From ec6d5f47bfe36f46aa0de707e5beb2f58d96b76d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= Date: Sat, 9 Nov 2019 21:37:27 +0100 Subject: [PATCH 1/6] libbpf: Unpin auto-pinned maps if loading fails MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since the automatic map-pinning happens during load, it will leave pinned maps around if the load fails at a later stage. Fix this by unpinning any pinned maps on cleanup. To avoid unpinning pinned maps that were reused rather than newly pinned, add a new boolean property on struct bpf_map to keep track of whether that map was reused or not; and only unpin those maps that were not reused. Fixes: 57a00f41644f ("libbpf: Add auto-pinning of maps when loading BPF objects") Signed-off-by: Toke Høiland-Jørgensen Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Acked-by: David S. Miller Acked-by: Song Liu Link: https://lore.kernel.org/bpf/157333184731.88376.9992935027056165873.stgit@toke.dk --- tools/lib/bpf/libbpf.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c index fde6cb3e5d41..d8ead85903d1 100644 --- a/tools/lib/bpf/libbpf.c +++ b/tools/lib/bpf/libbpf.c @@ -229,6 +229,7 @@ struct bpf_map { enum libbpf_map_type libbpf_type; char *pin_path; bool pinned; + bool reused; }; struct bpf_secdata { @@ -1995,6 +1996,7 @@ int bpf_map__reuse_fd(struct bpf_map *map, int fd) map->def.map_flags = info.map_flags; map->btf_key_type_id = info.btf_key_type_id; map->btf_value_type_id = info.btf_value_type_id; + map->reused = true; return 0; @@ -4026,7 +4028,7 @@ int bpf_object__unload(struct bpf_object *obj) int bpf_object__load_xattr(struct bpf_object_load_attr *attr) { struct bpf_object *obj; - int err; + int err, i; if (!attr) return -EINVAL; @@ -4047,6 +4049,11 @@ int bpf_object__load_xattr(struct bpf_object_load_attr *attr) return 0; out: + /* unpin any maps that were auto-pinned during load */ + for (i = 0; i < obj->nr_maps; i++) + if (obj->maps[i].pinned && !obj->maps[i].reused) + bpf_map__unpin(&obj->maps[i], NULL); + bpf_object__unload(obj); pr_warn("failed to load object '%s'\n", obj->path); return err; From 9c4e395a1e8c1dc14189b88f24d111f80353b9a4 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= Date: Sat, 9 Nov 2019 21:37:28 +0100 Subject: [PATCH 2/6] selftests/bpf: Add tests for automatic map unpinning on load failure MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This add tests for the different variations of automatic map unpinning on load failure. Signed-off-by: Toke Høiland-Jørgensen Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Acked-by: David S. Miller Acked-by: Song Liu Link: https://lore.kernel.org/bpf/157333184838.88376.8243704248624814775.stgit@toke.dk --- .../selftests/bpf/prog_tests/pinning.c | 20 ++++++++++++++++--- .../selftests/bpf/progs/test_pinning.c | 2 +- 2 files changed, 18 insertions(+), 4 deletions(-) diff --git a/tools/testing/selftests/bpf/prog_tests/pinning.c b/tools/testing/selftests/bpf/prog_tests/pinning.c index 525388971e08..041952524c55 100644 --- a/tools/testing/selftests/bpf/prog_tests/pinning.c +++ b/tools/testing/selftests/bpf/prog_tests/pinning.c @@ -163,12 +163,15 @@ void test_pinning(void) goto out; } - /* swap pin paths of the two maps */ + /* set pin paths so that nopinmap2 will attempt to reuse the map at + * pinpath (which will fail), but not before pinmap has already been + * reused + */ bpf_object__for_each_map(map, obj) { if (!strcmp(bpf_map__name(map), "nopinmap")) + err = bpf_map__set_pin_path(map, nopinpath2); + else if (!strcmp(bpf_map__name(map), "nopinmap2")) err = bpf_map__set_pin_path(map, pinpath); - else if (!strcmp(bpf_map__name(map), "pinmap")) - err = bpf_map__set_pin_path(map, NULL); else continue; @@ -181,6 +184,17 @@ void test_pinning(void) if (CHECK(err != -EINVAL, "param mismatch load", "err %d errno %d\n", err, errno)) goto out; + /* nopinmap2 should have been pinned and cleaned up again */ + err = stat(nopinpath2, &statbuf); + if (CHECK(!err || errno != ENOENT, "stat nopinpath2", + "err %d errno %d\n", err, errno)) + goto out; + + /* pinmap should still be there */ + err = stat(pinpath, &statbuf); + if (CHECK(err, "stat pinpath", "err %d errno %d\n", err, errno)) + goto out; + bpf_object__close(obj); /* test auto-pinning at custom path with open opt */ diff --git a/tools/testing/selftests/bpf/progs/test_pinning.c b/tools/testing/selftests/bpf/progs/test_pinning.c index f69a4a50d056..f20e7e00373f 100644 --- a/tools/testing/selftests/bpf/progs/test_pinning.c +++ b/tools/testing/selftests/bpf/progs/test_pinning.c @@ -21,7 +21,7 @@ struct { } nopinmap SEC(".maps"); struct { - __uint(type, BPF_MAP_TYPE_ARRAY); + __uint(type, BPF_MAP_TYPE_HASH); __uint(max_entries, 1); __type(key, __u32); __type(value, __u64); From 4f33ddb4e3e28ade924fbf05ec8bcd7d5c022fee Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= Date: Sat, 9 Nov 2019 21:37:29 +0100 Subject: [PATCH 3/6] libbpf: Propagate EPERM to caller on program load MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When loading an eBPF program, libbpf overrides the return code for EPERM errors instead of returning it to the caller. This makes it hard to figure out what went wrong on load. In particular, EPERM is returned when the system rlimit is too low to lock the memory required for the BPF program. Previously, this was somewhat obscured because the rlimit error would be hit on map creation (which does return it correctly). However, since maps can now be reused, object load can proceed all the way to loading programs without hitting the error; propagating it even in this case makes it possible for the caller to react appropriately (and, e.g., attempt to raise the rlimit before retrying). Signed-off-by: Toke Høiland-Jørgensen Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Acked-by: David S. Miller Acked-by: Song Liu Link: https://lore.kernel.org/bpf/157333184946.88376.11768171652794234561.stgit@toke.dk --- tools/lib/bpf/libbpf.c | 25 ++++++++++--------------- 1 file changed, 10 insertions(+), 15 deletions(-) diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c index d8ead85903d1..847b62895f9b 100644 --- a/tools/lib/bpf/libbpf.c +++ b/tools/lib/bpf/libbpf.c @@ -3721,7 +3721,7 @@ load_program(struct bpf_program *prog, struct bpf_insn *insns, int insns_cnt, free(log_buf); goto retry_load; } - ret = -LIBBPF_ERRNO__LOAD; + ret = -errno; cp = libbpf_strerror_r(errno, errmsg, sizeof(errmsg)); pr_warn("load bpf program failed: %s\n", cp); @@ -3734,23 +3734,18 @@ load_program(struct bpf_program *prog, struct bpf_insn *insns, int insns_cnt, pr_warn("Program too large (%zu insns), at most %d insns\n", load_attr.insns_cnt, BPF_MAXINSNS); ret = -LIBBPF_ERRNO__PROG2BIG; - } else { + } else if (load_attr.prog_type != BPF_PROG_TYPE_KPROBE) { /* Wrong program type? */ - if (load_attr.prog_type != BPF_PROG_TYPE_KPROBE) { - int fd; + int fd; - load_attr.prog_type = BPF_PROG_TYPE_KPROBE; - load_attr.expected_attach_type = 0; - fd = bpf_load_program_xattr(&load_attr, NULL, 0); - if (fd >= 0) { - close(fd); - ret = -LIBBPF_ERRNO__PROGTYPE; - goto out; - } + load_attr.prog_type = BPF_PROG_TYPE_KPROBE; + load_attr.expected_attach_type = 0; + fd = bpf_load_program_xattr(&load_attr, NULL, 0); + if (fd >= 0) { + close(fd); + ret = -LIBBPF_ERRNO__PROGTYPE; + goto