| Message ID | 20210111182027.1448538-3-qais.yousef@arm.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | BPF |
| Headers | show |
| Series | Allow attaching to bare tracepoints | expand |
On Mon, Jan 11, 2021 at 10:20 AM Qais Yousef <qais.yousef@arm.com> wrote: > > Reuse module_attach infrastructure to add a new bare tracepoint to check > we can attach to it as a raw tracepoint. > > Signed-off-by: Qais Yousef <qais.yousef@arm.com> > --- > > Andrii > > I was getting the error below when I was trying to run the test. > I had to comment out all related fentry* code to be able to test the raw_tp > stuff. Not sure something I've done wrong or it's broken for some reason. > I was on v5.11-rc2. Check that you have all the required Kconfig options from tools/testing/selftests/bpf/config. And also you will need to build pahole from master, 1.19 doesn't have some fixes that add kernel module support. I think pahole is the reasons why you have the failure below. > > $ sudo ./test_progs -v -t module_attach use -vv when debugging stuff like that with test_progs, it will output libbpf detailed logs, that often are very helpful > bpf_testmod.ko is already unloaded. > Loading bpf_testmod.ko... > Successfully loaded bpf_testmod.ko. > test_module_attach:PASS:skel_open 0 nsec > test_module_attach:PASS:set_attach_target 0 nsec > test_module_attach:PASS:skel_load 0 nsec > libbpf: prog 'handle_fentry': failed to attach: ERROR: strerror_r(-524)=22 > libbpf: failed to auto-attach program 'handle_fentry': -524 > test_module_attach:FAIL:skel_attach skeleton attach failed: -524 > #58 module_attach:FAIL > Successfully unloaded bpf_testmod.ko. > Summary: 0/0 PASSED, 0 SKIPPED, 1 FAILED > But even apart from test failure, there seems to be kernel build failure. See [0] for what fails in kernel-patches CI. [0] https://travis-ci.com/github/kernel-patches/bpf/builds/212730017 > > .../selftests/bpf/bpf_testmod/bpf_testmod-events.h | 6 ++++++ > tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c | 2 ++ > tools/testing/selftests/bpf/prog_tests/module_attach.c | 1 + > tools/testing/selftests/bpf/progs/test_module_attach.c | 10 ++++++++++ > 4 files changed, 19 insertions(+) > > diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > index b83ea448bc79..e1ada753f10c 100644 > --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > @@ -28,6 +28,12 @@ TRACE_EVENT(bpf_testmod_test_read, > __entry->pid, __entry->comm, __entry->off, __entry->len) > ); > > +/* A bare tracepoint with no event associated with it */ > +DECLARE_TRACE(bpf_testmod_test_read_bare, > + TP_PROTO(struct task_struct *task, struct bpf_testmod_test_read_ctx *ctx), > + TP_ARGS(task, ctx) > +); > + > #endif /* _BPF_TESTMOD_EVENTS_H */ > > #undef TRACE_INCLUDE_PATH > diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > index 2df19d73ca49..d63cebdaca44 100644 > --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > @@ -22,6 +22,8 @@ bpf_testmod_test_read(struct file *file, struct kobject *kobj, > }; > > trace_bpf_testmod_test_read(current, &ctx); > + ctx.len++; > + trace_bpf_testmod_test_read_bare(current, &ctx); It's kind of boring to have two read tracepoints :) Do you mind adding a write tracepoint and use bare tracepoint there? You won't need this ctx.len++ hack as well. Feel free to add identical bpf_testmod_test_write_ctx (renaming it is more of a pain). > > return -EIO; /* always fail */ > } > diff --git a/tools/testing/selftests/bpf/prog_tests/module_attach.c b/tools/testing/selftests/bpf/prog_tests/module_attach.c > index 50796b651f72..7085a118f38c 100644 > --- a/tools/testing/selftests/bpf/prog_tests/module_attach.c > +++ b/tools/testing/selftests/bpf/prog_tests/module_attach.c > @@ -50,6 +50,7 @@ void test_module_attach(void) > ASSERT_OK(trigger_module_test_read(READ_SZ), "trigger_read"); > > ASSERT_EQ(bss->raw_tp_read_sz, READ_SZ, "raw_tp"); > + ASSERT_EQ(bss->raw_tp_bare_read_sz, READ_SZ+1, "raw_tp_bare"); > ASSERT_EQ(bss->tp_btf_read_sz, READ_SZ, "tp_btf"); > ASSERT_EQ(bss->fentry_read_sz, READ_SZ, "fentry"); > ASSERT_EQ(bss->fentry_manual_read_sz, READ_SZ, "fentry_manual"); > diff --git a/tools/testing/selftests/bpf/progs/test_module_attach.c b/tools/testing/selftests/bpf/progs/test_module_attach.c > index efd1e287ac17..08aa157afa1d 100644 > --- a/tools/testing/selftests/bpf/progs/test_module_attach.c > +++ b/tools/testing/selftests/bpf/progs/test_module_attach.c > @@ -17,6 +17,16 @@ int BPF_PROG(handle_raw_tp, > return 0; > } > > +__u32 raw_tp_bare_read_sz = 0; > + > +SEC("raw_tp/bpf_testmod_test_read_bare") > +int BPF_PROG(handle_raw_tp_bare, > + struct task_struct *task, struct bpf_testmod_test_read_ctx *read_ctx) > +{ > + raw_tp_bare_read_sz = BPF_CORE_READ(read_ctx, len); > + return 0; > +} > + > __u32 tp_btf_read_sz = 0; > > SEC("tp_btf/bpf_testmod_test_read") > -- > 2.25.1 >
On 01/11/21 23:26, Andrii Nakryiko wrote: > On Mon, Jan 11, 2021 at 10:20 AM Qais Yousef <qais.yousef@arm.com> wrote: > > > > Reuse module_attach infrastructure to add a new bare tracepoint to check > > we can attach to it as a raw tracepoint. > > > > Signed-off-by: Qais Yousef <qais.yousef@arm.com> > > --- > > > > Andrii > > > > I was getting the error below when I was trying to run the test. > > I had to comment out all related fentry* code to be able to test the raw_tp > > stuff. Not sure something I've done wrong or it's broken for some reason. > > I was on v5.11-rc2. > > Check that you have all the required Kconfig options from > tools/testing/selftests/bpf/config. And also you will need to build Yep I have merged this config snippet using merge_config.sh script. > pahole from master, 1.19 doesn't have some fixes that add kernel > module support. I think pahole is the reasons why you have the failure > below. I am using pahole 1.19. I have built it from tip of master though. /trying using v1.19 tag Still fails the same. > > > > > $ sudo ./test_progs -v -t module_attach > > use -vv when debugging stuff like that with test_progs, it will output > libbpf detailed logs, that often are very helpful I tried that but it didn't help me. Full output is here https://paste.debian.net/1180846 > > > bpf_testmod.ko is already unloaded. > > Loading bpf_testmod.ko... > > Successfully loaded bpf_testmod.ko. > > test_module_attach:PASS:skel_open 0 nsec > > test_module_attach:PASS:set_attach_target 0 nsec > > test_module_attach:PASS:skel_load 0 nsec > > libbpf: prog 'handle_fentry': failed to attach: ERROR: strerror_r(-524)=22 > > libbpf: failed to auto-attach program 'handle_fentry': -524 > > test_module_attach:FAIL:skel_attach skeleton attach failed: -524 > > #58 module_attach:FAIL > > Successfully unloaded bpf_testmod.ko. > > Summary: 0/0 PASSED, 0 SKIPPED, 1 FAILED > > > > But even apart from test failure, there seems to be kernel build > failure. See [0] for what fails in kernel-patches CI. > > [0] https://travis-ci.com/github/kernel-patches/bpf/builds/212730017 Sorry about that. I did a last minute change because of checkpatch.pl error and it seems I either forgot to rebuild or missed that the rebuild failed :/ > > > > > > .../selftests/bpf/bpf_testmod/bpf_testmod-events.h | 6 ++++++ > > tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c | 2 ++ > > tools/testing/selftests/bpf/prog_tests/module_attach.c | 1 + > > tools/testing/selftests/bpf/progs/test_module_attach.c | 10 ++++++++++ > > 4 files changed, 19 insertions(+) > > > > diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > > index b83ea448bc79..e1ada753f10c 100644 > > --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > > +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > > @@ -28,6 +28,12 @@ TRACE_EVENT(bpf_testmod_test_read, > > __entry->pid, __entry->comm, __entry->off, __entry->len) > > ); > > > > +/* A bare tracepoint with no event associated with it */ > > +DECLARE_TRACE(bpf_testmod_test_read_bare, > > + TP_PROTO(struct task_struct *task, struct bpf_testmod_test_read_ctx *ctx), > > + TP_ARGS(task, ctx) > > +); > > + > > #endif /* _BPF_TESTMOD_EVENTS_H */ > > > > #undef TRACE_INCLUDE_PATH > > diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > > index 2df19d73ca49..d63cebdaca44 100644 > > --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > > +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > > @@ -22,6 +22,8 @@ bpf_testmod_test_read(struct file *file, struct kobject *kobj, > > }; > > > > trace_bpf_testmod_test_read(current, &ctx); > > + ctx.len++; > > + trace_bpf_testmod_test_read_bare(current, &ctx); > > It's kind of boring to have two read tracepoints :) Do you mind adding Hehe boring is good :p > a write tracepoint and use bare tracepoint there? You won't need this > ctx.len++ hack as well. Feel free to add identical > bpf_testmod_test_write_ctx (renaming it is more of a pain). It was easy to get this done. So I think it should be easy to make it a write too :) Thanks -- Qais Yousef > > > > > return -EIO; /* always fail */ > > } > > diff --git a/tools/testing/selftests/bpf/prog_tests/module_attach.c b/tools/testing/selftests/bpf/prog_tests/module_attach.c > > index 50796b651f72..7085a118f38c 100644 > > --- a/tools/testing/selftests/bpf/prog_tests/module_attach.c > > +++ b/tools/testing/selftests/bpf/prog_tests/module_attach.c > > @@ -50,6 +50,7 @@ void test_module_attach(void) > > ASSERT_OK(trigger_module_test_read(READ_SZ), "trigger_read"); > > > > ASSERT_EQ(bss->raw_tp_read_sz, READ_SZ, "raw_tp"); > > + ASSERT_EQ(bss->raw_tp_bare_read_sz, READ_SZ+1, "raw_tp_bare"); > > ASSERT_EQ(bss->tp_btf_read_sz, READ_SZ, "tp_btf"); > > ASSERT_EQ(bss->fentry_read_sz, READ_SZ, "fentry"); > > ASSERT_EQ(bss->fentry_manual_read_sz, READ_SZ, "fentry_manual"); > > diff --git a/tools/testing/selftests/bpf/progs/test_module_attach.c b/tools/testing/selftests/bpf/progs/test_module_attach.c > > index efd1e287ac17..08aa157afa1d 100644 > > --- a/tools/testing/selftests/bpf/progs/test_module_attach.c > > +++ b/tools/testing/selftests/bpf/progs/test_module_attach.c > > @@ -17,6 +17,16 @@ int BPF_PROG(handle_raw_tp, > > return 0; > > } > > > > +__u32 raw_tp_bare_read_sz = 0; > > + > > +SEC("raw_tp/bpf_testmod_test_read_bare") > > +int BPF_PROG(handle_raw_tp_bare, > > + struct task_struct *task, struct bpf_testmod_test_read_ctx *read_ctx) > > +{ > > + raw_tp_bare_read_sz = BPF_CORE_READ(read_ctx, len); > > + return 0; > > +} > > + > > __u32 tp_btf_read_sz = 0; > > > > SEC("tp_btf/bpf_testmod_test_read") > > -- > > 2.25.1 > >
On Tue, Jan 12, 2021 at 11:27 AM Qais Yousef <qais.yousef@arm.com> wrote: > > On 01/11/21 23:26, Andrii Nakryiko wrote: > > On Mon, Jan 11, 2021 at 10:20 AM Qais Yousef <qais.yousef@arm.com> wrote: > > > > > > Reuse module_attach infrastructure to add a new bare tracepoint to check > > > we can attach to it as a raw tracepoint. > > > > > > Signed-off-by: Qais Yousef <qais.yousef@arm.com> > > > --- > > > > > > Andrii > > > > > > I was getting the error below when I was trying to run the test. > > > I had to comment out all related fentry* code to be able to