Don't leak file descriptors

We're leaking file descriptors to child processes. Set FD_CLOEXEC on file
descriptors that don't need to be passed to children to stop this misbehaviour.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
diff --git a/block/raw-posix.c b/block/raw-posix.c
index 706799f..5a6a22b 100644
--- a/block/raw-posix.c
+++ b/block/raw-posix.c
@@ -153,7 +153,7 @@
         s->open_flags |= O_DSYNC;
 
     s->fd = -1;
-    fd = open(filename, s->open_flags, 0644);
+    fd = qemu_open(filename, s->open_flags, 0644);
     if (fd < 0) {
         ret = -errno;
         if (ret == -EROFS)
diff --git a/configure b/configure
index 8b8b94c..ef15948 100755
--- a/configure
+++ b/configure
@@ -1570,6 +1570,23 @@
   pipe2=yes
 fi
 
+# check if accept4 is there
+accept4=no
+cat > $TMPC << EOF
+#define _GNU_SOURCE
+#include <sys/socket.h>
+#include <stddef.h>
+
+int main(void)
+{
+    accept4(0, NULL, NULL, SOCK_CLOEXEC);
+    return 0;
+}
+EOF
+if compile_prog "" "" ; then
+  accept4=yes
+fi
+
 # check if tee/splice is there. vmsplice was added same time.
 splice=no
 cat > $TMPC << EOF
@@ -2014,6 +2031,9 @@
 if test "$pipe2" = "yes" ; then
   echo "CONFIG_PIPE2=y" >> $config_host_mak
 fi
+if test "$accept4" = "yes" ; then
+  echo "CONFIG_ACCEPT4=y" >> $config_host_mak
+fi
 if test "$splice" = "yes" ; then
   echo "CONFIG_SPLICE=y" >> $config_host_mak
 fi
diff --git a/gdbstub.c b/gdbstub.c
index 055093f..5a4f7d4 100644
--- a/gdbstub.c
+++ b/gdbstub.c
@@ -2356,6 +2356,9 @@
             perror("accept");
             return;
         } else if (fd >= 0) {
+#ifndef _WIN32
+            fcntl(fd, F_SETFD, FD_CLOEXEC);
+#endif
             break;
         }
     }
@@ -2385,6 +2388,9 @@
         perror("socket");
         return -1;
     }
+#ifndef _WIN32
+    fcntl(fd, F_SETFD, FD_CLOEXEC);
+#endif
 
     /* allow fast reuse */
     val = 1;
diff --git a/kvm-all.c b/kvm-all.c
index b605caa..b8ffe54 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -416,7 +416,7 @@
         s->slots[i].slot = i;
 
     s->vmfd = -1;
-    s->fd = open("/dev/kvm", O_RDWR);
+    s->fd = qemu_open("/dev/kvm", O_RDWR);
     if (s->fd == -1) {
         fprintf(stderr, "Could not access KVM kernel module: %m\n");
         ret = -errno;
diff --git a/migration-tcp.c b/migration-tcp.c
index b77ed87..2cfa8cb 100644
--- a/migration-tcp.c
+++ b/migration-tcp.c
@@ -105,7 +105,7 @@
     s->state = MIG_STATE_ACTIVE;
     s->mon = NULL;
     s->bandwidth_limit = bandwidth_limit;
-    s->fd = socket(PF_INET, SOCK_STREAM, 0);
+    s->fd = qemu_socket(PF_INET, SOCK_STREAM, 0);
     if (s->fd == -1) {
         qemu_free(s);
         return NULL;
@@ -146,7 +146,7 @@
     int c, ret;
 
     do {
-        c = accept(s, (struct sockaddr *)&addr, &addrlen);
+        c = qemu_accept(s, (struct sockaddr *)&addr, &addrlen);
     } while (c == -1 && socket_error() == EINTR);
 
     dprintf("accepted migration\n");
@@ -193,7 +193,7 @@
         return -EINVAL;
     }
 
