00d7d6cef

Denis Vlasenko

2006-09-11

SOURCE BUG Duration: 40s

Subject

nfsmount: sanitize it. It had a rather peculiar idea of implementing "bg" option - it was going to return a special flag back to caller and expecting caller to call it again with special parameter! Also caller was charged with calling mount() syscall...

SOURCE BUG

This commit was genuinely broken at the time (not related to our build system): scripts/config/mkconfigs.c (2005-08-01 - 2006-10-05 arasi) uretilen bbconfigopts.h icine .config satirlarini kacissiz, ham newline'larla gömüyordu - gecerli C degildi, busybox'un kendi bb_mkdep araci bile parse edemiyordu (7d219aab7 ile duzeltildi, mkconfigs shell scripte cevrildi)

→ Bu hatanin tam hikayesini ve duzeltme diffini gor

Binary

No binary (build failed)

Architecture & Compiler Flags

Architecture: x86_64-linux-gnu Compiler: gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0

—

Build Log

(log yok)

Commit Diff GitHub →

include/libbb.h +2 -2 util-linux/mount.c +14 -14 util-linux/nfsmount.c +207 -173

include/libbb.h+2 -2

@@ -303,8 +303,8 @@ extern int set_loop(char *device, const char file, int offset);
	extern int vdprintf(int d, const char *format, va_list ap);
	#endif

-	int nfsmount(const char spec, const char node, int *flags,
-	char **mount_opts, int running_bg);
+	int mount_it_now(struct mntent mp, int vfsflags, char filteropts);
+	int nfsmount(struct mntent mp, int vfsflags, char filteropts);

	/* Include our own copy of struct sysinfo to avoid binary compatibility
	* problems with Linux 2.4, which changed things. Grumble, grumble. */

util-linux/mount.c+14 -14

@@ -102,7 +102,7 @@ static void append_mount_options(char *oldopts, char newopts)
	}

	/* Use the mount_options list to parse options into flags.
-	* Return list of unrecognized options in strflags if strflags!=NULL /
+	* Also return list of unrecognized options if unrecognized!=NULL */
	static int parse_mount_options(char options, char *unrecognized)
	{
	int flags = MS_SILENT;
@@ -188,7 +188,7 @@ void delete_block_backed_filesystems(void);
	#endif

	#if ENABLE_FEATURE_MTAB_SUPPORT
-	static int useMtab;
+	static int useMtab = 1;
	static int fakeIt;
	#else
	#define useMtab 0
@@ -196,8 +196,7 @@ static int fakeIt;
	#endif

	// Perform actual mount of specific filesystem at specific location.
-
-	static int mount_it_now(struct mntent mp, int vfsflags, char filteropts)
+	int mount_it_now(struct mntent mp, int vfsflags, char filteropts)
	{
	int rc;

@@ -228,7 +227,7 @@ static int mount_it_now(struct mntent mp, int vfsflags, char filteropts)
	int i;

	if(!mountTable)
-	bb_error_msg("No %s",bb_path_mtab_file);
+	bb_error_msg("no %s",bb_path_mtab_file);

	// Add vfs string flags

@@ -244,7 +243,13 @@ static int mount_it_now(struct mntent mp, int vfsflags, char filteropts)
	// Write and close.

	if(!mp->mnt_type \|\| !*mp->mnt_type) mp->mnt_type="--bind";
-	addmntent(mountTable, mp);
+	// addmntent(mountTable, mp);
+	if(0) bb_error_msg("buggy: addmntent(fsname='%s' dir='%s' type='%s' opts='%s')",
+	mp->mnt_fsname,
+	mp->mnt_dir,
+	mp->mnt_type,
+	mp->mnt_opts
+	);
	endmntent(mountTable);
	if (ENABLE_FEATURE_CLEAN_UP)
	if(strcmp(mp->mnt_type,"--bind")) mp->mnt_type = 0;
@@ -319,13 +324,7 @@ static int singlemount(struct mntent *mp, int ignore_busy)
	(!mp->mnt_type \|\| !strcmp(mp->mnt_type,"nfs")) &&
	strchr(mp->mnt_fsname, ':') != NULL)
	{
-	if (nfsmount(mp->mnt_fsname, mp->mnt_dir, &vfsflags, &filteropts, 1)) {
-	bb_perror_msg("nfsmount failed");
-	} else {
-	// Strangely enough, nfsmount() doesn't actually mount() anything.
-	mp->mnt_type = "nfs";
-	rc = mount_it_now(mp, vfsflags, filteropts);
-	}
+	rc = nfsmount(mp, vfsflags, filteropts);
	goto report_error;
	}

