| Message ID | 20220422100025.1469207-5-jolsa@kernel.org (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | BPF |
| Headers | show |
| Series | perf tools: Fix prologue generation | expand |
| Context | Check | Description |
|---|---|---|
| bpf/vmtest-bpf-PR | fail | PR summary |
| bpf/vmtest-bpf-VM_Test-1 | fail | Logs for Kernel LATEST on ubuntu-latest + selftests |
| bpf/vmtest-bpf-VM_Test-2 | fail | Logs for Kernel LATEST on z15 + selftests |
| netdev/tree_selection | success | Not a local patch, async |
On Fri, Apr 22, 2022 at 3:01 AM Jiri Olsa <jolsa@kernel.org> wrote: > > Perf is using section name to declare special kprobe arguments, > which no longer works with current libbpf, that either requires > certain form of the section name or allows to register custom > handler. > > Adding support for 'perfkprobe/' section name handler to take > care of perf kprobe programs. > > The handler servers two purposes: > - allows perf programs to have special arguments in section name > - allows perf to use pre-load callback where we can attach init > code (zeroing all argument registers) to each perf program > > The second is essential part of new prologue generation code, > that's coming in following patch. > > Signed-off-by: Jiri Olsa <jolsa@kernel.org> > --- > tools/perf/util/bpf-loader.c | 50 ++++++++++++++++++++++++++++++++++++ > 1 file changed, 50 insertions(+) > > diff --git a/tools/perf/util/bpf-loader.c b/tools/perf/util/bpf-loader.c > index f8ad581ea247..92dd8cc18edb 100644 > --- a/tools/perf/util/bpf-loader.c > +++ b/tools/perf/util/bpf-loader.c > @@ -86,6 +86,7 @@ bpf_perf_object__next(struct bpf_perf_object *prev) > (perf_obj) = (tmp), (tmp) = bpf_perf_object__next(tmp)) > > static bool libbpf_initialized; > +static int libbpf_sec_handler; > > static int bpf_perf_object__add(struct bpf_object *obj) > { > @@ -99,12 +100,61 @@ static int bpf_perf_object__add(struct bpf_object *obj) > return perf_obj ? 0 : -ENOMEM; > } > > +static struct bpf_insn prologue_init_insn[] = { > + BPF_MOV64_IMM(BPF_REG_0, 0), > + BPF_MOV64_IMM(BPF_REG_1, 0), > + BPF_MOV64_IMM(BPF_REG_2, 0), > + BPF_MOV64_IMM(BPF_REG_3, 0), > + BPF_MOV64_IMM(BPF_REG_4, 0), > + BPF_MOV64_IMM(BPF_REG_5, 0), > +}; > + > +#define LIBBPF_SEC_PREFIX "perfkprobe/" libbpf allows to register fallback handler that will handle any SEC() definition besides the ones that libbpf handles. Would that work in this case instead of adding a custom prefix handler here? See prog_tests/custom_sec_handlers.c for example: fallback_id = libbpf_register_prog_handler(NULL, BPF_PROG_TYPE_SYSCALL, 0, &opts); > + [...]
On Mon, Apr 25, 2022 at 11:22:54PM -0700, Andrii Nakryiko wrote: > On Fri, Apr 22, 2022 at 3:01 AM Jiri Olsa <jolsa@kernel.org> wrote: > > > > Perf is using section name to declare special kprobe arguments, > > which no longer works with current libbpf, that either requires > > certain form of the section name or allows to register custom > > handler. > > > > Adding support for 'perfkprobe/' section name handler to take > > care of perf kprobe programs. > > > > The handler servers two purposes: > > - allows perf programs to have special arguments in section name > > - allows perf to use pre-load callback where we can attach init > > code (zeroing all argument registers) to each perf program > > > > The second is essential part of new prologue generation code, > > that's coming in following patch. > > > > Signed-off-by: Jiri Olsa <jolsa@kernel.org> > > --- > > tools/perf/util/bpf-loader.c | 50 ++++++++++++++++++++++++++++++++++++ > > 1 file changed, 50 insertions(+) > > > > diff --git a/tools/perf/util/bpf-loader.c b/tools/perf/util/bpf-loader.c > > index f8ad581ea247..92dd8cc18edb 100644 > > --- a/tools/perf/util/bpf-loader.c > > +++ b/tools/perf/util/bpf-loader.c > > @@ -86,6 +86,7 @@ bpf_perf_object__next(struct bpf_perf_object *prev) > > (perf_obj) = (tmp), (tmp) = bpf_perf_object__next(tmp)) > > > > static bool libbpf_initialized; > > +static int libbpf_sec_handler; > > > > static int bpf_perf_object__add(struct bpf_object *obj) > > { > > @@ -99,12 +100,61 @@ static int bpf_perf_object__add(struct bpf_object *obj) > > return perf_obj ? 