summaryrefslogtreecommitdiff
path: root/kernel
diff options
context:
space:
mode:
authorSong Liu <songliubraving@fb.com>2020-04-30 00:15:04 -0700
committerAlexei Starovoitov <ast@kernel.org>2020-05-01 10:36:32 -0700
commitd46edd671a147032e22cfeb271a5734703093649 (patch)
treeeee3bc7ad6740e4115f5df017b375ea8ea974829 /kernel
parentc321022244708aec4675de4f032ef1ba9ff0c640 (diff)
bpf: Sharing bpf runtime stats with BPF_ENABLE_STATS
Currently, sysctl kernel.bpf_stats_enabled controls BPF runtime stats. Typical userspace tools use kernel.bpf_stats_enabled as follows: 1. Enable kernel.bpf_stats_enabled; 2. Check program run_time_ns; 3. Sleep for the monitoring period; 4. Check program run_time_ns again, calculate the difference; 5. Disable kernel.bpf_stats_enabled. The problem with this approach is that only one userspace tool can toggle this sysctl. If multiple tools toggle the sysctl at the same time, the measurement may be inaccurate. To fix this problem while keep backward compatibility, introduce a new bpf command BPF_ENABLE_STATS. On success, this command enables stats and returns a valid fd. BPF_ENABLE_STATS takes argument "type". Currently, only one type, BPF_STATS_RUN_TIME, is supported. We can extend the command to support other types of stats in the future. With BPF_ENABLE_STATS, user space tool would have the following flow: 1. Get a fd with BPF_ENABLE_STATS, and make sure it is valid; 2. Check program run_time_ns; 3. Sleep for the monitoring period; 4. Check program run_time_ns again, calculate the difference; 5. Close the fd. Signed-off-by: Song Liu <songliubraving@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Link: https://lore.kernel.org/bpf/20200430071506.1408910-2-songliubraving@fb.com
Diffstat (limited to 'kernel')
-rw-r--r--kernel/bpf/syscall.c57
-rw-r--r--kernel/sysctl.c36
2 files changed, 92 insertions, 1 deletions
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index c75b2dd2459c..4f34eecec9ce 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -3872,6 +3872,60 @@ static int bpf_link_get_fd_by_id(const union bpf_attr *attr)
return fd;
}
+DEFINE_MUTEX(bpf_stats_enabled_mutex);
+
+static int bpf_stats_release(struct inode *inode, struct file *file)
+{
+ mutex_lock(&bpf_stats_enabled_mutex);
+ static_key_slow_dec(&bpf_stats_enabled_key.key);
+ mutex_unlock(&bpf_stats_enabled_mutex);
+ return 0;
+}
+
+static const struct file_operations bpf_stats_fops = {
+ .release = bpf_stats_release,
+};
+
+static int bpf_enable_runtime_stats(void)
+{
+ int fd;
+
+ mutex_lock(&bpf_stats_enabled_mutex);
+
+ /* Set a very high limit to avoid overflow */
+ if (static_key_count(&bpf_stats_enabled_key.key) > INT_MAX / 2) {
+ mutex_unlock(&bpf_stats_enabled_mutex);
+ return -EBUSY;
+ }
+
+ fd = anon_inode_getfd("bpf-stats", &bpf_stats_fops, NULL, O_CLOEXEC);
+ if (fd >= 0)
+ static_key_slow_inc(&bpf_stats_enabled_key.key);
+
+ mutex_unlock(&bpf_stats_enabled_mutex);
+ return fd;
+}
+
+#define BPF_ENABLE_STATS_LAST_FIELD enable_stats.type
+
+static int bpf_enable_stats(union bpf_attr *attr)
+{
+
+ if (CHECK_ATTR(BPF_ENABLE_STATS))
+ return -EINVAL;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EPERM;
+
+ switch (attr->enable_stats.type) {
+ case BPF_STATS_RUN_TIME:
+ return bpf_enable_runtime_stats();
+ default:
+ break;
+ }
+ return -EINVAL;
+}
+
SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, size)
{
union bpf_attr attr;
@@ -3996,6 +4050,9 @@ SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, siz
err = bpf_obj_get_next_id(&attr, uattr,
&link_idr, &link_idr_lock);
break;
+ case BPF_ENABLE_STATS:
+ err = bpf_enable_stats(&attr);
+ break;
default:
err = -EINVAL;
break;
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index e961286d0e14..7adfe5dbce9d 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -201,6 +201,40 @@ static int max_extfrag_threshold = 1000;
#endif /* CONFIG_SYSCTL */
+#ifdef CONFIG_BPF_SYSCALL
+static int bpf_stats_handler(struct ctl_table *table, int write,
+ void __user *buffer, size_t *lenp,
+ loff_t *ppos)
+{
+ struct static_key *key = (struct static_key *)table->data;
+ static int saved_val;
+ int val, ret;
+ struct ctl_table tmp = {
+ .data = &val,
+ .maxlen = sizeof(val),
+ .mode = table->mode,
+ .extra1 = SYSCTL_ZERO,
+ .extra2 = SYSCTL_ONE,
+ };
+
+ if (write && !capable(CAP_SYS_ADMIN))
+ return -EPERM;
+
+ mutex_lock(&bpf_stats_enabled_mutex);
+ val = saved_val;
+ ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
+ if (write && !ret && val != saved_val) {
+ if (val)
+ static_key_slow_inc(key);
+ else
+ static_key_slow_dec(key);
+ saved_val = val;
+ }
+ mutex_unlock(&bpf_stats_enabled_mutex);
+ return ret;
+}
+#endif
+
/*
* /proc/sys support
*/
@@ -2549,7 +2583,7 @@ static struct ctl_table kern_table[] = {
.data = &bpf_stats_enabled_key.key,
.maxlen = sizeof(bpf_stats_enabled_key),
.mode = 0644,
- .proc_handler = proc_do_static_key,
+ .proc_handler = bpf_stats_handler,
},
#endif
#if defined(CONFIG_TREE_RCU)