] [k] ext4_claim_free_clusters > 0.00% perf [kernel.vmlinux] [k] generic_perform_write > 0.00% perf [kernel.vmlinux] [k] mem_cgroup_commit_charge > 0.00% perf [kernel.vmlinux] [k] file_modified > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051e779 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000053974c > 0.00% perf [kernel.vmlinux] [k] ext4_da_write_begin > 0.00% perf [kernel.vmlinux] [k] filemap_add_folio > 0.00% perf libc.so.6 [.] write > 0.00% perf [kernel.vmlinux] [k] __es_remove_extent > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000052244a > 0.00% swapper [kernel.vmlinux] [k] ct_nmi_exit > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000034bee5 > 0.00% kworker/2:2-eve [kernel.vmlinux] [k] __rcu_read_lock > 0.00% perf [kernel.vmlinux] [k] __dquot_alloc_space > 0.00% swapper [kernel.vmlinux] [k] native_read_msr > 0.00% perf [kernel.vmlinux] [k] folio_unlock > 0.00% perf [kernel.vmlinux] [k] radix_tree_node_ctor > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000006c5b82 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000527259 > 0.00% perf [kernel.vmlinux] [k] __mod_node_page_state > 0.00% perf [kernel.vmlinux] [k] __radix_tree_lookup > 0.00% kworker/u65:2-e [kernel.vmlinux] [k] __rcu_read_unlock > 0.00% perf [kernel.vmlinux] [k] ext4_get_reserved_space > 0.00% swapper [kernel.vmlinux] [k] hrtimer_cancel > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004ba8ce > 0.00% swapper [amdgpu] [k] 0x0000000000126f3a > 0.00% perf perf [.] perf_cpu_map__idx > 0.00% swapper [kernel.vmlinux] [k] sched_balance_update_blocked_averages > 0.00% systemd-journal [kernel.vmlinux] [k] do_timerfd_settime > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097ddc5 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097e5a4 > 0.00% perf [kernel.vmlinux] [k] folio_batch_move_lru > 0.00% swapper [amdgpu] [k] 0x000000000046f8a5 > 0.00% perf [kernel.vmlinux] [k] xas_nomem > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051d24c > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000097df7c > 0.00% tuned [kernel.vmlinux] [k] psi_task_switch > 0.00% swapper [kernel.vmlinux] [k] __dequeue_entity > 0.00% swapper [amdgpu] [k] 0x00000000006e7c5d > 0.00% swapper [kernel.vmlinux] [k] rcu_preempt_deferred_qs > 0.00% kworker/7:1-eve [kernel.vmlinux] [k] __free_one_page > 0.00% swapper [kernel.vmlinux] [k] memchr_inv > 0.00% perf [kernel.vmlinux] [k] psi_task_switch > 0.00% swapper [kernel.vmlinux] [k] _nohz_idle_balance.isra.0 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000009a6fbd > 0.00% swapper [kernel.vmlinux] [k] acpi_idle_enter > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000456a9f > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004ac642 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000477888 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000529a52 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004b36d2 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000525b33 > 0.00% swapper [kernel.vmlinux] [k] tmigr_active_up > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000452444 > 0.00% swapper [amdgpu] [k] 0x000000000046f865 > 0.00% swapper [kernel.vmlinux] [k] next_timer_interrupt > 0.00% kworker/6:1-eve [kernel.vmlinux] [k] __free_pages_ok > 0.00% perf [kernel.vmlinux] [k] exc_page_fault > 0.00% swapper [kernel.vmlinux] [k] __schedule > 0.00% irqbalance [kernel.vmlinux] [k] __kmalloc_cache_noprof > 0.00% swapper [kernel.vmlinux] [k] cpuidle_not_available > 0.00% containerd [kernel.vmlinux] [k] pick_next_task_fair > 0.00% kworker/6:1-eve [amdgpu] [k] 0x00000000009a70b7 > 0.00% swapper [kernel.vmlinux] [k] tick_nohz_next_event > 0.00% swapper [kernel.vmlinux] [k] tick_check_oneshot_broadcast_this_cpu > 0.00% swapper [kernel.vmlinux] [k] irq_chip_ack_parent > 0.00% irqbalance [kernel.vmlinux] [k] show_stat > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000004b8295 > 0.00% swapper [amdgpu] [k] 0x000000000045d385 > 0.00% swapper [kernel.vmlinux] [k] tick_nohz_idle_stop_tick > 0.00% containerd [kernel.vmlinux] [k] __schedule > 0.00% swapper [kernel.vmlinux] [k] switch_mm_irqs_off > 0.00% swapper [amdgpu] [k] 0x0000000000461a21 > 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000051d7c2 > 0.00% swapper [kernel.vmlinux] [k] try_to_wake_up > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000051e7cf > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064ed4e > 0.00% perf [kernel.vmlinux] [k] perf_mmap_to_page > 0.00% migration/7 [kernel.vmlinux] [k] update_sd_lb_stats.constprop.0 > 0.00% irqbalance [kernel.vmlinux] [k] __rcu_read_lock > 0.00% migration/3 [kernel.vmlinux] [k] __update_load_avg_cfs_rq > 0.00% swapper [br_netfilter] [k] 0x0000000000000775 > 0.00% bench [kernel.vmlinux] [k] remove_vma > 0.00% swapper [kernel.vmlinux] [k] sched_clock_noinstr > 0.00% kworker/6:1-eve [amdgpu] [k] 0x0000000000529169 > 0.00% swapper [amdgpu] [k] 0x000000000000df07 > 0.00% swapper [kernel.vmlinux] [k] __switch_to > 0.00% swapper [kernel.vmlinux] [k] place_entity > 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000097dd9a > 0.00% swapper [amdgpu] [k] 0x00000000006e7c5a > 0.00% irqbalance [kernel.vmlinux] [k] __rcu_read_unlock > 0.00% migration/6 [kernel.vmlinux] [k] enqueue_entity > 0.00% swapper [kernel.vmlinux] [k] __update_load_avg_cfs_rq > 0.00% perf [kernel.vmlinux] [k] event_function > 0.00% perf [kernel.vmlinux] [k] do_user_addr_fault > 0.00% kworker/2:2-eve [amdgpu] [k] 0x000000000064eaee > 0.00% migration/4 [kernel.vmlinux] [k] __update_load_avg_cfs_rq > 0.00% swapper [amdgpu] [k] 0x00000000006e79ff > 0.00% containerd [kernel.vmlinux] [k] schedule_hrtimeout_range_clock > 0.00% irqbalance [kernel.vmlinux] [k] security_file_open > 0.00% perf [kernel.vmlinux] [k] _find_next_and_bit > 0.00% multipathd [kernel.vmlinux] [k] __rseq_handle_notify_resume > 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_r12 > 0.00% swapper [kernel.vmlinux] [k] rcu_note_context_switch > 0.00% bench [kernel.vmlinux] [k] free_unref_folios > 0.00% migration/0 [kernel.vmlinux] [k] migration_cpu_stop > 0.00% migration/1 [kernel.vmlinux] [k] rcu_note_context_switch > 0.00% kworker/u66:1-e [kernel.vmlinux] [k] n_tty_receive_buf2 > 0.00% swapper [kernel.vmlinux] [k] __common_interrupt > 0.00% swapper [drm] [k] 0x000000000003d658 > 0.00% migration/5 [kernel.vmlinux] [k] psi_group_change > 0.00% swapper [kernel.vmlinux] [k] clockevents_program_event > 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_array > 0.00% perf [kernel.vmlinux] [k] set_pte_range > 0.00% perf [kernel.vmlinux] [k] __schedule > 0.00% swapper [kernel.vmlinux] [k] __update_load_avg_se > 0.00% swapper [kernel.vmlinux] [k] asm_sysvec_reschedule_ipi > 0.00% swapper [kernel.vmlinux] [k] local_clock_noinstr > 0.00% swapper [kernel.vmlinux] [k] tick_nohz_stop_idle > 0.00% swapper [amdgpu] [k] 0x000000000044a039 > 0.00% kworker/4:1-mm_ [kernel.vmlinux] [k] native_queued_spin_lock_slowpath > 0.00% rcu_preempt [kernel.vmlinux] [k] __update_idle_core > 0.00% rcu_preempt [kernel.vmlinux] [k] pick_next_task_fair > 0.00% rcu_preempt [kernel.vmlinux] [k] prepare_to_swait_event > 0.00% rcu_preempt [kernel.vmlinux] [k] sched_balance_rq > 0.00% rcu_preempt [kernel.vmlinux] [k] update_rq_clock > 0.00% rs:main Q:Reg [kernel.vmlinux] [k] ext4_inode_csum > 0.00% swapper [kernel.vmlinux] [k] tmigr_inactive_up > 0.00% swapper [kernel.vmlinux] [k] hrtimer_forward > 0.00% swapper [kernel.vmlinux] [k] cpuidle_reflect > 0.00% perf [kernel.vmlinux] [k] percpu_counter_add_batch > 0.00% swapper [amdgpu] [k] 0x000000000031e745 > 0.00% swapper [kernel.vmlinux] [k] sched_idle_set_state > 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000045675c > 0.00% kworker/6:1-eve [amdgpu] [k] 0x000000000051cbb1 > 0.00% swapper [amdgpu] [k] 0x000000000002dce5 > 0.00% systemd-journal [kernel.vmlinux] [k] __ext4_ext_check > 0.00% swapper [kernel.vmlinux] [k] sched_core_idle_cpu > 0.00% swapper [drm] [k] 0x000000000003de52 > 0.00% swapper [kernel.vmlinux] [k] __msecs_to_jiffies > 0.00% swapper [amdgpu] [k] 0x000000000046f9c4 > 0.00% swapper [kernel.vmlinux] [k] call_cpuidle > 0.00% swapper [kernel.vmlinux] [k] cpuidle_governor_latency_req > 0.00% kworker/7:1-eve [kernel.vmlinux] [k] __switch_to_asm > 0.00% dockerd [kernel.vmlinux] [k] apparmor_file_permission > 0.00% swapper [amdgpu] [k] 0x0000000000126f37 > 0.00% swapper [amdgpu] [k] 0x00000000003218b6 > 0.00% swapper [amdgpu] [k] 0x00000000006e8d65 > 0.00% swapper [amdgpu] [k] 0x000000000096f62d > 0.00% systemd-journal [kernel.vmlinux] [k] __es_tree_search.isra.0 > 0.00% swapper [kernel.vmlinux] [k] update_group_capacity > 0.00% kworker/2:2-eve [amdgpu] [k] 0x00000000005225c3 > 0.00% kworker/2:2-eve [amdgpu] [k] 0x0000000000990c4f > 0.00% swapper [kernel.vmlinux] [k] timer_base_try_to_set_idle > 0.00% swapper [kernel.vmlinux] [k] ct_idle_exit > 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_rbx > 0.00% swapper [kernel.vmlinux] [k] fast_mix > 0.00% swapper [kernel.vmlinux] [k] sched_clock_idle_sleep_event > 0.00% swapper [kernel.vmlinux] [k] touch_softlockup_watchdog_sched > 0.00% containerd [kernel.vmlinux] [k] psi_task_switch > 0.00% swapper [amdgpu] [k] 0x0000000000123820 > 0.00% swapper [kernel.vmlinux] [k] sched_clock > 0.00% swapper [amdgpu] [k] 0x00000000002f46a5 > 0.00% swapper [amdgpu] [k] 0x0000000000449ff5 > 0.00% containerd [kernel.vmlinux] [k] _copy_to_user > 0.00% swapper [kernel.vmlinux] [k] enqueue_task > 0.00% swapper [amdgpu] [k] 0x000000000096f625 > 0.00% swapper [amdgpu] [k] 0x00000000006e7c66 > 0.00% swapper [kernel.vmlinux] [k] hrtimer_interrupt > 0.00% swapper [kernel.vmlinux] [k] irq_enter_rcu > 0.00% perf [kernel.vmlinux] [k] _raw_spin_lock_irqsave > 0.00% perf [kernel.vmlinux] [k] poll_freewait > 0.00% rcu_preempt [kernel.vmlinux] [k] __rcu_read_lock > 0.00% swapper [kernel.vmlinux] [k] wake_q_add > 0.00% kworker/7:1H-kb [kernel.vmlinux] [k] __switch_to_asm > 0.00% swapper [amdgpu] [k] 0x00000000002f8058 > 0.00% swapper [kernel.vmlinux] [k] sched_clock_tick > 0.00% swapper [kernel.vmlinux] [k] slab_update_freelist.constprop.0.isra.0 > 0.00% swapper [kernel.vmlinux] [k] tick_nohz_get_next_hrtimer > 0.00% swapper [kernel.vmlinux] [k] ct_kernel_enter_state > 0.00% swapper [kernel.vmlinux] [k] hrtimer_next_event_without > 0.00% swapper [kernel.vmlinux] [k] resched_curr > 0.00% swapper [amdgpu] [k] 0x0000000000003a55 > 0.00% swapper [amdgpu] [k] 0x000000000045d3df > 0.00% swapper [amdgpu] [k] 0x000000000098aeb9 > 0.00% containerd [kernel.vmlinux] [k] __switch_to > 0.00% swapper [kernel.vmlinux] [k] rcu_sched_clock_irq > 0.00% kworker/9:1-eve [kernel.vmlinux] [k] sched_clock_cpu > 0.00% swapper [amdgpu] [k] 0x0000000000124ef5 > 0.00% containerd [kernel.vmlinux] [k] update_curr > 0.00% swapper [drm] [k] 0x0000000000035f4e > 0.00% swapper [kernel.vmlinux] [k] tmigr_update_events > 0.00% swapper [kernel.vmlinux] [k] profile_pc > 0.00% swapper [nvme] [k] 0x0000000000004315 > 0.00% swapper [kernel.vmlinux] [k] raw_spin_rq_unlock > 0.00% containerd [kernel.vmlinux] [k] posix_get_monotonic_timespec > 0.00% swapper [amdgpu] [k] 0x0000000000002f4b > 0.00% swapper [amdgpu] [k] 0x000000000034bf25 > 0.00% perf [kernel.vmlinux] [k] get_page_from_freelist > 0.00% rcu_preempt [kernel.vmlinux] [k] __mod_timer > 0.00% perf [kernel.vmlinux] [k] mutex_unlock > 0.00% swapper [amdgpu] [k] 0x0000000000460f85 > 0.00% swapper [amdgpu] [k] 0x0000000000977ec5 > 0.00% swapper [amdgpu] [k] 0x0000000000977ed6 > 0.00% swapper [kernel.vmlinux] [k] sched_clock_idle_wakeup_event > 0.00% swapper [amdgpu] [k] 0x0000000000124016 > 0.00% swapper [kernel.vmlinux] [k] pick_next_task_fair > 0.00% swapper [raid6_pq] [k] 0x000000000001acf5 > 0.00% swapper [amdgpu] [k] 0x0000000000977ec6 > 0.00% systemd-journal [kernel.vmlinux] [k] security_inode_setattr > 0.00% swapper [kernel.vmlinux] [k] __sysvec_apic_timer_interrupt > 0.00% swapper [amdgpu] [k] 0x0000000000461985 > 0.00% swapper [amdgpu] [k] 0x0000000000123825 > 0.00% swapper [amdgpu] [k] 0x0000000000126ee9 > 0.00% swapper [amdgpu] [k] 0x000000000096f6cd > 0.00% perf [kernel.vmlinux] [k] do_syscall_64 > 0.00% systemd-journal [kernel.vmlinux] [k] unmap_mapping_range > 0.00% systemd-journal libc.so.6 [.] 0x0000000000090a64 > 0.00% containerd [kernel.vmlinux] [k] dequeue_entity > 0.00% systemd-journal [kernel.vmlinux] [k] dax_layout_busy_page > 0.00% swapper [kernel.vmlinux] [k] apic_ack_irq > 0.00% swapper [drm] [k] 0x000000000003d83b > 0.00% containerd [kernel.vmlinux] [k] __futex_queue > 0.00% swapper [kernel.vmlinux] [k] complete_all > 0.00% swapper [amdgpu] [k] 0x000000000098b0d6 > 0.00% swapper [kernel.vmlinux] [k] can_stop_idle_tick > 0.00% swapper [amdgpu] [k] 0x000000000000dee5 > 0.00% swapper [amdgpu] [k] 0x000000000044ecef > 0.00% swapper [amdgpu] [k] 0x000000000000dee6 > 0.00% migration/14 [kernel.vmlinux] [k] update_curr_common > 0.00% rs:main Q:Reg [kernel.vmlinux] [k] ext4_get_group_desc > 0.00% swapper [kernel.vmlinux] [k] tick_do_update_jiffies64 > 0.00% swapper [kernel.vmlinux] [k] tmigr_cpu_new_timer > 0.00% kworker/7:1-eve [kernel.vmlinux] [k] free_pcppages_bulk > 0.00% dockerd [kernel.vmlinux] [k] srso_untrain_ret > 0.00% dockerd [kernel.vmlinux] [k] update_curr > 0.00% swapper [amdgpu] [k] 0x0000000000124f0c > 0.00% swapper [drm] [k] 0x000000000003f455 > 0.00% swapper [amdgpu] [k] 0x00000000003218b5 > 0.00% swapper [kernel.vmlinux] [k] ct_irq_enter > 0.00% swapper [kernel.vmlinux] [k] ct_irq_exit > 0.00% swapper [kernel.vmlinux] [k] blk_stat_add > 0.00% swapper [amdgpu] [k] 0x00000000006e8dd0 > 0.00% swapper [amdgpu] [k] 0x0000000000124f4c > 0.00% swapper [amdgpu] [k] 0x000000000098aef6 > 0.00% swapper [drm] [k] 0x000000000003f468 > 0.00% swapper [kernel.vmlinux] [k] nohz_balance_enter_idle > 0.00% swapper [kernel.vmlinux] [k] __wake_up_common > 0.00% swapper [amdgpu] [k] 0x00000000001240ea > 0.00% swapper [kernel.vmlinux] [k] __wake_up > 0.00% swapper [amdgpu] [k] 0x00000000002f7f81 > 0.00% swapper [amdgpu] [k] 0x000000000046f97f > 0.00% swapper [drm] [k] 0x000000000003f45c > 0.00% swapper [kernel.vmlinux] [k] __mod_memcg_lruvec_state > 0.00% swapper [kernel.vmlinux] [k] mempool_free_slab > 0.00% swapper [amdgpu] [k] 0x0000000000977ef0 > 0.00% swapper [kernel.vmlinux] [k] x86_pmu_disable_all > 0.00% swapper [amdgpu] [k] 0x00000000002f7f71 > 0.00% kworker/3:1H-kb [kernel.vmlinux] [k] read_hpet > 0.00% swapper [kernel.vmlinux] [k] psi_task_change > 0.00% swapper [kernel.vmlinux] [k] perf_adjust_freq_unthr_context > 0.00% swapper [kernel.vmlinux] [k] account_idle_ticks > 