-    s = socket(PF_INET, SOCK_STREAM, 0);
+    s = qemu_socket(PF_INET, SOCK_STREAM, 0);
     if (s == -1)
         return -socket_error();
 
diff --git a/migration-unix.c b/migration-unix.c
index 7dd787c..783228b 100644
--- a/migration-unix.c
+++ b/migration-unix.c
@@ -104,7 +104,7 @@
     s->state = MIG_STATE_ACTIVE;
     s->mon = NULL;
     s->bandwidth_limit = bandwidth_limit;
-    s->fd = socket(PF_UNIX, SOCK_STREAM, 0);
+    s->fd = qemu_socket(PF_UNIX, SOCK_STREAM, 0);
     if (s->fd < 0) {
         dprintf("Unable to open socket");
         goto err_after_alloc;
@@ -150,7 +150,7 @@
     int c, ret;
 
     do {
-        c = accept(s, (struct sockaddr *)&addr, &addrlen);
+        c = qemu_accept(s, (struct sockaddr *)&addr, &addrlen);
     } while (c == -1 && socket_error() == EINTR);
 
     dprintf("accepted migration\n");
@@ -191,7 +191,7 @@
 
     dprintf("Attempting to start an incoming migration\n");
 
-    sock = socket(PF_UNIX, SOCK_STREAM, 0);
+    sock = qemu_socket(PF_UNIX, SOCK_STREAM, 0);
     if (sock < 0) {
         fprintf(stderr, "Could not open unix socket: %s\n", strerror(errno));
         return -EINVAL;
diff --git a/net/socket.c b/net/socket.c
index 7331d87..5533737 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -161,7 +161,7 @@
 	return -1;
 
     }
-    fd = socket(PF_INET, SOCK_DGRAM, 0);
+    fd = qemu_socket(PF_INET, SOCK_DGRAM, 0);
     if (fd < 0) {
         perror("socket(PF_INET, SOCK_DGRAM)");
         return -1;
@@ -355,7 +355,7 @@
 
     for(;;) {
         len = sizeof(saddr);
-        fd = accept(s->fd, (struct sockaddr *)&saddr, &len);
+        fd = qemu_accept(s->fd, (struct sockaddr *)&saddr, &len);
         if (fd < 0 && errno != EINTR) {
             return;
         } else if (fd >= 0) {
@@ -386,7 +386,7 @@
 
     s = qemu_mallocz(sizeof(NetSocketListenState));
 
-    fd = socket(PF_INET, SOCK_STREAM, 0);
+    fd = qemu_socket(PF_INET, SOCK_STREAM, 0);
     if (fd < 0) {
         perror("socket");
         return -1;
@@ -427,7 +427,7 @@
     if (parse_host_port(&saddr, host_str) < 0)
         return -1;
 
-    fd = socket(PF_INET, SOCK_STREAM, 0);
+    fd = qemu_socket(PF_INET, SOCK_STREAM, 0);
     if (fd < 0) {
         perror("socket");
         return -1;
diff --git a/osdep.c b/osdep.c
index fd8bbd7..7509c5b 100644
--- a/osdep.c
+++ b/osdep.c
@@ -121,7 +121,7 @@
 #ifndef _WIN32
     int fd;
 
-    fd = open(filename, O_RDWR | O_CREAT, 0600);
+    fd = qemu_open(filename, O_RDWR | O_CREAT, 0600);
     if (fd == -1)
         return -1;
 
