[PATCH bpf-next] bpf: Print error message for bpftool cgroup show

[PATCH bpf-next] bpf: Print error message for bpftool cgroup show

[Hechao Li]

Currently, when bpftool cgroup show <path> has an error, no error
message is printed. This is confusing because the user may think the
result is empty.

Before the change:

$ bpftool cgroup show /sys/fs/cgroup
ID       AttachType      AttachFlags     Name
$ echo $?
255

After the change:
$ ./bpftool cgroup show /sys/fs/cgroup
Error: can't query bpf programs attached to /sys/fs/cgroup: Operation
not permitted

Signed-off-by: Hechao Li <hechaol@fb.com>
---
 tools/bpf/bpftool/cgroup.c | 60 ++++++++++++++++++++++++++------------
 1 file changed, 42 insertions(+), 18 deletions(-)

diff --git a/tools/bpf/bpftool/cgroup.c b/tools/bpf/bpftool/cgroup.c
index 1ef45e55039e..74b57e2d7524 100644
--- a/tools/bpf/bpftool/cgroup.c
+++ b/tools/bpf/bpftool/cgroup.c
@@ -117,6 +117,28 @@ static int count_attached_bpf_progs(int cgroup_fd, enum bpf_attach_type type)
 	return prog_cnt;
 }
 
+#define QUERY_CGROUP_ERR (-1)
+#define QUERY_CGROUP_NO_PROG (0)
+#define QUERY_CGROUP_SUCCESS (1)
+static int check_query_cgroup_progs(int cgroup_fd)
+{
+	enum bpf_attach_type type;
+	bool no_prog = true;
+
+	for (type = 0; type < __MAX_BPF_ATTACH_TYPE; type++) {
+		int count = count_attached_bpf_progs(cgroup_fd, type);
+
+		if (count < 0 && errno != EINVAL)
+			return QUERY_CGROUP_ERR;
+
+		if (count > 0) {
+			no_prog = false;
+			break;
+		}
+	}
+
+	return no_prog ? QUERY_CGROUP_NO_PROG : QUERY_CGROUP_SUCCESS;
+}
 static int show_attached_bpf_progs(int cgroup_fd, enum bpf_attach_type type,
 				   int level)
 {
@@ -162,6 +184,7 @@ static int do_show(int argc, char **argv)
 {
 	enum bpf_attach_type type;
 	const char *path;
+	int query_check;
 	int cgroup_fd;
 	int ret = -1;
 
@@ -192,6 +215,16 @@ static int do_show(int argc, char **argv)
 		goto exit;
 	}
 
+	query_check = check_query_cgroup_progs(cgroup_fd);
+	if (query_check == QUERY_CGROUP_ERR) {
+		p_err("can't query bpf programs attached to %s: %s",
+		      path, strerror(errno));
+		goto exit_cgroup;
+	} else if (query_check == QUERY_CGROUP_NO_PROG) {
+		ret = 0;
+		goto exit_cgroup;
+	}
+
 	if (json_output)
 		jsonw_start_array(json_wtr);
 	else
@@ -212,8 +245,8 @@ static int do_show(int argc, char **argv)
 	if (json_output)
 		jsonw_end_array(json_wtr);
 
+exit_cgroup:
 	close(cgroup_fd);
-exit:
 	return ret;
 }
 
@@ -228,7 +261,7 @@ static int do_show_tree_fn(const char *fpath, const struct stat *sb,
 			   int typeflag, struct FTW *ftw)
 {
 	enum bpf_attach_type type;
-	bool skip = true;
+	int query_check;
 	int cgroup_fd;
 
 	if (typeflag != FTW_D)
@@ -240,22 +273,13 @@ static int do_show_tree_fn(const char *fpath, const struct stat *sb,
 		return SHOW_TREE_FN_ERR;
 	}
 
