| Message ID | 20220328083703.2880079-1-jolsa@kernel.org (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | ef8a257b4e499a979364b1f9caf25a325f6ee8b8 |
| Delegated to: | BPF |
| Headers | show |
| Series | bpftool: Fix generated code in codegen_asserts | expand |
| Context | Check | Description |
|---|---|---|
| bpf/vmtest-bpf-next-PR | fail | PR summary |
| netdev/tree_selection | success | Not a local patch |
Em Mon, Mar 28, 2022 at 10:37:03AM +0200, Jiri Olsa escreveu: > Arnaldo reported perf compilation fail with: > > $ make -k BUILD_BPF_SKEL=1 CORESIGHT=1 PYTHON=python3 > ... > In file included from util/bpf_counter.c:28: > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h: In function ‘bperf_leader_bpf__assert’: > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h:351:51: error: unused parameter ‘s’ [-Werror=unused-parameter] > 351 | bperf_leader_bpf__assert(struct bperf_leader_bpf *s) > | ~~~~~~~~~~~~~~~~~~~~~~~~~^ > cc1: all warnings being treated as errors > > If there's nothing to generate in the new assert function, > we will get unused 's' warn/error, adding 'unused' attribute to it. We need some Makefile dependency to trigger the skels to be regenerated when bpftool is updated. And before that to notice that the bpftool source code changed and thus the bootstrap bpftool needs to be rebuilt. So, after a full tools/perf/ build from scratch it works: Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com> Ah, and to wire up a tools/perf/ build when doing a selftests build perhaps so that libbpf developers can check if in-kernel tools linking against it still builds when they change libbpf/bpftool? Regards, - Arnaldo > Cc: Delyan Kratunov <delyank@fb.com> > Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com> > Fixes: 08d4dba6ae77 ("bpftool: Bpf skeletons assert type sizes") > Signed-off-by: Jiri Olsa <jolsa@kernel.org> > --- > tools/bpf/bpftool/gen.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c > index 7ba7ff55d2ea..91af2850b505 100644 > --- a/tools/bpf/bpftool/gen.c > +++ b/tools/bpf/bpftool/gen.c > @@ -477,7 +477,7 @@ static void codegen_asserts(struct bpf_object *obj, const char *obj_name) > codegen("\ > \n\ > __attribute__((unused)) static void \n\ > - %1$s__assert(struct %1$s *s) \n\ > + %1$s__assert(struct %1$s *s __attribute__((unused))) \n\ > { \n\ > #ifdef __cplusplus \n\ > #define _Static_assert static_assert \n\ > -- > 2.35.1
On 3/28/22 1:37 AM, Jiri Olsa wrote: > Arnaldo reported perf compilation fail with: > > $ make -k BUILD_BPF_SKEL=1 CORESIGHT=1 PYTHON=python3 > ... > In file included from util/bpf_counter.c:28: > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h: In function ‘bperf_leader_bpf__assert’: > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h:351:51: error: unused parameter ‘s’ [-Werror=unused-parameter] > 351 | bperf_leader_bpf__assert(struct bperf_leader_bpf *s) > | ~~~~~~~~~~~~~~~~~~~~~~~~~^ > cc1: all warnings being treated as errors > > If there's nothing to generate in the new assert function, > we will get unused 's' warn/error, adding 'unused' attribute to it. If there is nothing to generate, should we avoid generating the assert function itself? > > Cc: Delyan Kratunov <delyank@fb.com> > Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com> > Fixes: 08d4dba6ae77 ("bpftool: Bpf skeletons assert type sizes") > Signed-off-by: Jiri Olsa <jolsa@kernel.org> > --- > tools/bpf/bpftool/gen.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c > index 7ba7ff55d2ea..91af2850b505 100644 > --- a/tools/bpf/bpftool/gen.c > +++ b/tools/bpf/bpftool/gen.c > @@ -477,7 +477,7 @@ static void codegen_asserts(struct bpf_object *obj, const char *obj_name) > codegen("\ > \n\ > __attribute__((unused)) static void \n\ > - %1$s__assert(struct %1$s *s) \n\ > + %1$s__assert(struct %1$s *s __attribute__((unused))) \n\ > { \n\ > #ifdef __cplusplus \n\ > #define _Static_assert static_assert \n\