@@ -400,7 +399,8 @@ report_error:

	if (rc && errno == EBUSY && ignore_busy) rc = 0;
	if (rc < 0)
-	bb_perror_msg("mounting %s on %s failed", mp->mnt_fsname, mp->mnt_dir);
+	/* perror here sometimes says "mounting ... on ... failed: Success" */
+	bb_error_msg("mounting %s on %s failed", mp->mnt_fsname, mp->mnt_dir);

	return rc;
	}

util-linux/nfsmount.c+207 -173

@@ -22,6 +22,8 @@
	*/

	#include "busybox.h"
+	#include <syslog.h>
+	#include <mntent.h>
	#include <sys/utsname.h>
	#undef TRUE
	#undef FALSE
@@ -332,15 +334,8 @@ static bool_t xdr_mountres3(XDR xdrs, mountres3 objp)
	return TRUE;
	}

-
	#define MAX_NFSPROT ((nfs_mount_version >= 4) ? 3 : 2)

-	enum {
-	EX_FAIL = 32, /* mount failure */
-	EX_BG = 256 /* retry in background (internal only) */
-	};
-
-
	/*
	* nfs_mount_version according to the sources seen at compile time.
	*/
@@ -426,50 +421,62 @@ next:
	return &p;
	}

-	int nfsmount(const char spec, const char node, int *flags,
-	char **mount_opts, int running_bg)
+	static int daemonize(void)
	{
-	/* prev_bg_host is a -o bg support:
-	* "bg: if the first NFS mount attempt times out,
-	* retry the mount in the background.
-	* After a mount operation is backgrounded,
-	* all subsequent mounts on the same NFS server
-	* will be backgrounded immediately, without first
-	* attempting the mount. A missing mount point is treated
-	* as a timeout, to allow for nested NFS mounts."
-	*
-	* But current implementation is a dirty hack -
-	* it works only if all mounts are to one host!
-	* IOW: think what will happen if you have this sequence:
-	* mount a.a.a.a:/dir /mnt/a -o bg
-	* mount b.b.b.b:/dir /mnt/b -o bg
-	* mount a.a.a.a:/dir /mnt/a -o bg
-	* mount b.b.b.b:/dir /mnt/b -o bg
-	*/
+	int fd;
+	int pid = fork();
+	if (pid < 0) /* error */
+	return -errno;
+	if (pid > 0) /* parent */
+	return 0;
+	/* child */
+	fd = xopen(bb_dev_null, O_RDWR);
+	dup2(fd, 0);
+	dup2(fd, 1);
+	dup2(fd, 2);
+	if (fd > 2) close(fd);
+	setsid();
+	openlog(bb_applet_name, LOG_PID, LOG_DAEMON);
+	logmode = LOGMODE_SYSLOG;
+	return 1;
+	}
+
+	// TODO
+	static inline int we_saw_this_host_before(const char *hostname)
+	{
+	return 0;
+	}
+
+	/* RPC strerror analogs are terminally idiotic:
+	* mandatory prefix and \n at end.
+	* This hopefully helps. Usage:
+	* error_msg_rpc(clnt_error(" ")) */
+	static void error_msg_rpc(const char *msg)
+	{
+	size_t len;
+	while (msg[0] == ' ' \|\| msg[0] == ':') msg++;
+	len = strlen(msg);
+	while (len && msg[len-1] == '\n') len--;
+	bb_error_msg("%.*s", len, msg);
+	}

