diff -up nfs-utils-1.2.3/configure.ac.orig nfs-utils-1.2.3/configure.ac
--- nfs-utils-1.2.3/configure.ac.orig 2010-10-04 15:47:06.581340251 -0400
+++ nfs-utils-1.2.3/configure.ac 2010-10-04 15:48:00.049475138 -0400
@@ -72,11 +72,15 @@ AC_ARG_ENABLE(nfsv4,
enable_nfsv4=yes)
if test "$enable_nfsv4" = yes; then
AC_DEFINE(NFS4_SUPPORTED, 1, [Define this if you want NFSv4 support compiled in])
+ BLKMAPD=blkmapd
IDMAPD=idmapd
+ SPNFSD=spnfsd
else
enable_nfsv4=
+ BLKMAPD=
IDMAPD=
fi
+ AC_SUBST(BLKMAPD)
AC_SUBST(IDMAPD)
AC_SUBST(enable_nfsv4)
AM_CONDITIONAL(CONFIG_NFSV4, [test "$enable_nfsv4" = "yes"])
@@ -439,9 +443,11 @@ AC_CONFIG_FILES([
tools/mountstats/Makefile
tools/nfs-iostat/Makefile
utils/Makefile
+ utils/blkmapd/Makefile
utils/exportfs/Makefile
utils/gssd/Makefile
utils/idmapd/Makefile
+ utils/spnfsd/Makefile
utils/mount/Makefile
utils/mountd/Makefile
utils/nfsd/Makefile
diff -up nfs-utils-1.2.3/support/include/nfslib.h.orig nfs-utils-1.2.3/support/include/nfslib.h
--- nfs-utils-1.2.3/support/include/nfslib.h.orig 2010-09-28 08:24:16.000000000 -0400
+++ nfs-utils-1.2.3/support/include/nfslib.h 2010-10-04 15:48:00.069391711 -0400
@@ -89,6 +89,7 @@ struct exportent {
char * e_fslocdata;
char * e_uuid;
struct sec_entry e_secinfo[SECFLAVOR_COUNT+1];
+ int e_pnfs;
};
struct rmtabent {
diff -up nfs-utils-1.2.3/support/include/nfs/nfs.h.orig nfs-utils-1.2.3/support/include/nfs/nfs.h
--- nfs-utils-1.2.3/support/include/nfs/nfs.h.orig 2010-09-28 08:24:16.000000000 -0400
+++ nfs-utils-1.2.3/support/include/nfs/nfs.h 2010-10-04 15:48:00.050474882 -0400
@@ -47,6 +47,7 @@ struct nfs_fh_old {
#define NFSCTL_GETFH 6 /* get an fh (used by mountd) */
#define NFSCTL_GETFD 7 /* get an fh by path (used by mountd) */
#define NFSCTL_GETFS 8 /* get an fh by path with max size (used by mountd) */
+#define NFSCTL_FD2FH 9 /* get a fh from a fd */
#define NFSCTL_UDPBIT (1 << (17 - 1))
#define NFSCTL_TCPBIT (1 << (18 - 1))
@@ -136,6 +137,11 @@ struct nfsctl_fsparm {
int gd_maxlen;
};
+/* FD2FH */
+struct nfsctl_fd2fh {
+ int fd;
+};
+
/*
* This is the argument union.
*/
@@ -149,6 +155,7 @@ struct nfsctl_arg {
struct nfsctl_fhparm u_getfh;
struct nfsctl_fdparm u_getfd;
struct nfsctl_fsparm u_getfs;
+ struct nfsctl_fd2fh u_fd2fh;
} u;
#define ca_svc u.u_svc
#define ca_client u.u_client
@@ -157,6 +164,7 @@ struct nfsctl_arg {
#define ca_getfh u.u_getfh
#define ca_getfd u.u_getfd
#define ca_getfs u.u_getfs
+#define ca_fd2fh u.u_fd2fh
#define ca_authd u.u_authd
};
diff -up nfs-utils-1.2.3/support/nfs/exports.c.orig nfs-utils-1.2.3/support/nfs/exports.c
--- nfs-utils-1.2.3/support/nfs/exports.c.orig 2010-10-04 15:47:06.589340404 -0400
+++ nfs-utils-1.2.3/support/nfs/exports.c 2010-10-04 15:48:00.070350101 -0400
@@ -107,6 +107,7 @@ static void init_exportent (struct expor
ee->e_nsquids = 0;
ee->e_nsqgids = 0;
ee->e_uuid = NULL;
+ ee->e_pnfs = 0;
}
struct exportent *
@@ -299,6 +300,8 @@ putexportent(struct exportent *ep)
}
fprintf(fp, "anonuid=%d,anongid=%d", ep->e_anonuid, ep->e_anongid);
secinfo_show(fp, ep);
+ if (ep->e_pnfs)
+ fprintf(fp, ",pnfs");
fprintf(fp, ")\n");
}
@@ -557,6 +560,10 @@ parseopts(char *cp, struct exportent *ep
clearflags(NFSEXP_NOACL, active, ep);
else if (strcmp(opt, "no_acl") == 0)
setflags(NFSEXP_NOACL, active, ep);
+ else if (strcmp(opt, "pnfs") == 0)
+ ep->e_pnfs = 1;
+ else if (strcmp(opt, "no_pnfs") == 0)
+ ep->e_pnfs = 0;
else if (strncmp(opt, "anonuid=", 8) == 0) {
char *oe;
ep->e_anonuid = strtol(opt+8, &oe, 10);
diff -up nfs-utils-1.2.3/utils/blkmapd/atomicio.c.orig nfs-utils-1.2.3/utils/blkmapd/atomicio.c
--- nfs-utils-1.2.3/utils/blkmapd/atomicio.c.orig 2010-10-04 15:48:00.071418550 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/atomicio.c 2010-10-04 15:48:00.071418550 -0400
@@ -0,0 +1,54 @@
+/*
+ * Copyright (c) 2002 Marius Aamodt Eriksen <marius@monkey.org>
+ * Copyright (c) 1995,1999 Theo de Raadt. All rights reserved.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/types.h>
+#include <unistd.h>
+#include <errno.h>
+
+/*
+ * ensure all of data on socket comes through. f==read || f==write
+ */
+ssize_t atomicio(ssize_t(*f) (int, void *, size_t), int fd, void *_s, size_t n)
+{
+ char *s = _s;
+ ssize_t res, pos = 0;
+
+ while (n > pos) {
+ res = (f) (fd, s + pos, n - pos);
+ switch (res) {
+ case -1:
+ if (errno == EINTR || errno == EAGAIN)
+ continue;
+ case 0:
+ if (pos != 0)
+ return pos;
+ return res;
+ default:
+ pos += res;
+ }
+ }
+ return pos;
+}
diff -up nfs-utils-1.2.3/utils/blkmapd/cfg.c.orig nfs-utils-1.2.3/utils/blkmapd/cfg.c
--- nfs-utils-1.2.3/utils/blkmapd/cfg.c.orig 2010-10-04 15:48:00.072379352 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/cfg.c 2010-10-04 15:48:00.072379352 -0400
@@ -0,0 +1,248 @@
+/*
+ * Copyright (c) 2010 EMC Corporation, Haiying Tang <Tang_Haiying@emc.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/param.h>
+#include <sys/stat.h>
+#include <linux/errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <string.h>
+#include <fcntl.h>
+#include <ctype.h>
+
+#include "device-discovery.h"
+#include "cfg.h"
+
+char *conf_path = "/etc/blkmapd.conf";
+
+struct scan_root_list *scan_root_list_head;
+
+void bl_release_list(void)
+{
+ struct scan_root_list *root = scan_root_list_head;
+ struct scan_device_list *disk;
+
+ while (root) {
+ disk = root->disk;
+ while (disk) {
+ root->disk = disk->next;
+ free(disk->name);
+ free(disk);
+ disk = root->disk;
+ }
+ scan_root_list_head = root->next;
+ free(root->name);
+ free(root);
+ root = scan_root_list_head;
+ }
+}
+
+struct scan_root_list *bl_alloc_root_list(char *name, int all_disk)
+{
+ struct scan_root_list *root;
+
+ root = malloc(sizeof(struct scan_root_list));
+ if (!root)
+ goto nomem;
+
+ root->name = strdup(name);
+ if (!root->name)
+ goto nomem;
+ root->next = scan_root_list_head;
+ root->all_disk = all_disk;
+ scan_root_list_head = root;
+ return root;
+
+ nomem:
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ if (root)
+ free(root);
+ return NULL;
+}
+
+struct scan_device_list *bl_alloc_device_list(struct scan_root_list *root,
+ char *name)
+{
+ struct scan_device_list *device;
+
+ device = malloc(sizeof(struct scan_device_list));
+ if (!device)
+ goto nomem;
+
+ device->name = strdup(name);
+ if (!device->name)
+ goto nomem;
+ device->next = root->disk;
+ root->disk = device;
+ return device;
+
+ nomem:
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ if (device)
+ free(device);
+ return NULL;
+}
+
+struct scan_device_list *bl_insert_device_list(struct scan_root_list *root,
+ char *name)
+{
+ struct scan_device_list *device = root->disk;
+
+ /* Check whether this device has been inserted */
+ while (device) {
+ if (device->name && !strcmp(device->name, name))
+ return device;
+ device = device->next;
+ }
+
+ return bl_alloc_device_list(root, name);
+}
+
+struct scan_root_list *bl_insert_root_list(char *name, int all_disk)
+{
+ struct scan_root_list *root = scan_root_list_head;
+
+ /* Check whether this root has been inserted */
+ while (root) {
+ if (root->name && !strcmp(root->name, name))
+ return root;
+ root = root->next;
+ }
+
+ return bl_alloc_root_list(name, all_disk);
+}
+
+int bl_parse_line(char *line, struct scan_root_list **bl_root)
+{
+ char *root, *device, *end;
+
+ root = strdup(line);
+ end = root + strlen(line);
+
+ /* Skip comments */
+ if (*root == '#')
+ return 0;
+
+ /* Trim leading space */
+ while (*root != '\0' && isspace(*root))
+ root++;
+ if (*root == '\0')
+ return 0;
+
+ /* Trim trailing space and set "end" to last char */
+ while ((isspace(*end) || (*end == '\0')) && (end > root))
+ end--;
+
+ /* For lines ending with '/' or '/','*': add as a dir root */
+ if ((*end == '/') ||
+ ((*end == '*') && (end - root >= 1) && (*(end - 1) == '/'))) {
+ if (*end == '*')
+ end--;
+ if (*end == '/')
+ end--;
+ *(end + 1) = '\0';
+ *bl_root = bl_insert_root_list(root, 1);
+ return 0;
+ }
+
+ /* Other lines: add as a device */
+ device = end;
+ while ((*device != '/') && (device > root))
+ device--;
+ if (device == root) {
+ BL_LOG_ERR("%s: invalid config line\n", __func__);
+ return -1;
+ }
+ *device = '\0';
+ *bl_root = bl_insert_root_list(root, 0);
+ if (*bl_root == NULL)
+ return -ENOMEM;
+ if (*end == '*')
+ end--;
+ *(end + 1) = '\0';
+ if (bl_insert_device_list(*bl_root, device + 1) == NULL)
+ return -ENOMEM;
+
+ return 0;
+}
+
+int bl_set_default_conf(void)
+{
+ struct scan_root_list *root = NULL;
+ int rv;
+
+ bl_release_list();
+ rv = bl_parse_line("/dev/sd*", &root);
+ if (rv < 0)
+ return rv;
+ rv = bl_parse_line("/dev/mapper/", &root);
+ return rv;
+}
+
+int bl_parse_conf(char *buf)
+{
+ char *tmp = buf, *line = buf, *end = buf + strlen(buf);
+ struct scan_root_list *bl_root = NULL;
+ int rv;
+
+ while (tmp < end) {
+ if (*tmp == '\n') {
+ *tmp = '\0';
+ rv = bl_parse_line(line, &bl_root);
+ if (rv < 0)
+ return rv;
+ line = tmp + 1;
+ }
+ tmp++;
+ }
+
+ return 0;
+}
+
+int bl_cfg_init(void)
+{
+ struct scan_root_list *root = NULL;
+ FILE *f = NULL;
+ char buf[PATH_MAX];
+ int rv = 0;
+
+ f = fopen(conf_path, "r");
+ if (f == NULL)
+ rv = bl_set_default_conf();
+ else {
+ while (fgets(buf, sizeof buf, f) != NULL) {
+ rv = bl_parse_line(buf, &root);
+ if (rv < 0)
+ break;
+ }
+ }
+ if (!scan_root_list_head)
+ rv = -EINVAL;
+
+ if (f)
+ fclose(f);
+ return rv;
+}
diff -up nfs-utils-1.2.3/utils/blkmapd/cfg.h.orig nfs-utils-1.2.3/utils/blkmapd/cfg.h
--- nfs-utils-1.2.3/utils/blkmapd/cfg.h.orig 2010-10-04 15:48:00.072379352 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/cfg.h 2010-10-04 15:48:00.072379352 -0400
@@ -0,0 +1,47 @@
+/*
+ * bl-cfg.h
+ *
+ * Copyright (c) 2010 EMC Corporation, Haiying Tang <Tang_Haiying@emc.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+#ifndef BL_CFG_H
+#define BL_CFG_H
+
+extern char *conf_path;
+extern struct scan_root_list *scan_root_list_head;
+
+struct scan_device_list {
+ struct scan_device_list *next;
+ char *name;
+};
+
+struct scan_root_list {
+ struct scan_root_list *next;
+ unsigned int all_disk;
+ char *name;
+ struct scan_device_list *disk;
+};
+
+int bl_cfg_init(void);
+
+#endif
diff -up nfs-utils-1.2.3/utils/blkmapd/device-discovery.c.orig nfs-utils-1.2.3/utils/blkmapd/device-discovery.c
--- nfs-utils-1.2.3/utils/blkmapd/device-discovery.c.orig 2010-10-04 15:48:00.073350775 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/device-discovery.c 2010-10-04 15:48:00.073350775 -0400
@@ -0,0 +1,502 @@
+/*
+ * device-discovery.c: main function, discovering device and processing
+ * pipe request from kernel.
+ *
+ * Copyright (c) 2010 EMC Corporation, Haiying Tang <Tang_Haiying@emc.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <dirent.h>
+#include <ctype.h>
+#include <linux/kdev_t.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/ioctl.h>
+#include <sys/mount.h>
+#include <sys/select.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <libgen.h>
+#include <errno.h>
+#include <scsi/scsi.h>
+#include <scsi/scsi_ioctl.h>
+#include <scsi/sg.h>
+#include "device-discovery.h"
+#include "cfg.h"
+
+#define BL_PIPE_FILE "/var/lib/nfs/rpc_pipefs/bl_device_pipe"
+#define PID_FILE "/var/run/blkmapd.pid"
+
+struct bl_disk *visible_disk_list;
+
+struct bl_disk_path *bl_get_path(const char *filepath,
+ struct bl_disk_path *paths)
+{
+ struct bl_disk_path *tmp = paths;
+ while (tmp) {
+ if (!strcmp(tmp->full_path, filepath))
+ break;
+ tmp = tmp->next;
+ }
+ return tmp;
+}
+
+/* Check whether valid_path is a substring(partition) of path */
+int bl_is_partition(struct bl_disk_path *valid_path, struct bl_disk_path *path)
+{
+ if (!strncmp(valid_path->full_path, path->full_path,
+ strlen(valid_path->full_path)))
+ return 1;
+
+ return 0;
+}
+
+/*
+ * For multipath devices, devices state could be PASSIVE/ACTIVE/PSEUDO,
+ * where PSEUDO > ACTIVE > PASSIVE. Device with highest state is used to
+ * create pseudo device. So if state is higher, the device path needs to
+ * be updated.
+ * If device-mapper multipath support is a must, pseudo devices should
+ * exist for each multipath device. If not, active device path will be
+ * chosen for device creation.
+ * Treat partition as invalid path.