On Mon, Mar 28, 2022 at 08:41:18AM -0700, Yonghong Song wrote: > > > On 3/28/22 1:37 AM, Jiri Olsa wrote: > > Arnaldo reported perf compilation fail with: > > > > $ make -k BUILD_BPF_SKEL=1 CORESIGHT=1 PYTHON=python3 > > ... > > In file included from util/bpf_counter.c:28: > > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h: In function ‘bperf_leader_bpf__assert’: > > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h:351:51: error: unused parameter ‘s’ [-Werror=unused-parameter] > > 351 | bperf_leader_bpf__assert(struct bperf_leader_bpf *s) > > | ~~~~~~~~~~~~~~~~~~~~~~~~~^ > > cc1: all warnings being treated as errors > > > > If there's nothing to generate in the new assert function, > > we will get unused 's' warn/error, adding 'unused' attribute to it. > > If there is nothing to generate, should we avoid generating > the assert function itself? good point, will check jirka > > > > > Cc: Delyan Kratunov <delyank@fb.com> > > Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com> > > Fixes: 08d4dba6ae77 ("bpftool: Bpf skeletons assert type sizes") > > Signed-off-by: Jiri Olsa <jolsa@kernel.org> > > --- > > tools/bpf/bpftool/gen.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c > > index 7ba7ff55d2ea..91af2850b505 100644 > > --- a/tools/bpf/bpftool/gen.c > > +++ b/tools/bpf/bpftool/gen.c > > @@ -477,7 +477,7 @@ static void codegen_asserts(struct bpf_object *obj, const char *obj_name) > > codegen("\ > > \n\ > > __attribute__((unused)) static void \n\ > > - %1$s__assert(struct %1$s *s) \n\ > > + %1$s__assert(struct %1$s *s __attribute__((unused))) \n\ > > { \n\ > > #ifdef __cplusplus \n\ > > #define _Static_assert static_assert \n\
On Mon, Mar 28, 2022 at 11:50 AM Jiri Olsa <olsajiri@gmail.com> wrote: > > On Mon, Mar 28, 2022 at 08:41:18AM -0700, Yonghong Song wrote: > > > > > > On 3/28/22 1:37 AM, Jiri Olsa wrote: > > > Arnaldo reported perf compilation fail with: > > > > > > $ make -k BUILD_BPF_SKEL=1 CORESIGHT=1 PYTHON=python3 > > > ... > > > In file included from util/bpf_counter.c:28: > > > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h: In function ‘bperf_leader_bpf__assert’: > > > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h:351:51: error: unused parameter ‘s’ [-Werror=unused-parameter] > > > 351 | bperf_leader_bpf__assert(struct bperf_leader_bpf *s) > > > | ~~~~~~~~~~~~~~~~~~~~~~~~~^ > > > cc1: all warnings being treated as errors > > > > > > If there's nothing to generate in the new assert function, > > > we will get unused 's' warn/error, adding 'unused' attribute to it. > > > > If there is nothing to generate, should we avoid generating > > the assert function itself? > > good point, will check we can use this function for some more assertions in the future, so instead of trying to be smart about generating or not of this function, I think unused attribute is a more robust solution. > > jirka > > > > > > > > > Cc: Delyan Kratunov <delyank@fb.com> > > > Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com> > > > Fixes: 08d4dba6ae77 ("bpftool: Bpf skeletons assert type sizes") > > > Signed-off-by: Jiri Olsa <jolsa@kernel.org> > > > --- > > > tools/bpf/bpftool/gen.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c > > > index 7ba7ff55d2ea..91af2850b505 100644 > > > --- a/tools/bpf/bpftool/gen.c > > > +++ b/tools/bpf/bpftool/gen.c > > > @@ -477,7 +477,7 @@ static void codegen_asserts(struct bpf_object *obj, const char *obj_name) > > > codegen("\ > > > \n\ > > > __attribute__((unused)) static void \n\ > > > - %1$s__assert(struct %1$s *s) \n\ > > > + %1$s__assert(struct %1$s *s __attribute__((unused))) \n\ > > > { \n\ > > > #ifdef __cplusplus \n\ > > > #define _Static_assert static_assert \n\