0 : -ENOMEM; > > } > > > > +static struct bpf_insn prologue_init_insn[] = { > > + BPF_MOV64_IMM(BPF_REG_0, 0), > > + BPF_MOV64_IMM(BPF_REG_1, 0), > > + BPF_MOV64_IMM(BPF_REG_2, 0), > > + BPF_MOV64_IMM(BPF_REG_3, 0), > > + BPF_MOV64_IMM(BPF_REG_4, 0), > > + BPF_MOV64_IMM(BPF_REG_5, 0), > > +}; > > + > > +#define LIBBPF_SEC_PREFIX "perfkprobe/" > > libbpf allows to register fallback handler that will handle any SEC() > definition besides the ones that libbpf handles. Would that work in > this case instead of adding a custom prefix handler here? See > prog_tests/custom_sec_handlers.c for example: > > fallback_id = libbpf_register_prog_handler(NULL, > BPF_PROG_TYPE_SYSCALL, 0, &opts); nice, I did not see that.. that should be better, no need for the prefix thanks, jirka
diff --git a/tools/perf/util/bpf-loader.c b/tools/perf/util/bpf-loader.c index f8ad581ea247..92dd8cc18edb 100644 --- a/tools/perf/util/bpf-loader.c +++ b/tools/perf/util/bpf-loader.c @@ -86,6 +86,7 @@ bpf_perf_object__next(struct bpf_perf_object *prev) (perf_obj) = (tmp), (tmp) = bpf_perf_object__next(tmp)) static bool libbpf_initialized; +static int libbpf_sec_handler; static int bpf_perf_object__add(struct bpf_object *obj) { @@ -99,12 +100,61 @@ static int bpf_perf_object__add(struct bpf_object *obj) return perf_obj ? 0 : -ENOMEM; } +static struct bpf_insn prologue_init_insn[] = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_MOV64_IMM(BPF_REG_1, 0), + BPF_MOV64_IMM(BPF_REG_2, 0), + BPF_MOV64_IMM(BPF_REG_3, 0), + BPF_MOV64_IMM(BPF_REG_4, 0), + BPF_MOV64_IMM(BPF_REG_5, 0), +}; + +#define LIBBPF_SEC_PREFIX "perfkprobe/" + +static int libbpf_prog_prepare_load_fn(struct bpf_program *prog, + struct bpf_prog_load_opts *opts __maybe_unused, + long cookie __maybe_unused) +{ + size_t init_size_cnt = ARRAY_SIZE(prologue_init_insn); + size_t orig_insn_cnt, insn_cnt, init_size, orig_size; + const struct bpf_insn *orig_insn; + struct bpf_insn *insn; + + /* prepend initialization code to program instructions */ + orig_insn = bpf_program__insns(prog); + orig_insn_cnt = bpf_program__insn_cnt(prog); + init_size = init_size_cnt * sizeof(*insn); + orig_size = orig_insn_cnt * sizeof(*insn); + + insn_cnt = orig_insn_cnt + init_size_cnt; + insn = malloc(insn_cnt * sizeof(*insn)); + if (!insn) + return -ENOMEM; + + memcpy(insn, prologue_init_insn, init_size); + memcpy((char *) insn + init_size, orig_insn, orig_size); + bpf_program__set_insns(prog, insn, insn_cnt); + return 0; +} + static int libbpf_init(void) { + LIBBPF_OPTS(libbpf_prog_handler_opts, handler_opts, + .prog_prepare_load_fn = libbpf_prog_prepare_load_fn, + ); + if (libbpf_initialized) return 0; libbpf_set_print(libbpf_perf_print); + libbpf_sec_handler = libbpf_register_prog_handler(LIBBPF_SEC_PREFIX, + BPF_PROG_TYPE_KPROBE, + 0, &handler_opts); + if (libbpf_sec_handler < 0) { + pr_debug("bpf: failed to register libbpf section handler: %d\n", + libbpf_sec_handler); + return -BPF_LOADER_ERRNO__INTERNAL; + } libbpf_initialized = true; return 0; }
Perf is using section name to declare special kprobe arguments, which no longer works with current libbpf, that either requires certain form of the section name or allows to register custom handler. Adding support for 'perfkprobe/' section name handler to take care of perf kprobe programs. The handler servers two purposes: - allows perf programs to have special arguments in section name - allows perf to use pre-load callback where we can attach init code (zeroing all argument registers) to each perf program The second is essential part of new prologue generation code, that's coming in following patch. Signed-off-by: Jiri Olsa <jolsa@kernel.org> --- tools/perf/util/bpf-loader.c | 50 ++++++++++++++++++++++++++++++++++++ 1 file changed, 50 insertions(+)