summaryrefslogtreecommitdiffstats
path: root/utils/mount/nfs4mount.c
diff options
context:
space:
mode:
Diffstat (limited to 'utils/mount/nfs4mount.c')
-rw-r--r--utils/mount/nfs4mount.c481
1 files changed, 481 insertions, 0 deletions
diff --git a/utils/mount/nfs4mount.c b/utils/mount/nfs4mount.c
new file mode 100644
index 0000000..3e4f1e2
--- /dev/null
+++ b/utils/mount/nfs4mount.c
@@ -0,0 +1,481 @@
+/*
+ * nfs4mount.c -- Linux NFS mount
+ * Copyright (C) 2002 Trond Myklebust <trond.myklebust@fys.uio.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * Note: this file based on the original nfsmount.c
+ *
+ * 2006-06-06 Amit Gud <agud@redhat.com>
+ * - Moved to nfs-utils/utils/mount from util-linux/mount.
+ */
+
+#ifdef HAVE_CONFIG_H
+#include <config.h>
+#endif
+
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <netdb.h>
+#include <time.h>
+#include <sys/stat.h>
+#include <sys/mount.h>
+#include <netinet/in.h>
+#include <arpa/inet.h>
+#include <rpc/auth.h>
+#include <rpc/rpc.h>
+
+#ifdef HAVE_RPCSVC_NFS_PROT_H
+#include <rpcsvc/nfs_prot.h>
+#else
+#include <linux/nfs.h>
+#define nfsstat nfs_stat
+#endif
+
+#include "pseudoflavors.h"
+#include "nls.h"
+#include "xcommon.h"
+
+#include "mount.h"
+#include "mount_constants.h"
+#include "nfs4_mount.h"
+#include "nfs_mount.h"
+#include "error.h"
+#include "network.h"
+
+#if defined(VAR_LOCK_DIR)
+#define DEFAULT_DIR VAR_LOCK_DIR
+#else
+#define DEFAULT_DIR "/var/lock/subsys"
+#endif
+
+extern char *progname;
+extern int verbose;
+extern int sloppy;
+
+char *IDMAPLCK = DEFAULT_DIR "/rpcidmapd";
+#define idmapd_check() do { \
+ if (access(IDMAPLCK, F_OK)) { \
+ printf(_("Warning: rpc.idmapd appears not to be running.\n" \
+ " All uids will be mapped to the nobody uid.\n")); \
+ } \
+} while(0);
+
+char *GSSDLCK = DEFAULT_DIR "/rpcgssd";
+#define gssd_check() do { \
+ if (access(GSSDLCK, F_OK)) { \
+ printf(_("Warning: rpc.gssd appears not to be running.\n")); \
+ } \
+} while(0);
+
+#ifndef NFS_PORT
+#define NFS_PORT 2049
+#endif
+
+#define MAX_USER_FLAVOUR 16
+
+static int parse_sec(char *sec, int *pseudoflavour)
+{
+ int i, num_flavour = 0;
+
+ for (sec = strtok(sec, ":"); sec; sec = strtok(NULL, ":")) {
+ if (num_flavour >= MAX_USER_FLAVOUR) {
+ nfs_error(_("%s: maximum number of security flavors "
+ "exceeded"), progname);
+ return 0;
+ }
+ for (i = 0; i < flav_map_size; i++) {
+ if (strcmp(sec, flav_map[i].flavour) == 0) {
+ pseudoflavour[num_flavour++] = flav_map[i].fnum;
+ break;
+ }
+ }
+ if (i == flav_map_size) {
+ nfs_error(_("%s: unknown security type %s\n"),
+ progname, sec);
+ return 0;
+ }
+ }
+ if (!num_flavour)
+ nfs_error(_("%s: no security flavors passed to sec= option"),
+ progname);
+ return num_flavour;
+}
+
+static int parse_devname(char *hostdir, char **hostname, char **dirname)
+{
+ char *s;
+
+ if (!(s = strchr(hostdir, ':'))) {