-	static char *prev_bg_host;
-	char *hostdir;
+	int nfsmount(struct mntent mp, int vfsflags, char filteropts)
+	{
	CLIENT *mclient;
	char *hostname;
	char *pathname;
-	char *old_opts;
	char *mounthost;
-	struct timeval total_timeout;
-	enum clnt_stat clnt_stat;
	struct nfs_mount_data data;
-	char opt, opteq;
-	int val;
+	char *opt;
	struct hostent *hp;
	struct sockaddr_in server_addr;
	struct sockaddr_in mount_server_addr;
-	struct pmap* pm_mnt;
	int msock, fsock;
-	struct timeval retry_timeout;
	union {
	struct fhstatus nfsv2;
	struct mountres3 nfsv3;
	} status;
-	struct stat statbuf;
+	int daemonized;
	char *s;
	int port;
	int mountport;
@@ -488,29 +495,27 @@ int nfsmount(const char spec, const char node, int *flags,
	int nfsprog;
	int nfsvers;
	int retval;
-	time_t t;
-	time_t prevt;
-	time_t timeout;

	find_kernel_nfs_mount_version();

-	/* NB: old_opts and hostdir must be free()d prior to return! */
-
-	old_opts = NULL;
+	daemonized = 0;
	mounthost = NULL;
-	retval = EX_FAIL;
+	retval = ETIMEDOUT;
	msock = fsock = -1;
	mclient = NULL;

-	hostdir = xstrdup(spec);
+	/* NB: hostname, mounthost, filteropts must be free()d prior to return */
+
+	filteropts = xstrdup(filteropts); /* going to trash it later... */
+
+	hostname = xstrdup(mp->mnt_fsname);
	/* mount_main() guarantees that ':' is there */
-	s = strchr(hostdir, ':');
-	hostname = hostdir;
+	s = strchr(hostname, ':');
	pathname = s + 1;
	*s = '\0';
	/* Ignore all but first hostname in replicated mounts
	until they can be fully supported. ([email protected]) */
-	s = strchr(hostdir, ',');
+	s = strchr(hostname, ',');
	if (s) {
	*s = '\0';
	bb_error_msg("warning: multiple hostnames not supported");
@@ -538,12 +543,15 @@ int nfsmount(const char spec, const char node, int *flags,

	/* add IP address to mtab options for use when unmounting */

-	s = inet_ntoa(server_addr.sin_addr);
-	old_opts = *mount_opts;
-	if (!old_opts)
-	old_opts = xstrdup("");
-	*mount_opts = xasprintf("%s%saddr=%s",
-	old_opts, *old_opts ? "," : "", s);
+	if (!mp->mnt_opts) { /* TODO: actually mp->mnt_opts is never NULL */
+	mp->mnt_opts = xasprintf("addr=%s", inet_ntoa(server_addr.sin_addr));
+	} else {
+	char *tmp = xasprintf("%s%saddr=%s", mp->mnt_opts,
+	mp->mnt_opts[0] ? "," : "",
+	inet_ntoa(server_addr.sin_addr));
+	free(mp->mnt_opts);
+	mp->mnt_opts = tmp;
+	}

	/* Set default options.
	* rsize/wsize (and bsize, for ver >= 3) are left 0 in order to
@@ -576,10 +584,10 @@ int nfsmount(const char spec, const char node, int *flags,

	/* parse options */

-	for (opt = strtok(old_opts, ","); opt; opt = strtok(NULL, ",")) {
-	opteq = strchr(opt, '=');
+	for (opt = strtok(filteropts, ","); opt; opt = strtok(NULL, ",")) {
+	char *opteq = strchr(opt, '=');
	if (opteq) {
-	val = atoi(opteq + 1);
+	int val = atoi(opteq + 1);
	*opteq = '\0';
	if (!strcmp(opt, "rsize"))
	data.rsize = val;
@@ -641,7 +649,7 @@ int nfsmount(const char spec, const char node, int *flags,
	}
	}
	else {
-	val = 1;
+	int val = 1;
	if (!strncmp(opt, "no", 2)) {
	val = 0;
	opt += 2;
@@ -723,19 +731,20 @@ int nfsmount(const char spec, const char node, int *flags,

	data.version = nfs_mount_version;

