netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Andrii Nakryiko <andriin@fb.com>
To: <bpf@vger.kernel.org>, <netdev@vger.kernel.org>, <ast@fb.com>,
	<daniel@iogearbox.net>
Cc: <andrii.nakryiko@gmail.com>, <kernel-team@fb.com>,
	Andrii Nakryiko <andriin@fb.com>, Matthew Lim <matthewlim@fb.com>
Subject: [PATCH bpf-next 5/6] libbpf: handle missing BPF_OBJ_GET_INFO_BY_FD gracefully in perf_buffer
Date: Tue, 7 Jul 2020 18:53:17 -0700	[thread overview]
Message-ID: <20200708015318.3827358-6-andriin@fb.com> (raw)
In-Reply-To: <20200708015318.3827358-1-andriin@fb.com>

perf_buffer__new() is relying on BPF_OBJ_GET_INFO_BY_FD availability for few
sanity checks. OBJ_GET_INFO for maps is actually much more recent feature than
perf_buffer support itself, so this causes unnecessary problems on old kernels
before BPF_OBJ_GET_INFO_BY_FD was added.

This patch makes those sanity checks optional and just assumes best if command
is not supported. If user specified something incorrectly (e.g., wrong map
type), kernel will reject it later anyway, except user won't get a nice
explanation as to why it failed. This seems like a good trade off for
supporting perf_buffer on old kernels.

Signed-off-by: Andrii Nakryiko <andriin@fb.com>
---
 tools/lib/bpf/libbpf.c | 31 ++++++++++++++++++++-----------
 1 file changed, 20 insertions(+), 11 deletions(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index b4f50d12f51f..6f86b57a7e24 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -8587,7 +8587,7 @@ static struct perf_buffer *__perf_buffer__new(int map_fd, size_t page_cnt,
 					      struct perf_buffer_params *p)
 {
 	const char *online_cpus_file = "/sys/devices/system/cpu/online";
-	struct bpf_map_info map = {};
+	struct bpf_map_info map;
 	char msg[STRERR_BUFSIZE];
 	struct perf_buffer *pb;
 	bool *online = NULL;
@@ -8600,19 +8600,28 @@ static struct perf_buffer *__perf_buffer__new(int map_fd, size_t page_cnt,
 		return ERR_PTR(-EINVAL);
 	}
 
+	/* best-effort sanity checks */
+	memset(&map, 0, sizeof(map));
 	map_info_len = sizeof(map);
 	err = bpf_obj_get_info_by_fd(map_fd, &map, &map_info_len);
 	if (err) {
 		err = -errno;
-		pr_warn("failed to get map info for map FD %d: %s\n",
-			map_fd, libbpf_strerror_r(err, msg, sizeof(msg)));
-		return ERR_PTR(err);
-	}
-
-	if (map.type != BPF_MAP_TYPE_PERF_EVENT_ARRAY) {
-		pr_warn("map '%s' should be BPF_MAP_TYPE_PERF_EVENT_ARRAY\n",
-			map.name);
-		return ERR_PTR(-EINVAL);
+		/* if BPF_OBJ_GET_INFO_BY_FD is supported, will return
+		 * -EBADFD, -EFAULT, or -E2BIG on real error
+		 */
+		if (err != -EINVAL) {
+			pr_warn("failed to get map info for map FD %d: %s\n",
+				map_fd, libbpf_strerror_r(err, msg, sizeof(msg)));
+			return ERR_PTR(err);
+		}
+		pr_debug("failed to get map info for FD %d; API not supported? Ignoring...\n",
+			 map_fd);
+	} else {
+		if (map.type != BPF_MAP_TYPE_PERF_EVENT_ARRAY) {
+			pr_warn("map '%s' should be BPF_MAP_TYPE_PERF_EVENT_ARRAY\n",
+				map.name);
+			return ERR_PTR(-EINVAL);
+		}
 	}
 
 	pb = calloc(1, sizeof(*pb));
@@ -8644,7 +8653,7 @@ static struct perf_buffer *__perf_buffer__new(int map_fd, size_t page_cnt,
 			err = pb->cpu_cnt;
 			goto error;
 		}
-		if (map.max_entries < pb->cpu_cnt)
+		if (map.max_entries && map.max_entries < pb->cpu_cnt)
 			pb->cpu_cnt = map.max_entries;
 	}
 
-- 
2.24.1


  parent reply	other threads:[~2020-07-08  1:53 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-08  1:53 [PATCH bpf-next 0/6] Improve libbpf support of old kernels Andrii Nakryiko
2020-07-08  1:53 ` [PATCH bpf-next 1/6] libbpf: make BTF finalization strict Andrii Nakryiko
2020-07-08  1:53 ` [PATCH bpf-next 2/6] libbpf: add btf__set_fd() for more control over loaded BTF FD Andrii Nakryiko
2020-07-08  1:53 ` [PATCH bpf-next 3/6] libbpf: improve BTF sanitization handling Andrii Nakryiko
2020-07-08  1:53 ` [PATCH bpf-next 4/6] selftests/bpf: add test relying only on CO-RE and no recent kernel features Andrii Nakryiko
2020-07-08  1:53 ` Andrii Nakryiko [this message]
2020-07-08  1:53 ` [PATCH bpf-next 6/6] selftests/bpf: switch perf_buffer test to tracepoint and skeleton Andrii Nakryiko
2020-07-08 22:58 ` [PATCH bpf-next 0/6] Improve libbpf support of old kernels Daniel Borkmann

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200708015318.3827358-6-andriin@fb.com \
    --to=andriin@fb.com \
    --cc=andrii.nakryiko@gmail.com \
    --cc=ast@fb.com \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=kernel-team@fb.com \
    --cc=matthewlim@fb.com \
    --cc=netdev@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
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).