@@ -201,11 +201,113 @@
     ia->s_addr = addr;
     return 1;
 }
+
+void qemu_set_cloexec(int fd)
+{
+}
+
 #else
+
 void socket_set_nonblock(int fd)
 {
     int f;
     f = fcntl(fd, F_GETFL);
     fcntl(fd, F_SETFL, f | O_NONBLOCK);
 }
+
+void qemu_set_cloexec(int fd)
+{
+    int f;
+    f = fcntl(fd, F_GETFD);
+    fcntl(fd, F_SETFD, f | FD_CLOEXEC);
+}
+
 #endif
+
+/*
+ * Opens a file with FD_CLOEXEC set
+ */
+int qemu_open(const char *name, int flags, ...)
+{
+    int ret;
+    int mode = 0;
+
+    if (flags & O_CREAT) {
+        va_list ap;
+
+        va_start(ap, flags);
+        mode = va_arg(ap, int);
+        va_end(ap);
+    }
+
+#ifdef O_CLOEXEC
+    ret = open(name, flags | O_CLOEXEC, mode);
+#else
+    ret = open(name, flags, mode);
+    if (ret >= 0) {
+        qemu_set_cloexec(ret);
+    }
+#endif
+
+    return ret;
+}
+
+#ifndef _WIN32
+/*
+ * Creates a pipe with FD_CLOEXEC set on both file descriptors
+ */
+int qemu_pipe(int pipefd[2])
+{
+    int ret;
+
+#ifdef CONFIG_PIPE2
+    ret = pipe2(pipefd, O_CLOEXEC);
+#else
+    ret = pipe(pipefd);
+    if (ret == 0) {
+        qemu_set_cloexec(pipefd[0]);
+        qemu_set_cloexec(pipefd[1]);
+    }
+#endif
+
+    return ret;
+}
+#endif
+
+/*
+ * Opens a socket with FD_CLOEXEC set
+ */
+int qemu_socket(int domain, int type, int protocol)
+{
+    int ret;
+
+#ifdef SOCK_CLOEXEC
+    ret = socket(domain, type | SOCK_CLOEXEC, protocol);
+#else
+    ret = socket(domain, type, protocol);
+    if (ret >= 0) {
+        qemu_set_cloexec(ret);
+    }
+#endif
+
+    return ret;
+}
+
+/*
+ * Accept a connection and set FD_CLOEXEC
+ */
+int qemu_accept(int s, struct sockaddr *addr, socklen_t *addrlen)
+{
+    int ret;
+
+#ifdef CONFIG_ACCEPT4
+    ret = accept4(s, addr, addrlen, SOCK_CLOEXEC);
+#else
+    ret = accept(s, addr, addrlen);
+    if (ret >= 0) {
+        qemu_set_cloexec(ret);
+    }
+#endif
+
+    return ret;
+}
diff --git a/posix-aio-compat.c b/posix-aio-compat.c
index 7f391c9..ac247e1 100644
--- a/posix-aio-compat.c
+++ b/posix-aio-compat.c
@@ -625,7 +625,7 @@
     sigaction(SIGUSR2, &act, NULL);
 
     s->first_aio = NULL;
-    if (pipe(fds) == -1) {
+    if (qemu_pipe(fds) == -1) {
         fprintf(stderr, "failed to create pipe\n");
         return -1;
     }
diff --git a/qemu-char.c b/qemu-char.c
index e202585..da5c15c 100644
--- a/qemu-char.c
+++ b/qemu-char.c
@@ -620,7 +620,7 @@
 {
     int fd_out;
 
-    TFR(fd_out = open(qemu_opt_get(opts, "path"),
+    TFR(fd_out = qemu_open(qemu_opt_get(opts, "path"),
                       O_WRONLY | O_TRUNC | O_CREAT | O_BINARY, 0666));
     if (fd_out < 0)
         return NULL;
@@ -640,8 +640,8 @@
 
     snprintf(filename_in, 256, "%s.in", filename);
     snprintf(filename_out, 256, "%s.out", filename);
-    TFR(fd_in = open(filename_in, O_RDWR | O_BINARY));
-    TFR(fd_out = open(filename_out, O_RDWR | O_BINARY));
+    TFR(fd_in = qemu_open(filename_in, O_RDWR | O_BINARY));
+    TFR(fd_out = qemu_open(filename_out, O_RDWR | O_BINARY));
     if (fd_in < 0 || fd_out < 0) {
 	if (fd_in >= 0)
 	    close(fd_in);
@@ -2101,7 +2101,7 @@
 	    len = sizeof(saddr);
 	    addr = (struct sockaddr *)&saddr;
 	}
-        fd = accept(s->listen_fd, addr, &len);
+        fd = qemu_accept(s->listen_fd, addr, &len);
         if (fd < 0 && errno != EINTR) {
             return;
         } else if (fd >= 0) {
diff --git a/qemu-common.h b/qemu-common.h
index 57af677..8630f8c 100644
--- a/qemu-common.h
+++ b/qemu-common.h
@@ -159,6 +159,13 @@
 void qemu_mutex_lock_iothread(void);
 void qemu_mutex_unlock_iothread(void);
 