out; } - - if (log_buf) - ret = -LIBBPF_ERRNO__KVER; } out: From b6e99b010ecf829fd8453a7a77e389501bb81990 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= Date: Sat, 9 Nov 2019 21:37:30 +0100 Subject: [PATCH 4/6] libbpf: Use pr_warn() when printing netlink errors MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The netlink functions were using fprintf(stderr, ) directly to print out error messages, instead of going through the usual logging macros. This makes it impossible for the calling application to silence or redirect those error messages. Fix this by switching to pr_warn() in nlattr.c and netlink.c. Signed-off-by: Toke Høiland-Jørgensen Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Acked-by: David S. Miller Acked-by: Song Liu Link: https://lore.kernel.org/bpf/157333185055.88376.15999360127117901443.stgit@toke.dk --- tools/lib/bpf/netlink.c | 3 ++- tools/lib/bpf/nlattr.c | 10 +++++----- 2 files changed, 7 insertions(+), 6 deletions(-) diff --git a/tools/lib/bpf/netlink.c b/tools/lib/bpf/netlink.c index ce3ec81b71c0..a261df9cb488 100644 --- a/tools/lib/bpf/netlink.c +++ b/tools/lib/bpf/netlink.c @@ -12,6 +12,7 @@ #include "bpf.h" #include "libbpf.h" +#include "libbpf_internal.h" #include "nlattr.h" #ifndef SOL_NETLINK @@ -43,7 +44,7 @@ int libbpf_netlink_open(__u32 *nl_pid) if (setsockopt(sock, SOL_NETLINK, NETLINK_EXT_ACK, &one, sizeof(one)) < 0) { - fprintf(stderr, "Netlink error reporting not supported\n"); + pr_warn("Netlink error reporting not supported\n"); } if (bind(sock, (struct sockaddr *)&sa, sizeof(sa)) < 0) { diff --git a/tools/lib/bpf/nlattr.c b/tools/lib/bpf/nlattr.c index 1e69c0c8d413..8db44bbfc66d 100644 --- a/tools/lib/bpf/nlattr.c +++ b/tools/lib/bpf/nlattr.c @@ -8,6 +8,7 @@ #include #include "nlattr.h" +#include "libbpf_internal.h" #include #include #include @@ -121,8 +122,8 @@ int libbpf_nla_parse(struct nlattr *tb[], int maxtype, struct nlattr *head, } if (tb[type]) - fprintf(stderr, "Attribute of type %#x found multiple times in message, " - "previous attribute is being ignored.\n", type); + pr_warn("Attribute of type %#x found multiple times in message, " + "previous attribute is being ignored.\n", type); tb[type] = nla; } @@ -181,15 +182,14 @@ int libbpf_nla_dump_errormsg(struct nlmsghdr *nlh) if (libbpf_nla_parse(tb, NLMSGERR_ATTR_MAX, attr, alen, extack_policy) != 0) { - fprintf(stderr, - "Failed to parse extended error attributes\n"); + pr_warn("Failed to parse extended error attributes\n"); return 0; } if (tb[NLMSGERR_ATTR_MSG]) errmsg = (char *) libbpf_nla_data(tb[NLMSGERR_ATTR_MSG]); - fprintf(stderr, "Kernel error message: %s\n", errmsg); + pr_warn("Kernel error message: %s\n", errmsg); return 0; } From 473f4e133a12dd083bae044ba1782be4767177c1 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= Date: Sat, 9 Nov 2019 21:37:31 +0100 Subject: [PATCH 5/6] libbpf: Add bpf_get_link_xdp_info() function to get more XDP information MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Currently, libbpf only provides a function to get a single ID for the XDP program attached to the interface. However, it can be useful to get the full set of program IDs attached, along with the attachment mode, in one go. Add a new getter function to support this, using an extendible structure to carry the information. Express the old bpf_get_link_id() function in terms of the new function. Signed-off-by: Toke Høiland-Jørgensen Signed-off-by: Alexei Starovoitov Acked-by: David S. Miller Acked-by: Song Liu Acked-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/157333185164.88376.7520653040667637246.stgit@toke.dk --- tools/lib/bpf/libbpf.h | 10 +++++ tools/lib/bpf/libbpf.map | 1 + tools/lib/bpf/netlink.c | 84 ++++++++++++++++++++++++++-------------- 3 files changed, 67 insertions(+), 28 deletions(-) diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h index 6ddc0419337b..f0947cc949d2 100644 --- a/tools/lib/bpf/libbpf.h +++ b/tools/lib/bpf/libbpf.h @@ -427,8 +427,18 @@ LIBBPF_API int bpf_prog_load_xattr(const struct bpf_prog_load_attr *attr, LIBBPF_API int bpf_prog_load(const char *file, enum bpf_prog_type type, struct bpf_object **pobj, int *prog_fd); +struct xdp_link_info { + __u32 prog_id; + __u32 drv_prog_id; + __u32 hw_prog_id; + __u32 skb_prog_id; + __u8 attach_mode; +}; + LIBBPF_API int bpf_set_link_xdp_fd(int ifindex, int fd, __u32 flags); LIBBPF_API int bpf_get_link_xdp_id(int ifindex, __u32 *prog_id, __u32 flags); +LIBBPF_API int bpf_get_link_xdp_info(int ifindex, struct xdp_link_info *info, + size_t info_size, __u32 flags); struct perf_buffer; diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map index 86173cbb159d..d1a782a3a58d 100644 --- a/tools/lib/bpf/libbpf.map +++ b/tools/lib/bpf/libbpf.map @@ -193,6 +193,7 @@ LIBBPF_0.0.5 { LIBBPF_0.0.6 { global: + bpf_get_link_xdp_info; bpf_map__get_pin_path; bpf_map__is_pinned; bpf_map__set_pin_path; diff --git a/tools/lib/bpf/netlink.c b/tools/lib/bpf/netlink.c index a261df9cb488..5065c1aa1061 100644 --- a/tools/lib/bpf/netlink.c +++ b/tools/lib/bpf/netlink.c @@ -25,7 +25,7 @@ typedef int (*__dump_nlmsg_t)(struct nlmsghdr *nlmsg, libbpf_dump_nlmsg_t, struct xdp_id_md { int ifindex; __u32 flags; - __u32 id; + struct xdp_link_info info; }; int libbpf_netlink_open(__u32 *nl_pid) @@ -203,26 +203,11 @@ static int __dump_link_nlmsg(struct nlmsghdr *nlh, return dump_link_nlmsg(cookie, ifi, tb); } -static unsigned char get_xdp_id_attr(unsigned char mode, __u32 flags) -{ - if (mode != XDP_ATTACHED_MULTI) - return IFLA_XDP_PROG_ID; - if (flags & XDP_FLAGS_DRV_MODE) - return IFLA_XDP_DRV_PROG_ID; - if (flags & XDP_FLAGS_HW_MODE) - return IFLA_XDP_HW_PROG_ID; - if (flags & XDP_FLAGS_SKB_MODE) - return IFLA_XDP_SKB_PROG_ID; - - return IFLA_XDP_UNSPEC; -} - -static int get_xdp_id(void *cookie, void *msg, struct nlattr **tb) +static int get_xdp_info(void *cookie, void *msg, struct nlattr **tb) { struct nlattr *xdp_tb[IFLA_XDP_MAX + 1]; struct xdp_id_md *xdp_id = cookie; struct ifinfomsg *ifinfo = msg; - unsigned char mode, xdp_attr; int ret; if (xdp_id->ifindex && xdp_id->ifindex != ifinfo->ifi_index) @@ -238,27 +223,40 @@ static int get_xdp_id(void *cookie, void *msg, struct nlattr **tb) if (!xdp_tb[IFLA_XDP_ATTACHED]) return 0; - mode = libbpf_nla_getattr_u8(xdp_tb[IFLA_XDP_ATTACHED]); - if (mode == XDP_ATTACHED_NONE) + xdp_id->info.attach_mode = libbpf_nla_getattr_u8( + xdp_tb[IFLA_XDP_ATTACHED]); + + if (xdp_id->info.attach_mode == XDP_ATTACHED_NONE) return 0; - xdp_attr = get_xdp_id_attr(mode, xdp_id->flags); - if (!xdp_attr || !xdp_tb[xdp_attr]) - return 0; + if (xdp_tb[IFLA_XDP_PROG_ID]) + xdp_id->info.prog_id = libbpf_nla_getattr_u32( + xdp_tb[IFLA_XDP_PROG_ID]); - xdp_id->id = libbpf_nla_getattr_u32(xdp_tb[xdp_attr]); + if (xdp_tb[IFLA_XDP_SKB_PROG_ID]) + xdp_id->info.skb_prog_id = libbpf_nla_getattr_u32( + xdp_tb[IFLA_XDP_SKB_PROG_ID]); + + if (xdp_tb[IFLA_XDP_DRV_PROG_ID]) + xdp_id->info.drv_prog_id = libbpf_nla_getattr_u32( + xdp_tb[IFLA_XDP_DRV_PROG_ID]); + + if (xdp_tb[IFLA_XDP_HW_PROG_ID]) + xdp_id->info.hw_prog_id = libbpf_nla_getattr_u32( + xdp_tb[IFLA_XDP_HW_PROG_ID]); return 0; } -int bpf_get_link_xdp_id(int ifindex, __u32 *prog_id, __u32 flags) +int bpf_get_link_xdp_info(int ifindex, struct xdp_link_info *info, + size_t info_size, __u32 