Merge tag 'seccomp-v5.14-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux

Pull seccomp updates from Kees Cook:

 - Add "atomic addfd + send reply" mode to SECCOMP_USER_NOTIF to better
   handle EINTR races visible to seccomp monitors. (Rodrigo Campos,
   Sargun Dhillon)

 - Improve seccomp selftests for readability in CI systems. (Kees Cook)

* tag 'seccomp-v5.14-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  selftests/seccomp: Avoid using "sysctl" for report
  selftests/seccomp: Flush benchmark output
  selftests/seccomp: More closely track fds being assigned
  selftests/seccomp: Add test for atomic addfd+send
  seccomp: Support atomic "addfd + send reply"

2 files changed, 55 insertions, 6 deletions

diff --git a/tools/testing/selftests/seccomp/seccomp_benchmark.c b/tools/testing/selftests/seccomp/seccomp_benchmark.c
index fcc806585266..6e5102a7d7c9 100644
--- a/tools/testing/selftests/seccomp/seccomp_benchmark.c
+++ b/tools/testing/selftests/seccomp/seccomp_benchmark.c
@@ -143,9 +143,15 @@ int main(int argc, char *argv[])
 	unsigned long long native, filter1, filter2, bitmap1, bitmap2;
 	unsigned long long entry, per_filter1, per_filter2;
 
+	setbuf(stdout, NULL);
+
+	printf("Running on:\n");
+	system("uname -a");
+
 	printf("Current BPF sysctl settings:\n");
-	system("sysctl net.core.bpf_jit_enable");
-	system("sysctl net.core.bpf_jit_harden");
+	/* Avoid using "sysctl" which may not be installed. */
+	system("grep -H . /proc/sys/net/core/bpf_jit_enable");
+	system("grep -H . /proc/sys/net/core/bpf_jit_harden");
 
 	if (argc > 1)
 		samples = strtoull(argv[1], NULL, 0);
diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c
index e3d5c77a8612..1d64891e6492 100644
--- a/tools/testing/selftests/seccomp/seccomp_bpf.c
+++ b/tools/testing/selftests/seccomp/seccomp_bpf.c
@@ -235,6 +235,10 @@ struct seccomp_notif_addfd {
 };
 #endif
 
+#ifndef SECCOMP_ADDFD_FLAG_SEND
+#define SECCOMP_ADDFD_FLAG_SEND	(1UL << 1) /* Addfd and return it, atomically */
+#endif
+
 struct seccomp_notif_addfd_small {
 	__u64 id;
 	char weird[4];
@@ -3959,7 +3963,7 @@ TEST(user_notification_addfd)
 {
 	pid_t pid;
 	long ret;
-	int status, listener, memfd, fd;
+	int status, listener, memfd, fd, nextfd;
 	struct seccomp_notif_addfd addfd = {};
 	struct seccomp_notif_addfd_small small = {};
 	struct seccomp_notif_addfd_big big = {};
@@ -3968,25 +3972,34 @@ TEST(user_notification_addfd)
 	/* 100 ms */
 	struct timespec delay = { .tv_nsec = 100000000 };
 
+	/* There may be arbitrary already-open fds at test start. */
 	memfd = memfd_create("test", 0);
 	ASSERT_GE(memfd, 0);
+	nextfd = memfd + 1;
 
 	ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0);
 	ASSERT_EQ(0, ret) {
 		TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!");
 	}
 
+	/* fd: 4 */
 	/* Check that the basic notification machinery works */
 	listener = user_notif_syscall(__NR_getppid,
 				      SECCOMP_FILTER_FLAG_NEW_LISTENER);
-	ASSERT_GE(listener, 0);
+	ASSERT_EQ(listener, nextfd++);
 
 	pid = fork();
 	ASSERT_GE(pid, 0);
 
 	if (pid == 0) {
+		/* fds will be added and this value is expected */
 		if (syscall(__NR_getppid) != USER_NOTIF_MAGIC)
 			exit(1);
+
+		/* Atomic addfd+send is received here. Check it is a valid fd */
+		if (fcntl(syscall(__NR_getppid), F_GETFD) == -1)
+			exit(1);
+
 		exit(syscall(__NR_getppid) != USER_NOTIF_MAGIC);
 	}
 
@@ -4028,14 +4041,14 @@ TEST(user_notification_addfd)
 
 	/* Verify we can set an arbitrary remote fd */
 	fd = ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd);
-	EXPECT_GE(fd, 0);
+	EXPECT_EQ(fd, nextfd++);
 	EXPECT_EQ(filecmp(getpid(), pid, memfd, fd), 0);
 
 	/* Verify we can set an arbitrary remote fd with large size */
 	memset(&big, 0x0, sizeof(big));
 	big.addfd = addfd;
 	fd = ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD_BIG, &big);
-	EXPECT_GE(fd, 0);
+	EXPECT_EQ(fd, nextfd++);
 
 	/* Verify we can set a specific remote fd */
 	addfd.newfd = 42;
@@ -4065,6 +4078,32 @@ TEST(user_notification_addfd)
 	ASSERT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0);
 	ASSERT_EQ(addfd.id, req.id);
 
+	/* Verify we can do an atomic addfd and send */
+	addfd.newfd = 0;
+	addfd.flags = SECCOMP_ADDFD_FLAG_SEND;
+	fd = ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd);
+	/*
+	 * Child has earlier "low" fds and now 42, so we expect the next
+	 * lowest available fd to be assigned here.
+	 */
+	EXPECT_EQ(fd, nextfd++);
+	EXPECT_EQ(filecmp(getpid(), pid, memfd, fd), 0);
+
+	/*
+	 * This sets the ID of the ADD FD to the last request plus 1. The
+	 * notification ID increments 1 per notification.
+	 */
+	addfd.id = req.id + 1;
+
+	/* This spins until the underlying notification is generated */
+	while (ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd) != -1 &&
+	       errno != -EINPROGRESS)
+		nanosleep(&delay, NULL);
+
+	memset(&req, 0, sizeof(req));
+	ASSERT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0);
+	ASSERT_EQ(addfd.id, req.id);
+
 	resp.id = req.id;
 	resp.error = 0;
 	resp.val = USER_NOTIF_MAGIC;
@@ -4125,6 +4164,10 @@ TEST(user_notification_addfd_rlimit)
 	EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd), -1);
 	EXPECT_EQ(errno, EMFILE);
 
+	addfd.flags = SECCOMP_ADDFD_FLAG_SEND;
+	EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd), -1);
+	EXPECT_EQ(errno, EMFILE);
+
 	addfd.newfd = 100;
 	addfd.flags = SECCOMP_ADDFD_FLAG_SETFD;
 	EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd), -1);