* [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link @ 2019-08-06 23:42 Daniel Xu 2019-08-06 23:42 ` [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE Daniel Xu ` (2 more replies) 0 siblings, 3 replies; 7+ messages in thread From: Daniel Xu @ 2019-08-06 23:42 UTC (permalink / raw) To: songliubraving, yhs, andriin; +Cc: Daniel Xu, bpf, kernel-team It is sometimes necessary to perform ioctl's on the underlying perf fd. There is not currently a way to extract the fd given a bpf_link, so add a helper for it. --- tools/lib/bpf/libbpf.c | 13 +++++++++++++ tools/lib/bpf/libbpf.h | 1 + tools/lib/bpf/libbpf.map | 5 +++++ 3 files changed, 19 insertions(+) diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c index ead915aec349..8469d69448ae 100644 --- a/tools/lib/bpf/libbpf.c +++ b/tools/lib/bpf/libbpf.c @@ -4004,6 +4004,19 @@ static int bpf_link__destroy_perf_event(struct bpf_link *link) return err; } +int bpf_link__get_perf_fd(struct bpf_link *link) +{ + struct bpf_link_fd *l = (void *)link; + + if (!link) + return -1; + + if (link->destroy != &bpf_link__destroy_perf_event) + return -1; + + return l->fd; +} + struct bpf_link *bpf_program__attach_perf_event(struct bpf_program *prog, int pfd) { diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h index 8a9d462a6f6d..5391ac95e4fa 100644 --- a/tools/lib/bpf/libbpf.h +++ b/tools/lib/bpf/libbpf.h @@ -168,6 +168,7 @@ LIBBPF_API void bpf_program__unload(struct bpf_program *prog); struct bpf_link; LIBBPF_API int bpf_link__destroy(struct bpf_link *link); +LIBBPF_API int bpf_link__get_perf_fd(struct bpf_link *link); LIBBPF_API struct bpf_link * bpf_program__attach_perf_event(struct bpf_program *prog, int pfd); diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map index f9d316e873d8..0f844ce29b04 100644 --- a/tools/lib/bpf/libbpf.map +++ b/tools/lib/bpf/libbpf.map @@ -184,3 +184,8 @@ LIBBPF_0.0.4 { perf_buffer__new_raw; perf_buffer__poll; } LIBBPF_0.0.3; + +LIBBPF_0.0.5 { + global: + bpf_link__get_perf_fd; +} LIBBPF_0.0.4; -- 2.20.1 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE 2019-08-06 23:42 [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link Daniel Xu @ 2019-08-06 23:42 ` Daniel Xu 2019-08-07 6:01 ` Yonghong Song 2019-08-07 21:49 ` Andrii Nakryiko 2019-08-07 18:19 ` [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link Song Liu 2019-08-07 18:57 ` Andrii Nakryiko 2 siblings, 2 replies; 7+ messages in thread From: Daniel Xu @ 2019-08-06 23:42 UTC (permalink / raw) To: songliubraving, yhs, andriin; +Cc: Daniel Xu, bpf, kernel-team --- tools/include/uapi/linux/perf_event.h | 23 ++++++++++ .../selftests/bpf/prog_tests/attach_probe.c | 43 +++++++++++++++++++ 2 files changed, 66 insertions(+) diff --git a/tools/include/uapi/linux/perf_event.h b/tools/include/uapi/linux/perf_event.h index 7198ddd0c6b1..4a5e18606baf 100644 --- a/tools/include/uapi/linux/perf_event.h +++ b/tools/include/uapi/linux/perf_event.h @@ -447,6 +447,28 @@ struct perf_event_query_bpf { __u32 ids[0]; }; +/* + * Structure used by below PERF_EVENT_IOC_QUERY_KPROE command + * to query information about the kprobe attached to the perf + * event. + */ +struct perf_event_query_kprobe { + /* + * Size of structure for forward/backward compatibility + */ + __u32 size; + /* + * Set by the kernel to indicate number of times this kprobe + * was