+ */
+int bl_update_path(struct bl_disk_path *path, enum bl_path_state_e state,
+ struct bl_disk *disk)
+{
+ struct bl_disk_path *valid_path = disk->valid_path;
+
+ if (valid_path) {
+ if (valid_path->state >= state) {
+ if (bl_is_partition(valid_path, path))
+ return 0;
+ }
+ }
+ return 1;
+}
+
+void bl_release_disk(void)
+{
+ struct bl_disk *disk;
+ struct bl_disk_path *path = NULL;
+
+ while (visible_disk_list) {
+ disk = visible_disk_list;
+ path = disk->paths;
+ while (path) {
+ disk->paths = path->next;
+ free(path->full_path);
+ free(path);
+ path = disk->paths;
+ }
+ if (disk->serial)
+ free(disk->serial);
+ visible_disk_list = disk->next;
+ free(disk);
+ }
+}
+
+void bl_add_disk(char *filepath)
+{
+ struct bl_disk *disk = NULL;
+ int fd = 0;
+ struct stat sb;
+ off_t size = 0;
+ struct bl_serial *serial = NULL;
+ enum bl_path_state_e ap_state = BL_PATH_STATE_PASSIVE;
+ struct bl_disk_path *diskpath = NULL, *path = NULL;
+ dev_t dev;
+
+ BL_LOG_ERR("%s: %s\n", __func__, filepath);
+
+ fd = open(filepath, O_RDONLY | O_LARGEFILE);
+ if (fd < 0)
+ return;
+
+ if (fstat(fd, &sb)) {
+ close(fd);
+ return;
+ }
+
+ if (!sb.st_size)
+ ioctl(fd, BLKGETSIZE, &size);
+ else
+ size = sb.st_size;
+
+ if (!size) {
+ close(fd);
+ return;
+ }
+
+ dev = sb.st_rdev;
+ serial = bldev_read_serial(fd, filepath);
+
+ for (disk = visible_disk_list; disk != NULL; disk = disk->next) {
+ /* Already scanned or a partition?
+ * XXX: if released each time, maybe not need to compare
+ */
+ if ((serial->len == disk->serial->len) &&
+ !memcmp(serial->data, disk->serial->data, serial->len)) {
+ diskpath = bl_get_path(filepath, disk->paths);
+ break;
+ }
+ }
+
+ if (disk && diskpath) {
+ close(fd);
+ return;
+ }
+
+ bldev_read_ap_state(fd, &ap_state);
+ close(fd);
+
+ /*
+ * Not sure how to identify a pseudo device created by
+ * device-mapper, so leave /dev/mapper for now.
+ */
+ if (strncmp(filepath, "/dev/mapper", 11) == 0)
+ ap_state = BL_PATH_STATE_PSEUDO;
+
+ /* add path */
+ path = malloc(sizeof(struct bl_disk_path));
+ if (!path) {
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ goto out_err;
+ }
+ path->next = NULL;
+ path->state = ap_state;
+ path->full_path = strdup(filepath);
+ if (!path->full_path)
+ goto out_err;
+
+ if (!disk) { /* add disk */
+ disk = malloc(sizeof(struct bl_disk));
+ if (!disk) {
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ goto out_err;
+ }
+ disk->next = visible_disk_list;
+ disk->dev = dev;
+ disk->size = size;
+ disk->serial = serial;
+ disk->valid_path = path;
+ disk->paths = path;
+ visible_disk_list = disk;
+ } else {
+ path->next = disk->paths;
+ disk->paths = path;
+ /* check whether we need to update disk info */
+ if (bl_update_path(path, path->state, disk)) {
+ disk->dev = dev;
+ disk->size = size;
+ disk->valid_path = path;
+ }
+ }
+ return;
+
+ out_err:
+ if (path) {
+ if (path->full_path)
+ free(path->full_path);
+ free(path);
+ }
+ return;
+}
+
+void bl_devicescan(const char *filename, struct scan_root_list *root)
+{
+ /* scan all disks */
+ char filepath[PATH_MAX];
+ struct scan_device_list *device;
+
+ if (!strcmp(filename, ".") || !strcmp(filename, ".."))
+ return;
+
+ memset(filepath, 0, sizeof(filepath));
+ if (strlen(filename) < (PATH_MAX - strlen(root->name) - 2))
+ sprintf(filepath, "%s/%s", root->name, filename);
+ else {
+ BL_LOG_ERR("%s: name too long\n", __func__);
+ return;
+ }
+ if (root->all_disk)
+ goto valid;
+
+ device = root->disk;
+ while (device) {
+ /* If device->name is a subset of filename, this disk should be
+ * valid for scanning.
+ * For example, device->name is "sd", filename is "sda".
+ */
+ if (device->name
+ && !memcmp(filename, device->name, strlen(device->name)))
+ goto valid;
+ device = device->next;
+ }
+
+ return;
+
+ valid:
+ /*
+ * sg device is not a real device, but a device created according
+ * to each scsi device. It won't be used for pseudo device creation.
+ * I moved it here, so that sg devices will not be scanned.
+ */
+ if (!strncmp(filepath, "/dev/sg", 7))
+ return;
+ bl_add_disk(filepath);
+ return;
+}
+
+int bl_discover_devices(void)
+{
+ DIR *dir;
+ struct dirent *dp;
+ struct scan_root_list *root = scan_root_list_head;
+
+ /* release previous list */
+ bl_release_disk();
+
+ /* scan all disks */
+ while (root) {
+ dir = opendir(root->name);
+ if (dir == NULL) {
+ root = root->next;
+ continue;
+ }
+
+ while ((dp = readdir(dir)) != NULL)
+ bl_devicescan(dp->d_name, root);
+
+ root = root->next;
+ closedir(dir);
+ }
+
+ return 0;
+}
+
+/* process kernel request
+ * return 0: request processed, and no more request waiting;
+ * return 1: request processed, and more requests waiting;
+ * return < 0: error
+ */
+int bl_disk_inquiry_process(int fd)
+{
+ int ret = 0;
+ struct pipefs_hdr *head = NULL, *tmp;
+ char *buf = NULL;
+ uint32_t major, minor;
+ uint16_t buflen;
+ unsigned int len = 0;
+
+ head = calloc(1, sizeof(struct pipefs_hdr));
+ if (!head) {
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ return -ENOMEM;
+ }
+
+ /* read request */
+ if (atomicio(read, fd, head, sizeof(*head)) != sizeof(*head)) {
+ /* Note that an error in this or the next read is pretty
+ * catastrophic, as there is no good way to resync into
+ * the pipe's stream.
+ */
+ BL_LOG_ERR("Read pipefs head error!\n");
+ ret = -EIO;
+ goto out;
+ }
+
+ buflen = head->totallen - sizeof(*head);
+ buf = malloc(buflen);
+ if (!buf) {
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ if (atomicio(read, fd, buf, buflen) != buflen) {
+ BL_LOG_ERR("Read pipefs content error!\n");
+ ret = -EIO;
+ goto out;
+ }
+
+ head->status = BL_DEVICE_REQUEST_PROC;
+ switch (head->type) {
+ case BL_DEVICE_MOUNT:
+ if (!process_deviceinfo(buf, buflen, &major, &minor)) {
+ head->status = BL_DEVICE_REQUEST_ERR;
+ goto out;
+ }
+ tmp = realloc(head, sizeof(major) + sizeof(minor) +
+ sizeof(struct pipefs_hdr));
+ if (!tmp) {
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ ret = -ENOMEM;
+ goto out;
+ }
+ head = tmp;
+ memcpy((void *)head + sizeof(struct pipefs_hdr),
+ &major, sizeof(major));
+ memcpy((void *)head + sizeof(struct pipefs_hdr) + sizeof(major),
+ &minor, sizeof(minor));
+ len = sizeof(major) + sizeof(minor);
+ break;
+ case BL_DEVICE_UMOUNT:
+ if (!dm_device_remove_all((uint64_t *) buf))
+ head->status = BL_DEVICE_REQUEST_ERR;
+ bl_discover_devices();
+ break;
+ default:
+ head->status = BL_DEVICE_REQUEST_ERR;
+ }
+
+ head->totallen = sizeof(struct pipefs_hdr) + len;
+ /* write to pipefs */
+ if (atomicio((void *)write, fd, head, head->totallen)
+ != head->totallen) {
+ BL_LOG_ERR("Write pipefs error!\n");
+ ret = -EIO;
+ }
+
+ out:
+ if (buf)
+ free(buf);
+ if (head)
+ free(head);
+ return ret;
+}
+
+/* TODO: set bl_process_stop to 1 in command */
+unsigned int bl_process_stop;
+
+int bl_run_disk_inquiry_process(int fd)
+{
+ fd_set rset;
+ struct timeval tv;
+ int ret;
+
+ bl_process_stop = 0;
+
+ for (;;) {
+ if (bl_process_stop)
+ return 1;
+ FD_ZERO(&rset);
+ FD_SET(fd, &rset);
+ ret = 0;
+ tv.tv_sec = BL_DEVICE_DISCOVERY_INTERVAL;
+ switch (select(fd + 1, &rset, NULL, NULL, &tv)) {
+ case -1:
+ if (errno == EINTR)
+ continue;
+ else {
+ ret = -errno;
+ goto out;
+ }
+ case 0:
+ goto out;
+ default:
+ if (FD_ISSET(fd, &rset))
+ ret = bl_disk_inquiry_process(fd);
+ }
+ }
+ out:
+ return ret;
+}
+
+/* Daemon */
+int main(int argc, char **argv)
+{
+ int fd, opt, fg = 0, ret = 1;
+ struct stat statbuf;
+ char pidbuf[64];
+
+ while ((opt = getopt(argc, argv, "c:f")) != -1) {
+ switch (opt) {
+ case 'c':
+ conf_path = optarg;
+ break;
+ case 'f':
+ fg = 1;
+ break;
+ }
+ }
+
+ if (!stat(PID_FILE, &statbuf)) {
+ fprintf(stderr, "Pid file already existed\n");
+ return -1;
+ }
+
+ if (!fg && daemon(0, 0) != 0) {
+ fprintf(stderr, "Daemonize failed\n");
+ return -1;
+ }
+
+ openlog("blkmapd", LOG_PID, 0);
+ fd = open(PID_FILE, O_WRONLY | O_CREAT, 0644);
+ if (fd < 0) {
+ BL_LOG_ERR("Create pid file failed\n");
+ return -1;
+ }
+
+ if (lockf(fd, F_TLOCK, 0) < 0) {
+ BL_LOG_ERR("Lock pid file failed\n");
+ close(fd);
+ return -1;
+ }
+ ftruncate(fd, 0);
+ sprintf(pidbuf, "%d\n", getpid());
+ write(fd, pidbuf, strlen(pidbuf));
+
+ /* open pipe file */
+ fd = open(BL_PIPE_FILE, O_RDWR);
+ if (fd < 0) {
+ BL_LOG_ERR("open pipe file error\n");
+ return -1;
+ }
+
+ ret = bl_cfg_init();
+ if (ret < 0) {
+ if (ret == -ENOENT)
+ BL_LOG_WARNING("Config file not exist, use default\n");
+ else {
+ BL_LOG_ERR("Open/read Block pNFS config file error\n");
+ return -1;
+ }
+ }
+
+ while (1) {
+ /* discover device when needed */
+ bl_discover_devices();
+
+ ret = bl_run_disk_inquiry_process(fd);
+ if (ret < 0) {
+ /* what should we do with process error? */
+ BL_LOG_ERR("inquiry process return %d\n", ret);
+ }
+ }
+ close(fd);
+ return ret;
+}
diff -up nfs-utils-1.2.3/utils/blkmapd/device-discovery.h.orig nfs-utils-1.2.3/utils/blkmapd/device-discovery.h
--- nfs-utils-1.2.3/utils/blkmapd/device-discovery.h.orig 2010-10-04 15:48:00.073350775 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/device-discovery.h 2010-10-04 15:48:00.073350775 -0400
@@ -0,0 +1,162 @@
+/*
+ * bl-device-discovery.h
+ *
+ * Copyright (c) 2010 EMC Corporation, Haiying Tang <Tang_Haiying@emc.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+#ifndef BL_DEVICE_DISCOVERY_H
+#define BL_DEVICE_DISCOVERY_H
+
+#define BL_DEVICE_DISCOVERY_INTERVAL 60
+
+#include <stdint.h>
+#include <syslog.h>
+
+enum blk_vol_type {
+ BLOCK_VOLUME_SIMPLE = 0, /* maps to a single LU */
+ BLOCK_VOLUME_SLICE = 1, /* slice of another volume */
+ BLOCK_VOLUME_CONCAT = 2, /* concatenation of multiple volumes */
+ BLOCK_VOLUME_STRIPE = 3, /* striped across multiple volumes */
+ BLOCK_VOLUME_PSEUDO = 4,
+};
+
+/* All disk offset/lengths are stored in 512-byte sectors */
+struct bl_volume {
+ uint32_t bv_type;
+ off_t bv_size;
+ struct bl_volume **bv_vols;
+ int bv_vol_n;
+ union {
+ dev_t bv_dev; /* for BLOCK_VOLUME_SIMPLE(PSEUDO) */
+ off_t bv_stripe_unit; /* for BLOCK_VOLUME_STRIPE(CONCAT) */
+ off_t bv_offset; /* for BLOCK_VOLUME_SLICE */
+ } param;
+};
+
+struct bl_sig_comp {
+ int64_t bs_offset; /* In bytes */
+ uint32_t bs_length; /* In bytes */
+ char *bs_string;
+};
+
+/* Maximum number of signatures components in a simple volume */
+# define BLOCK_MAX_SIG_COMP 16
+
+struct bl_sig {
+ int si_num_comps;
+ struct bl_sig_comp si_comps[BLOCK_MAX_SIG_COMP];
+};
+
+/*
+ * Multipath support: ACTIVE or PSEUDO device is valid,
+ * PASSIVE is a standby for ACTIVE.
+ */
+enum bl_path_state_e {
+ BL_PATH_STATE_PASSIVE = 1,
+ BL_PATH_STATE_ACTIVE = 2,
+ BL_PATH_STATE_PSEUDO = 3,
+};
+
+struct bl_serial {
+ int len;
+ char *data;
+};
+
+struct bl_disk_path {
+ struct bl_disk_path *next;
+ char *full_path;
+ enum bl_path_state_e state;
+};
+
+struct bl_disk {
+ struct bl_disk *next;
+ struct bl_serial *serial;
+ dev_t dev;
+ off_t size;
+ struct bl_disk_path *valid_path;
+ struct bl_disk_path *paths;
+};
+
+struct bl_dev_id {
+ unsigned char type;
+ unsigned char ids;
+ unsigned char reserve;
+ unsigned char len;
+ char data[0];
+};
+
+struct pipefs_hdr {
+ uint32_t msgid;
+ uint8_t type;
+ uint8_t flags;
+ uint16_t totallen; /* length of entire message, including hdr */
+ uint32_t status;
+};
+
+#define BL_DEVICE_UMOUNT 0x0 /* Umount--delete devices */
+#define BL_DEVICE_MOUNT 0x1 /* Mount--create devices */
+#define BL_DEVICE_REQUEST_INIT 0x0 /* Start request */
+#define BL_DEVICE_REQUEST_PROC 0x1 /* User process succeeds */
+#define BL_DEVICE_REQUEST_ERR 0x2 /* User process fails */
+
+uint32_t *blk_overflow(uint32_t * p, uint32_t * end, size_t nbytes);
+
+#define BLK_READBUF(p, e, nbytes) do { \
+ p = blk_overflow(p, e, nbytes); \
+ if (!p) {\
+ goto out_err;\
+ } \
+} while (0)
+
+#define READ32(x) (x) = ntohl(*p++)
+
+#define READ64(x) do { \
+ (x) = (uint64_t)ntohl(*p++) << 32; \
+ (x) |= ntohl(*p++); \
+} while (0)
+
+#define READ_SECTOR(x) do { \
+ READ64(tmp); \
+ if (tmp & 0x1ff) { \
+ goto out_err; \
+ } \
+ (x) = tmp >> 9; \
+} while (0)
+
+extern struct bl_disk *visible_disk_list;
+uint64_t dm_device_create(struct bl_volume *vols, int num_vols);
+int dm_device_remove_all(uint64_t *dev);
+uint64_t process_deviceinfo(const char *dev_addr_buf,
+ unsigned int dev_addr_len,
+ uint32_t *major, uint32_t *minor);
+
+extern ssize_t atomicio(ssize_t(*f) (int, void *, size_t),
+ int fd, void *_s, size_t n);
+extern struct bl_serial *bldev_read_serial(int fd, const char *filename);
+extern void bldev_read_ap_state(int fd, enum bl_path_state_e *ap_state_out);
+extern int bl_discover_devices(void);
+
+#define BL_LOG_WARNING(fmt...) syslog(LOG_WARNING, fmt)
+#define BL_LOG_ERR(fmt...) syslog(LOG_ERR, fmt)
+#define BL_LOG_DEBUG(fmt...) syslog(LOG_DEBUG, fmt)
+#endif
diff -up nfs-utils-1.2.3/utils/blkmapd/device-inq.c.orig nfs-utils-1.2.3/utils/blkmapd/device-inq.c
--- nfs-utils-1.2.3/utils/blkmapd/device-inq.c.orig 2010-10-04 15:48:00.074563030 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/device-inq.c 2010-10-04 15:48:00.074563030 -0400
@@ -0,0 +1,235 @@
+/*
+ * device-inq.c: inquire SCSI device information.
