UBUNTU: SAUCE: security,perf: Allow further restriction of perf_event_open

https://lkml.org/lkml/2016/1/11/587

The GRKERNSEC_PERF_HARDEN feature extracted from grsecurity.  Adds the
option to disable perf_event_open() entirely for unprivileged users.
This standalone version doesn't include making the variable read-only
(or renaming it).

When kernel.perf_event_open is set to 3 (or greater), disallow all
access to performance events by users without CAP_SYS_ADMIN.
Add a Kconfig symbol CONFIG_SECURITY_PERF_EVENTS_RESTRICT that
makes this value the default.

This is based on a similar feature in grsecurity
(CONFIG_GRKERNSEC_PERF_HARDEN).  This version doesn't include making
the variable read-only.  It also allows enabling further restriction
at run-time regardless of whether the default is changed.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
[ saf: resolve conflicts with v5.8-rc1 ]
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
[ arighi: resolve conflicts with v6.2-rc2 ]
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>

include/linux/perf_event.h

@@ -1612,6 +1612,12 @@ int perf_event_max_stack_handler(struct ctl_table *table, int write,
 #define PERF_SECURITY_CPU		1
 #define PERF_SECURITY_KERNEL		2
 #define PERF_SECURITY_TRACEPOINT	3
+#define PERF_SECURITY_MAX		4
+
+static inline bool perf_paranoid_any(void)
+{
+	return sysctl_perf_event_paranoid >= PERF_SECURITY_MAX;
+}
 
 static inline int perf_is_paranoid(void)
 {

kernel/events/core.c

@@ -415,8 +415,13 @@ static struct kmem_cache *perf_event_cache;
  *   0 - disallow raw tracepoint access for unpriv
  *   1 - disallow cpu events for unpriv
  *   2 - disallow kernel profiling for unpriv
+ *   4 - disallow all unpriv perf event use
  */
+#ifdef CONFIG_SECURITY_PERF_EVENTS_RESTRICT
+int sysctl_perf_event_paranoid __read_mostly = PERF_SECURITY_MAX;
+#else
 int sysctl_perf_event_paranoid __read_mostly = 2;
+#endif
 
 /* Minimum for 512 kiB + 1 user control page */
 int sysctl_perf_event_mlock __read_mostly = 512 + (PAGE_SIZE / 1024); /* 'free' kiB per user */
@@ -12440,6 +12445,9 @@ SYSCALL_DEFINE5(perf_event_open,
 	if (flags & ~PERF_FLAG_ALL)
 		return -EINVAL;
 
+	if (perf_paranoid_any() && !capable(CAP_SYS_ADMIN))
+		return -EACCES;
+
 	err = perf_copy_attr(attr_uptr, &attr);
 	if (err)
 		return err;

security/Kconfig

@@ -19,6 +19,15 @@ config SECURITY_DMESG_RESTRICT
 
 	  If you are unsure how to answer this question, answer N.
 
+config SECURITY_PERF_EVENTS_RESTRICT
+	bool "Restrict unprivileged use of performance events"
+	depends on PERF_EVENTS
+	help
+	  If you say Y here, the kernel.perf_event_paranoid sysctl
+	  will be set to 3 by default, and no unprivileged use of the
+	  perf_event_open syscall will be permitted unless it is
+	  changed.
+
 config SECURITY
 	bool "Enable different security models"
 	depends on SYSFS