0.00% swapper [amdgpu] [k] 0x000000000045d365 > 0.00% swapper [kernel.vmlinux] [k] ttwu_do_activate > 0.00% swapper [raid6_pq] [k] 0x000000000001acc5 > 0.00% swapper [kernel.vmlinux] [k] enqueue_entity > 0.00% wpa_supplicant [kernel.vmlinux] [k] read_hpet > 0.00% swapper [amdgpu] [k] 0x0000000000012205 > 0.00% swapper [drm] [k] 0x000000000003d649 > 0.00% containerd [kernel.vmlinux] [k] __update_load_avg_se > 0.00% rs:main Q:Reg [kernel.vmlinux] [k] futex_wake > 0.00% swapper [kernel.vmlinux] [k] __x86_indirect_thunk_r13 > 0.00% swapper [kernel.vmlinux] [k] sched_balance_softirq > 0.00% swapper [kernel.vmlinux] [k] timerqueue_del > 0.00% swapper [kernel.vmlinux] [k] cpu_util > > > # > # (Tip: Limit to show entries above 5% only: perf report --percent-limit 5) > #
On Mon, 19 Aug 2024 19:48:23 -0700 Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > +bpf > > On Mon, Aug 19, 2024 at 6:17 PM Masami Hiramatsu <mhiramat@kernel.org> wrote: > > > > On Mon, 19 Aug 2024 15:56:39 +0000 > > Daniel Xu <dlxu@meta.com> wrote: > > > > > [sorry for outlook top post] > > > > > > Hi Masami, > > > > > > test_progs is checked into kernel tree. There should be source files in selftests > > > for the test names. For example, for fill_link_info/kprobe_multi_invalid_ubuff > > > failure: > > > > > > $ find . -name "*fill_link_info*" > > > ./tools/testing/selftests/bpf/prog_tests/fill_link_info.c > > > ./tools/testing/selftests/bpf/progs/test_fill_link_info.c > > > > > > veristat I'm less famiiar with. My understanding is that it checks for verifier > > > regressions. Skimming your patchset, it's not obvious to me why verifier > > > would regress. If you have issues debugging that, we can poke Andrii for > > > help. > > > > Thanks for the information! Hmm, maybe kprobe_multi_testmod_check might check the > > register which is not supported on the ftrace_regs. But I also don't have any idea > > This test is getting IP of the function using bpf_get_func_ip() > helper. If that somehow started returning wrong value on arm64/s390x, > then the test will basically not find expected addresses Ah, I got a clue. For the kprobe_multi (fprobe) uses bpf_kprobe_multi_cookie() to get func_ip, it is implemented as below; static u64 bpf_kprobe_multi_cookie(struct bpf_run_ctx *ctx) { struct bpf_kprobe_multi_run_ctx *run_ctx; struct bpf_kprobe_multi_link *link; u64 *cookie, entry_ip; unsigned long *addr; if (WARN_ON_ONCE(!ctx)) return 0; run_ctx = container_of(current->bpf_ctx, struct bpf_kprobe_multi_run_ctx, session_ctx.run_ctx); link = run_ctx->link; if (!link->cookies) return 0; entry_ip = run_ctx->entry_ip; addr = bsearch(&entry_ip, link->addrs, link->cnt, sizeof(entry_ip), bpf_kprobe_multi_addrs_cmp); if (!addr) return 0; cookie = link->cookies + (addr - link->addrs); return *cookie; } The problem is that the `run_ctx->entry_ip` may not be a real function entry but the ftrace entry address, which is not the same address of the function entry on MANY architectures. X86 uses -fentry and it is able to call another function from the first instruction, but Even on x86, IBT can change it, so it is fixed by get_entry_ip(). On other architectures, usually they need to prepare ftrace stub call by - saving return address to the frame pointer or stack - set ftrace stub address to a register (optional?) - call ftrace stub. Thus the ftrace's entry_ip is a bit shift by some instruction size. To solve this issue, To solve this problem, I think we can fix it by modifying bpf_kprobe_multi_addrs_cmp(). static int bpf_kprobe_multi_addrs_cmp(const void *a, const void *b) { const unsigned long *addr_a = a, *addr_b = b; if (*addr_a >= *addr_b && *addr_a < *addr_b + SOME_OFFSET) return 0; return *addr_a < *addr_b ? -1 : 1; } SOME_OFFSET depends on architecture implmentation. E.g. arm64 is 4 or 8 (if BTI is enabled). Thank you, > > > about veristat. Is that also checks all pt_regs? Andrii, do you have any idea? > > I wouldn't worry about veristat, your changes shouldn't regress BPF > verifier logic, so it's probably just an artifact of our BPF CI setup. > The above test regression seems much more worrying. > > > > > > Thank you, > > > > > > > > Thanks, > > > Daniel > > > > > > > > > ________________________________ > > > From: Masami Hiramatsu <mhiramat@kernel.org> > > > Sent: Sunday, August 18, 2024 5:58 PM > > > To: bot+bpf-ci@kernel.org <bot+bpf-ci@kernel.org> > > > Cc: kernel-ci <kernel-ci@meta.com>; andrii@kernel.org <andrii@kernel.org>; daniel@iogearbox.net <daniel@iogearbox.net>; martin.lau@linux.dev <martin.lau@linux.dev> > > > Subject: Re: [PATCH v13 00/20] tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph > > > > > > Hi, > > > > > > Where can I get the test programs? I would like to check what the programs > > > actually expected. > > > > > > On Sun, 18 Aug 2024 13:51:30 +0000 (UTC) > > > bot+bpf-ci@kernel.org wrote: > > > > > > > Dear patch submitter, > > > > > > > > CI has tested the following submission: > > > > Status: FAILURE > > > > Name: [v13,00/20] tracing: fprobe: function_graph: Multi-function graph and fprobe on fgraph > > > > Patchwork: https://patchwork.kernel.org/project/netdevbpf/list/?series=880630&state=* > > > > Matrix: https://github.com/kernel-patches/bpf/actions/runs/10440799833 > > > > > > > > Failed jobs: > > > > test_progs-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439106 > > > > test_progs_no_alu32-aarch64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911439234 > > > > test_progs-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911405063 > > > > test_progs_no_alu32-s390x-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911404959 > > > > veristat-x86_64-gcc: https://github.com/kernel-patches/bpf/actions/runs/10440799833/job/28911401263 > > > > > > > > First test_progs failure (test_progs-aarch64-gcc): > > > > #126 kprobe_multi_testmod_test > > > > serial_test_kprobe_multi_testmod_test:PASS:load_kallsyms_local 0 nsec > > > > #126/1 kprobe_multi_testmod_test/testmod_attach_api_syms > > > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > > > #126/2 kprobe_multi_testmod_test/testmod_attach_api_addrs > > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > > test_testmod_attach_api_addrs:PASS:ksym_get_addr_local 0 nsec > > > > test_testmod_attach_api:PASS:fentry_raw_skel_load 0 nsec > > > > trigger_module_test_read:PASS:testmod_file_open 0 nsec > > > > test_testmod_attach_api:PASS:trigger_read 0 nsec > > > > kprobe_multi_testmod_check:FAIL:kprobe_test1_result unexpected kprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test2_result unexpected kprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kprobe_test3_result unexpected kprobe_test3_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test1_result unexpected kretprobe_test1_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test2_result unexpected kretprobe_test2_result: actual 0 != expected 1 > > > > kprobe_multi_testmod_check:FAIL:kretprobe_test3_result unexpected kretprobe_test3_result: actual 0 != expected 1 > > > > > > > > > > > > Please note: this email is coming from an unmonitored mailbox. If you have > > > > questions or feedback, please reach out to the Meta Kernel CI team at > > > > kernel-ci@meta.com. > > > > > > > > > -- > > > Masami Hiramatsu (Google) <mhiramat@kernel.org> > > > > > > -- > > Masami Hiramatsu (Google) <mhiramat@kernel.org>