+ *
+ * Copyright (c) 2010 EMC Corporation, Haiying Tang <Tang_Haiying@emc.com>
+ * All rights reserved.
+ *
+ * This program refers to "SCSI Primary Commands - 3 (SPC-3)
+ * at http://www.t10.org and sg_inq.c in sg3_utils-1.26 for
+ * Linux OS SCSI subsystem, by D. Gilbert.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <dirent.h>
+#include <ctype.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/ioctl.h>
+#include <sys/mount.h>
+#include <sys/select.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <libgen.h>
+#include <errno.h>
+#include <scsi/scsi.h>
+#include <scsi/scsi_ioctl.h>
+#include <scsi/sg.h>
+#include "device-discovery.h"
+
+#define DEF_ALLOC_LEN 255
+#define MX_ALLOC_LEN (0xc000 + 0x80)
+
+struct bl_serial *bl_create_scsi_string(int len, const char *bytes)
+{
+ struct bl_serial *s;
+ s = malloc(sizeof(*s) + len);
+ if (s) {
+ s->data = (char *)&s[1];
+ s->len = len;
+ memcpy(s->data, bytes, len);
+ }
+ return s;
+}
+
+void bl_free_scsi_string(struct bl_serial *str)
+{
+ if (str)
+ free(str);
+}
+
+#define sg_io_ok(io_hdr) \
+ ((((io_hdr).status & 0x7e) == 0) && \
+ ((io_hdr).host_status == 0) && \
+ (((io_hdr).driver_status & 0x0f) == 0))
+
+static int sg_timeout = 1 * 1000;
+
+static int bldev_inquire_page(int fd, int page, char *buffer, int len)
+{
+ unsigned char cmd[] = { INQUIRY, 0, 0, 0, 0, 0 };
+ unsigned char sense_b[28];
+ struct sg_io_hdr io_hdr;
+ if (page >= 0) {
+ cmd[1] = 1;
+ cmd[2] = page;
+ }
+ cmd[3] = (unsigned char)((len >> 8) & 0xff);
+ cmd[4] = (unsigned char)(len & 0xff);
+
+ memset(&io_hdr, 0, sizeof(struct sg_io_hdr));
+ io_hdr.interface_id = 'S';
+ io_hdr.cmd_len = sizeof(cmd);
+ io_hdr.mx_sb_len = sizeof(sense_b);
+ io_hdr.dxfer_direction = SG_DXFER_FROM_DEV;
+ io_hdr.dxfer_len = len;
+ io_hdr.dxferp = buffer;
+ io_hdr.cmdp = cmd;
+ io_hdr.sbp = sense_b;
+ io_hdr.timeout = sg_timeout;
+ if (ioctl(fd, SG_IO, &io_hdr) < 0)
+ return -1;
+
+ if (sg_io_ok(io_hdr))
+ return 0;
+ return -1;
+}
+
+int bldev_inquire_pages(int fd, int page, char **buffer)
+{
+ int status = 0;
+ char *tmp;
+ int len;
+
+ *buffer = calloc(DEF_ALLOC_LEN, sizeof(char));
+ if (!*buffer) {
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ return -ENOMEM;
+ }
+
+ status = bldev_inquire_page(fd, page, *buffer, DEF_ALLOC_LEN);
+ if (status)
+ goto out;
+
+ status = -1;
+ if ((*(*buffer + 1) & 0xff) != page)
+ goto out;
+
+ len = (*(*buffer + 2) << 8) + *(*buffer + 3) + 4;
+ if (len > MX_ALLOC_LEN) {
+ BL_LOG_ERR("SCSI response length too long: %d\n", len);
+ goto out;
+ }
+ if (len > DEF_ALLOC_LEN) {
+ tmp = realloc(*buffer, len);
+ if (!tmp) {
+ BL_LOG_ERR("%s: Out of memory!\n", __func__);
+ status = -ENOMEM;
+ goto out;
+ }
+ *buffer = tmp;
+ status = bldev_inquire_page(fd, page, *buffer, len);
+ if (status)
+ goto out;
+ }
+ status = 0;
+ out:
+ return status;
+}
+
+/* For EMC multipath devices, use VPD page (0xc0) to get status.
+ * For other devices, return ACTIVE for now
+ */
+void bldev_read_ap_state(int fd, enum bl_path_state_e *ap_state_out)
+{
+ int status = 0;
+ char *buffer;
+
+ *ap_state_out = BL_PATH_STATE_ACTIVE;
+
+ status = bldev_inquire_pages(fd, 0xc0, &buffer);
+ if (status)
+ goto out;
+
+ if (buffer[4] < 0x02)
+ *ap_state_out = BL_PATH_STATE_PASSIVE;
+ out:
+ if (buffer)
+ free(buffer);
+ return;
+}
+
+struct bl_serial *bldev_read_serial(int fd, const char *filename)
+{
+ struct bl_serial *serial_out = NULL;
+ int status = 0, pos, len;
+ char *buffer;
+ struct bl_dev_id *dev_root, *dev_id;
+ unsigned int current_id = 0;
+
+ status = bldev_inquire_pages(fd, 0x83, &buffer);
+ if (status)
+ goto out;
+
+ dev_root = (struct bl_dev_id *)buffer;
+
+ pos = 0;
+ current_id = 0;
+ len = dev_root->len;
+ while (pos < (len - sizeof(struct bl_dev_id) + sizeof(unsigned char))) {
+ dev_id = (struct bl_dev_id *)&(dev_root->data[pos]);
+ if ((dev_id->ids & 0xf) < current_id)
+ continue;
+ switch (dev_id->ids & 0xf) {
+ /* We process SCSI ID with four ID cases: 0, 1, 2 and 3.
+ * When more than one ID is available, priority is
+ * 3>2>1>0.
+ */
+ case 2: /* EUI-64 based */
+ if ((dev_id->len != 8) && (dev_id->len != 12) &&
+ (dev_id->len != 16)) {
+ BL_LOG_ERR("EUI-64 only decodes 8, "
+ "12 and 16\n");
+ break;
+ }
+ case 3: /* NAA */
+ /* TODO: NAA validity judgement too complicated,
+ * so just ingore it here.
+ */
+ if ((dev_id->type & 0xf) != 1) {
+ BL_LOG_ERR("Binary code_set expected\n");
+ break;
+ }
+ case 0: /* vendor specific */
+ case 1: /* T10 vendor identification */
+ current_id = dev_id->ids & 0xf;
+ if (serial_out)
+ bl_free_scsi_string(serial_out);
+ serial_out = bl_create_scsi_string(dev_id->len,
+ dev_id->data);
+ break;
+ default:
+ break;
+ }
+ if (current_id == 3)
+ break;
+ pos += (dev_id->len + sizeof(struct bl_dev_id) -
+ sizeof(unsigned char));
+ }
+ out:
+ if (!serial_out)
+ serial_out = bl_create_scsi_string(strlen(filename), filename);
+ if (buffer)
+ free(buffer);
+ return serial_out;
+}
diff -up nfs-utils-1.2.3/utils/blkmapd/device-process.c.orig nfs-utils-1.2.3/utils/blkmapd/device-process.c
--- nfs-utils-1.2.3/utils/blkmapd/device-process.c.orig 2010-10-04 15:48:00.075557769 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/device-process.c 2010-10-04 15:48:00.075557769 -0400
@@ -0,0 +1,394 @@
+/*
+ * device-process.c: detailed processing of device information sent
+ * from kernel.
+ *
+ * Copyright (c) 2006 The Regents of the University of Michigan.
+ * All rights reserved.
+ *
+ * Andy Adamson <andros@citi.umich.edu>
+ * Fred Isaman <iisaman@umich.edu>
+ *
+ * Copyright (c) 2010 EMC Corporation, Haiying Tang <Tang_Haiying@emc.com>
+ *
+ * Used codes in linux/fs/nfs/blocklayout/blocklayoutdev.c.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <libdevmapper.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/user.h>
+#include <fcntl.h>
+#include <errno.h>
+#include <arpa/inet.h>
+#include <linux/kdev_t.h>
+#include "device-discovery.h"
+
+uint32_t *blk_overflow(uint32_t * p, uint32_t * end, size_t nbytes)
+{
+ uint32_t *q = p + ((nbytes + 3) >> 2);
+ if (q > end || q < p)
+ return NULL;
+ return p;
+}
+
+static int decode_blk_signature(uint32_t **pp, uint32_t *end,
+ struct bl_sig *sig)
+{
+ int i, tmp;
+ uint32_t *p = *pp;
+
+ BLK_READBUF(p, end, 4);
+ READ32(sig->si_num_comps);
+ if (sig->si_num_comps == 0) {
+ BL_LOG_ERR("0 components in sig\n");
+ goto out_err;
+ }
+ if (sig->si_num_comps >= BLOCK_MAX_SIG_COMP) {
+ BL_LOG_ERR("number of sig comps %i >= BLOCK_MAX_SIG_COMP\n",
+ sig->si_num_comps);
+ goto out_err;
+ }
+ for (i = 0; i < sig->si_num_comps; i++) {
+ BLK_READBUF(p, end, 12);
+ READ64(sig->si_comps[i].bs_offset);
+ READ32(tmp);
+ sig->si_comps[i].bs_length = tmp;
+ BLK_READBUF(p, end, tmp);
+ /* Note we rely here on fact that sig is used immediately
+ * for mapping, then thrown away.
+ */
+ sig->si_comps[i].bs_string = (char *)p;
+ BL_LOG_ERR("%s: si_comps[%d]: bs_length %d, bs_string %s\n",
+ __func__, i, sig->si_comps[i].bs_length,
+ sig->si_comps[i].bs_string);
+ p += ((tmp + 3) >> 2);
+ }
+ *pp = p;
+ return 0;
+ out_err:
+ return -EIO;
+}
+
+/* Read signature from device
+ * return 0: read successfully
+ * return -1: error
+ */
+int
+read_cmp_blk_sig(const char *dev_name, struct bl_sig_comp *comp,
+ int64_t bs_offset)
+{
+ int fd, ret = -1;
+ char *sig = NULL;
+
+ fd = open(dev_name, O_RDONLY | O_LARGEFILE);
+ if (fd < 0) {
+ BL_LOG_ERR("%s could not be opened for read\n", dev_name);
+ goto error;
+ }
+
+ sig = (char *)malloc(comp->bs_length);
+ if (!sig) {
+ BL_LOG_ERR("%s: Out of memory\n", __func__);
+ goto error;
+ }
+
+ if (lseek64(fd, bs_offset, SEEK_SET) == -1) {
+ BL_LOG_ERR("File %s lseek error\n", dev_name);
+ goto error;
+ }
+
+ if (atomicio(read, fd, sig, comp->bs_length) != comp->bs_length) {
+ BL_LOG_ERR("File %s read error\n", dev_name);
+ goto error;
+ }
+
+ BL_LOG_ERR
+ ("%s: %s sig: %s, bs_string: %s, bs_length: %d, bs_offset: %lld\n",
+ __func__, dev_name, sig, comp->bs_string, comp->bs_length,
+ (long long)bs_offset);
+ ret = memcmp(sig, comp->bs_string, comp->bs_length);
+
+ error:
+ if (sig)
+ free(sig);
+ if (fd >= 0)
+ close(fd);
+ return ret;
+}
+
+/*
+ * All signatures in sig must be found on disk for verification.
+ * Returns True if sig matches, False otherwise.
+ */
+static int verify_sig(struct bl_disk *disk, struct bl_sig *sig)
+{
+ struct bl_sig_comp *comp;
+ int i, ret;
+ int64_t bs_offset;
+
+ for (i = 0; i < sig->si_num_comps; i++) {
+ comp = &sig->si_comps[i];
+ bs_offset = comp->bs_offset;
+ if (bs_offset < 0)
+ bs_offset += (((int64_t) disk->size) << 9);
+ BL_LOG_ERR("%s: bs_offset: %lld\n",
+ __func__, (long long) bs_offset);
+ ret = read_cmp_blk_sig(disk->valid_path->full_path,
+ comp, bs_offset);
+ if (ret)
+ return 0;
+ }
+ return 1;
+}
+
+/*
+ * map_sig_to_device()
+ * Given a signature, walk the list of visible disks searching for
+ * a match. Returns True if mapping was done, False otherwise.
+ *
+ * While we're at it, fill in the vol->bv_size.
+ */
+static int map_sig_to_device(struct bl_sig *sig, struct bl_volume *vol)
+{
+ int mapped = 0;
+ struct bl_disk *disk = visible_disk_list;
+ char *filepath = 0;
+ struct bl_disk *lolDisk = disk;
+
+ while (lolDisk) {
+ BL_LOG_ERR("%s: visible_disk_list: %s\n", __func__,
+ lolDisk->valid_path->full_path);
+ lolDisk = lolDisk->next;
+ }
+
+ /* scan disk list to find out match device */
+ while (disk) {
+ /* FIXME: should we use better algorithm for disk scan? */
+ mapped = verify_sig(disk, sig);
+ if (mapped) {
+ vol->param.bv_dev = disk->dev;
+ filepath = disk->valid_path->full_path;
+ vol->bv_size = disk->size;
+ break;
+ }
+ disk = disk->next;
+ }
+ return mapped;
+}
+
+/* We are given an array of XDR encoded array indices, each of which should
+ * refer to a previously decoded device. Translate into a list of pointers
+ * to the appropriate pnfs_blk_volume's.