+ nfs_error(_("%s: directory to mount not in host:dir format"),
+ progname);
+ return -1;
+ }
+ *hostname = hostdir;
+ *dirname = s + 1;
+ *s = '\0';
+ /* Ignore all but first hostname in replicated mounts
+ until they can be fully supported. (mack@sgi.com) */
+ if ((s = strchr(hostdir, ','))) {
+ *s = '\0';
+ nfs_error(_("%s: warning: multiple hostnames not supported"),
+ progname);
+ }
+ return 0;
+}
+
+static int fill_ipv4_sockaddr(const char *hostname, struct sockaddr_in *addr)
+{
+ struct hostent *hp;
+ addr->sin_family = AF_INET;
+
+ if (inet_aton(hostname, &addr->sin_addr))
+ return 0;
+ if ((hp = gethostbyname(hostname)) == NULL) {
+ nfs_error(_("%s: can't get address for %s\n"),
+ progname, hostname);
+ return -1;
+ }
+ if (hp->h_length > (int)sizeof(struct in_addr)) {
+ nfs_error(_("%s: got bad hp->h_length"), progname);
+ hp->h_length = sizeof(struct in_addr);
+ }
+ memcpy(&addr->sin_addr, hp->h_addr, hp->h_length);
+ return 0;
+}
+
+static int get_my_ipv4addr(char *ip_addr, int len)
+{
+ char myname[1024];
+ struct sockaddr_in myaddr;
+
+ if (gethostname(myname, sizeof(myname))) {
+ nfs_error(_("%s: can't determine client address\n"),
+ progname);
+ return -1;
+ }
+ if (fill_ipv4_sockaddr(myname, &myaddr))
+ return -1;
+ snprintf(ip_addr, len, "%s", inet_ntoa(myaddr.sin_addr));
+ ip_addr[len-1] = '\0';
+ return 0;
+}
+
+int nfs4mount(const char *spec, const char *node, int flags,
+ char **extra_opts, int fake, int running_bg)
+{
+ static struct nfs4_mount_data data;
+ static char hostdir[1024];
+ static char ip_addr[16] = "127.0.0.1";
+ static struct sockaddr_in server_addr, client_addr;
+ static int pseudoflavour[MAX_USER_FLAVOUR];
+ int num_flavour = 0;
+ int ip_addr_in_opts = 0;
+
+ char *hostname, *dirname, *old_opts;
+ char new_opts[1024];
+ char *opt, *opteq;
+ char *s;
+ int val;
+ int bg, soft, intr;
+ int nocto, noac, unshared;
+ int retry;
+ int retval = EX_FAIL;
+ time_t timeout, t;
+
+ if (strlen(spec) >= sizeof(hostdir)) {
+ nfs_error(_("%s: excessively long host:dir argument\n"),
+ progname);
+ goto fail;
+ }
+ strcpy(hostdir, spec);
+ if (parse_devname(hostdir, &hostname, &dirname))
+ goto fail;
+
+ if (fill_ipv4_sockaddr(hostname, &server_addr))
+ goto fail;
+ if (get_my_ipv4addr(ip_addr, sizeof(ip_addr)))
+ goto fail;
+
+ /* add IP address to mtab options for use when unmounting */
+ s = inet_ntoa(server_addr.sin_addr);
+ old_opts = *extra_opts;
+ if (!old_opts)
+ old_opts = "";
+ if (strlen(old_opts) + strlen(s) + 10 >= sizeof(new_opts)) {
+ nfs_error(_("%s: excessively long option argument\n"),
+ progname);
+ goto fail;
+ }
+ if (running_bg)
+ strncpy(new_opts, old_opts, sizeof(new_opts)-1);
+ else
+ snprintf(new_opts, sizeof(new_opts), "%s%saddr=%s",
+ old_opts, *old_opts ? "," : "", s);
+ *extra_opts = xstrdup(new_opts);
+
+ /* Set default options.
+ * rsize/wsize and timeo are left 0 in order to
+ * let the kernel decide.