test the raw_tp > > > stuff. Not sure something I've done wrong or it's broken for some reason. > > > I was on v5.11-rc2. > > > > Check that you have all the required Kconfig options from > > tools/testing/selftests/bpf/config. And also you will need to build > > Yep I have merged this config snippet using merge_config.sh script. > > > pahole from master, 1.19 doesn't have some fixes that add kernel > > module support. I think pahole is the reasons why you have the failure > > below. > > I am using pahole 1.19. I have built it from tip of master though. > > /trying using v1.19 tag > > Still fails the same. > > > > > > > > > $ sudo ./test_progs -v -t module_attach > > > > use -vv when debugging stuff like that with test_progs, it will output > > libbpf detailed logs, that often are very helpful > > I tried that but it didn't help me. Full output is here > > https://paste.debian.net/1180846 > It did help a bit for me to make sure that you have bpf_testmod properly loaded and its BTF was found, so the problem is somewhere else. Also, given load succeeded and attach failed with OPNOTSUPP, I suspect you are missing some of FTRACE configs, which seems to be missing from selftests's config as well. Check that you have CONFIG_FTRACE=y and CONFIG_DYNAMIC_FTRACE=y, and you might need some more. See [0] for a real config we are using to run all tests in libbpf CI. If you figure out what you were missing, please also contribute a patch to selftests' config. [0] https://github.com/libbpf/libbpf/blob/master/travis-ci/vmtest/configs/latest.config > > > > > bpf_testmod.ko is already unloaded. > > > Loading bpf_testmod.ko... > > > Successfully loaded bpf_testmod.ko. > > > test_module_attach:PASS:skel_open 0 nsec > > > test_module_attach:PASS:set_attach_target 0 nsec > > > test_module_attach:PASS:skel_load 0 nsec > > > libbpf: prog 'handle_fentry': failed to attach: ERROR: strerror_r(-524)=22 > > > libbpf: failed to auto-attach program 'handle_fentry': -524 > > > test_module_attach:FAIL:skel_attach skeleton attach failed: -524 > > > #58 module_attach:FAIL > > > Successfully unloaded bpf_testmod.ko. > > > Summary: 0/0 PASSED, 0 SKIPPED, 1 FAILED > > > > > > > But even apart from test failure, there seems to be kernel build > > failure. See [0] for what fails in kernel-patches CI. > > > > [0] https://travis-ci.com/github/kernel-patches/bpf/builds/212730017 > > Sorry about that. I did a last minute change because of checkpatch.pl error and > it seems I either forgot to rebuild or missed that the rebuild failed :/ > no worries, just fix and re-submit. Good that we have CI that caught this early on. > > > > > > > > > > .../selftests/bpf/bpf_testmod/bpf_testmod-events.h | 6 ++++++ > > > tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c | 2 ++ > > > tools/testing/selftests/bpf/prog_tests/module_attach.c | 1 + > > > tools/testing/selftests/bpf/progs/test_module_attach.c | 10 ++++++++++ > > > 4 files changed, 19 insertions(+) > > > > > > diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > > > index b83ea448bc79..e1ada753f10c 100644 > > > --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > > > +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h > > > @@ -28,6 +28,12 @@ TRACE_EVENT(bpf_testmod_test_read, > > > __entry->pid, __entry->comm, __entry->off, __entry->len) > > > ); > > > > > > +/* A bare tracepoint with no event associated with it */ > > > +DECLARE_TRACE(bpf_testmod_test_read_bare, > > > + TP_PROTO(struct task_struct *task, struct bpf_testmod_test_read_ctx *ctx), > > > + TP_ARGS(task, ctx) > > > +); > > > + > > > #endif /* _BPF_TESTMOD_EVENTS_H */ > > > > > > #undef TRACE_INCLUDE_PATH > > > diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > > > index 2df19d73ca49..d63cebdaca44 100644 > > > --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > > > +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c > > > @@ -22,6 +22,8 @@ bpf_testmod_test_read(struct file *file, struct kobject *kobj, > > > }; > > > > > > trace_bpf_testmod_test_read(current, &ctx); > > > + ctx.len++; > > > + trace_bpf_testmod_test_read_bare(current, &ctx); > > > > It's kind of boring to have two read tracepoints :) Do you mind adding > > Hehe boring is good :p > > > a write tracepoint and use bare tracepoint there? You won't need this > > ctx.len++ hack as well. Feel free to add identical > > bpf_testmod_test_write_ctx (renaming it is more of a pain). > > It was easy to get this done. So I think it should be easy to make it a write > too :) yep, having two tracepoints allow more flexibility over longer term, so I think it's good to do (regardless of boring or not ;) ) > > Thanks > > -- > Qais Yousef > > > > > > > > > return -EIO; /* always fail */ > > > } > > > diff --git a/tools/testing/selftests/bpf/prog_tests/module_attach.c b/tools/testing/selftests/bpf/prog_tests/module_attach.c > > > index 50796b651f72..7085a118f38c 100644 > > > --- a/tools/testing/selftests/bpf/prog_tests/module_attach.c > > > +++ b/tools/testing/selftests/bpf/prog_tests/module_attach.c > > > @@ -50,6 +50,7 @@ void test_module_attach(void) > > > ASSERT_OK(trigger_module_test_read(READ_SZ), "trigger_read"); > > > > > > ASSERT_EQ(bss->raw_tp_read_sz, READ_SZ, "raw_tp"); > > > + ASSERT_EQ(bss->raw_tp_bare_read_sz, READ_SZ+1, "raw_tp_bare"); > > > ASSERT_EQ(bss->tp_btf_read_sz, READ_SZ, "tp_btf"); > > > ASSERT_EQ(bss->fentry_read_sz, READ_SZ, "fentry"); > > > ASSERT_EQ(bss->fentry_manual_read_sz, READ_SZ, "fentry_manual"); > > > diff --git a/tools/testing/selftests/bpf/progs/test_module_attach.c b/tools/testing/selftests/bpf/progs/test_module_attach.c > > > index efd1e287ac17..08aa157afa1d 100644 > > > --- a/tools/testing/selftests/bpf/progs/test_module_attach.c > > > +++ b/tools/testing/selftests/bpf/progs/test_module_attach.c > > > @@ -17,6 +17,16 @@ int BPF_PROG(handle_raw_tp, > > > return 0; > > > } > > > > > > +__u32 raw_tp_bare_read_sz = 0; > > > + > > > +SEC("raw_tp/bpf_testmod_test_read_bare") > > > +int BPF_PROG(handle_raw_tp_bare, > > > + struct task_struct *task, struct bpf_testmod_test_read_ctx *read_ctx) > > > +{ > > > + raw_tp_bare_read_sz = BPF_CORE_READ(read_ctx, len); > > > + return 0; > > > +} > > > + > > > __u32 tp_btf_read_sz = 0; > > > > > > SEC("tp_btf/bpf_testmod_test_read") > > > -- > > > 2.25.1 > > >
On 01/12/21 12:07, Andrii Nakryiko wrote: > > > > $ sudo ./test_progs -v -t module_attach > > > > > > use -vv when debugging stuff like that with test_progs, it will output > > > libbpf detailed logs, that often are very helpful > > > > I tried that but it didn't help me. Full output is here > > > > https://paste.debian.net/1180846 > > > > It did help a bit for me to make sure that you have bpf_testmod > properly loaded and its BTF was found, so the problem is somewhere > else. Also, given load succeeded and attach failed with OPNOTSUPP, I > suspect you are missing some of FTRACE configs, which seems to be > missing from selftests's config as well. Check that you have > CONFIG_FTRACE=y and CONFIG_DYNAMIC_FTRACE=y, and you might need some > more. See [0] for a real config we are using to run all tests in > libbpf CI. If you figure out what you were missing, please also > contribute a patch to selftests' config. > > [0] https://github.com/libbpf/libbpf/blob/master/travis-ci/vmtest/configs/latest.config Yeah that occurred to me too. I do have all necessary FTRACE options enabled, including DYNAMIC_FTRACE. I think I did try enabling fault injection too just in case. I have CONFIG_FAULT_INJECTION=y and CONFIG_FUNCTION_ERROR_INJECTION=y. I will look at the CI config and see if I can figure it out. I will likely get a chance to look at all of this and send v2 over the weekend. Thanks -- Qais Yousef