+ */
+static int set_vol_array(uint32_t **pp, uint32_t *end,
+ struct bl_volume *vols, int working)
+{
+ int i, index;
+ uint32_t *p = *pp;
+ struct bl_volume **array = vols[working].bv_vols;
+ for (i = 0; i < vols[working].bv_vol_n; i++) {
+ BLK_READBUF(p, end, 4);
+ READ32(index);
+ if ((index < 0) || (index >= working)) {
+ BL_LOG_ERR("set_vol_array: Id %i out of range\n",
+ index);
+ goto out_err;
+ }
+ array[i] = &vols[index];
+ }
+ *pp = p;
+ return 0;
+ out_err:
+ return -EIO;
+}
+
+static uint64_t sum_subvolume_sizes(struct bl_volume *vol)
+{
+ int i;
+ uint64_t sum = 0;
+ for (i = 0; i < vol->bv_vol_n; i++)
+ sum += vol->bv_vols[i]->bv_size;
+ return sum;
+}
+
+static int decode_blk_volume(uint32_t **pp, uint32_t *end,
+ struct bl_volume *vols, int i, int *array_cnt)
+{
+ int status = 0, j;
+ struct bl_sig sig;
+ uint32_t *p = *pp;
+ struct bl_volume *vol = &vols[i];
+ uint64_t tmp, tmp_size;
+ div_t d;
+
+ BLK_READBUF(p, end, 4);
+ READ32(vol->bv_type);
+ switch (vol->bv_type) {
+ case BLOCK_VOLUME_SIMPLE:
+ *array_cnt = 0;
+ status = decode_blk_signature(&p, end, &sig);
+ if (status)
+ return status;
+ status = map_sig_to_device(&sig, vol);
+ if (!status) {
+ BL_LOG_ERR("Could not find disk for device\n");
+ return -ENXIO;
+ }
+ status = 0;
+ break;
+ case BLOCK_VOLUME_SLICE:
+ BLK_READBUF(p, end, 16);
+ READ_SECTOR(vol->param.bv_offset);
+ READ_SECTOR(vol->bv_size);
+ *array_cnt = vol->bv_vol_n = 1;
+ status = set_vol_array(&p, end, vols, i);
+ break;
+ case BLOCK_VOLUME_STRIPE:
+ BLK_READBUF(p, end, 8);
+ READ_SECTOR(vol->param.bv_stripe_unit);
+ off_t chunksize = vol->param.bv_stripe_unit;
+ if ((chunksize == 0) ||
+ ((chunksize & (chunksize - 1)) != 0) ||
+ (chunksize < (PAGE_SIZE >> 9)))
+ return -EIO;
+ BLK_READBUF(p, end, 4);
+ READ32(vol->bv_vol_n);
+ if (!vol->bv_vol_n)
+ return -EIO;
+ *array_cnt = vol->bv_vol_n;
+ status = set_vol_array(&p, end, vols, i);
+ if (status)
+ return status;
+ for (j = 1; j < vol->bv_vol_n; j++) {
+ if (vol->bv_vols[j]->bv_size !=
+ vol->bv_vols[0]->bv_size) {
+ BL_LOG_ERR("varying subvol size\n");
+ return -EIO;
+ }
+ }
+ /* Make sure total size only includes addressable areas */
+ tmp_size = vol->bv_vols[0]->bv_size;
+ d = div(tmp_size, (uint32_t) vol->param.bv_stripe_unit);
+ tmp_size = d.quot;
+ vol->bv_size = tmp_size * vol->param.bv_stripe_unit;
+ break;
+ case BLOCK_VOLUME_CONCAT:
+ BLK_READBUF(p, end, 4);
+ READ32(vol->bv_vol_n);
+ if (!vol->bv_vol_n)
+ return -EIO;
+ *array_cnt = vol->bv_vol_n;
+ status = set_vol_array(&p, end, vols, i);
+ if (status)
+ return status;
+ vol->bv_size = sum_subvolume_sizes(vol);
+ break;
+ default:
+ BL_LOG_ERR("Unknown volume type %i\n", vol->bv_type);
+ out_err:
+ return -EIO;
+ }
+ *pp = p;
+ return status;
+}
+
+uint64_t process_deviceinfo(const char *dev_addr_buf,
+ unsigned int dev_addr_len,
+ uint32_t *major, uint32_t *minor)
+{
+ int num_vols, i, status, count;
+ uint32_t *p, *end;
+ struct bl_volume *vols = NULL, **arrays = NULL, **arrays_ptr = NULL;
+ uint64_t dev = 0;
+ int tried = 0;
+
+ restart:
+ p = (uint32_t *) dev_addr_buf;
+ end = (uint32_t *) ((char *)p + dev_addr_len);
+ /* Decode block volume */
+ BLK_READBUF(p, end, 4);
+ READ32(num_vols);
+ if (num_vols <= 0) {
+ BL_LOG_WARNING("Error: number of vols: %d\n", num_vols);
+ goto out_err;
+ }
+
+ vols = (struct bl_volume *)malloc(num_vols * sizeof(struct bl_volume));
+ if (!vols) {
+ BL_LOG_ERR("%s: Out of memory\n", __func__);
+ goto out_err;
+ }
+
+ /* Each volume in vols array needs its own array. Save time by
+ * allocating them all in one large hunk. Because each volume
+ * array can only reference previous volumes, and because once
+ * a concat or stripe references a volume, it may never be
+ * referenced again, the volume arrays are guaranteed to fit
+ * in the suprisingly small space allocated.
+ */
+ arrays =
+ (struct bl_volume **)malloc(num_vols * 2 *
+ sizeof(struct bl_volume *));
+ if (!arrays) {
+ BL_LOG_ERR("%s: Out of memory\n", __func__);
+ goto out_err;
+ }
+
+ arrays_ptr = arrays;
+
+ for (i = 0; i < num_vols; i++) {
+ vols[i].bv_vols = arrays_ptr;
+ status = decode_blk_volume(&p, end, vols, i, &count);
+ if (status == -ENXIO && (tried <= 5)) {
+ sleep(1);
+ BL_LOG_DEBUG("%s: discover again!\n", __func__);
+ bl_discover_devices();
+ tried++;
+ free(vols);
+ free(arrays);
+ goto restart;
+ }
+ if (status)
+ goto out_err;
+ arrays_ptr += count;
+ }
+
+ if (p != end) {
+ BL_LOG_ERR("p is not equal to end!\n");
+ goto out_err;
+ }
+
+ dev = dm_device_create(vols, num_vols);
+ *major = MAJOR(dev);
+ *minor = MINOR(dev);
+ out_err:
+ if (vols)
+ free(vols);
+ if (arrays)
+ free(arrays);
+ return dev;
+}
diff -up nfs-utils-1.2.3/utils/blkmapd/dm-device.c.orig nfs-utils-1.2.3/utils/blkmapd/dm-device.c
--- nfs-utils-1.2.3/utils/blkmapd/dm-device.c.orig 2010-10-04 15:48:00.075557769 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/dm-device.c 2010-10-04 15:48:00.076394449 -0400
@@ -0,0 +1,509 @@
+/*
+ * dm-device.c: create or remove device via device mapper API.
+ *
+ * Copyright (c) 2010 EMC Corporation, Haiying Tang <Tang_Haiying@emc.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+#include <libdevmapper.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <errno.h>
+#include <linux/kdev_t.h>
+#include "device-discovery.h"
+
+#define DM_DEV_NAME_LEN 256
+
+#ifndef DM_MAX_TYPE_NAME
+#define DM_MAX_TYPE_NAME 16
+#endif
+
+#define DM_PARAMS_LEN 512 /* XXX: is this enough for target? */
+#define DM_DIR "/dev/mapper"
+#define DM_DIR_LEN12
+#define TYPE_HAS_DEV(type) ((type == BLOCK_VOLUME_SIMPLE) || \
+ (type == BLOCK_VOLUME_PSEUDO))
+
+struct bl_dm_table {
+ uint64_t offset;
+ uint64_t size;
+ char target_type[DM_MAX_TYPE_NAME];
+ char params[DM_PARAMS_LEN];
+ struct bl_dm_table *next;
+};
+
+struct bl_dm_tree {
+ uint64_t dev;
+ struct dm_tree *tree;
+ struct bl_dm_tree *next;
+};
+
+static inline struct bl_dm_table *bl_dm_table_alloc(void)
+{
+ return (struct bl_dm_table *)calloc(1, sizeof(struct bl_dm_table));
+}
+
+void bl_dm_table_free(struct bl_dm_table *bl_table_head)
+{
+ struct bl_dm_table *p = bl_table_head;
+ while (bl_table_head) {
+ p = bl_table_head->next;
+ free(bl_table_head);
+ bl_table_head = p;
+ }
+}
+
+void add_to_bl_dm_table(struct bl_dm_table **bl_table_head,
+ struct bl_dm_table *table)
+{
+ struct bl_dm_table *pre;
+ if (!*bl_table_head) {
+ *bl_table_head = table;
+ return;
+ }
+ pre = *bl_table_head;
+ while (pre->next)
+ pre = pre->next;
+ pre->next = table;
+ return;
+}
+
+struct bl_dm_tree *bl_tree_head;
+
+struct bl_dm_tree *find_bl_dm_tree(uint64_t dev)
+{
+ struct bl_dm_tree *p = bl_tree_head;
+ while (p) {
+ if (p->dev == dev)
+ return p;
+ p = p->next;
+ }
+ return NULL;
+}
+
+void del_from_bl_dm_tree(uint64_t dev)
+{
+ struct bl_dm_tree *pre = bl_tree_head;
+ struct bl_dm_tree *p;
+
+ p = pre;
+ while (p) {
+ if (p->dev == dev) {
+ pre->next = p->next;
+ if (p == bl_tree_head)
+ bl_tree_head = bl_tree_head->next;
+ free(p);
+ break;
+ }
+ pre = p;
+ p = pre->next;
+ }
+}
+
+void add_to_bl_dm_tree(struct bl_dm_tree *tree)
+{
+ struct bl_dm_tree *pre;
+ if (!bl_tree_head) {
+ bl_tree_head = tree;
+ return;
+ }
+ pre = bl_tree_head;
+ while (pre->next)
+ pre = pre->next;
+ pre->next = tree;
+ return;
+}
+
+/* Create device via device mapper
+ * return 0 when creation failed
+ * return dev no for created device
+ */
+uint64_t dm_single_device_create(const char *dev_name, struct bl_dm_table * p)
+{
+ struct dm_task *dmt;
+ struct dm_info dminfo;
+ int ret = 0;
+
+ dmt = dm_task_create(DM_DEVICE_CREATE);
+ if (!dmt) {
+ BL_LOG_ERR("Create dm_task for %s failed\n", dev_name);
+ return 0;
+ }
+ ret = dm_task_set_name(dmt, dev_name);
+ if (!ret)
+ goto err_out;
+
+ while (p) {
+ ret = dm_task_add_target(dmt, p->offset, p->size,
+ p->target_type, p->params);
+ if (!ret)
+ goto err_out;
+ p = p->next;
+ }
+
+ ret = dm_task_run(dmt) &&
+ dm_task_get_info(dmt, &dminfo) && dminfo.exists;
+
+ if (!ret)
+ goto err_out;
+
+ dm_task_update_nodes();
+
+ err_out:
+ dm_task_destroy(dmt);
+
+ if (!ret) {
+ BL_LOG_ERR("Create device %s failed\n", dev_name);
+ return 0;
+ }
+ return MKDEV(dminfo.major, dminfo.minor);
+}
+
+int dm_device_remove_byname(const char *dev_name)
+{
+ struct dm_task *dmt;
+ int ret = 0;
+
+ dmt = dm_task_create(DM_DEVICE_REMOVE);
+ if (!dmt)
+ return -ENODEV;
+
+ ret = dm_task_set_name(dmt, dev_name) && dm_task_run(dmt);
+
+ dm_task_update_nodes();
+
+ if (dmt)
+ dm_task_destroy(dmt);
+
+ return ret;
+}
+
+int dm_device_remove(uint64_t dev)
+{
+ struct dm_task *dmt;
+ struct dm_names *dmnames;
+ char *names = NULL;
+ int ret = -1;
+
+ /* Look for dev_name via dev, if dev_name could be transferred here,
+ we could jump to DM_DEVICE_REMOVE directly */
+ dmt = dm_task_create(DM_DEVICE_LIST);
+ if (!dmt) {
+ BL_LOG_ERR("dm_task creation failed\n");
+ return -ENODEV;
+ }
+
+ ret = dm_task_run(dmt);
+ if (!ret) {
+ BL_LOG_ERR("dm_task_run failed\n");
+ goto error;
+ }
+
+ dmnames = dm_task_get_names(dmt);
+ if (!dmnames || !dmnames->dev) {
+ BL_LOG_ERR("dm_task_get_names failed\n");
+ goto error;
+ }
+
+ do {
+ if (dmnames->dev == dev) {
+ names = dmnames->name;
+ break;
+ }
+ dmnames = (void *)dmnames + dmnames->next;
+ } while (dmnames);
+
+ if (!names) {
+ BL_LOG_ERR("Could not find device\n");
+ goto error;
+ }
+
+ dm_task_update_nodes();
+
+ error:
+ dm_task_destroy(dmt);
+
+ /* Start to remove device */
+ if (names)
+ ret = dm_device_remove_byname(names);
+ return ret;
+}
+
+static unsigned long dev_count;
+
+void dm_devicelist_remove(unsigned long start, unsigned long end)
+{
+ char dev_name[DM_DEV_NAME_LEN];
+ unsigned long count;
+
+ if ((start >= dev_count) || (end <= 1) || (start >= end - 1))
+ return;
+
+ for (count = end - 1; count > start; count--) {
+ sprintf(dev_name, "pnfs_vol_%lu", count - 1);
+ dm_device_remove_byname(dev_name);
+ }
+
+ return;
+}
+
+void bl_dm_remove_tree(uint64_t dev)
+{
+ struct bl_dm_tree *p;
+
+ p = find_bl_dm_tree(dev);
+ if (!p)
+ return;
+
+ dm_tree_free(p->tree);
+ del_from_bl_dm_tree(dev);
+}
+
+void bl_dm_create_tree(uint64_t dev)
+{
+ struct dm_tree *tree;
+ struct bl_dm_tree *bl_tree;
+
+ bl_tree = find_bl_dm_tree(dev);
+ if (bl_tree)
+ return; /* XXX: error? */
+
+ tree = dm_tree_create();
+ if (!tree)
+ return;
+
+ if (!dm_tree_add_dev(tree, MAJOR(dev), MINOR(dev))) {
+ dm_tree_free(tree);
+ return;
+ }
+
+ bl_tree = malloc(sizeof(struct bl_dm_tree));
+ if (!bl_tree) {
+ dm_tree_free(tree);
+ return;
+ }
+
+ bl_tree->dev = dev;
+ bl_tree->tree = tree;
+ bl_tree->next = NULL;
+ add_to_bl_dm_tree(bl_tree);
+
+ return;
+}
+
+uint64_t dm_device_nametodev(char *dev_name)
+{
+ struct dm_task *dmt;
+ int ret = 0;
+ struct dm_info dminfo;
+
+ dmt = dm_task_create(DM_DEVICE_INFO);
+ if (!dmt)
+ return -ENODEV;
+
+ ret = dm_task_set_name(dmt, dev_name) &&
+ dm_task_run(dmt) && dm_task_get_info(dmt, &dminfo);
+
+ if (dmt)
+ dm_task_destroy(dmt);
+
+ if (!ret)
+ return 0;
+
+ return MKDEV(dminfo.major, dminfo.minor);
+}
+
+int dm_device_remove_all(uint64_t *dev)
+{
+ struct bl_dm_tree *p;
+ struct dm_tree_node *node;
+ const char *uuid;
+ int ret = 0;
+ uint32_t major, minor;
+ uint64_t bl_dev;
+
+ memcpy(&major, dev, sizeof(uint32_t));
+ memcpy(&minor, (void *)dev + sizeof(uint32_t), sizeof(uint32_t));
+ bl_dev = MKDEV(major, minor);
+ p = find_bl_dm_tree(bl_dev);
+ if (!p)
+ return ret;
+
+ node = dm_tree_find_node(p->tree, MAJOR(bl_dev), MINOR(bl_dev));
+ if (!node)
+ return ret;
+
+ uuid = dm_tree_node_get_uuid(node);
+ if (!uuid)
+ return ret;
+
+ dm_device_remove(bl_dev);
+ ret = dm_tree_deactivate_children(node, uuid, strlen(uuid));
+ dm_task_update_nodes();
+ bl_dm_remove_tree(bl_dev);
+ return ret;
+}
+
+/* TODO: check the value for DM_DEV_NAME_LEN, DM_TYPE_LEN, DM_PARAMS_LEN */
+uint64_t dm_device_create(struct bl_volume *vols, int num_vols)
+{
+ uint64_t size, dev = 0;
+ unsigned long count = dev_count;
+ int number = 0, i, pos;
+ struct bl_volume *node;
+ char *tmp;
+ struct bl_dm_table *table = NULL;
+ struct bl_dm_table *bl_table_head = NULL;
+ unsigned int len;
+ char *dev_name = NULL;
+ /* Create pseudo device here */
+ while (number < num_vols) {
+ node = &vols[number];
+ switch (node->bv_type) {
+ case BLOCK_VOLUME_SIMPLE:
+ /* Do not need to create device here */
+ dev = node->param.bv_dev;
+ goto continued;
+ case BLOCK_VOLUME_SLICE:
+ table = bl_dm_table_alloc();
+ if (!table)
+ goto out;
+ table->offset = 0;
+ table->size = node->bv_size;
+ strcpy(table->target_type, "linear");
+ if (!TYPE_HAS_DEV(node->bv_vols[0]->bv_type)) {
+ free(table);
+ goto out;
+ }
+ dev = node->bv_vols[0]->param.bv_dev;
+ tmp = table->params;
+ if (!dm_format_dev(tmp, DM_PARAMS_LEN,
+ MAJOR(dev), MINOR(dev))) {
+ free(table);
+ goto out;
+ }
+ tmp += strlen(tmp);
+ sprintf(tmp, " %lu", node->param.bv_offset);
+ add_to_bl_dm_table(&bl_table_head, table);
+ break;
+ case BLOCK_VOLUME_STRIPE:
+ table = bl_dm_table_alloc();
+ if (!table)
+ goto out;
+ table->offset = 0;
+ table->size = node->bv_size;
+ strcpy(table->target_type, "striped");
+ sprintf(table->params, "%d %lu %n", node->bv_vol_n,
+ node->param.bv_stripe_unit, &pos);
+ /* Repeatedly copy subdev to params */
+ tmp = table->params + pos;
+ len = DM_PARAMS_LEN - pos;
+ for (i = 0; i < node->bv_vol_n; i++) {
+ if (!TYPE_HAS_DEV(node->bv_vols[i]->bv_type)) {
+ free(table);
+ goto out;
+ }
+ dev = node->bv_vols[i]->param.bv_dev;
+ if (!dm_format_dev(tmp, len, MAJOR(dev),
+ MINOR(dev))) {
+ free(table);
+ goto out;
+ }
+ pos = strlen(tmp);
+ tmp += pos;
+ len -= pos;
+ sprintf(tmp, " %d ", 0);
+ tmp += 3;
+ len -= 3;
+ }
+ add_to_bl_dm_table(&bl_table_head, table);
+ break;
+ case BLOCK_VOLUME_CONCAT:
+ size = 0;
+ for (i = 0; i < node->bv_vol_n; i++) {
+ table = bl_dm_table_alloc();
+ if (!table)
+ goto out;
+ table->offset = size;
+ table->size = node->bv_vols[i]->bv_size;
+ if (!TYPE_HAS_DEV(node->bv_vols[i]->bv_type)) {
+ free(table);
+ goto out;
+ }
+ strcpy(table->target_type, "linear");
+ tmp = table->params;
+ dev = node->bv_vols[i]->param.bv_dev;
+ if (!dm_format_dev(tmp, DM_PARAMS_LEN,
+ MAJOR(dev), MINOR(dev))) {
+ free(table);
+ goto out;
+ }
+ tmp += strlen(tmp);
+ sprintf(tmp, " %d", 0);
+ size += table->size;
+ add_to_bl_dm_table(&bl_table_head, table);
+ }
+ break;
+ default:
+ /* Delete previous temporary devices */
+ dm_devicelist_remove(count, dev_count);
+ goto out;
+ } /* end of swtich */
+ /* Create dev_name here. Name of device is pnfs_vol_XXX */
+ if (dev_name)
+ free(dev_name);
+ dev_name = (char *)calloc(DM_DEV_NAME_LEN, sizeof(char));
+ if (!dev_name) {
+ BL_LOG_ERR("%s: Out of memory\n", __func__);
+ goto out;
+ }
+ sprintf(dev_name, "pnfs_vol_%lu", dev_count++);
+
+ dev = dm_single_device_create(dev_name, bl_table_head);
+ if (!dev) {
+ /* Delete previous temporary devices */
+ dm_devicelist_remove(count, dev_count);
+ goto out;
+ }
+ node->param.bv_dev = dev;
+ /* TODO: extend use with PSEUDO later */
+ node->bv_type = BLOCK_VOLUME_PSEUDO;
+ continued:
+ number++;
+ if (bl_table_head)
+ bl_dm_table_free(bl_table_head);
+ bl_table_head = NULL;
+ }
+ out:
+ if (bl_table_head)
+ bl_dm_table_free(bl_table_head);
+ bl_table_head = NULL;
+ if (dev)
+ bl_dm_create_tree(dev);
+ if (dev_name)
+ free(dev_name);
+ return dev;
+}
diff -up nfs-utils-1.2.3/utils/blkmapd/etc/blkmapd.conf.orig nfs-utils-1.2.3/utils/blkmapd/etc/blkmapd.conf
--- nfs-utils-1.2.3/utils/blkmapd/etc/blkmapd.conf.orig 2010-10-04 15:48:00.076394449 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/etc/blkmapd.conf 2010-10-04 15:48:00.076394449 -0400
@@ -0,0 +1,10 @@
+# This is an example config file
+
+# Look at all /dev/sd* devices
+# /dev/sd or /dev/sd*
+/dev/sd*
+
+# Look at all /dev/mapper/* devices
+# /dev/mapper/* or
+# /dev/mapper/
+/dev/mapper/*
diff -up nfs-utils-1.2.3/utils/blkmapd/etc/initd/initd.redhat.orig nfs-utils-1.2.3/utils/blkmapd/etc/initd/initd.redhat
--- nfs-utils-1.2.3/utils/blkmapd/etc/initd/initd.redhat.orig 2010-10-04 15:48:00.076394449 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/etc/initd/initd.redhat 2010-10-04 15:48:00.077339793 -0400
@@ -0,0 +1,76 @@
+#!/bin/sh
+#
+# description: Starts and stops the iSCSI initiator
+#
+# processname: blkmapd
+# pidfile: /var/run/blkmapd.pid
+# config: /etc/blkmapd.conf
+
+# Source function library.