Hi, Here is the 13th version of the series to re-implement the fprobe on function-graph tracer. The previous version is; https://lore.kernel.org/all/172000134410.63468.13742222887213469474.stgit@devnote2/ This version is based on v6.11-rc3. In this version, I added a bugfix as [1/20], which should go to urgent branch, and dropped the performance improvement patch which was introduced in v12 because I found that does not work with new kernel. Overview -------- This series rewrites the fprobe on this function-graph. The purposes of this change are; 1) Remove dependency of the rethook from fprobe so that we can reduce the return hook code and shadow stack. 2) Make 'ftrace_regs' the common trace interface for the function boundary. 1) Currently we have 2(or 3) different function return hook codes, the function-graph tracer and rethook (and legacy kretprobe). But since this is redundant and needs double maintenance cost, I would like to unify those. From the user's viewpoint, function- graph tracer is very useful to grasp the execution path. For this purpose, it is hard to use the rethook in the function-graph tracer, but the opposite is possible. (Strictly speaking, kretprobe can not use it because it requires 'pt_regs' for historical reasons.) 2) Now the fprobe provides the 'pt_regs' for its handler, but that is wrong for the function entry and exit. Moreover, depending on the architecture, there is no way to accurately reproduce 'pt_regs' outside of interrupt or exception handlers. This means fprobe should not use 'pt_regs' because it does not use such exceptions. (Conversely, kprobe should use 'pt_regs' because it is an abstract interface of the software breakpoint exception.) This series changes fprobe to use function-graph tracer for tracing function entry and exit, instead of mixture of ftrace and rethook. Unlike the rethook which is a per-task list of system-wide allocated nodes, the function graph's ret_stack is a per-task shadow stack. Thus it does not need to set 'nr_maxactive' (which is the number of pre-allocated nodes). Also the handlers will get the 'ftrace_regs' instead of 'pt_regs'. Since eBPF mulit_kprobe/multi_kretprobe events still use 'pt_regs' as their register interface, this changes it to convert 'ftrace_regs' to 'pt_regs'. Of course this conversion makes an incomplete 'pt_regs', so users must access only registers for function parameters or return value. Design ------ Instead of using ftrace's function entry hook directly, the new fprobe is built on top of the function-graph's entry and return callbacks with 'ftrace_regs'. Since the fprobe requires access to 'ftrace_regs', the architecture must support CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS and CONFIG_HAVE_FTRACE_GRAPH_FUNC, which enables to call function-graph entry callback with 'ftrace_regs', and also CONFIG_HAVE_FUNCTION_GRAPH_FREGS, which passes the ftrace_regs to return_to_handler. All fprobes share a single function-graph ops (means shares a common ftrace filter) similar to the kprobe-on-ftrace. This needs another layer to find corresponding fprobe in the common function-graph callbacks, but has much better scalability, since the number of registered function-graph ops is limited. In the entry callback, the fprobe runs its entry_handler and saves the address of 'fprobe' on the function-graph's shadow stack as data. The return callback decodes the data to get the 'fprobe' address, and runs the exit_handler. The fprobe introduces two hash-tables, one is for entry callback which searches fprobes related