temporarily disabled + */ + __u64 nmissed; + /* + * Set by the kernel to indicate number of times this kprobe + * was hit + */ + __u64 nhit; +}; + /* * Ioctls that can be done on a perf event fd: */ @@ -462,6 +484,7 @@ struct perf_event_query_bpf { #define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW('$', 9, __u32) #define PERF_EVENT_IOC_QUERY_BPF _IOWR('$', 10, struct perf_event_query_bpf *) #define PERF_EVENT_IOC_MODIFY_ATTRIBUTES _IOW('$', 11, struct perf_event_attr *) +#define PERF_EVENT_IOC_QUERY_KPROBE _IOWR('$', 12, struct perf_event_query_kprobe *) enum perf_event_ioc_flags { PERF_IOC_FLAG_GROUP = 1U << 0, diff --git a/tools/testing/selftests/bpf/prog_tests/attach_probe.c b/tools/testing/selftests/bpf/prog_tests/attach_probe.c index 5ecc267d98b0..5f118e9a1469 100644 --- a/tools/testing/selftests/bpf/prog_tests/attach_probe.c +++ b/tools/testing/selftests/bpf/prog_tests/attach_probe.c @@ -38,9 +38,12 @@ void test_attach_probe(void) struct bpf_link *kretprobe_link = NULL; struct bpf_link *uprobe_link = NULL; struct bpf_link *uretprobe_link = NULL; + int kprobe_fd, kretprobe_fd; int results_map_fd; size_t uprobe_offset; ssize_t base_addr; + struct perf_event_query_kprobe kprobe_query; + struct perf_event_query_kprobe kretprobe_query; base_addr = get_base_addr(); if (CHECK(base_addr < 0, "get_base_addr", @@ -116,6 +119,46 @@ void test_attach_probe(void) /* trigger & validate kprobe && kretprobe */ usleep(1); + kprobe_fd = bpf_link__get_perf_fd(kprobe_link); + if (CHECK(kprobe_fd < 0, "kprobe_get_perf_fd", + "failed to get perf fd from kprobe link\n")) + goto cleanup; + + kretprobe_fd = bpf_link__get_perf_fd(kretprobe_link); + if (CHECK(kprobe_fd < 0, "kprobe_get_perf_fd", + "failed to get perf fd from kprobe link\n")) + goto cleanup; + + memset(&kprobe_query, 0, sizeof(kprobe_query)); + kprobe_query.size = sizeof(kprobe_query); + err = ioctl(kprobe_fd, PERF_EVENT_IOC_QUERY_KPROBE, &kprobe_query); + if (CHECK(err, "get_kprobe_ioctl", + "failed to issue kprobe query ioctl\n")) + goto cleanup; + if (CHECK(kprobe_query.nmissed > 0, "get_kprobe_ioctl", + "read incorect nmissed from kprobe_ioctl: %llu\n", + kprobe_query.nmissed)) + goto cleanup; + if (CHECK(kprobe_query.nhit <= 0, "get_kprobe_ioctl", + "read incorect nhit from kprobe_ioctl: %llu\n", + kprobe_query.nhit)) + goto cleanup; + + memset(&kretprobe_query, 0, sizeof(kretprobe_query)); + kretprobe_query.size = sizeof(kretprobe_query); + err = ioctl(kretprobe_fd, PERF_EVENT_IOC_QUERY_KPROBE, &kretprobe_query); + if (CHECK(err, "get_kretprobe_ioctl", + "failed to issue kprobe query ioctl\n")) + goto cleanup; + if (CHECK(kretprobe_query.nmissed > 0, "get_kretprobe_ioctl", + "read incorect nmissed from kretprobe_ioctl: %llu\n", + kretprobe_query.nmissed)) + goto cleanup; + if (CHECK(kretprobe_query.nhit <= 0, "get_kretprobe_ioctl", + "read incorect nhit from kretprobe_ioctl: %llu\n", + kretprobe_query.nhit)) + goto cleanup; + err = bpf_map_lookup_elem(results_map_fd, &kprobe_idx, &res); if (CHECK(err, "get_kprobe_res", "failed to get kprobe res: %d\n", err)) -- 2.20.1 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE 2019-08-06 23:42 ` [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE Daniel Xu @ 2019-08-07 6:01 ` Yonghong Song 2019-08-07 21:49 ` Andrii Nakryiko 1 sibling, 0 replies; 7+ messages in thread From: Yonghong Song @ 2019-08-07 6:01 UTC (permalink / raw) To: Daniel Xu, Song Liu, Andrii Nakryiko; +Cc: bpf, Kernel Team On 8/6/19 4:42 PM, Daniel Xu wrote: > --- > tools/include/uapi/linux/perf_event.h | 23 ++++++++++ > .../selftests/bpf/prog_tests/attach_probe.c | 43 +++++++++++++++++++ Could you separate this into two patches? one patch for uapi update tools/include/uapi/linux/perf_event.h and the other for .../selftests/bpf/prog_tests/attach_probe.c The reason is to help libbpf mirror sync'ing. > 2 files changed, 66 insertions(+) > > diff --git a/tools/include/uapi/linux/perf_event.h b/tools/include/uapi/linux/perf_event.h > index 7198ddd0c6b1..4a5e18606baf 100644 > --- a/tools/include/uapi/linux/perf_event.h > +++ b/tools/include/uapi/linux/perf_event.h > @@ -447,6 +447,28 @@ struct perf_event_query_bpf { > __u32 ids[0]; > }; > > +/* > + * Structure used by below PERF_EVENT_IOC_QUERY_KPROE command > + * to query information about the kprobe attached to the perf > + * event. > + */ > +struct perf_event_query_kprobe { > + /* > + * Size of structure for forward/backward compatibility > + */ > + __u32 size; > + /* > + * Set by the kernel to indicate number of times this kprobe > + * was temporarily disabled > + */ > + __u64 nmissed; > + /* > + * Set by the kernel to indicate number of times this kprobe > + * was hit > + */ > + __u64 nhit; > +}; > + > /* > * Ioctls that can be done on a perf event fd: > */ > @@ -462,6 +484,7 @@ struct perf_event_query_bpf { > #define PERF_EVENT_IOC_PAUSE_OUTPUT _IOW('$', 9, __u32) > #define PERF_EVENT_IOC_QUERY_BPF _IOWR('$', 10, struct perf_event_query_bpf *) > #define PERF_EVENT_IOC_MODIFY_ATTRIBUTES _IOW('$', 11, struct perf_event_attr *) > +#define PERF_EVENT_IOC_QUERY_KPROBE _IOWR('$', 12, struct perf_event_query_kprobe *) > > enum perf_event_ioc_flags { > PERF_IOC_FLAG_GROUP = 1U << 0, > diff --git a/tools/testing/selftests/bpf/prog_tests/attach_probe.c b/tools/testing/selftests/bpf/prog_tests/attach_probe.c > index 5ecc267d98b0..5f118e9a1469 100644 > --- a/tools/testing/selftests/bpf/prog_tests/attach_probe.c > +++ b/tools/testing/selftests/bpf/prog_tests/attach_probe.c > @@ -38,9 +38,12 @@ void test_attach_probe(void) > struct bpf_link *kretprobe_link = NULL; > struct bpf_link *uprobe_link = NULL; > struct bpf_link *uretprobe_link = NULL; > + int kprobe_fd, kretprobe_fd; > int results_map_fd; > size_t uprobe_offset; > ssize_t base_addr; > + struct perf_event_query_kprobe kprobe_query; > + struct perf_event_query_kprobe kretprobe_query; Since you are adding new fields, could you try to maintain reverse Christmas coding style? > > base_addr = get_base_addr(); > if (CHECK(base_addr < 0, "get_base_addr", > @@ -116,6 +119,46 @@ void test_attach_probe(void) > /* trigger & validate kprobe && kretprobe */ > usleep(1); > > + kprobe_fd = bpf_link__get_perf_fd(kprobe_link); > + if (CHECK(kprobe_fd < 0, "kprobe_get_perf_fd", > + "failed to get perf fd from kprobe link\n")) > + goto cleanup; > + > + kretprobe_fd = bpf_link__get_perf_fd(kretprobe_link); > + if (CHECK(kprobe_fd < 0, "kprobe_get_perf_fd", typo: kprobe_fd => kretprobe_fd, kretprobe_get_perf_fd? > + "failed to get perf fd from kprobe link\n")) kretprobe link? > + goto cleanup; > + > + memset(&kprobe_query, 0, sizeof(kprobe_query)); > + kprobe_query.size = sizeof(kprobe_query); > + err = ioctl(kprobe_fd, PERF_EVENT_IOC_QUERY_KPROBE, &kprobe_query); > + if (CHECK(err, "get_kprobe_ioctl", > + "failed to issue kprobe query ioctl\n")) > + goto cleanup; > + if (CHECK(kprobe_query.nmissed > 0, "get_kprobe_ioctl", > + "read incorect nmissed from kprobe_ioctl: %llu\n", > + kprobe_query.nmissed)) > + goto cleanup; > + if (CHECK(kprobe_query.nhit <= 0, "get_kprobe_ioctl", nhit is __u64, it cannot be less than 0. > + "read incorect nhit from kprobe_ioctl: %llu\n", > + kprobe_query.nhit)) > + goto cleanup; > + > + memset(&kretprobe_query, 0, sizeof(kretprobe_query)); > + kretprobe_query.size = sizeof(kretprobe_query); > + err = ioctl(kretprobe_fd, PERF_EVENT_IOC_QUERY_KPROBE, &kretprobe_query); > + if (CHECK(err, "get_kretprobe_ioctl", > + "failed to issue kprobe query ioctl\n")) > + goto cleanup; > + if (CHECK(kretprobe_query.nmissed > 0, "get_kretprobe_ioctl", > + "read incorect nmissed from kretprobe_ioctl: %llu\n", > + kretprobe_query.nmissed)) > + goto cleanup; > + if (CHECK(kretprobe_query.nhit <= 0, "get_kretprobe_ioctl", <= 0 => == 0? > + "read incorect nhit from kretprobe_ioctl: %llu\n", > + kretprobe_query.nhit)) > + goto cleanup; > + > err = bpf_map_lookup_elem(results_map_fd, &kprobe_idx, &res); > if (CHECK(err, "get_kprobe_res", > "failed to get kprobe res: %d\n", err)) > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE 2019-08-06 23:42 ` [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE Daniel Xu 2019-08-07 6:01 ` Yonghong Song @ 2019-08-07 21:49 ` Andrii Nakryiko 1 sibling, 0 replies; 7+ messages in thread From: Andrii Nakryiko @ 2019-08-07 21:49 UTC (permalink / raw) To: Daniel Xu; +Cc: Song Liu, Yonghong Song, Andrii Nakryiko, bpf, Kernel Team On Tue, Aug 6, 2019 at 4:42 PM Daniel Xu <dxu@dxuuu.xyz> wrote: > > --- > tools/include/uapi/linux/perf_event.h | 23 ++++++++++ > .../selftests/bpf/prog_tests/attach_probe.c | 43 +++++++++++++++++++ > 2 files changed, 66 insertions(+) > > diff --git a/tools/include/uapi/linux/perf_event.h b/tools/include/uapi/linux/perf_event.h > index 7198ddd0c6b1..4a5e18606baf 100644 > --- a/tools/include/uapi/linux/perf_event.h > +++ b/tools/include/uapi/linux/perf_event.h Please split header sync into separate patch, this will cause problems syncing libbpf to Github. > @@ -447,6 +447,28 @@ struct perf_event_query_bpf { > __u32 ids[0]; > }; > [...] > ssize_t base_addr; > + struct perf_event_query_kprobe kprobe_query; > + struct perf_event_query_kprobe kretprobe_query; Add = {} and avoid using memsets below. > > base_addr = get_base_addr(); > if (CHECK(base_addr < 0, "get_base_addr", > @@ -116,6 +119,46 @@ void test_attach_probe(void) > /* trigger & validate kprobe && kretprobe */ > usleep(1); > [...] ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link 2019-08-06 23:42 [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link Daniel Xu 2019-08-06 23:42 ` [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE Daniel Xu @ 2019-08-07 18:19 ` Song Liu 2019-08-07 18:57 ` Andrii Nakryiko 2 siblings, 0 replies; 7+ messages in thread From: Song Liu @ 2019-08-07 18:19 UTC (permalink / raw) To: Daniel Xu; +Cc: Song Liu, Yonghong Song, Andrii Nakryiko, bpf, Kernel Team On Tue, Aug 6, 2019 at 4:42 PM Daniel Xu <dxu@dxuuu.xyz> wrote: > > It is sometimes necessary to perform ioctl's on the underlying perf fd. > There is not currently a