+if [ -f /etc/init.d/functions ] ; then
+ . /etc/init.d/functions
+elif [ -f /etc/rc.d/init.d/functions ] ; then
+ . /etc/rc.d/init.d/functions
+else
+ exit 0
+fi
+
+PATH=/sbin:/bin:/usr/sbin:/usr/bin
+
+RETVAL=0
+
+start()
+{
+ echo -n $"Starting pNFS block-layout device discovery service: "
+ modprobe -q blocklayoutdriver
+ daemon /usr/sbin/blkmapd
+ RETVAL=$?
+ if [ $RETVAL -eq 0 ]; then
+ touch /var/lock/subsys/blkmapd
+ fi
+ echo
+ return $RETVAL
+}
+
+stop()
+{
+ echo -n $"Stopping pNFS block-layout device discovery service: "
+ killproc blkmapd 2> /dev/null
+ rm -f /var/run/blkmapd.pid
+ RETVAL=$?
+ [ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/blkmapd
+ if [ $RETVAL -eq 0 ]; then
+ echo_success
+ else
+ echo_failure
+ fi
+ echo
+ return $RETVAL
+}
+
+restart()
+{
+ stop
+ start
+}
+
+case "$1" in
+ start)
+ start
+ ;;
+ stop)
+ stop
+ ;;
+ restart)
+ stop
+ start
+ ;;
+ status)
+ status blkmapd
+ ;;
+ *)
+ echo $"Usage: $0 {start|stop|restart|status}"
+ exit 1
+esac
+
+exit $RETVAL
diff -up nfs-utils-1.2.3/utils/blkmapd/Makefile.am.orig nfs-utils-1.2.3/utils/blkmapd/Makefile.am
--- nfs-utils-1.2.3/utils/blkmapd/Makefile.am.orig 2010-10-04 15:48:00.071418550 -0400
+++ nfs-utils-1.2.3/utils/blkmapd/Makefile.am 2010-10-04 15:48:00.071418550 -0400
@@ -0,0 +1,22 @@
+## Process this file with automake to produce Makefile.in
+
+#man8_MANS = blkmapd.man
+
+AM_CFLAGS += -D_LARGEFILE64_SOURCE
+sbin_PROGRAMS = blkmapd
+
+blkmapd_SOURCES = \
+ atomicio.c \
+ cfg.c \
+ device-discovery.c \
+ device-inq.c \
+ device-process.c \
+ dm-device.c \
+ \
+ cfg.h \
+ device-discovery.h
+
+blkmapd_LDADD = -ldevmapper ../../support/nfs/libnfs.a
+
+MAINTAINERCLEANFILES = Makefile.in
+
diff -up nfs-utils-1.2.3/utils/exportfs/exportfs.c.orig nfs-utils-1.2.3/utils/exportfs/exportfs.c
--- nfs-utils-1.2.3/utils/exportfs/exportfs.c.orig 2010-09-28 08:24:16.000000000 -0400
+++ nfs-utils-1.2.3/utils/exportfs/exportfs.c 2010-10-04 15:48:00.077339793 -0400
@@ -577,6 +577,8 @@ dump(int verbose)
#endif
}
secinfo_show(stdout, ep);
+ if (ep->e_pnfs)
+ c = dumpopt(c, ",pnfs");
printf("%c\n", (c != '(')? ')' : ' ');
}
}
diff -up nfs-utils-1.2.3/utils/Makefile.am.orig nfs-utils-1.2.3/utils/Makefile.am
--- nfs-utils-1.2.3/utils/Makefile.am.orig 2010-09-28 08:24:16.000000000 -0400
+++ nfs-utils-1.2.3/utils/Makefile.am 2010-10-04 15:48:00.070350101 -0400
@@ -4,6 +4,8 @@ OPTDIRS =
if CONFIG_NFSV4
OPTDIRS += idmapd
+OPTDIRS += spnfsd
+OPTDIRS += blkmapd
endif
if CONFIG_GSS
diff -up nfs-utils-1.2.3/utils/mountd/cache.c.orig nfs-utils-1.2.3/utils/mountd/cache.c
--- nfs-utils-1.2.3/utils/mountd/cache.c.orig 2010-09-28 08:24:16.000000000 -0400
+++ nfs-utils-1.2.3/utils/mountd/cache.c 2010-10-04 15:48:00.078357497 -0400
@@ -653,6 +653,8 @@ static int dump_to_cache(FILE *f, char *
qword_printint(f, exp->e_anonuid);
qword_printint(f, exp->e_anongid);
qword_printint(f, exp->e_fsid);
+ if (exp->e_pnfs == 1)
+ qword_print(f, "pnfs");
write_fsloc(f, exp);
write_secinfo(f, exp, flag_mask);
if (exp->e_uuid == NULL || different_fs) {
diff -up nfs-utils-1.2.3/utils/spnfsd/atomicio.c.orig nfs-utils-1.2.3/utils/spnfsd/atomicio.c
--- nfs-utils-1.2.3/utils/spnfsd/atomicio.c.orig 2010-10-04 15:48:00.079379348 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/atomicio.c 2010-10-04 15:48:00.079379348 -0400
@@ -0,0 +1,63 @@
+/*
+ * Copyright (c) 2002 Marius Aamodt Eriksen <marius@monkey.org>
+ * Copyright (c) 1995,1999 Theo de Raadt. All rights reserved.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/types.h>
+#include <unistd.h>
+#include <errno.h>
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif /* HAVE_CONFIG_H */
+
+/*
+ * ensure all of data on socket comes through. f==read || f==write
+ */
+ssize_t
+atomicio(f, fd, _s, n)
+ ssize_t (*f) ();
+ int fd;
+ void *_s;
+ size_t n;
+{
+ char *s = _s;
+ ssize_t res, pos = 0;
+
+ while (n > pos) {
+ res = (f) (fd, s + pos, n - pos);
+ switch (res) {
+ case -1:
+ if (errno == EINTR || errno == EAGAIN)
+ continue;
+ case 0:
+ if (pos != 0)
+ return (pos);
+ return (res);
+ default:
+ pos += res;
+ }
+ }
+ return (pos);
+}
diff -up nfs-utils-1.2.3/utils/spnfsd/cfg.c.orig nfs-utils-1.2.3/utils/spnfsd/cfg.c
--- nfs-utils-1.2.3/utils/spnfsd/cfg.c.orig 2010-10-04 15:48:00.080414616 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/cfg.c 2010-10-04 15:48:00.080414616 -0400
@@ -0,0 +1,893 @@
+/* $OpenBSD: conf.c,v 1.55 2003/06/03 14:28:16 ho Exp $ */
+/* $EOM: conf.c,v 1.48 2000/12/04 02:04:29 angelos Exp $ */
+
+/*
+ * Copyright (c) 1998, 1999, 2000, 2001 Niklas Hallqvist. All rights reserved.
+ * Copyright (c) 2000, 2001, 2002 Håkan Olsson. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ * This code was written under funding by Ericsson Radio Systems.
+ */
+
+#include <sys/param.h>
+#include <sys/mman.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <netinet/in.h>
+#include <arpa/inet.h>
+#include <ctype.h>
+#include <fcntl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <errno.h>
+#include <err.h>
+
+#include "cfg.h"
+
+static void conf_load_defaults (int);
+#if 0
+static int conf_find_trans_xf (int, char *);
+#endif
+
+size_t strlcpy(char *, const char *, size_t);
+
+struct conf_trans {
+ TAILQ_ENTRY (conf_trans) link;
+ int trans;
+ enum conf_op { CONF_SET, CONF_REMOVE, CONF_REMOVE_SECTION } op;
+ char *section;
+ char *tag;
+ char *value;
+ int override;
+ int is_default;
+};
+
+TAILQ_HEAD (conf_trans_head, conf_trans) conf_trans_queue;
+
+/*
+ * Radix-64 Encoding.
+ */
+const u_int8_t bin2asc[]
+ = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
+
+const u_int8_t asc2bin[] =
+{
+ 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 62, 255, 255, 255, 63,
+ 52, 53, 54, 55, 56, 57, 58, 59,
+ 60, 61, 255, 255, 255, 255, 255, 255,
+ 255, 0, 1, 2, 3, 4, 5, 6,
+ 7, 8, 9, 10, 11, 12, 13, 14,
+ 15, 16, 17, 18, 19, 20, 21, 22,
+ 23, 24, 25, 255, 255, 255, 255, 255,
+ 255, 26, 27, 28, 29, 30, 31, 32,
+ 33, 34, 35, 36, 37, 38, 39, 40,
+ 41, 42, 43, 44, 45, 46, 47, 48,
+ 49, 50, 51, 255, 255, 255, 255, 255
+};
+
+struct conf_binding {
+ LIST_ENTRY (conf_binding) link;
+ char *section;
+ char *tag;
+ char *value;
+ int is_default;
+};
+
+char *conf_path;
+LIST_HEAD (conf_bindings, conf_binding) conf_bindings[256];
+
+static char *conf_addr;
+
+static __inline__ u_int8_t
+conf_hash (char *s)
+{
+ u_int8_t hash = 0;
+
+ while (*s)
+ {
+ hash = ((hash << 1) | (hash >> 7)) ^ tolower (*s);
+ s++;
+ }
+ return hash;
+}
+
+/*
+ * Insert a tag-value combination from LINE (the equal sign is at POS)
+ */
+static int
+conf_remove_now (char *section, char *tag)
+{
+ struct conf_binding *cb, *next;
+
+ for (cb = LIST_FIRST (&conf_bindings[conf_hash (section)]); cb; cb = next)
+ {
+ next = LIST_NEXT (cb, link);
+ if (strcasecmp (cb->section, section) == 0
+ && strcasecmp (cb->tag, tag) == 0)
+ {
+ LIST_REMOVE (cb, link);
+ warnx("[%s]:%s->%s removed", section, tag, cb->value);
+ free (cb->section);
+ free (cb->tag);
+ free (cb->value);
+ free (cb);
+ return 0;
+ }
+ }
+ return 1;
+}
+
+static int
+conf_remove_section_now (char *section)
+{
+ struct conf_binding *cb, *next;
+ int unseen = 1;
+
+ for (cb = LIST_FIRST (&conf_bindings[conf_hash (section)]); cb; cb = next)
+ {
+ next = LIST_NEXT (cb, link);
+ if (strcasecmp (cb->section, section) == 0)
+ {
+ unseen = 0;
+ LIST_REMOVE (cb, link);
+ warnx("[%s]:%s->%s removed", section, cb->tag, cb->value);
+ free (cb->section);
+ free (cb->tag);
+ free (cb->value);
+ free (cb);
+ }
+ }
+ return unseen;
+}
+
+/*
+ * Insert a tag-value combination from LINE (the equal sign is at POS)
+ * into SECTION of our configuration database.
+ */
+static int
+conf_set_now (char *section, char *tag, char *value, int override,
+ int is_default)
+{
+ struct conf_binding *node = 0;
+
+ if (override)
+ conf_remove_now (section, tag);
+ else if (conf_get_str (section, tag))
+ {
+ if (!is_default)
+ warnx("conf_set: duplicate tag [%s]:%s, ignoring...\n", section, tag);
+ return 1;
+ }
+
+ node = calloc (1, sizeof *node);
+ if (!node)
+ {
+ warnx("conf_set: calloc (1, %lu) failed", (unsigned long)sizeof *node);
+ return 1;
+ }
+ node->section = strdup (section);
+ node->tag = strdup (tag);
+ node->value = strdup (value);
+ node->is_default = is_default;
+
+ LIST_INSERT_HEAD (&conf_bindings[conf_hash (section)], node, link);
+ return 0;
+}
+
+/*
+ * Parse the line LINE of SZ bytes. Skip Comments, recognize section
+ * headers and feed tag-value pairs into our configuration database.