+ */
+ memset(&data, 0, sizeof(data));
+ data.retrans = 3;
+ data.acregmin = 3;
+ data.acregmax = 60;
+ data.acdirmin = 30;
+ data.acdirmax = 60;
+ data.proto = IPPROTO_TCP;
+
+ bg = 0;
+ soft = 0;
+ intr = NFS4_MOUNT_INTR;
+ nocto = 0;
+ noac = 0;
+ unshared = 0;
+ retry = -1;
+
+ /*
+ * NFSv4 specifies that the default port should be 2049
+ */
+ server_addr.sin_port = htons(NFS_PORT);
+
+ /* parse options */
+
+ for (opt = strtok(old_opts, ","); opt; opt = strtok(NULL, ",")) {
+ if ((opteq = strchr(opt, '='))) {
+ val = atoi(opteq + 1);
+ *opteq = '\0';
+ if (!strcmp(opt, "rsize"))
+ data.rsize = val;
+ else if (!strcmp(opt, "wsize"))
+ data.wsize = val;
+ else if (!strcmp(opt, "timeo"))
+ data.timeo = val;
+ else if (!strcmp(opt, "retrans"))
+ data.retrans = val;
+ else if (!strcmp(opt, "acregmin"))
+ data.acregmin = val;
+ else if (!strcmp(opt, "acregmax"))
+ data.acregmax = val;
+ else if (!strcmp(opt, "acdirmin"))
+ data.acdirmin = val;
+ else if (!strcmp(opt, "acdirmax"))
+ data.acdirmax = val;
+ else if (!strcmp(opt, "actimeo")) {
+ data.acregmin = val;
+ data.acregmax = val;
+ data.acdirmin = val;
+ data.acdirmax = val;
+ }
+ else if (!strcmp(opt, "retry"))
+ retry = val;
+ else if (!strcmp(opt, "port"))
+ server_addr.sin_port = htons(val);
+ else if (!strcmp(opt, "proto")) {
+ if (!strncmp(opteq+1, "tcp", 3))
+ data.proto = IPPROTO_TCP;
+ else if (!strncmp(opteq+1, "udp", 3))
+ data.proto = IPPROTO_UDP;
+ else
+ printf(_("Warning: Unrecognized proto= option.\n"));
+ } else if (!strcmp(opt, "clientaddr")) {
+ if (strlen(opteq+1) >= sizeof(ip_addr))
+ printf(_("Invalid client address %s"),
+ opteq+1);
+ strncpy(ip_addr,opteq+1, sizeof(ip_addr));
+ ip_addr[sizeof(ip_addr)-1] = '\0';
+ ip_addr_in_opts = 1;
+ } else if (!strcmp(opt, "sec")) {
+ num_flavour = parse_sec(opteq+1, pseudoflavour);
+ if (!num_flavour)
+ goto fail;
+ } else if (!strcmp(opt, "addr") || sloppy) {
+ /* ignore */;
+ } else {
+ printf(_("unknown nfs mount parameter: "
+ "%s=%d\n"), opt, val);
+ goto fail;
+ }
+ } else {
+ val = 1;
+ if (!strncmp(opt, "no", 2)) {
+ val = 0;
+ opt += 2;
+ }
+ if (!strcmp(opt, "bg"))
+ bg = val;
+ else if (!strcmp(opt, "fg"))
+ bg = !val;
+ else if (!strcmp(opt, "soft"))
+ soft = val;
+ else if (!strcmp(opt, "hard"))
+ soft = !val;
+ else if (!strcmp(opt, "intr"))
+ intr = val;
+ else if (!strcmp(opt, "cto"))
+ nocto = !val;
+ else if (!strcmp(opt, "ac"))
+ noac = !val;
+ else if (!strcmp(opt, "sharecache"))
+ unshared = !val;
+ else if (!sloppy) {
+ printf(_("unknown nfs mount option: %s%s\n"),
+ val ? "" : "no", opt);
+ goto fail;
+ }
+ }
+ }
+
+ /* if retry is still -1, then it wasn't set via an option */
+ if (retry == -1) {
+ if (bg)
+ retry = 10000; /* 10000 mins == ~1 week */
+ else
+ retry = 2; /* 2 min default on fg mounts */
+ }
+
+ data.flags = (soft ? NFS4_MOUNT_SOFT : 0)
+ | (intr ? NFS4_MOUNT_INTR : 0)
+ | (nocto ? NFS4_MOUNT_NOCTO : 0)
+ | (noac ? NFS4_MOUNT_NOAC : 0)
+ | (unshared ? NFS4_MOUNT_UNSHARED : 0);
+
+ /*
+ * Give a warning if the rpc.idmapd daemon is not running
+ */
+#if 0