way to extract the fd given a bpf_link, so add a > helper for it. Missing "Signed-off-by" tag. Please run scripts/checkpatch.pl on the patches. Otherwise, looks good to me. Acked-by: Song Liu <songliubraving@fb.com> > --- > tools/lib/bpf/libbpf.c | 13 +++++++++++++ > tools/lib/bpf/libbpf.h | 1 + > tools/lib/bpf/libbpf.map | 5 +++++ > 3 files changed, 19 insertions(+) > > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c > index ead915aec349..8469d69448ae 100644 > --- a/tools/lib/bpf/libbpf.c > +++ b/tools/lib/bpf/libbpf.c > @@ -4004,6 +4004,19 @@ static int bpf_link__destroy_perf_event(struct bpf_link *link) > return err; > } > > +int bpf_link__get_perf_fd(struct bpf_link *link) > +{ > + struct bpf_link_fd *l = (void *)link; > + > + if (!link) > + return -1; > + > + if (link->destroy != &bpf_link__destroy_perf_event) > + return -1; > + > + return l->fd; > +} > + > struct bpf_link *bpf_program__attach_perf_event(struct bpf_program *prog, > int pfd) > { > diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h > index 8a9d462a6f6d..5391ac95e4fa 100644 > --- a/tools/lib/bpf/libbpf.h > +++ b/tools/lib/bpf/libbpf.h > @@ -168,6 +168,7 @@ LIBBPF_API void bpf_program__unload(struct bpf_program *prog); > struct bpf_link; > > LIBBPF_API int bpf_link__destroy(struct bpf_link *link); > +LIBBPF_API int bpf_link__get_perf_fd(struct bpf_link *link); > > LIBBPF_API struct bpf_link * > bpf_program__attach_perf_event(struct bpf_program *prog, int pfd); > diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map > index f9d316e873d8..0f844ce29b04 100644 > --- a/tools/lib/bpf/libbpf.map > +++ b/tools/lib/bpf/libbpf.map > @@ -184,3 +184,8 @@ LIBBPF_0.0.4 { > perf_buffer__new_raw; > perf_buffer__poll; > } LIBBPF_0.0.3; > + > +LIBBPF_0.0.5 { > + global: > + bpf_link__get_perf_fd; > +} LIBBPF_0.0.4; > -- > 2.20.1 > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link 2019-08-06 23:42 [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link Daniel Xu 2019-08-06 23:42 ` [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE Daniel Xu 2019-08-07 18:19 ` [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link Song Liu @ 2019-08-07 18:57 ` Andrii Nakryiko 2019-08-09 1:31 ` Daniel Xu 2 siblings, 1 reply; 7+ messages in thread From: Andrii Nakryiko @ 2019-08-07 18:57 UTC (permalink / raw) To: Daniel Xu; +Cc: Song Liu, Yonghong Song, Andrii Nakryiko, bpf, Kernel Team On Tue, Aug 6, 2019 at 4:42 PM Daniel Xu <dxu@dxuuu.xyz> wrote: > > It is sometimes necessary to perform ioctl's on the underlying perf fd. > There is not currently a way to extract the fd given a bpf_link, so add a > helper for it. > --- So I've been going back and forth with this approach and the alternative one, and I think I'm leaning towards the alternative one still. I think it's better to have a broad "categories" of bpf_links, e.g.: - FD-based bpf_link (which is the only one we have right now): bpf_link_fd. It's not just for perf FD-based ones, raw tracepoint is not, but it's still FD-based; - for cgroup-related links (once they are added), it will be bpf_link_cg (or something along the lines); - there probably should be separate XDP-related bpf_link with device ID/name inside; - etc, whatever we'll need. Then we can have a set of casting APIs and getter APIs that extract useful information from specific type of bpf_link. We