+ */
+static void
+conf_parse_line (int trans, char *line, size_t sz)
+{
+ char *val;
+ size_t i;
+ int j;
+ static char *section = 0;
+ static int ln = 0;
+
+ ln++;
+
+ /* Lines starting with '#' or ';' are comments. */
+ if (*line == '#' || *line == ';')
+ return;
+
+ /* '[section]' parsing... */
+ if (*line == '[')
+ {
+ for (i = 1; i < sz; i++)
+ if (line[i] == ']')
+ break;
+ if (section)
+ free (section);
+ if (i == sz)
+ {
+ warnx("conf_parse_line: %d:"
+ "non-matched ']', ignoring until next section", ln);
+ section = 0;
+ return;
+ }
+ section = malloc (i);
+ if (!section)
+ {
+ warnx("conf_parse_line: %d: malloc (%lu) failed", ln,
+ (unsigned long)i);
+ return;
+ }
+ strlcpy (section, line + 1, i);
+ return;
+ }
+
+ /* Deal with assignments. */
+ for (i = 0; i < sz; i++)
+ if (line[i] == '=')
+ {
+ /* If no section, we are ignoring the lines. */
+ if (!section)
+ {
+ warnx("conf_parse_line: %d: ignoring line due to no section", ln);
+ return;
+ }
+ line[strcspn (line, " \t=")] = '\0';
+ val = line + i + 1 + strspn (line + i + 1, " \t");
+ /* Skip trailing whitespace, if any */
+ for (j = sz - (val - line) - 1; j > 0 && isspace (val[j]); j--)
+ val[j] = '\0';
+ /* XXX Perhaps should we not ignore errors? */
+ conf_set (trans, section, line, val, 0, 0);
+ return;
+ }
+
+ /* Other non-empty lines are weird. */
+ i = strspn (line, " \t");
+ if (line[i])
+ warnx("conf_parse_line: %d: syntax error", ln);
+
+ return;
+}
+
+/* Parse the mapped configuration file. */
+static void
+conf_parse (int trans, char *buf, size_t sz)
+{
+ char *cp = buf;
+ char *bufend = buf + sz;
+ char *line;
+
+ line = cp;
+ while (cp < bufend)
+ {
+ if (*cp == '\n')
+ {
+ /* Check for escaped newlines. */
+ if (cp > buf && *(cp - 1) == '\\')
+ *(cp - 1) = *cp = ' ';
+ else
+ {
+ *cp = '\0';
+ conf_parse_line (trans, line, cp - line);
+ line = cp + 1;
+ }
+ }
+ cp++;
+ }
+ if (cp != line)
+ warnx("conf_parse: last line non-terminated, ignored.");
+}
+
+static void
+conf_load_defaults (int tr)
+{
+ /* No defaults */
+ return;
+}
+
+void
+conf_init (void)
+{
+ unsigned int i;
+
+ for (i = 0; i < sizeof conf_bindings / sizeof conf_bindings[0]; i++)
+ LIST_INIT (&conf_bindings[i]);
+ TAILQ_INIT (&conf_trans_queue);
+ conf_reinit ();
+}
+
+/* Open the config file and map it into our address space, then parse it. */
+void
+conf_reinit (void)
+{
+ struct conf_binding *cb = 0;
+ int fd, trans;
+ unsigned int i;
+ size_t sz;
+ char *new_conf_addr = 0;
+ struct stat sb;
+
+ if ((stat (conf_path, &sb) == 0) || (errno != ENOENT))
+ {
+ sz = sb.st_size;
+ fd = open (conf_path, O_RDONLY, 0);
+ if (fd == -1)
+ {
+ warnx("conf_reinit: open (\"%s\", O_RDONLY) failed", conf_path);
+ return;
+ }
+
+ new_conf_addr = malloc (sz);
+ if (!new_conf_addr)
+ {
+ warnx("conf_reinit: malloc (%lu) failed", (unsigned long)sz);
+ goto fail;
+ }
+
+ /* XXX I assume short reads won't happen here. */
+ if (read (fd, new_conf_addr, sz) != (int)sz)
+ {
+ warnx("conf_reinit: read (%d, %p, %lu) failed",
+ fd, new_conf_addr, (unsigned long)sz);
+ goto fail;
+ }
+ close (fd);
+
+ trans = conf_begin ();
+
+ /* XXX Should we not care about errors and rollback? */
+ conf_parse (trans, new_conf_addr, sz);
+ }
+ else
+ trans = conf_begin ();
+
+ /* Load default configuration values. */
+ conf_load_defaults (trans);
+
+ /* Free potential existing configuration. */
+ if (conf_addr)
+ {
+ for (i = 0; i < sizeof conf_bindings / sizeof conf_bindings[0]; i++)
+ for (cb = LIST_FIRST (&conf_bindings[i]); cb;
+ cb = LIST_FIRST (&conf_bindings[i]))
+ conf_remove_now (cb->section, cb->tag);
+ free (conf_addr);
+ }
+
+ conf_end (trans, 1);
+ conf_addr = new_conf_addr;
+ return;
+
+ fail:
+ if (new_conf_addr)
+ free (new_conf_addr);
+ close (fd);
+}
+
+/*
+ * Return the numeric value denoted by TAG in section SECTION or DEF
+ * if that tag does not exist.
+ */
+int
+conf_get_num (char *section, char *tag, int def)
+{
+ char *value = conf_get_str (section, tag);
+
+ if (value)
+ return atoi (value);
+ return def;
+}
+
+/* Validate X according to the range denoted by TAG in section SECTION. */
+int
+conf_match_num (char *section, char *tag, int x)
+{
+ char *value = conf_get_str (section, tag);
+ int val, min, max, n;
+
+ if (!value)
+ return 0;
+ n = sscanf (value, "%d,%d:%d", &val, &min, &max);
+ switch (n)
+ {
+ case 1:
+ warnx("conf_match_num: %s:%s %d==%d?", section, tag, val, x);
+ return x == val;
+ case 3:
+ warnx("conf_match_num: %s:%s %d<=%d<=%d?", section, tag, min, x, max);
+ return min <= x && max >= x;
+ default:
+ warnx("conf_match_num: section %s tag %s: invalid number spec %s",
+ section, tag, value);
+ }
+ return 0;
+}
+
+/* Return the string value denoted by TAG in section SECTION. */
+char *
+conf_get_str (char *section, char *tag)
+{
+ struct conf_binding *cb;
+
+ for (cb = LIST_FIRST (&conf_bindings[conf_hash (section)]); cb;
+ cb = LIST_NEXT (cb, link))
+ if (strcasecmp (section, cb->section) == 0
+ && strcasecmp (tag, cb->tag) == 0)
+ {
+ return cb->value;
+ }
+ return 0;
+}
+
+/*
+ * Build a list of string values out of the comma separated value denoted by
+ * TAG in SECTION.
+ */
+struct conf_list *
+conf_get_list (char *section, char *tag)
+{
+ char *liststr = 0, *p, *field, *t;
+ struct conf_list *list = 0;
+ struct conf_list_node *node;
+
+ list = malloc (sizeof *list);
+ if (!list)
+ goto cleanup;
+ TAILQ_INIT (&list->fields);
+ list->cnt = 0;
+ liststr = conf_get_str (section, tag);
+ if (!liststr)
+ goto cleanup;
+ liststr = strdup (liststr);
+ if (!liststr)
+ goto cleanup;
+ p = liststr;
+ while ((field = strsep (&p, ",")) != NULL)
+ {
+ /* Skip leading whitespace */
+ while (isspace (*field))
+ field++;
+ /* Skip trailing whitespace */
+ if (p)
+ for (t = p - 1; t > field && isspace (*t); t--)
+ *t = '\0';
+ if (*field == '\0')
+ {
+ warnx("conf_get_list: empty field, ignoring...");
+ continue;
+ }
+ list->cnt++;
+ node = calloc (1, sizeof *node);
+ if (!node)
+ goto cleanup;
+ node->field = strdup (field);
+ if (!node->field) {
+ free(node);
+ goto cleanup;
+ }
+ TAILQ_INSERT_TAIL (&list->fields, node, link);
+ }
+ free (liststr);
+ return list;
+
+ cleanup:
+ if (list)
+ conf_free_list (list);
+ if (liststr)
+ free (liststr);
+ return 0;
+}
+
+struct conf_list *
+conf_get_tag_list (char *section)
+{
+ struct conf_list *list = 0;
+ struct conf_list_node *node;
+ struct conf_binding *cb;
+
+ list = malloc (sizeof *list);
+ if (!list)
+ goto cleanup;
+ TAILQ_INIT (&list->fields);
+ list->cnt = 0;
+ for (cb = LIST_FIRST (&conf_bindings[conf_hash (section)]); cb;
+ cb = LIST_NEXT (cb, link))
+ if (strcasecmp (section, cb->section) == 0)
+ {
+ list->cnt++;
+ node = calloc (1, sizeof *node);
+ if (!node)
+ goto cleanup;
+ node->field = strdup (cb->tag);
+ if (!node->field) {
+ free(node);
+ goto cleanup;
+ }
+ TAILQ_INSERT_TAIL (&list->fields, node, link);
+ }
+ return list;
+
+ cleanup:
+ if (list)
+ conf_free_list (list);
+ return 0;
+}
+
+/* Decode a PEM encoded buffer. */
+int
+conf_decode_base64 (u_int8_t *out, u_int32_t *len, u_char *buf)
+{
+ u_int32_t c = 0;
+ u_int8_t c1, c2, c3, c4;
+
+ while (*buf)
+ {
+ if (*buf > 127 || (c1 = asc2bin[*buf]) == 255)
+ return 0;
+ buf++;
+
+ if (*buf > 127 || (c2 = asc2bin[*buf]) == 255)
+ return 0;
+ buf++;
+
+ if (*buf == '=')
+ {
+ c3 = c4 = 0;
+ c++;
+
+ /* Check last four bit */
+ if (c2 & 0xF)
+ return 0;
+
+ if (strcmp ((char *)buf, "==") == 0)
+ buf++;
+ else
+ return 0;
+ }
+ else if (*buf > 127 || (c3 = asc2bin[*buf]) == 255)
+ return 0;
+ else
+ {
+ if (*++buf == '=')
+ {
+ c4 = 0;
+ c += 2;
+
+ /* Check last two bit */
+ if (c3 & 3)
+ return 0;
+
+ if (strcmp ((char *)buf, "="))
+ return 0;
+
+ }
+ else if (*buf > 127 || (c4 = asc2bin[*buf]) == 255)
+ return 0;
+ else
+ c += 3;
+ }
+
+ buf++;
+ *out++ = (c1 << 2) | (c2 >> 4);
+ *out++ = (c2 << 4) | (c3 >> 2);
+ *out++ = (c3 << 6) | c4;
+ }
+
+ *len = c;
+ return 1;
+
+}
+
+void
+conf_free_list (struct conf_list *list)
+{
+ struct conf_list_node *node = TAILQ_FIRST (&list->fields);
+
+ while (node)
+ {
+ TAILQ_REMOVE (&list->fields, node, link);
+ if (node->field)
+ free (node->field);
+ free (node);
+ node = TAILQ_FIRST (&list->fields);
+ }
+ free (list);
+}
+
+int
+conf_begin (void)
+{
+ static int seq = 0;
+
+ return ++seq;
+}
+
+static struct conf_trans *
+conf_trans_node (int transaction, enum conf_op op)
+{
+ struct conf_trans *node;
+
+ node = calloc (1, sizeof *node);
+ if (!node)
+ {
+ warnx("conf_trans_node: calloc (1, %lu) failed",
+ (unsigned long)sizeof *node);
+ return 0;
+ }
+ node->trans = transaction;
+ node->op = op;
+ TAILQ_INSERT_TAIL (&conf_trans_queue, node, link);
+ return node;
+}
+
+/* Queue a set operation. */
+int
+conf_set (int transaction, char *section, char *tag, char *value, int override,
+ int is_default)
+{
+ struct conf_trans *node;
+
+ node = conf_trans_node (transaction, CONF_SET);
+ if (!node)
+ return 1;
+ node->section = strdup (section);
+ if (!node->section)
+ {
+ warnx("conf_set: strdup (\"%s\") failed", section);
+ goto fail;
+ }
+ node->tag = strdup (tag);
+ if (!node->tag)
+ {
+ warnx("conf_set: strdup (\"%s\") failed", tag);
+ goto fail;
+ }
+ node->value = strdup (value);
+ if (!node->value)
+ {
+ warnx("conf_set: strdup (\"%s\") failed", value);
+ goto fail;
+ }
+ node->override = override;
+ node->is_default = is_default;
+ return 0;
+
+ fail:
+ if (node->tag)
+ free (node->tag);
+ if (node->section)
+ free (node->section);
+ if (node)
+ free (node);
+ return 1;
+}
+
+/* Queue a remove operation. */
+int
+conf_remove (int transaction, char *section, char *tag)
+{
+ struct conf_trans *node;
+
+ node = conf_trans_node (transaction, CONF_REMOVE);
+ if (!node)
+ goto fail;
+ node->section = strdup (section);
+ if (!node->section)
+ {
+ warnx("conf_remove: strdup (\"%s\") failed", section);
+ goto fail;
+ }
+ node->tag = strdup (tag);
+ if (!node->tag)
+ {
+ warnx("conf_remove: strdup (\"%s\") failed", tag);
+ goto fail;
+ }
+ return 0;
+
+ fail:
+ if (node && node->section)
+ free (node->section);
+ if (node)
+ free (node);
+ return 1;
+}
+
+/* Queue a remove section operation. */
+int
+conf_remove_section (int transaction, char *section)
+{
+ struct conf_trans *node;
+
+ node = conf_trans_node (transaction, CONF_REMOVE_SECTION);
+ if (!node)
+ goto fail;
+ node->section = strdup (section);
+ if (!node->section)
+ {
+ warnx("conf_remove_section: strdup (\"%s\") failed", section);
+ goto fail;
+ }
+ return 0;
+
+ fail:
+ if (node)
+ free (node);
+ return 1;
+}
+
+/* Execute all queued operations for this transaction. Cleanup. */
+int
+conf_end (int transaction, int commit)
+{
+ struct conf_trans *node, *next;
+
+ for (node = TAILQ_FIRST (&conf_trans_queue); node; node = next)
+ {
+ next = TAILQ_NEXT (node, link);
+ if (node->trans == transaction)
+ {
+ if (commit)
+ switch (node->op)
+ {
+ case CONF_SET:
+ conf_set_now (node->section, node->tag, node->value,
+ node->override, node->is_default);
+ break;
+ case CONF_REMOVE:
+ conf_remove_now (node->section, node->tag);
+ break;
+ case CONF_REMOVE_SECTION:
+ conf_remove_section_now (node->section);
+ break;
+ default:
+ warnx("conf_end: unknown operation: %d", node->op);
+ }
+ TAILQ_REMOVE (&conf_trans_queue, node, link);
+ if (node->section)
+ free (node->section);
+ if (node->tag)
+ free (node->tag);
+ if (node->value)
+ free (node->value);
+ free (node);
+ }
+ }
+ return 0;
+}
+
+/*
+ * Dump running configuration upon SIGUSR1.
+ * Configuration is "stored in reverse order", so reverse it again.
+ */
+struct dumper {
+ char *s, *v;
+ struct dumper *next;
+};
+
+static void
+conf_report_dump (struct dumper *node)
+{
+ /* Recursive, cleanup when we're done. */
+
+ if (node->next)
+ conf_report_dump (node->next);
+
+ if (node->v)
+ warnx("%s=\t%s", node->s, node->v);
+ else if (node->s)
+ {
+ warnx("%s", node->s);
+ if (strlen (node->s) > 0)
+ free (node->s);
+ }
+
+ free (node);
+}
+
+void
+conf_report (void)
+{
+ struct conf_binding *cb, *last = 0;
+ unsigned int i, len;
+ char *current_section = (char *)0;
+ struct dumper *dumper, *dnode;
+
+ dumper = dnode = (struct dumper *)calloc (1, sizeof *dumper);
+ if (!dumper)
+ goto mem_fail;
+
+ warnx("conf_report: dumping running configuration");
+
+ for (i = 0; i < sizeof conf_bindings / sizeof conf_bindings[0]; i++)
+ for (cb = LIST_FIRST (&conf_bindings[i]); cb;
+ cb = LIST_NEXT (cb, link))
+ {
+ if (!cb->is_default)
+ {
+ /* Dump this entry. */
+ if (!current_section || strcmp (cb->section, current_section))
+ {
+ if (current_section)
+ {
+ len = strlen (current_section) + 3;
+ dnode->s = malloc (len);
+ if (!dnode->s)
+ goto mem_fail;
+
+ snprintf (dnode->s, len, "[%s]", current_section);
+ dnode->next
+ = (struct dumper *)calloc (1, sizeof (struct dumper));
+ dnode = dnode->next;
+ if (!dnode)
+ goto mem_fail;
+
+ dnode->s = "";
+ dnode->next
+ = (struct dumper *)calloc (1, sizeof (struct dumper));
+ dnode = dnode->next;
+ if (!dnode)
+ goto mem_fail;
+ }
+ current_section = cb->section;
+ }
+ dnode->s = cb->tag;
+ dnode->v = cb->value;
+ dnode->next = (struct dumper *)calloc (1, sizeof (struct dumper));
+ dnode = dnode->next;
+ if (!dnode)
+ goto mem_fail;
+ last = cb;
+ }
+ }
+
+ if (last)
+ {
+ len = strlen (last->section) + 3;
+ dnode->s = malloc (len);
+ if (!dnode->s)
+ goto mem_fail;
+ snprintf (dnode->s, len, "[%s]", last->section);
+ }
+
+ conf_report_dump (dumper);
+
+ return;
+
+ mem_fail:
+ warnx("conf_report: malloc/calloc failed");
+ while ((dnode = dumper) != 0)
+ {
+ dumper = dumper->next;
+ if (dnode->s)
+ free (dnode->s);
+ free (dnode);
+ }
+ return;
+}
diff -up nfs-utils-1.2.3/utils/spnfsd/cfg.h.orig nfs-utils-1.2.3/utils/spnfsd/cfg.h
--- nfs-utils-1.2.3/utils/spnfsd/cfg.h.orig 2010-10-04 15:48:00.080414616 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/cfg.h 2010-10-04 15:48:00.080414616 -0400
@@ -0,0 +1,67 @@
+/* $OpenBSD: conf.h,v 1.30 2004/06/25 20:25:34 hshoexer Exp $ */
+/* $EOM: conf.h,v 1.13 2000/09/18 00:01:47 ho Exp $ */
+
+/*
+ * Copyright (c) 1998, 1999, 2001 Niklas Hallqvist. All rights reserved.