+int qemu_open(const char *name, int flags, ...);
+void qemu_set_cloexec(int fd);
+
+#ifndef _WIN32
+int qemu_pipe(int pipefd[2]);
+#endif
+
 /* Error handling.  */
 
 void QEMU_NORETURN hw_error(const char *fmt, ...)
diff --git a/qemu-sockets.c b/qemu-sockets.c
index 8801453..8850516 100644
--- a/qemu-sockets.c
+++ b/qemu-sockets.c
@@ -160,7 +160,7 @@
         getnameinfo((struct sockaddr*)e->ai_addr,e->ai_addrlen,
 		        uaddr,INET6_ADDRSTRLEN,uport,32,
 		        NI_NUMERICHOST | NI_NUMERICSERV);
-        slisten = socket(e->ai_family, e->ai_socktype, e->ai_protocol);
+        slisten = qemu_socket(e->ai_family, e->ai_socktype, e->ai_protocol);
         if (slisten < 0) {
             fprintf(stderr,"%s: socket(%s): %s\n", __FUNCTION__,
                     inet_strfamily(e->ai_family), strerror(errno));
@@ -258,7 +258,7 @@
             fprintf(stderr,"%s: getnameinfo: oops\n", __FUNCTION__);
             continue;
         }
-        sock = socket(e->ai_family, e->ai_socktype, e->ai_protocol);
+        sock = qemu_socket(e->ai_family, e->ai_socktype, e->ai_protocol);
         if (sock < 0) {
             fprintf(stderr,"%s: socket(%s): %s\n", __FUNCTION__,
             inet_strfamily(e->ai_family), strerror(errno));
@@ -351,7 +351,7 @@
     }
 
     /* create socket */
-    sock = socket(peer->ai_family, peer->ai_socktype, peer->ai_protocol);
+    sock = qemu_socket(peer->ai_family, peer->ai_socktype, peer->ai_protocol);
     if (sock < 0) {
         fprintf(stderr,"%s: socket(%s): %s\n", __FUNCTION__,
                 inet_strfamily(peer->ai_family), strerror(errno));
@@ -505,7 +505,7 @@
     const char *path = qemu_opt_get(opts, "path");
     int sock, fd;
 
-    sock = socket(PF_UNIX, SOCK_STREAM, 0);
+    sock = qemu_socket(PF_UNIX, SOCK_STREAM, 0);
     if (sock < 0) {
         perror("socket(unix)");
         return -1;
@@ -560,7 +560,7 @@
         return -1;
     }
 
-    sock = socket(PF_UNIX, SOCK_STREAM, 0);
+    sock = qemu_socket(PF_UNIX, SOCK_STREAM, 0);
     if (sock < 0) {
         perror("socket(unix)");
         return -1;
diff --git a/qemu_socket.h b/qemu_socket.h
index c253b32..86bdbf5 100644
--- a/qemu_socket.h
+++ b/qemu_socket.h
@@ -32,6 +32,8 @@
 #include "qemu-option.h"
 
 /* misc helpers */
+int qemu_socket(int domain, int type, int protocol);
+int qemu_accept(int s, struct sockaddr *addr, socklen_t *addrlen);
 void socket_set_nonblock(int fd);
 int send_all(int fd, const void *buf, int len1);
 
diff --git a/slirp/misc.c b/slirp/misc.c
index e9f08fd..c76ad8f 100644
--- a/slirp/misc.c
+++ b/slirp/misc.c
@@ -132,7 +132,7 @@
 		addr.sin_port = 0;
 		addr.sin_addr.s_addr = INADDR_ANY;
 
-		if ((s = socket(AF_INET, SOCK_STREAM, 0)) < 0 ||
+		if ((s = qemu_socket(AF_INET, SOCK_STREAM, 0)) < 0 ||
 		    bind(s, (struct sockaddr *)&addr, addrlen) < 0 ||
 		    listen(s, 1) < 0) {
 			lprint("Error: inet socket: %s\n", strerror(errno));
@@ -165,7 +165,7 @@
 			 * Connect to the socket
 			 * XXX If any of these fail, we're in trouble!
 	 		 */
-			s = socket(AF_INET, SOCK_STREAM, 0);
+			s = qemu_socket(AF_INET, SOCK_STREAM, 0);
 			addr.sin_addr = loopback_addr;
                         do {
                             ret = connect(s, (struct sockaddr *)&addr, addrlen);
diff --git a/slirp/slirp.h b/slirp/slirp.h
index 9ef57ea..98a2644 100644
--- a/slirp/slirp.h
+++ b/slirp/slirp.h
@@ -197,6 +197,10 @@
 #include "bootp.h"
 #include "tftp.h"
 