can also add direct bpf_link creation API (e.g., from known FD), for cases where it makes sense. So something like (in libbpf.h): struct bpf_link_fd; struct bpf_link_cg; /* casting APIs */ const struct bpf_link_fd *bpf_link__as_fd(const struct bpf_link *link); const struct bpf_link_cg *bpf_link__as_cg(const struct bpf_link *link); /* getters APIs */ int bpf_link_fd__fd(const struct bpf_link_fd *link); int bpf_link_cg__cgroup_fd(const struct bpf_link_cg *link); /* link factories (in addition to attach APIs) */ const struct bpf_link_fd *bpf_link__from_fd(int fd); const struct bpf_link_cg *bpf_link__from_cg(int cg_fd, /* whatever else necessary */); I think this way it becomes obvious what you can expect to get of each possible type of bpf_link and you'll have to explicitly cast to the right type. Yet we still hide implementation details, allow no-brainer bpf_link__destroy regardless of specific type of link (which probably will be a common case). Thoughts? > tools/lib/bpf/libbpf.c | 13 +++++++++++++ > tools/lib/bpf/libbpf.h | 1 + > tools/lib/bpf/libbpf.map | 5 +++++ > 3 files changed, 19 insertions(+) > > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c > index ead915aec349..8469d69448ae 100644 > --- a/tools/lib/bpf/libbpf.c > +++ b/tools/lib/bpf/libbpf.c > @@ -4004,6 +4004,19 @@ static int bpf_link__destroy_perf_event(struct bpf_link *link) > return err; > } > > +int bpf_link__get_perf_fd(struct bpf_link *link) this seems like a bit too specific name (and we should avoid "get" words, as we do in a bunch of other libbpf APIs for getters). Maybe just `bpf_link__fd`? This especially makes sense with a "file-based bpf_link" abstraction I proposed above. > +{ > + struct bpf_link_fd *l = (void *)link; > + > + if (!link) > + return -1; > + > + if (link->destroy != &bpf_link__destroy_perf_event) > + return -1; > + > + return l->fd; > +} > + > struct bpf_link *bpf_program__attach_perf_event(struct bpf_program *prog, > int pfd) > { > diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h > index 8a9d462a6f6d..5391ac95e4fa 100644 > --- a/tools/lib/bpf/libbpf.h > +++ b/tools/lib/bpf/libbpf.h > @@ -168,6 +168,7 @@ LIBBPF_API void bpf_program__unload(struct bpf_program *prog); > struct bpf_link; > > LIBBPF_API int bpf_link__destroy(struct bpf_link *link); > +LIBBPF_API int bpf_link__get_perf_fd(struct bpf_link *link); > > LIBBPF_API struct bpf_link * > bpf_program__attach_perf_event(struct bpf_program *prog, int pfd); > diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map > index f9d316e873d8..0f844ce29b04 100644 > --- a/tools/lib/bpf/libbpf.map > +++ b/tools/lib/bpf/libbpf.map > @@ -184,3 +184,8 @@ LIBBPF_0.0.4 { > perf_buffer__new_raw; > perf_buffer__poll; > } LIBBPF_0.0.3; > + > +LIBBPF_0.0.5 { > + global: > + bpf_link__get_perf_fd; > +} LIBBPF_0.0.4; > -- > 2.20.1 > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link 2019-08-07 18:57 ` Andrii Nakryiko @ 2019-08-09 1:31 ` Daniel Xu 0 siblings, 0 replies; 7+ messages in thread From: Daniel Xu @ 2019-08-09 1:31 UTC (permalink / raw) To: Andrii Nakryiko Cc: Song Liu, Yonghong Song, Andrii Nakryiko, bpf, Kernel Team On Wed, Aug 7, 2019, at 11:57 AM, Andrii Nakryiko wrote: > On Tue, Aug 6, 2019 at 4:42 PM Daniel Xu <dxu@dxuuu.xyz> wrote: > > > > It is sometimes necessary to perform ioctl's on the underlying perf fd. > > There is