+ * Copyright (c) 2000, 2003 Håkan Olsson. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ * This code was written under funding by Ericsson Radio Systems.
+ */
+
+#ifndef _CONF_H_
+#define _CONF_H_
+
+#include "queue.h"
+
+struct conf_list_node {
+ TAILQ_ENTRY(conf_list_node) link;
+ char *field;
+};
+
+struct conf_list {
+ size_t cnt;
+ TAILQ_HEAD(conf_list_fields_head, conf_list_node) fields;
+};
+
+extern char *conf_path;
+
+extern int conf_begin(void);
+extern int conf_decode_base64(u_int8_t *, u_int32_t *, u_char *);
+extern int conf_end(int, int);
+extern void conf_free_list(struct conf_list *);
+extern struct sockaddr *conf_get_address(char *, char *);
+extern struct conf_list *conf_get_list(char *, char *);
+extern struct conf_list *conf_get_tag_list(char *);
+extern int conf_get_num(char *, char *, int);
+extern char *conf_get_str(char *, char *);
+extern void conf_init(void);
+extern int conf_match_num(char *, char *, int);
+extern void conf_reinit(void);
+extern int conf_remove(int, char *, char *);
+extern int conf_remove_section(int, char *);
+extern int conf_set(int, char *, char *, char *, int, int);
+extern void conf_report(void);
+
+#endif /* _CONF_H_ */
diff -up nfs-utils-1.2.3/utils/spnfsd/Makefile.am.orig nfs-utils-1.2.3/utils/spnfsd/Makefile.am
--- nfs-utils-1.2.3/utils/spnfsd/Makefile.am.orig 2010-10-04 15:48:00.078357497 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/Makefile.am 2010-10-04 15:48:00.078357497 -0400
@@ -0,0 +1,30 @@
+## Process this file with automake to produce Makefile.in
+
+man5_MANS =
+man8_MANS =
+
+RPCPREFIX =
+KPREFIX = @kprefix@
+sbin_PROGRAMS = spnfsd
+
+EXTRA_DIST = \
+ $(man5_MANS) \
+ $(man8_MANS) \
+ spnfsd.conf
+
+spnfsd_SOURCES = \
+ atomicio.c \
+ spnfsd_ops.c \
+ cfg.c \
+ spnfsd.c \
+ strlcat.c \
+ strlcpy.c \
+ \
+ cfg.h \
+ spnfsd.h \
+ spnfsd_queue.h \
+ queue.h
+
+spnfsd_LDADD = -levent ../../support/nfs/libnfs.a
+
+MAINTAINERCLEANFILES = Makefile.in
diff -up nfs-utils-1.2.3/utils/spnfsd/makestatic.orig nfs-utils-1.2.3/utils/spnfsd/makestatic
--- nfs-utils-1.2.3/utils/spnfsd/makestatic.orig 2010-10-04 15:48:00.081429548 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/makestatic 2010-10-04 15:48:00.081429548 -0400
@@ -0,0 +1,23 @@
+/******************************************************************************
+
+(c) 2007 Network Appliance, Inc. All Rights Reserved.
+
+Network Appliance provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+gcc -static -Wall -pipe -g -O2 -o spnfsd atomicio.o spnfsd_ops.o cfg.o spnfsd.o strlcat.o strlcpy.o /usr/lib/libevent.a ../../support/nfs/libnfs.a
diff -up nfs-utils-1.2.3/utils/spnfsd/nfsd4_spnfs.h.orig nfs-utils-1.2.3/utils/spnfsd/nfsd4_spnfs.h
--- nfs-utils-1.2.3/utils/spnfsd/nfsd4_spnfs.h.orig 2010-10-04 15:48:00.081429548 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/nfsd4_spnfs.h 2010-10-04 15:48:00.081429548 -0400
@@ -0,0 +1,334 @@
+/*
+ * include/linux/nfsd4_spnfs.h
+ *
+ * spNFS - simple pNFS implementation with userspace daemon
+ *
+ */
+
+/******************************************************************************
+
+(c) 2007 Network Appliance, Inc. All Rights Reserved.
+
+Network Appliance provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#ifndef NFS_SPNFS_H
+#define NFS_SPNFS_H
+
+
+#ifdef __KERNEL__
+#include "exportfs.h"
+#include "sunrpc/svc.h"
+#include "nfsd/nfsfh.h"
+#else
+#include <sys/types.h>
+#endif /* __KERNEL__ */
+
+#define SPNFS_STATUS_INVALIDMSG 0x01
+#define SPNFS_STATUS_AGAIN 0x02
+#define SPNFS_STATUS_FAIL 0x04
+#define SPNFS_STATUS_SUCCESS 0x08
+
+#define SPNFS_TYPE_LAYOUTGET 0x01
+#define SPNFS_TYPE_LAYOUTCOMMIT 0x02
+#define SPNFS_TYPE_LAYOUTRETURN 0x03
+#define SPNFS_TYPE_GETDEVICEITER 0x04
+#define SPNFS_TYPE_GETDEVICEINFO 0x05
+#define SPNFS_TYPE_SETATTR 0x06
+#define SPNFS_TYPE_OPEN 0x07
+#define SPNFS_TYPE_CLOSE 0x08
+#define SPNFS_TYPE_CREATE 0x09
+#define SPNFS_TYPE_REMOVE 0x0a
+#define SPNFS_TYPE_COMMIT 0x0b
+#define SPNFS_TYPE_READ 0x0c
+#define SPNFS_TYPE_WRITE 0x0d
+
+#define SPNFS_MAX_DEVICES 1
+#define SPNFS_MAX_DATA_SERVERS 16
+#define SPNFS_MAX_IO 512
+
+/* layout */
+struct spnfs_msg_layoutget_args {
+ unsigned long inode;
+ unsigned long generation;
+};
+
+struct spnfs_filelayout_list {
+ u_int32_t fh_len;
+ unsigned char fh_val[128]; /* DMXXX fix this const */
+};
+
+struct spnfs_msg_layoutget_res {
+ int status;
+ u_int64_t devid;
+ u_int64_t stripe_size;
+ u_int32_t stripe_type;
+ u_int32_t stripe_count;
+ struct spnfs_filelayout_list flist[SPNFS_MAX_DATA_SERVERS];
+};
+
+/* layoutcommit */
+struct spnfs_msg_layoutcommit_args {
+ unsigned long inode;
+ unsigned long generation;
+ u_int64_t file_size;
+};
+
+struct spnfs_msg_layoutcommit_res {
+ int status;
+};
+
+/* layoutreturn */
+/* No op for the daemon */
+/*
+struct spnfs_msg_layoutreturn_args {
+};
+
+struct spnfs_msg_layoutreturn_res {
+};
+*/
+
+/* getdeviceiter */
+struct spnfs_msg_getdeviceiter_args {
+ unsigned long inode;
+ u_int64_t cookie;
+ u_int64_t verf;
+};
+
+struct spnfs_msg_getdeviceiter_res {
+ int status;
+ u_int64_t devid;
+ u_int64_t cookie;
+ u_int64_t verf;
+ u_int32_t eof;
+};
+
+/* getdeviceinfo */
+struct spnfs_data_server {
+ u_int32_t dsid;
+ char netid[5];
+ char addr[29];
+};
+
+struct spnfs_device {
+ u_int64_t devid;
+ int dscount;
+ struct spnfs_data_server dslist[SPNFS_MAX_DATA_SERVERS];
+};
+
+struct spnfs_msg_getdeviceinfo_args {
+ u_int64_t devid;
+};
+
+struct spnfs_msg_getdeviceinfo_res {
+ int status;
+ struct spnfs_device devinfo;
+};
+
+/* setattr */
+struct spnfs_msg_setattr_args {
+ unsigned long inode;
+ unsigned long generation;
+ int file_size;
+};
+
+struct spnfs_msg_setattr_res {
+ int status;
+};
+
+/* open */
+struct spnfs_msg_open_args {
+ unsigned long inode;
+ unsigned long generation;
+ int create;
+ int createmode;
+ int truncate;
+};
+
+struct spnfs_msg_open_res {
+ int status;
+};
+
+/* close */
+/* No op for daemon */
+struct spnfs_msg_close_args {
+ int x;
+};
+
+struct spnfs_msg_close_res {
+ int y;
+};
+
+/* create */
+/*
+struct spnfs_msg_create_args {
+ int x;
+};
+
+struct spnfs_msg_create_res {
+ int y;
+};
+*/
+
+/* remove */
+struct spnfs_msg_remove_args {
+ unsigned long inode;
+ unsigned long generation;
+};
+
+struct spnfs_msg_remove_res {
+ int status;
+};
+
+/* commit */
+/*
+struct spnfs_msg_commit_args {
+ int x;
+};
+
+struct spnfs_msg_commit_res {
+ int y;
+};
+*/
+
+/* read */
+struct spnfs_msg_read_args {
+ unsigned long inode;
+ unsigned long generation;
+ loff_t offset;
+ unsigned long len;
+};
+
+struct spnfs_msg_read_res {
+ int status;
+ char data[SPNFS_MAX_IO];
+};
+
+/* write */
+struct spnfs_msg_write_args {
+ unsigned long inode;
+ unsigned long generation;
+ loff_t offset;
+ unsigned long len;
+ char data[SPNFS_MAX_IO];
+};
+
+struct spnfs_msg_write_res {
+ int status;
+};
+
+/* bundle args and responses */
+union spnfs_msg_args {
+ struct spnfs_msg_layoutget_args layoutget_args;
+ struct spnfs_msg_layoutcommit_args layoutcommit_args;
+/*
+ struct spnfs_msg_layoutreturn_args layoutreturn_args;
+*/
+ struct spnfs_msg_getdeviceiter_args getdeviceiter_args;
+ struct spnfs_msg_getdeviceinfo_args getdeviceinfo_args;
+ struct spnfs_msg_setattr_args setattr_args;
+ struct spnfs_msg_open_args open_args;
+ struct spnfs_msg_close_args close_args;
+/*
+ struct spnfs_msg_create_args create_args;
+*/
+ struct spnfs_msg_remove_args remove_args;
+/*
+ struct spnfs_msg_commit_args commit_args;
+*/
+ struct spnfs_msg_read_args read_args;
+ struct spnfs_msg_write_args write_args;
+};
+
+union spnfs_msg_res {
+ struct spnfs_msg_layoutget_res layoutget_res;
+ struct spnfs_msg_layoutcommit_res layoutcommit_res;
+/*
+ struct spnfs_msg_layoutreturn_res layoutreturn_res;
+*/
+ struct spnfs_msg_getdeviceiter_res getdeviceiter_res;
+ struct spnfs_msg_getdeviceinfo_res getdeviceinfo_res;
+ struct spnfs_msg_setattr_res setattr_res;
+ struct spnfs_msg_open_res open_res;
+ struct spnfs_msg_close_res close_res;
+/*
+ struct spnfs_msg_create_res create_res;
+*/
+ struct spnfs_msg_remove_res remove_res;
+/*
+ struct spnfs_msg_commit_res commit_res;
+*/
+ struct spnfs_msg_read_res read_res;
+ struct spnfs_msg_write_res write_res;
+};
+
+/* a spnfs message, args and response */
+struct spnfs_msg {
+ unsigned char im_type;
+ unsigned char im_status;
+ union spnfs_msg_args im_args;
+ union spnfs_msg_res im_res;
+};
+
+/* spnfs configuration info */
+struct spnfs_config {
+ unsigned char dense_striping;
+ int stripe_size;
+ int num_ds;
+ char ds_dir[SPNFS_MAX_DATA_SERVERS][80]; /* XXX */
+};
+
+#ifdef __KERNEL__
+
+/* pipe mgmt structure. messages flow through here */
+struct spnfs {
+ char spnfs_path[48]; /* path to pipe */
+ struct dentry *spnfs_dentry; /* dentry for pipe */
+ wait_queue_head_t spnfs_wq;
+ struct spnfs_msg spnfs_im; /* spnfs message */
+ struct mutex spnfs_lock; /* Serializes upcalls */
+ struct mutex spnfs_plock;
+};
+
+int spnfs_layout_type(void);
+int spnfs_layoutget(struct inode *, struct pnfs_layoutget_arg *);
+int spnfs_layoutcommit(void);
+int spnfs_layoutreturn(struct inode *, void *);
+int spnfs_getdeviceiter(struct super_block *, struct pnfs_deviter_arg *);
+int spnfs_getdeviceinfo(struct super_block *, struct pnfs_devinfo_arg *);
+int spnfs_setattr(void);
+int spnfs_open(struct inode *, void *);
+int spnfs_close(struct inode *);
+int spnfs_get_state(struct inode *, void *, void *);
+int spnfs_remove(unsigned long, unsigned long);
+int spnfs_read(struct inode *, loff_t, unsigned long *, int, struct svc_rqst *);
+int spnfs_write(struct inode *, loff_t, size_t, int, struct svc_rqst *);
+int spnfs_getfh(int, struct nfs_fh *);
+int spnfs_test_layoutrecall(char *);
+int spnfs_layoutrecall(struct inode *, int);
+
+int nfsd_spnfs_new(void);
+void nfsd_spnfs_delete(void);
+int spnfs_upcall(struct spnfs *, struct spnfs_msg *, union spnfs_msg_res *);
+int spnfs_enabled(void);
+int nfs4_spnfs_propagate_open(struct super_block *, struct svc_fh *, void *);
+int spnfs_init_proc(void);
+
+#endif /* __KERNEL__ */
+
+#endif /* NFS_SPNFS_H */
diff -up nfs-utils-1.2.3/utils/spnfsd/queue.h.orig nfs-utils-1.2.3/utils/spnfsd/queue.h
--- nfs-utils-1.2.3/utils/spnfsd/queue.h.orig 2010-10-04 15:48:00.082454595 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/queue.h 2010-10-04 15:48:00.082454595 -0400
@@ -0,0 +1,499 @@
+/* $OpenBSD: queue.h,v 1.22 2001/06/23 04:39:35 angelos Exp $ */
+/* $NetBSD: queue.h,v 1.11 1996/05/16 05:17:14 mycroft Exp $ */
+
+/*
+ * Copyright (c) 1991, 1993
+ * The Regents of the University of California. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. Neither the name of the University nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * @(#)queue.h 8.5 (Berkeley) 8/20/94
+ */
+
+#ifndef _SYS_QUEUE_H_
+#define _SYS_QUEUE_H_
+
+/*
+ * This file defines five types of data structures: singly-linked lists,
+ * lists, simple queues, tail queues, and circular queues.
+ *
+ *
+ * A singly-linked list is headed by a single forward pointer. The elements
+ * are singly linked for minimum space and pointer manipulation overhead at
+ * the expense of O(n) removal for arbitrary elements. New elements can be
+ * added to the list after an existing element or at the head of the list.
+ * Elements being removed from the head of the list should use the explicit
+ * macro for this purpose for optimum efficiency. A singly-linked list may
+ * only be traversed in the forward direction. Singly-linked lists are ideal
+ * for applications with large datasets and few or no removals or for
+ * implementing a LIFO queue.