-	for (type = 0; type < __MAX_BPF_ATTACH_TYPE; type++) {
-		int count = count_attached_bpf_progs(cgroup_fd, type);
-
-		if (count < 0 && errno != EINVAL) {
-			p_err("can't query bpf programs attached to %s: %s",
-			      fpath, strerror(errno));
-			close(cgroup_fd);
-			return SHOW_TREE_FN_ERR;
-		}
-		if (count > 0) {
-			skip = false;
-			break;
-		}
-	}
-
-	if (skip) {
+	query_check = check_query_cgroup_progs(cgroup_fd);
+	if (query_check == QUERY_CGROUP_ERR) {
+		p_err("can't query bpf programs attached to %s: %s",
+		      fpath, strerror(errno));
+		close(cgroup_fd);
+		return SHOW_TREE_FN_ERR;
+	} else if (query_check == QUERY_CGROUP_NO_PROG) {
 		close(cgroup_fd);
 		return 0;
 	}
-- 
2.17.1

Re: [PATCH bpf-next] bpf: Print error message for bpftool cgroup show

[Andrii Nakryiko]

On Fri, Dec 20, 2019 at 3:04 PM Hechao Li <hechaol@fb.com> wrote:
>

For next revision, please don't forget to add "PATCH" in patch prefix.
So you'll have: [PATCH v2 bpf-next] <subject here> for you v2.


> Currently, when bpftool cgroup show <path> has an error, no error
> message is printed. This is confusing because the user may think the
> result is empty.
>
> Before the change:
>
> $ bpftool cgroup show /sys/fs/cgroup
> ID       AttachType      AttachFlags     Name
> $ echo $?
> 255
>
> After the change:
> $ ./bpftool cgroup show /sys/fs/cgroup
> Error: can't query bpf programs attached to /sys/fs/cgroup: Operation
> not permitted
>
> Signed-off-by: Hechao Li <hechaol@fb.com>
> ---
>  tools/bpf/bpftool/cgroup.c | 60 ++++++++++++++++++++++++++------------
>  1 file changed, 42 insertions(+), 18 deletions(-)
>
> diff --git a/tools/bpf/bpftool/cgroup.c b/tools/bpf/bpftool/cgroup.c
> index 1ef45e55039e..74b57e2d7524 100644
> --- a/tools/bpf/bpftool/cgroup.c
> +++ b/tools/bpf/bpftool/cgroup.c
> @@ -117,6 +117,28 @@ static int count_attached_bpf_progs(int cgroup_fd, enum bpf_attach_type type)
>         return prog_cnt;
>  }
>
> +#define QUERY_CGROUP_ERR (-1)
> +#define QUERY_CGROUP_NO_PROG (0)
> +#define QUERY_CGROUP_SUCCESS (1)
> +static int check_query_cgroup_progs(int cgroup_fd)
> +{

How about calling it a bit differently and use typical bool + error
return values. E.g., "cgroup_has_attached_progs" (or something along
those lines), then 1 means "true, it has", 0 means "false, doesn't
have any BPF progs", <0 means error, as usual. No need for special
QUERY* constants.

> +       enum bpf_attach_type type;
> +       bool no_prog = true;
> +
> +       for (type = 0; type < __MAX_BPF_ATTACH_TYPE; type++) {
> +               int count = count_attached_bpf_progs(cgroup_fd, type);
> +
> +               if (count < 0 && errno != EINVAL)
> +                       return QUERY_CGROUP_ERR;
> +
> +               if (count > 0) {
> +                       no_prog = false;
> +                       break;
> +               }
> +       }
> +
> +       return no_prog ? QUERY_CGROUP_NO_PROG : QUERY_CGROUP_SUCCESS;
> +}
>  static int show_attached_bpf_progs(int cgroup_fd, enum bpf_attach_type type,
>                                    int level)
>  {
> @@ -162,6 +184,7 @@ static int do_show(int argc, char **argv)
>  {
>         enum bpf_attach_type type;
>         const char *path;
> +       int query_check;
>         int cgroup_fd;
>         int ret = -1;
>
> @@ -192,6 +215,16 @@ static int do_show(int argc, char **argv)
>                 goto exit;

you just removed exit label below, this causes compilation error here

>         }
>
> +       query_check = check_query_cgroup_progs(cgroup_fd);
> +       if (query_check == QUERY_CGROUP_ERR) {
> +               p_err("can't query bpf programs attached to %s: %s",
> +                     path, strerror(errno));
> +               goto exit_cgroup;
> +       } else if (query_check == QUERY_CGROUP_NO_PROG) {
> +               ret = 0;
> +               goto exit_cgroup;
> +       }
> +

[...]