not currently a way to extract the fd given a bpf_link, so add a > > helper for it. > > --- > > So I've been going back and forth with this approach and the > alternative one, and I think I'm leaning towards the alternative one > still. > > I think it's better to have a broad "categories" of bpf_links, e.g.: > > - FD-based bpf_link (which is the only one we have right now): > bpf_link_fd. It's not just for perf FD-based ones, raw tracepoint is > not, but it's still FD-based; > - for cgroup-related links (once they are added), it will be > bpf_link_cg (or something along the lines); > - there probably should be separate XDP-related bpf_link with device > ID/name inside; > - etc, whatever we'll need. > > Then we can have a set of casting APIs and getter APIs that extract > useful information from specific type of bpf_link. We can also add > direct bpf_link creation API (e.g., from known FD), for cases where it > makes sense. > > So something like (in libbpf.h): > > struct bpf_link_fd; > struct bpf_link_cg; > > /* casting APIs */ > const struct bpf_link_fd *bpf_link__as_fd(const struct bpf_link *link); > const struct bpf_link_cg *bpf_link__as_cg(const struct bpf_link *link); > > /* getters APIs */ > int bpf_link_fd__fd(const struct bpf_link_fd *link); > int bpf_link_cg__cgroup_fd(const struct bpf_link_cg *link); > > /* link factories (in addition to attach APIs) */ > const struct bpf_link_fd *bpf_link__from_fd(int fd); > const struct bpf_link_cg *bpf_link__from_cg(int cg_fd, /* whatever > else necessary */); > > I think this way it becomes obvious what you can expect to get of each > possible type of bpf_link and you'll have to explicitly cast to the > right type. Yet we still hide implementation details, allow no-brainer > bpf_link__destroy regardless of specific type of link (which probably > will be a common case). > > Thoughts? Makes sense to me. This would probably result in a more predictable API when new types are added. I'll make it this way in V2. > > > tools/lib/bpf/libbpf.c | 13 +++++++++++++ > > tools/lib/bpf/libbpf.h | 1 + > > tools/lib/bpf/libbpf.map | 5 +++++ > > 3 files changed, 19 insertions(+) > > > > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c > > index ead915aec349..8469d69448ae 100644 > > --- a/tools/lib/bpf/libbpf.c > > +++ b/tools/lib/bpf/libbpf.c > > @@ -4004,6 +4004,19 @@ static int bpf_link__destroy_perf_event(struct bpf_link *link) > > return err; > > } > > > > +int bpf_link__get_perf_fd(struct bpf_link *link) > > this seems like a bit too specific name (and we should avoid "get" > words, as we do in a bunch of other libbpf APIs for getters). Maybe > just `bpf_link__fd`? This especially makes sense with a "file-based > bpf_link" abstraction I proposed above. Ok. ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2019-08-09 1:31 UTC | newest] Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-08-06 23:42 [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link Daniel Xu 2019-08-06 23:42 ` [PATCH 3/3] tracing/kprobe: Add self test for PERF_EVENT_IOC_QUERY_KPROBE Daniel Xu 2019-08-07 6:01 ` Yonghong Song 2019-08-07 21:49 ` Andrii Nakryiko 2019-08-07 18:19 ` [PATCH 2/3] libbpf: Add helper to extract perf fd from bpf_link Song Liu 2019-08-07 18:57 ` Andrii Nakryiko 2019-08-09 1:31 ` Daniel Xu
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).