+/* osdep.c */
+int qemu_socket(int domain, int type, int protocol);
+
+
 struct Slirp {
     QTAILQ_ENTRY(Slirp) entry;
 
diff --git a/slirp/socket.c b/slirp/socket.c
index 207109c..cf6e6a9 100644
--- a/slirp/socket.c
+++ b/slirp/socket.c
@@ -626,7 +626,7 @@
 	addr.sin_addr.s_addr = haddr;
 	addr.sin_port = hport;
 
-	if (((s = socket(AF_INET,SOCK_STREAM,0)) < 0) ||
+	if (((s = qemu_socket(AF_INET,SOCK_STREAM,0)) < 0) ||
 	    (setsockopt(s,SOL_SOCKET,SO_REUSEADDR,(char *)&opt,sizeof(int)) < 0) ||
 	    (bind(s,(struct sockaddr *)&addr, sizeof(addr)) < 0) ||
 	    (listen(s,1) < 0)) {
diff --git a/slirp/tcp_subr.c b/slirp/tcp_subr.c
index 0417345..7851307 100644
--- a/slirp/tcp_subr.c
+++ b/slirp/tcp_subr.c
@@ -325,7 +325,7 @@
   DEBUG_CALL("tcp_fconnect");
   DEBUG_ARG("so = %lx", (long )so);
 
-  if( (ret=so->s=socket(AF_INET,SOCK_STREAM,0)) >= 0) {
+  if( (ret = so->s = qemu_socket(AF_INET,SOCK_STREAM,0)) >= 0) {
     int opt, s=so->s;
     struct sockaddr_in addr;
 
diff --git a/slirp/udp.c b/slirp/udp.c
index a88b645..d6c39b9 100644
--- a/slirp/udp.c
+++ b/slirp/udp.c
@@ -302,7 +302,7 @@
 int
 udp_attach(struct socket *so)
 {
-  if((so->s = socket(AF_INET,SOCK_DGRAM,0)) != -1) {
+  if((so->s = qemu_socket(AF_INET,SOCK_DGRAM,0)) != -1) {
     so->so_expire = curtime + SO_EXPIRE;
     insque(so, &so->slirp->udb);
   }
@@ -350,7 +350,7 @@
 	if (!so) {
 	    return NULL;
 	}
-	so->s = socket(AF_INET,SOCK_DGRAM,0);
+	so->s = qemu_socket(AF_INET,SOCK_DGRAM,0);
 	so->so_expire = curtime + SO_EXPIRE;
 	insque(so, &slirp->udb);
 
diff --git a/vl.c b/vl.c
index 0816ed7..d6f196c 100644
--- a/vl.c
+++ b/vl.c
@@ -1243,7 +1243,7 @@
     struct hpet_info info;
     int r, fd;
 
-    fd = open("/dev/hpet", O_RDONLY);
+    fd = qemu_open("/dev/hpet", O_RDONLY);
     if (fd < 0)
         return -1;
 
@@ -1292,7 +1292,7 @@
     int rtc_fd;
     unsigned long current_rtc_freq = 0;
 
-    TFR(rtc_fd = open("/dev/rtc", O_RDONLY));
+    TFR(rtc_fd = qemu_open("/dev/rtc", O_RDONLY));
     if (rtc_fd < 0)
         return -1;
     ioctl(rtc_fd, RTC_IRQP_READ, &current_rtc_freq);
@@ -3337,7 +3337,7 @@
     int err;
     int fds[2];
 
-    err = pipe(fds);
+    err = qemu_pipe(fds);
     if (err == -1)
         return -errno;
 
@@ -5507,6 +5507,9 @@
 	} else if (pid < 0)
             exit(1);
 
+	close(fds[0]);
+	qemu_set_cloexec(fds[1]);
+
 	setsid();
 
 	pid = fork();
@@ -5907,7 +5910,7 @@
 	    exit(1);
 
 	chdir("/");
-	TFR(fd = open("/dev/null", O_RDWR));
+	TFR(fd = qemu_open("/dev/null", O_RDWR));
 	if (fd == -1)
 	    exit(1);
     }
diff --git a/vnc.c b/vnc.c
index 2bb8024..32c4678 100644
--- a/vnc.c
+++ b/vnc.c
@@ -2247,7 +2247,7 @@
     /* Catch-up */
     vga_hw_update();
 
-    int csock = accept(vs->lsock, (struct sockaddr *)&addr, &addrlen);
+    int csock = qemu_accept(vs->lsock, (struct sockaddr *)&addr, &addrlen);
     if (csock != -1) {
         vnc_connect(vs, csock);
     }