+ /* We shouldn't have these checks as nothing in this package
+ * creates the files that are checked
+ */
+ idmapd_check();
+
+ if (num_flavour == 0)
+ pseudoflavour[num_flavour++] = AUTH_UNIX;
+ else {
+ /*
+ * ditto with rpc.gssd daemon
+ */
+ gssd_check();
+ }
+#endif
+ data.auth_flavourlen = num_flavour;
+ data.auth_flavours = pseudoflavour;
+
+ data.client_addr.data = ip_addr;
+ data.client_addr.len = strlen(ip_addr);
+
+ data.mnt_path.data = dirname;
+ data.mnt_path.len = strlen(dirname);
+
+ data.hostname.data = hostname;
+ data.hostname.len = strlen(hostname);
+ data.host_addr = (struct sockaddr *)&server_addr;
+ data.host_addrlen = sizeof(server_addr);
+
+#ifdef NFS_MOUNT_DEBUG
+ printf(_("rsize = %d, wsize = %d, timeo = %d, retrans = %d\n"),
+ data.rsize, data.wsize, data.timeo, data.retrans);
+ printf(_("acreg (min, max) = (%d, %d), acdir (min, max) = (%d, %d)\n"),
+ data.acregmin, data.acregmax, data.acdirmin, data.acdirmax);
+ printf(_("port = %d, bg = %d, retry = %d, flags = %.8x\n"),
+ ntohs(server_addr.sin_port), bg, retry, data.flags);
+ printf(_("soft = %d, intr = %d, nocto = %d, noac = %d, "
+ "nosharecache = %d\n"),
+ (data.flags & NFS4_MOUNT_SOFT) != 0,
+ (data.flags & NFS4_MOUNT_INTR) != 0,
+ (data.flags & NFS4_MOUNT_NOCTO) != 0,
+ (data.flags & NFS4_MOUNT_NOAC) != 0,
+ (data.flags & NFS4_MOUNT_UNSHARED) != 0);
+
+ if (num_flavour > 0) {
+ int pf_cnt, i;
+
+ printf(_("sec = "));
+ for (pf_cnt = 0; pf_cnt < num_flavour; pf_cnt++) {
+ for (i = 0; i < flav_map_size; i++) {
+ if (flav_map[i].fnum == pseudoflavour[pf_cnt]) {
+ printf("%s", flav_map[i].flavour);
+ break;
+ }
+ }
+ printf("%s", (pf_cnt < num_flavour-1) ? ":" : "\n");
+ }
+ }
+ printf(_("proto = %s\n"), (data.proto == IPPROTO_TCP) ? _("tcp") : _("udp"));
+#endif
+
+ timeout = time(NULL) + 60 * retry;
+ data.version = NFS4_MOUNT_VERSION;
+ for (;;) {
+ if (verbose) {
+ printf(_("%s: pinging: prog %d vers %d prot %s port %d\n"),
+ progname, NFS_PROGRAM, 4,
+ data.proto == IPPROTO_UDP ? "udp" : "tcp",
+ ntohs(server_addr.sin_port));
+ }
+ client_addr.sin_family = 0;
+ client_addr.sin_addr.s_addr = 0;
+ clnt_ping(&server_addr, NFS_PROGRAM, 4, data.proto, &client_addr);
+ if (rpc_createerr.cf_stat == RPC_SUCCESS) {
+ if (!ip_addr_in_opts &&
+ client_addr.sin_family != 0 &&
+ client_addr.sin_addr.s_addr != 0) {
+ snprintf(ip_addr, sizeof(ip_addr), "%s",
+ inet_ntoa(client_addr.sin_addr));
+ data.client_addr.len = strlen(ip_addr);
+ }
+ break;
+ }
+
+ if (!bg) {
+ switch(rpc_createerr.cf_stat) {
+ case RPC_TIMEDOUT:
+ break;
+ case RPC_SYSTEMERROR:
+ if (errno == ETIMEDOUT)
+ break;
+ /* FALLTHRU */
+ default:
+ rpc_mount_errors(hostname, 0, bg);
+ goto fail;
+ }
+ }
+
+ if (bg && !running_bg) {
+ if (retry > 0)
+ retval = EX_BG;
+ goto fail;
+ }
+
+ t = time(NULL);
+ if (t >= timeout) {
+ rpc_mount_errors(hostname, 0, bg);
+ goto fail;
+ }
+ rpc_mount_errors(hostname, 1, bg);
+ continue;
+ }
+
+ if (!fake) {
+ if (mount(spec, node, "nfs4",
+ flags & ~(MS_USER|MS_USERS), &data)) {
+ mount_error(spec, node, errno);
+ goto fail;
+ }
+ }
+
+ return EX_SUCCESS;
+
+fail:
+ return retval;
+}