flags) { struct xdp_id_md xdp_id = {}; int sock, ret; __u32 nl_pid; __u32 mask; - if (flags & ~XDP_FLAGS_MASK) + if (flags & ~XDP_FLAGS_MASK || !info_size) return -EINVAL; /* Check whether the single {HW,DRV,SKB} mode is set */ @@ -274,14 +272,44 @@ int bpf_get_link_xdp_id(int ifindex, __u32 *prog_id, __u32 flags) xdp_id.ifindex = ifindex; xdp_id.flags = flags; - ret = libbpf_nl_get_link(sock, nl_pid, get_xdp_id, &xdp_id); - if (!ret) - *prog_id = xdp_id.id; + ret = libbpf_nl_get_link(sock, nl_pid, get_xdp_info, &xdp_id); + if (!ret) { + size_t sz = min(info_size, sizeof(xdp_id.info)); + + memcpy(info, &xdp_id.info, sz); + memset((void *) info + sz, 0, info_size - sz); + } close(sock); return ret; } +static __u32 get_xdp_id(struct xdp_link_info *info, __u32 flags) +{ + if (info->attach_mode != XDP_ATTACHED_MULTI) + return info->prog_id; + if (flags & XDP_FLAGS_DRV_MODE) + return info->drv_prog_id; + if (flags & XDP_FLAGS_HW_MODE) + return info->hw_prog_id; + if (flags & XDP_FLAGS_SKB_MODE) + return info->skb_prog_id; + + return 0; +} + +int bpf_get_link_xdp_id(int ifindex, __u32 *prog_id, __u32 flags) +{ + struct xdp_link_info info; + int ret; + + ret = bpf_get_link_xdp_info(ifindex, &info, sizeof(info), flags); + if (!ret) + *prog_id = get_xdp_id(&info, flags); + + return ret; +} + int libbpf_nl_get_link(int sock, unsigned int nl_pid, libbpf_dump_nlmsg_t dump_link_nlmsg, void *cookie) { From 1a734efe06948c17122808f74f0c8cc550c10cf5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Toke=20H=C3=B8iland-J=C3=B8rgensen?= Date: Sat, 9 Nov 2019 21:37:32 +0100 Subject: [PATCH 6/6] libbpf: Add getter for program size MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This adds a new getter for the BPF program size (in bytes). This is useful for a caller that is trying to predict how much memory will be locked by loading a BPF object into the kernel. Signed-off-by: Toke Høiland-Jørgensen Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Acked-by: David S. Miller Acked-by: Song Liu Link: https://lore.kernel.org/bpf/157333185272.88376.10996937115395724683.stgit@toke.dk --- tools/lib/bpf/libbpf.c | 5 +++++ tools/lib/bpf/libbpf.h | 3 +++ tools/lib/bpf/libbpf.map | 1 + 3 files changed, 9 insertions(+) diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c index 847b62895f9b..96ef18cfeffb 100644 --- a/tools/lib/bpf/libbpf.c +++ b/tools/lib/bpf/libbpf.c @@ -4782,6 +4782,11 @@ int bpf_program__fd(const struct bpf_program *prog) return bpf_program__nth_fd(prog, 0); } +size_t bpf_program__size(const struct bpf_program *prog) +{ + return prog->insns_cnt * sizeof(struct bpf_insn); +} + int bpf_program__set_prep(struct bpf_program *prog, int nr_instances, bpf_program_prep_t prep) { diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h index f0947cc949d2..5aa27caad6c2 100644 --- a/tools/lib/bpf/libbpf.h +++ b/tools/lib/bpf/libbpf.h @@ -214,6 +214,9 @@ LIBBPF_API void bpf_program__set_ifindex(struct bpf_program *prog, LIBBPF_API const char *bpf_program__title(const struct bpf_program *prog, bool needs_copy); +/* returns program size in bytes */ +LIBBPF_API size_t bpf_program__size(const struct bpf_program *prog); + LIBBPF_API int bpf_program__load(struct bpf_program *prog, char *license, __u32 kern_version); LIBBPF_API int bpf_program__fd(const struct bpf_program *prog); diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map index d1a782a3a58d..9f39ee06b2d4 100644 --- a/tools/lib/bpf/libbpf.map +++ b/tools/lib/bpf/libbpf.map @@ -203,4 +203,5 @@ LIBBPF_0.0.6 { bpf_program__get_type; bpf_program__is_tracing; bpf_program__set_tracing; + bpf_program__size; } LIBBPF_0.0.5;