On Wed, Jan 13, 2021 at 10:21:31AM +0000, Qais Yousef wrote: > On 01/12/21 12:07, Andrii Nakryiko wrote: > > > > > $ sudo ./test_progs -v -t module_attach > > > > > > > > use -vv when debugging stuff like that with test_progs, it will output > > > > libbpf detailed logs, that often are very helpful > > > > > > I tried that but it didn't help me. Full output is here > > > > > > https://paste.debian.net/1180846 > > > > > > > It did help a bit for me to make sure that you have bpf_testmod > > properly loaded and its BTF was found, so the problem is somewhere > > else. Also, given load succeeded and attach failed with OPNOTSUPP, I > > suspect you are missing some of FTRACE configs, which seems to be > > missing from selftests's config as well. Check that you have > > CONFIG_FTRACE=y and CONFIG_DYNAMIC_FTRACE=y, and you might need some > > more. See [0] for a real config we are using to run all tests in > > libbpf CI. If you figure out what you were missing, please also > > contribute a patch to selftests' config. > > > > [0] https://github.com/libbpf/libbpf/blob/master/travis-ci/vmtest/configs/latest.config > > Yeah that occurred to me too. I do have all necessary FTRACE options enabled, > including DYNAMIC_FTRACE. I think I did try enabling fault injection too just > in case. I have CONFIG_FAULT_INJECTION=y and CONFIG_FUNCTION_ERROR_INJECTION=y. Could it come from lack of fentry support on arm64 (or are you testing on x86?) Since the arm64 JIT doesn't have trampoline support at the moment, a lot of bpf selftests fail with ENOTSUPP. Thanks, Jean
On 01/13/21 17:40, Jean-Philippe Brucker wrote: > On Wed, Jan 13, 2021 at 10:21:31AM +0000, Qais Yousef wrote: > > On 01/12/21 12:07, Andrii Nakryiko wrote: > > > > > > $ sudo ./test_progs -v -t module_attach > > > > > > > > > > use -vv when debugging stuff like that with test_progs, it will output > > > > > libbpf detailed logs, that often are very helpful > > > > > > > > I tried that but it didn't help me. Full output is here > > > > > > > > https://paste.debian.net/1180846 > > > > > > > > > > It did help a bit for me to make sure that you have bpf_testmod > > > properly loaded and its BTF was found, so the problem is somewhere > > > else. Also, given load succeeded and attach failed with OPNOTSUPP, I > > > suspect you are missing some of FTRACE configs, which seems to be > > > missing from selftests's config as well. Check that you have > > > CONFIG_FTRACE=y and CONFIG_DYNAMIC_FTRACE=y, and you might need some > > > more. See [0] for a real config we are using to run all tests in > > > libbpf CI. If you figure out what you were missing, please also > > > contribute a patch to selftests' config. > > > > > > [0] https://github.com/libbpf/libbpf/blob/master/travis-ci/vmtest/configs/latest.config > > > > Yeah that occurred to me too. I do have all necessary FTRACE options enabled, > > including DYNAMIC_FTRACE. I think I did try enabling fault injection too just > > in case. I have CONFIG_FAULT_INJECTION=y and CONFIG_FUNCTION_ERROR_INJECTION=y. > > Could it come from lack of fentry support on arm64 (or are you testing on > x86?) Since the arm64 JIT doesn't have trampoline support at the moment, a > lot of bpf selftests fail with ENOTSUPP. I am on arm64. I honestly have no clue about this. I'll try to dig out. Thanks -- Qais Yousef
diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h index b83ea448bc79..e1ada753f10c 100644 --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod-events.h @@ -28,6 +28,12 @@ TRACE_EVENT(bpf_testmod_test_read, __entry->pid, __entry->comm, __entry->off, __entry->len) ); +/* A bare tracepoint with no event associated with it */ +DECLARE_TRACE(bpf_testmod_test_read_bare, + TP_PROTO(struct task_struct *task, struct bpf_testmod_test_read_ctx *ctx), + TP_ARGS(task, ctx) +); + #endif /* _BPF_TESTMOD_EVENTS_H */ #undef TRACE_INCLUDE_PATH diff --git a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c index 2df19d73ca49..d63cebdaca44 100644 --- a/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c +++ b/tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c @@ -22,6 +22,8 @@ bpf_testmod_test_read(struct file *file, struct kobject *kobj, }; trace_bpf_testmod_test_read(current, &ctx); + ctx.len++; + trace_bpf_testmod_test_read_bare(current, &ctx); return -EIO; /* always fail */ } diff --git a/tools/testing/selftests/bpf/prog_tests/module_attach.c b/tools/testing/selftests/bpf/prog_tests/module_attach.c index 50796b651f72..7085a118f38c 100644 --- a/tools/testing/selftests/bpf/prog_tests/module_attach.c +++ b/tools/testing/selftests/bpf/prog_tests/module_attach.c @@ -50,6 +50,7 @@ void test_module_attach(void) ASSERT_OK(trigger_module_test_read(READ_SZ), "trigger_read"); ASSERT_EQ(bss->raw_tp_read_sz, READ_SZ, "raw_tp"); + ASSERT_EQ(bss->raw_tp_bare_read_sz, READ_SZ+1, "raw_tp_bare"); ASSERT_EQ(bss->tp_btf_read_sz, READ_SZ, "tp_btf"); ASSERT_EQ(bss->fentry_read_sz, READ_SZ, "fentry"); ASSERT_EQ(bss->fentry_manual_read_sz, READ_SZ, "fentry_manual"); diff --git a/tools/testing/selftests/bpf/progs/test_module_attach.c b/tools/testing/selftests/bpf/progs/test_module_attach.c index efd1e287ac17..08aa157afa1d 100644 --- a/tools/testing/selftests/bpf/progs/test_module_attach.c +++ b/tools/testing/selftests/bpf/progs/test_module_attach.c @@ -17,6 +17,16 @@ int BPF_PROG(handle_raw_tp, return 0; } +__u32 raw_tp_bare_read_sz = 0; + +SEC("raw_tp/bpf_testmod_test_read_bare") +int BPF_PROG(handle_raw_tp_bare, + struct task_struct *task, struct bpf_testmod_test_read_ctx *read_ctx) +{ + raw_tp_bare_read_sz = BPF_CORE_READ(read_ctx, len); + return 0; +} + __u32 tp_btf_read_sz = 0; SEC("tp_btf/bpf_testmod_test_read")
Reuse module_attach infrastructure to add a new bare tracepoint to check we can attach to it as a raw tracepoint. Signed-off-by: Qais Yousef <qais.yousef@arm.com> --- Andrii I was getting the error below when I was trying to run the test. I had to comment out all related fentry* code to be able to test the raw_tp stuff. Not sure something I've done wrong or it's broken for some reason. I was on v5.11-rc2. $ sudo ./test_progs -v -t module_attach bpf_testmod.ko is already unloaded. Loading bpf_testmod.ko... Successfully loaded bpf_testmod.ko. test_module_attach:PASS:skel_open 0 nsec test_module_attach:PASS:set_attach_target 0 nsec test_module_attach:PASS:skel_load 0 nsec libbpf: prog 'handle_fentry': failed to attach: ERROR: strerror_r(-524)=22 libbpf: failed to auto-attach program 'handle_fentry': -524 test_module_attach:FAIL:skel_attach skeleton attach failed: -524 #58 module_attach:FAIL Successfully unloaded bpf_testmod.ko. Summary: 0/0 PASSED, 0 SKIPPED, 1 FAILED .../selftests/bpf/bpf_testmod/bpf_testmod-events.h | 6 ++++++ tools/testing/selftests/bpf/bpf_testmod/bpf_testmod.c | 2 ++ tools/testing/selftests/bpf/prog_tests/module_attach.c | 1 + tools/testing/selftests/bpf/progs/test_module_attach.c | 10 ++++++++++ 4 files changed, 19 insertions(+)