On 3/28/22 2:31 PM, Andrii Nakryiko wrote: > On Mon, Mar 28, 2022 at 11:50 AM Jiri Olsa <olsajiri@gmail.com> wrote: >> >> On Mon, Mar 28, 2022 at 08:41:18AM -0700, Yonghong Song wrote: >>> >>> >>> On 3/28/22 1:37 AM, Jiri Olsa wrote: >>>> Arnaldo reported perf compilation fail with: >>>> >>>> $ make -k BUILD_BPF_SKEL=1 CORESIGHT=1 PYTHON=python3 >>>> ... >>>> In file included from util/bpf_counter.c:28: >>>> /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h: In function ‘bperf_leader_bpf__assert’: >>>> /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h:351:51: error: unused parameter ‘s’ [-Werror=unused-parameter] >>>> 351 | bperf_leader_bpf__assert(struct bperf_leader_bpf *s) >>>> | ~~~~~~~~~~~~~~~~~~~~~~~~~^ >>>> cc1: all warnings being treated as errors >>>> >>>> If there's nothing to generate in the new assert function, >>>> we will get unused 's' warn/error, adding 'unused' attribute to it. >>> >>> If there is nothing to generate, should we avoid generating >>> the assert function itself? >> >> good point, will check > > we can use this function for some more assertions in the future, so > instead of trying to be smart about generating or not of this > function, I think unused attribute is a more robust solution. Okay, if there are possibly more assertions down the road, I am fine to keep the function even if it is empty to avoid more and more conditions to decide whether the functions should be generated at all. > >> >> jirka >> >>> >>>> >>>> Cc: Delyan Kratunov <delyank@fb.com> >>>> Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com> >>>> Fixes: 08d4dba6ae77 ("bpftool: Bpf skeletons assert type sizes") >>>> Signed-off-by: Jiri Olsa <jolsa@kernel.org> >>>> --- >>>> tools/bpf/bpftool/gen.c | 2 +- >>>> 1 file changed, 1 insertion(+), 1 deletion(-) >>>> >>>> diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c >>>> index 7ba7ff55d2ea..91af2850b505 100644 >>>> --- a/tools/bpf/bpftool/gen.c >>>> +++ b/tools/bpf/bpftool/gen.c >>>> @@ -477,7 +477,7 @@ static void codegen_asserts(struct bpf_object *obj, const char *obj_name) >>>> codegen("\ >>>> \n\ >>>> __attribute__((unused)) static void \n\ >>>> - %1$s__assert(struct %1$s *s) \n\ >>>> + %1$s__assert(struct %1$s *s __attribute__((unused))) \n\ >>>> { \n\ >>>> #ifdef __cplusplus \n\ >>>> #define _Static_assert static_assert \n\
Hello: This patch was applied to bpf/bpf.git (master) by Alexei Starovoitov <ast@kernel.org>: On Mon, 28 Mar 2022 10:37:03 +0200 you wrote: > Arnaldo reported perf compilation fail with: > > $ make -k BUILD_BPF_SKEL=1 CORESIGHT=1 PYTHON=python3 > ... > In file included from util/bpf_counter.c:28: > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h: In function ‘bperf_leader_bpf__assert’: > /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h:351:51: error: unused parameter ‘s’ [-Werror=unused-parameter] > 351 | bperf_leader_bpf__assert(struct bperf_leader_bpf *s) > | ~~~~~~~~~~~~~~~~~~~~~~~~~^ > cc1: all warnings being treated as errors > > [...] Here is the summary with links: - bpftool: Fix generated code in codegen_asserts https://git.kernel.org/bpf/bpf/c/ef8a257b4e49 You are awesome, thank you!
diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c index 7ba7ff55d2ea..91af2850b505 100644 --- a/tools/bpf/bpftool/gen.c +++ b/tools/bpf/bpftool/gen.c @@ -477,7 +477,7 @@ static void codegen_asserts(struct bpf_object *obj, const char *obj_name) codegen("\ \n\ __attribute__((unused)) static void \n\ - %1$s__assert(struct %1$s *s) \n\ + %1$s__assert(struct %1$s *s __attribute__((unused))) \n\ { \n\ #ifdef __cplusplus \n\ #define _Static_assert static_assert \n\
Arnaldo reported perf compilation fail with: $ make -k BUILD_BPF_SKEL=1 CORESIGHT=1 PYTHON=python3 ... In file included from util/bpf_counter.c:28: /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h: In function ‘bperf_leader_bpf__assert’: /tmp/build/perf//util/bpf_skel/bperf_leader.skel.h:351:51: error: unused parameter ‘s’ [-Werror=unused-parameter] 351 | bperf_leader_bpf__assert(struct bperf_leader_bpf *s) | ~~~~~~~~~~~~~~~~~~~~~~~~~^ cc1: all warnings being treated as errors If there's nothing to generate in the new assert function, we will get unused 's' warn/error, adding 'unused' attribute to it. Cc: Delyan Kratunov <delyank@fb.com> Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com> Fixes: 08d4dba6ae77 ("bpftool: Bpf skeletons assert type sizes") Signed-off-by: Jiri Olsa <jolsa@kernel.org> --- tools/bpf/bpftool/gen.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)