to the given function address passed by entry callback. The other is for a return callback which checks if the given 'fprobe' data structure pointer is still valid. Note that it is possible to unregister fprobe before the return callback runs. Thus the address validation must be done before using it in the return callback. Download -------- This series can be applied against the ftrace/for-next branch in linux-trace tree. This series can also be found below branch. https://git.kernel.org/pub/scm/linux/kernel/git/mhiramat/linux.git/log/?h=topic/fprobe-on-fgraph Thank you, --- Masami Hiramatsu (Google) (20): tracing: fgraph: Fix to add new fgraph_ops to array after ftrace_startup_subops() tracing: Add a comment about ftrace_regs definition tracing: Rename ftrace_regs_return_value to ftrace_regs_get_return_value function_graph: Pass ftrace_regs to entryfunc function_graph: Replace fgraph_ret_regs with ftrace_regs function_graph: Pass ftrace_regs to retfunc fprobe: Use ftrace_regs in fprobe entry handler fprobe: Use ftrace_regs in fprobe exit handler tracing: Add ftrace_partial_regs() for converting ftrace_regs to pt_regs tracing: Add ftrace_fill_perf_regs() for perf event tracing/fprobe: Enable fprobe events with CONFIG_DYNAMIC_FTRACE_WITH_ARGS bpf: Enable kprobe_multi feature if CONFIG_FPROBE is enabled ftrace: Add CONFIG_HAVE_FTRACE_GRAPH_FUNC fprobe: Rewrite fprobe on function-graph tracer tracing: Fix function timing profiler to initialize hashtable tracing/fprobe: Remove nr_maxactive from fprobe selftests: ftrace: Remove obsolate maxactive syntax check selftests/ftrace: Add a test case for repeating register/unregister fprobe Documentation: probes: Update fprobe on function-graph tracer fgraph: Skip recording calltime/rettime if it is not nneeded Documentation/trace/fprobe.rst | 42 + arch/arm64/Kconfig | 2 arch/arm64/include/asm/ftrace.h | 47 + arch/arm64/kernel/asm-offsets.c | 12 arch/arm64/kernel/entry-ftrace.S | 32 + arch/arm64/kernel/ftrace.c | 20 + arch/loongarch/Kconfig | 4 arch/loongarch/include/asm/ftrace.h | 32 - arch/loongarch/kernel/asm-offsets.c | 12 arch/loongarch/kernel/ftrace_dyn.c | 10 arch/loongarch/kernel/mcount.S | 17 - arch/loongarch/kernel/mcount_dyn.S | 14 arch/powerpc/Kconfig | 1 arch/powerpc/include/asm/ftrace.h | 15 arch/powerpc/kernel/trace/ftrace.c | 2 arch/powerpc/kernel/trace/ftrace_64_pg.c | 10 arch/riscv/Kconfig | 3 arch/riscv/include/asm/ftrace.h | 26 - arch/riscv/kernel/ftrace.c | 17 + arch/riscv/kernel/mcount.S | 24 - arch/s390/Kconfig | 3 arch/s390/include/asm/ftrace.h | 39 + arch/s390/kernel/asm-offsets.c | 6 arch/s390/kernel/mcount.S | 9 arch/x86/Kconfig | 4 arch/x86/include/asm/ftrace.h | 37 + arch/x86/kernel/ftrace.c | 50 +- arch/x86/kernel/ftrace_32.S | 15 arch/x86/kernel/ftrace_64.S | 17 - include/linux/fprobe.h | 57 +- include/linux/ftrace.h | 136 ++++ kernel/trace/Kconfig | 23 + kernel/trace/bpf_trace.c | 19 - kernel/trace/fgraph.c | 127 +++- kernel/trace/fprobe.c | 638 ++++++++++++++------ kernel/trace/ftrace.c | 10 kernel/trace/trace.h | 6 kernel/trace/trace_fprobe.c | 147 ++--- kernel/trace/trace_functions_graph.c | 10 kernel/trace/trace_irqsoff.c | 6 kernel/trace/trace_probe_tmpl.h | 2 kernel/trace/trace_sched_wakeup.c | 6 kernel/trace/trace_selftest.c | 11 lib/test_fprobe.c | 51 -- samples/fprobe/fprobe_example.c | 4 .../test.d/dynevent/add_remove_fprobe_repeat.tc | 19 + .../ftrace/test.d/dynevent/fprobe_syntax_errors.tc | 4 47 files changed, 1168 insertions(+), 630 deletions(-) create mode 100644 tools/testing/selftests/ftrace/test.d/dynevent/add_remove_fprobe_repeat.tc -- Masami Hiramatsu (Google) <mhiramat@kernel.org>