-	if (*flags & MS_REMOUNT)
-	goto copy_data_and_return;
+	if (vfsflags & MS_REMOUNT)
+	goto do_mount;

	/*
	* If the previous mount operation on the same host was
	* backgrounded, and the "bg" for this mount is also set,
	* give up immediately, to avoid the initial timeout.
	*/
-	if (bg && !running_bg &&
-	prev_bg_host && strcmp(hostname, prev_bg_host) == 0) {
-	if (retry > 0)
-	retval = EX_BG;
-	goto ret;
+	if (bg && we_saw_this_host_before(hostname)) {
+	daemonized = daemonize(); /* parent or error */
+	if (daemonized <= 0) { /* parent or error */
+	retval = -daemonized;
+	goto ret;
+	}
	}

	/* create mount daemon client */
@@ -745,7 +754,8 @@ int nfsmount(const char spec, const char node, int *flags,
	mount_server_addr.sin_family = AF_INET;
	mount_server_addr.sin_addr.s_addr = inet_addr(hostname);
	} else {
-	if ((hp = gethostbyname(mounthost)) == NULL) {
+	hp = gethostbyname(mounthost);
+	if (hp == NULL) {
	bb_herror_msg("%s", mounthost);
	goto fail;
	} else {
@@ -761,11 +771,9 @@ int nfsmount(const char spec, const char node, int *flags,
	}

