lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date:   Mon, 16 Mar 2020 13:33:29 -0700
From:   Song Liu <songliubraving@...com>
To:     <linux-fsdevel@...r.kernel.org>, <netdev@...r.kernel.org>,
        <bpf@...r.kernel.org>
CC:     <kernel-team@...com>, <ast@...nel.org>, <daniel@...earbox.net>,
        <mcgrof@...nel.org>, <keescook@...omium.org>, <yzaikin@...gle.com>,
        Song Liu <songliubraving@...com>
Subject: [PATCH v2 bpf-next] bpf: sharing bpf runtime stats with /dev/bpf_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_RUNTIME_STATS. On success, this command enables
run_time_ns stats and returns a valid fd.

With BPF_ENABLE_RUNTIME_STATS, user space tool would have the following
flow:

  1. Get a fd with BPF_ENABLE_RUNTIME_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@...com>

---
Changes RFC => v2:
1. Add a new bpf command instead of /dev/bpf_stats;
2. Remove the jump_label patch, which is no longer needed;
3. Add a static variable to save previous value of the sysctl.
---
 include/linux/bpf.h            |  1 +
 include/uapi/linux/bpf.h       |  1 +
 kernel/bpf/syscall.c           | 43 ++++++++++++++++++++++++++++++++++
 kernel/sysctl.c                | 36 +++++++++++++++++++++++++++-
 tools/include/uapi/linux/bpf.h |  1 +
 5 files changed, 81 insertions(+), 1 deletion(-)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 4fd91b7c95ea..d542349771df 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -970,6 +970,7 @@ _out:							\
 
 #ifdef CONFIG_BPF_SYSCALL
 DECLARE_PER_CPU(int, bpf_prog_active);
+extern struct mutex bpf_stats_enabled_mutex;
 
 /*
  * Block execution of BPF programs attached to instrumentation (perf,
diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
index 40b2d9476268..8285ff37210c 100644
--- a/include/uapi/linux/bpf.h
+++ b/include/uapi/linux/bpf.h
@@ -111,6 +111,7 @@ enum bpf_cmd {
 	BPF_MAP_LOOKUP_AND_DELETE_BATCH,
 	BPF_MAP_UPDATE_BATCH,
 	BPF_MAP_DELETE_BATCH,
+	BPF_ENABLE_RUNTIME_STATS,
 };
 
 enum bpf_map_type {
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index b2f73ecacced..823dc9de7953 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -24,6 +24,9 @@
 #include <linux/ctype.h>
 #include <linux/nospec.h>
 #include <linux/audit.h>
+#include <linux/miscdevice.h>
+#include <linux/fs.h>
+#include <linux/jump_label.h>
 #include <uapi/linux/btf.h>
 
 #define IS_FD_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY || \
@@ -3550,6 +3553,43 @@ static int bpf_map_do_batch(const union bpf_attr *attr,
 	return err;
 }
 
+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;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	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, 0);
+	if (fd >= 0)
+		static_key_slow_inc(&bpf_stats_enabled_key.key);
+
+	mutex_unlock(&bpf_stats_enabled_mutex);
+	return fd;
+}
+
+
 SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, size)
 {
 	union bpf_attr attr = {};
@@ -3660,6 +3700,9 @@ SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, siz
 	case BPF_MAP_DELETE_BATCH:
 		err = bpf_map_do_batch(&attr, uattr, BPF_MAP_DELETE_BATCH);
 		break;
+	case BPF_ENABLE_RUNTIME_STATS:
+		err = bpf_enable_runtime_stats();
+		break;
 	default:
 		err = -EINVAL;
 		break;
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index ad5b88a53c5a..14613d1e0b88 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -316,6 +316,40 @@ static int min_extfrag_threshold;
 static int max_extfrag_threshold = 1000;
 #endif
 
+#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;
+	int val, ret;
+	static int saved_val;
+	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
+
 static struct ctl_table kern_table[] = {
 	{
 		.procname	= "sched_child_runs_first",
@@ -1256,7 +1290,7 @@ 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)
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
index 40b2d9476268..8285ff37210c 100644
--- a/tools/include/uapi/linux/bpf.h
+++ b/tools/include/uapi/linux/bpf.h
@@ -111,6 +111,7 @@ enum bpf_cmd {
 	BPF_MAP_LOOKUP_AND_DELETE_BATCH,
 	BPF_MAP_UPDATE_BATCH,
 	BPF_MAP_DELETE_BATCH,
+	BPF_ENABLE_RUNTIME_STATS,
 };
 
 enum bpf_map_type {
-- 
2.17.1

Powered by blists - more mailing lists