	/*
-	* The following loop implements the mount retries. On the first
-	* call, "running_bg" is 0. When the mount times out, and the
-	* "bg" option is set, the exit status EX_BG will be returned.
-	* For a backgrounded mount, there will be a second call by the
-	* child process with "running_bg" set to 1.
+	* The following loop implements the mount retries. When the mount
+	* times out, and the "bg" option is set, we background ourself
+	* and continue trying.
	*
	* The case where the mount point is not present and the "bg"
	* option is set, is treated as a timeout. This is done to
@@ -773,115 +781,122 @@ int nfsmount(const char spec, const char node, int *flags,
	*
	* The "retry" count specified by the user is the number of
	* minutes to retry before giving up.
-	*
-	* Only the first error message will be displayed.
	*/
-	retry_timeout.tv_sec = 3;
-	retry_timeout.tv_usec = 0;
-	total_timeout.tv_sec = 20;
-	total_timeout.tv_usec = 0;
-	timeout = time(NULL) + 60 * retry;
-	prevt = 0;
-	t = 30;
-	val = 1;
-	for (;;) {
-	if (bg && stat(node, &statbuf) == -1) {
-	if (running_bg) {
-	sleep(val); /* 1, 2, 4, 8, 16, 30, ... */
-	val *= 2;
-	if (val > 30)
-	val = 30;
-	}
-	} else {
-	/* be careful not to use too many CPU cycles */
-	if (t - prevt < 30)
-	sleep(30);
-
-	pm_mnt = get_mountport(&mount_server_addr,
-	mountprog,
-	mountvers,
-	proto,
-	mountport);
-
-	/* contact the mount daemon via TCP */
-	mount_server_addr.sin_port = htons(pm_mnt->pm_port);
-	msock = RPC_ANYSOCK;
-
-	switch (pm_mnt->pm_prot) {
-	case IPPROTO_UDP:
-	mclient = clntudp_create(&mount_server_addr,
+	{
+	struct timeval total_timeout;
+	struct timeval retry_timeout;
+	struct pmap* pm_mnt;
+	time_t t;
+	time_t prevt;
+	time_t timeout;
+
+	retry_timeout.tv_sec = 3;
+	retry_timeout.tv_usec = 0;
+	total_timeout.tv_sec = 20;
+	total_timeout.tv_usec = 0;
+	timeout = time(NULL) + 60 * retry;
+	prevt = 0;
+	t = 30;
+	retry:
+	/* be careful not to use too many CPU cycles */
+	if (t - prevt < 30)
+	sleep(30);
+
+	pm_mnt = get_mountport(&mount_server_addr,
+	mountprog,
+	mountvers,
+	proto,
+	mountport);
+	nfsvers = (pm_mnt->pm_vers < 2) ? 2 : pm_mnt->pm_vers;
+
+	/* contact the mount daemon via TCP */
+	mount_server_addr.sin_port = htons(pm_mnt->pm_port);
+	msock = RPC_ANYSOCK;
+
+	switch (pm_mnt->pm_prot) {
+	case IPPROTO_UDP:
+	mclient = clntudp_create(&mount_server_addr,
	pm_mnt->pm_prog,
	pm_mnt->pm_vers,
	retry_timeout,
	&msock);
-	if (mclient)
-	break;
-	mount_server_addr.sin_port = htons(pm_mnt->pm_port);
-	msock = RPC_ANYSOCK;
-	case IPPROTO_TCP:
-	mclient = clnttcp_create(&mount_server_addr,
+	if (mclient)
+	break;
+	mount_server_addr.sin_port = htons(pm_mnt->pm_port);
+	msock = RPC_ANYSOCK;
+	case IPPROTO_TCP:
+	mclient = clnttcp_create(&mount_server_addr,
	pm_mnt->pm_prog,
	pm_mnt->pm_vers,
	&msock, 0, 0);
-	break;
-	default:
-	mclient = 0;
-	}
-	if (mclient) {
-	/* try to mount hostname:pathname */
-	mclient->cl_auth = authunix_create_default();
-
-	/* make pointers in xdr_mountres3 NULL so
-	* that xdr_array allocates memory for us
-	*/
-	memset(&status, 0, sizeof(status));
-
-	if (pm_mnt->pm_vers == 3)
-	clnt_stat = clnt_call(mclient, MOUNTPROC3_MNT,
-	(xdrproc_t) xdr_dirpath,
-	(caddr_t) &pathname,
-	(xdrproc_t) xdr_mountres3,
-	(caddr_t) &status,
-	total_timeout);
-	else
-	clnt_stat = clnt_call(mclient, MOUNTPROC_MNT,
-	(xdrproc_t) xdr_dirpath,
-	(caddr_t) &pathname,
-	(xdrproc_t) xdr_fhstatus,
-	(caddr_t) &status,
-	total_timeout);
-
-	if (clnt_stat == RPC_SUCCESS)
-	break; /* we're done */
-	if (errno != ECONNREFUSED) {
-	clnt_perror(mclient, "mount");
-	goto fail; /* don't retry */
-	}
-	if (!running_bg && prevt == 0)
-	clnt_perror(mclient, "mount");
-	auth_destroy(mclient->cl_auth);
-	clnt_destroy(mclient);
-	mclient = 0;
-	close(msock);
-	} else {
-	if (!running_bg && prevt == 0)
-	clnt_pcreateerror("mount");
+	break;
+	default:
+	mclient = 0;
+	}
+	if (!mclient) {
+	if (!daemonized && prevt == 0)
+	error_msg_rpc(clnt_spcreateerror(" "));
+	} else {
+	enum clnt_stat clnt_stat;
+	/* try to mount hostname:pathname */
+	mclient->cl_auth = authunix_create_default();
+
+	/* make pointers in xdr_mountres3 NULL so
+	* that xdr_array allocates memory for us
+	*/
+	memset(&status, 0, sizeof(status));
+
+	if (pm_mnt->pm_vers == 3)
+	clnt_stat = clnt_call(mclient, MOUNTPROC3_MNT,
+	(xdrproc_t) xdr_dirpath,
+	(caddr_t) &pathname,
+	(xdrproc_t) xdr_mountres3,
+	(caddr_t) &status,
+	total_timeout);
+	else
+	clnt_stat = clnt_call(mclient, MOUNTPROC_MNT,
+	(xdrproc_t) xdr_dirpath,
+	(caddr_t) &pathname,
+	(xdrproc_t) xdr_fhstatus,
+	(caddr_t) &status,
+	total_timeout);
+
+	if (clnt_stat == RPC_SUCCESS)
+	goto prepare_kernel_data; /* we're done */
+	if (errno != ECONNREFUSED) {
+	error_msg_rpc(clnt_sperror(mclient, " "));
+	goto fail; /* don't retry */
	}
-	prevt = t;
+	/* Connection refused */
+	if (!daemonized && prevt == 0) /* print just once */
+	error_msg_rpc(clnt_sperror(mclient, " "));
+	auth_destroy(mclient->cl_auth);
+	clnt_destroy(mclient);
+	mclient = 0;
+	close(msock);
	}
+
+	/* Timeout. We are going to retry... maybe */
+
	if (!bg)
	goto fail;
-	if (!running_bg) {
-	prev_bg_host = xstrdup(hostname);
-	if (retry > 0)
-	retval = EX_BG;
-	goto fail;
+	if (!daemonized) {
+	daemonized = daemonize();
+	if (daemonized <= 0) { /* parent or error */
+	retval = -daemonized;
+	goto ret;
+	}
	}
+	prevt = t;
	t = time(NULL);
	if (t >= timeout)
+	/* TODO error message */
	goto fail;
+
+	goto retry;
	}
-	nfsvers = (pm_mnt->pm_vers < 2) ? 2 : pm_mnt->pm_vers;
+
+	prepare_kernel_data:

	if (nfsvers == 2) {
	if (status.nfsv2.fhs_status != 0) {
@@ -927,17 +942,17 @@ int nfsmount(const char spec, const char node, int *flags,
	} else
	fsock = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP);
	if (fsock < 0) {
-	perror("nfs socket");
+	bb_perror_msg("nfs socket");
	goto fail;
	}
	if (bindresvport(fsock, 0) < 0) {
-	perror("nfs bindresvport");
+	bb_perror_msg("nfs bindresvport");
	goto fail;
	}
	if (port == 0) {
	server_addr.sin_port = PMAPPORT;
	port = pmap_getport(&server_addr, nfsprog, nfsvers,
-	tcp ? IPPROTO_TCP : IPPROTO_UDP);
+	tcp ? IPPROTO_TCP : IPPROTO_UDP);
	if (port == 0)
	port = NFS_PORT;
	#ifdef NFS_MOUNT_DEBUG
@@ -961,16 +976,34 @@ int nfsmount(const char spec, const char node, int *flags,
	auth_destroy(mclient->cl_auth);
	clnt_destroy(mclient);
	close(msock);
-	copy_data_and_return:
-	mount_opts = xrealloc(mount_opts, sizeof(data));
-	memcpy(*mount_opts, &data, sizeof(data));

-	retval = 0;
+	if (bg) {
+	/* We must wait until mount directory is available */
+	struct stat statbuf;
+	int delay = 1;
+	while (stat(mp->mnt_dir, &statbuf) == -1) {
+	if (!daemonized) {
+	daemonized = daemonize();
+	if (daemonized <= 0) { /* parent or error */
+	retval = -daemonized;
+	goto ret;
+	}
+	}
+	sleep(delay); /* 1, 2, 4, 8, 16, 30, ... */
+	delay *= 2;
+	if (delay > 30)
+	delay = 30;
+	}
+	}
+
+	do_mount: /* perform actual mount */
+
+	mp->mnt_type = "nfs";
+	retval = mount_it_now(mp, vfsflags, (char*)&data);
	goto ret;

-	/* abort */
+	fail: /* abort */

-	fail:
	if (msock != -1) {
	if (mclient) {
	auth_destroy(mclient->cl_auth);
@@ -982,7 +1015,8 @@ fail:
	close(fsock);

	ret:
-	free(hostdir);
-	free(old_opts);
+	free(hostname);
+	free(mounthost);
+	free(filteropts);
	return retval;
	}