+ *
+ * A list is headed by a single forward pointer (or an array of forward
+ * pointers for a hash table header). The elements are doubly linked
+ * so that an arbitrary element can be removed without a need to
+ * traverse the list. New elements can be added to the list before
+ * or after an existing element or at the head of the list. A list
+ * may only be traversed in the forward direction.
+ *
+ * A simple queue is headed by a pair of pointers, one the head of the
+ * list and the other to the tail of the list. The elements are singly
+ * linked to save space, so elements can only be removed from the
+ * head of the list. New elements can be added to the list before or after
+ * an existing element, at the head of the list, or at the end of the
+ * list. A simple queue may only be traversed in the forward direction.
+ *
+ * A tail queue is headed by a pair of pointers, one to the head of the
+ * list and the other to the tail of the list. The elements are doubly
+ * linked so that an arbitrary element can be removed without a need to
+ * traverse the list. New elements can be added to the list before or
+ * after an existing element, at the head of the list, or at the end of
+ * the list. A tail queue may be traversed in either direction.
+ *
+ * A circle queue is headed by a pair of pointers, one to the head of the
+ * list and the other to the tail of the list. The elements are doubly
+ * linked so that an arbitrary element can be removed without a need to
+ * traverse the list. New elements can be added to the list before or after
+ * an existing element, at the head of the list, or at the end of the list.
+ * A circle queue may be traversed in either direction, but has a more
+ * complex end of list detection.
+ *
+ * For details on the use of these macros, see the queue(3) manual page.
+ */
+
+/*
+ * Singly-linked List definitions.
+ */
+#define SLIST_HEAD(name, type) \
+struct name { \
+ struct type *slh_first; /* first element */ \
+}
+
+#define SLIST_HEAD_INITIALIZER(head) \
+ { NULL }
+
+#define SLIST_ENTRY(type) \
+struct { \
+ struct type *sle_next; /* next element */ \
+}
+
+/*
+ * Singly-linked List access methods.
+ */
+#define SLIST_FIRST(head) ((head)->slh_first)
+#define SLIST_END(head) NULL
+#define SLIST_EMPTY(head) (SLIST_FIRST(head) == SLIST_END(head))
+#define SLIST_NEXT(elm, field) ((elm)->field.sle_next)
+
+#define SLIST_FOREACH(var, head, field) \
+ for((var) = SLIST_FIRST(head); \
+ (var) != SLIST_END(head); \
+ (var) = SLIST_NEXT(var, field))
+
+/*
+ * Singly-linked List functions.
+ */
+#define SLIST_INIT(head) { \
+ SLIST_FIRST(head) = SLIST_END(head); \
+}
+
+#define SLIST_INSERT_AFTER(slistelm, elm, field) do { \
+ (elm)->field.sle_next = (slistelm)->field.sle_next; \
+ (slistelm)->field.sle_next = (elm); \
+} while (0)
+
+#define SLIST_INSERT_HEAD(head, elm, field) do { \
+ (elm)->field.sle_next = (head)->slh_first; \
+ (head)->slh_first = (elm); \
+} while (0)
+
+#define SLIST_REMOVE_HEAD(head, field) do { \
+ (head)->slh_first = (head)->slh_first->field.sle_next; \
+} while (0)
+
+#define SLIST_REMOVE(head, elm, type, field) do { \
+ if ((head)->slh_first == (elm)) { \
+ SLIST_REMOVE_HEAD((head), field); \
+ } \
+ else { \
+ struct type *curelm = (head)->slh_first; \
+ while( curelm->field.sle_next != (elm) ) \
+ curelm = curelm->field.sle_next; \
+ curelm->field.sle_next = \
+ curelm->field.sle_next->field.sle_next; \
+ } \
+} while (0)
+
+/*
+ * List definitions.
+ */
+#define LIST_HEAD(name, type) \
+struct name { \
+ struct type *lh_first; /* first element */ \
+}
+
+#define LIST_HEAD_INITIALIZER(head) \
+ { NULL }
+
+#define LIST_ENTRY(type) \
+struct { \
+ struct type *le_next; /* next element */ \
+ struct type **le_prev; /* address of previous next element */ \
+}
+
+/*
+ * List access methods
+ */
+#define LIST_FIRST(head) ((head)->lh_first)
+#define LIST_END(head) NULL
+#define LIST_EMPTY(head) (LIST_FIRST(head) == LIST_END(head))
+#define LIST_NEXT(elm, field) ((elm)->field.le_next)
+
+#define LIST_FOREACH(var, head, field) \
+ for((var) = LIST_FIRST(head); \
+ (var)!= LIST_END(head); \
+ (var) = LIST_NEXT(var, field))
+
+/*
+ * List functions.
+ */
+#define LIST_INIT(head) do { \
+ LIST_FIRST(head) = LIST_END(head); \
+} while (0)
+
+#define LIST_INSERT_AFTER(listelm, elm, field) do { \
+ if (((elm)->field.le_next = (listelm)->field.le_next) != NULL) \
+ (listelm)->field.le_next->field.le_prev = \
+ &(elm)->field.le_next; \
+ (listelm)->field.le_next = (elm); \
+ (elm)->field.le_prev = &(listelm)->field.le_next; \
+} while (0)
+
+#define LIST_INSERT_BEFORE(listelm, elm, field) do { \
+ (elm)->field.le_prev = (listelm)->field.le_prev; \
+ (elm)->field.le_next = (listelm); \
+ *(listelm)->field.le_prev = (elm); \
+ (listelm)->field.le_prev = &(elm)->field.le_next; \
+} while (0)
+
+#define LIST_INSERT_HEAD(head, elm, field) do { \
+ if (((elm)->field.le_next = (head)->lh_first) != NULL) \
+ (head)->lh_first->field.le_prev = &(elm)->field.le_next;\
+ (head)->lh_first = (elm); \
+ (elm)->field.le_prev = &(head)->lh_first; \
+} while (0)
+
+#define LIST_REMOVE(elm, field) do { \
+ if ((elm)->field.le_next != NULL) \
+ (elm)->field.le_next->field.le_prev = \
+ (elm)->field.le_prev; \
+ *(elm)->field.le_prev = (elm)->field.le_next; \
+} while (0)
+
+#define LIST_REPLACE(elm, elm2, field) do { \
+ if (((elm2)->field.le_next = (elm)->field.le_next) != NULL) \
+ (elm2)->field.le_next->field.le_prev = \
+ &(elm2)->field.le_next; \
+ (elm2)->field.le_prev = (elm)->field.le_prev; \
+ *(elm2)->field.le_prev = (elm2); \
+} while (0)
+
+/*
+ * Simple queue definitions.
+ */
+#define SIMPLEQ_HEAD(name, type) \
+struct name { \
+ struct type *sqh_first; /* first element */ \
+ struct type **sqh_last; /* addr of last next element */ \
+}
+
+#define SIMPLEQ_HEAD_INITIALIZER(head) \
+ { NULL, &(head).sqh_first }
+
+#define SIMPLEQ_ENTRY(type) \
+struct { \
+ struct type *sqe_next; /* next element */ \
+}
+
+/*
+ * Simple queue access methods.
+ */
+#define SIMPLEQ_FIRST(head) ((head)->sqh_first)
+#define SIMPLEQ_END(head) NULL
+#define SIMPLEQ_EMPTY(head) (SIMPLEQ_FIRST(head) == SIMPLEQ_END(head))
+#define SIMPLEQ_NEXT(elm, field) ((elm)->field.sqe_next)
+
+#define SIMPLEQ_FOREACH(var, head, field) \
+ for((var) = SIMPLEQ_FIRST(head); \
+ (var) != SIMPLEQ_END(head); \
+ (var) = SIMPLEQ_NEXT(var, field))
+
+/*
+ * Simple queue functions.
+ */
+#define SIMPLEQ_INIT(head) do { \
+ (head)->sqh_first = NULL; \
+ (head)->sqh_last = &(head)->sqh_first; \
+} while (0)
+
+#define SIMPLEQ_INSERT_HEAD(head, elm, field) do { \
+ if (((elm)->field.sqe_next = (head)->sqh_first) == NULL) \
+ (head)->sqh_last = &(elm)->field.sqe_next; \
+ (head)->sqh_first = (elm); \
+} while (0)
+
+#define SIMPLEQ_INSERT_TAIL(head, elm, field) do { \
+ (elm)->field.sqe_next = NULL; \
+ *(head)->sqh_last = (elm); \
+ (head)->sqh_last = &(elm)->field.sqe_next; \
+} while (0)
+
+#define SIMPLEQ_INSERT_AFTER(head, listelm, elm, field) do { \
+ if (((elm)->field.sqe_next = (listelm)->field.sqe_next) == NULL)\
+ (head)->sqh_last = &(elm)->field.sqe_next; \
+ (listelm)->field.sqe_next = (elm); \
+} while (0)
+
+#define SIMPLEQ_REMOVE_HEAD(head, elm, field) do { \
+ if (((head)->sqh_first = (elm)->field.sqe_next) == NULL) \
+ (head)->sqh_last = &(head)->sqh_first; \
+} while (0)
+
+/*
+ * Tail queue definitions.
+ */
+#define TAILQ_HEAD(name, type) \
+struct name { \
+ struct type *tqh_first; /* first element */ \
+ struct type **tqh_last; /* addr of last next element */ \
+}
+
+#define TAILQ_HEAD_INITIALIZER(head) \
+ { NULL, &(head).tqh_first }
+
+#define TAILQ_ENTRY(type) \
+struct { \
+ struct type *tqe_next; /* next element */ \
+ struct type **tqe_prev; /* address of previous next element */ \
+}
+
+/*
+ * tail queue access methods
+ */
+#define TAILQ_FIRST(head) ((head)->tqh_first)
+#define TAILQ_END(head) NULL
+#define TAILQ_NEXT(elm, field) ((elm)->field.tqe_next)
+#define TAILQ_LAST(head, headname) \
+ (*(((struct headname *)((head)->tqh_last))->tqh_last))
+/* XXX */
+#define TAILQ_PREV(elm, headname, field) \
+ (*(((struct headname *)((elm)->field.tqe_prev))->tqh_last))
+#define TAILQ_EMPTY(head) \
+ (TAILQ_FIRST(head) == TAILQ_END(head))
+
+#define TAILQ_FOREACH(var, head, field) \
+ for((var) = TAILQ_FIRST(head); \
+ (var) != TAILQ_END(head); \
+ (var) = TAILQ_NEXT(var, field))
+
+#define TAILQ_FOREACH_REVERSE(var, head, field, headname) \
+ for((var) = TAILQ_LAST(head, headname); \
+ (var) != TAILQ_END(head); \
+ (var) = TAILQ_PREV(var, headname, field))
+
+/*
+ * Tail queue functions.
+ */
+#define TAILQ_INIT(head) do { \
+ (head)->tqh_first = NULL; \
+ (head)->tqh_last = &(head)->tqh_first; \
+} while (0)
+
+#define TAILQ_INSERT_HEAD(head, elm, field) do { \
+ if (((elm)->field.tqe_next = (head)->tqh_first) != NULL) \
+ (head)->tqh_first->field.tqe_prev = \
+ &(elm)->field.tqe_next; \
+ else \
+ (head)->tqh_last = &(elm)->field.tqe_next; \
+ (head)->tqh_first = (elm); \
+ (elm)->field.tqe_prev = &(head)->tqh_first; \
+} while (0)
+
+#define TAILQ_INSERT_TAIL(head, elm, field) do { \
+ (elm)->field.tqe_next = NULL; \
+ (elm)->field.tqe_prev = (head)->tqh_last; \
+ *(head)->tqh_last = (elm); \
+ (head)->tqh_last = &(elm)->field.tqe_next; \
+} while (0)
+
+#define TAILQ_INSERT_AFTER(head, listelm, elm, field) do { \
+ if (((elm)->field.tqe_next = (listelm)->field.tqe_next) != NULL)\
+ (elm)->field.tqe_next->field.tqe_prev = \
+ &(elm)->field.tqe_next; \
+ else \
+ (head)->tqh_last = &(elm)->field.tqe_next; \
+ (listelm)->field.tqe_next = (elm); \
+ (elm)->field.tqe_prev = &(listelm)->field.tqe_next; \
+} while (0)
+
+#define TAILQ_INSERT_BEFORE(listelm, elm, field) do { \
+ (elm)->field.tqe_prev = (listelm)->field.tqe_prev; \
+ (elm)->field.tqe_next = (listelm); \
+ *(listelm)->field.tqe_prev = (elm); \
+ (listelm)->field.tqe_prev = &(elm)->field.tqe_next; \
+} while (0)
+
+#define TAILQ_REMOVE(head, elm, field) do { \
+ if (((elm)->field.tqe_next) != NULL) \
+ (elm)->field.tqe_next->field.tqe_prev = \
+ (elm)->field.tqe_prev; \
+ else \
+ (head)->tqh_last = (elm)->field.tqe_prev; \
+ *(elm)->field.tqe_prev = (elm)->field.tqe_next; \
+} while (0)
+
+#define TAILQ_REPLACE(head, elm, elm2, field) do { \
+ if (((elm2)->field.tqe_next = (elm)->field.tqe_next) != NULL) \
+ (elm2)->field.tqe_next->field.tqe_prev = \
+ &(elm2)->field.tqe_next; \
+ else \
+ (head)->tqh_last = &(elm2)->field.tqe_next; \
+ (elm2)->field.tqe_prev = (elm)->field.tqe_prev; \
+ *(elm2)->field.tqe_prev = (elm2); \
+} while (0)
+
+/*
+ * Circular queue definitions.
+ */
+#define CIRCLEQ_HEAD(name, type) \
+struct name { \
+ struct type *cqh_first; /* first element */ \
+ struct type *cqh_last; /* last element */ \
+}
+
+#define CIRCLEQ_HEAD_INITIALIZER(head) \
+ { CIRCLEQ_END(&head), CIRCLEQ_END(&head) }
+
+#define CIRCLEQ_ENTRY(type) \
+struct { \
+ struct type *cqe_next; /* next element */ \
+ struct type *cqe_prev; /* previous element */ \
+}
+
+/*
+ * Circular queue access methods
+ */
+#define CIRCLEQ_FIRST(head) ((head)->cqh_first)
+#define CIRCLEQ_LAST(head) ((head)->cqh_last)
+#define CIRCLEQ_END(head) ((void *)(head))
+#define CIRCLEQ_NEXT(elm, field) ((elm)->field.cqe_next)
+#define CIRCLEQ_PREV(elm, field) ((elm)->field.cqe_prev)
+#define CIRCLEQ_EMPTY(head) \
+ (CIRCLEQ_FIRST(head) == CIRCLEQ_END(head))
+
+#define CIRCLEQ_FOREACH(var, head, field) \
+ for((var) = CIRCLEQ_FIRST(head); \
+ (var) != CIRCLEQ_END(head); \
+ (var) = CIRCLEQ_NEXT(var, field))
+
+#define CIRCLEQ_FOREACH_REVERSE(var, head, field) \
+ for((var) = CIRCLEQ_LAST(head); \
+ (var) != CIRCLEQ_END(head); \
+ (var) = CIRCLEQ_PREV(var, field))
+
+/*
+ * Circular queue functions.
+ */
+#define CIRCLEQ_INIT(head) do { \
+ (head)->cqh_first = CIRCLEQ_END(head); \
+ (head)->cqh_last = CIRCLEQ_END(head); \
+} while (0)
+
+#define CIRCLEQ_INSERT_AFTER(head, listelm, elm, field) do { \
+ (elm)->field.cqe_next = (listelm)->field.cqe_next; \
+ (elm)->field.cqe_prev = (listelm); \
+ if ((listelm)->field.cqe_next == CIRCLEQ_END(head)) \
+ (head)->cqh_last = (elm); \
+ else \
+ (listelm)->field.cqe_next->field.cqe_prev = (elm); \
+ (listelm)->field.cqe_next = (elm); \
+} while (0)
+
+#define CIRCLEQ_INSERT_BEFORE(head, listelm, elm, field) do { \
+ (elm)->field.cqe_next = (listelm); \
+ (elm)->field.cqe_prev = (listelm)->field.cqe_prev; \
+ if ((listelm)->field.cqe_prev == CIRCLEQ_END(head)) \
+ (head)->cqh_first = (elm); \
+ else \
+ (listelm)->field.cqe_prev->field.cqe_next = (elm); \
+ (listelm)->field.cqe_prev = (elm); \
+} while (0)
+
+#define CIRCLEQ_INSERT_HEAD(head, elm, field) do { \
+ (elm)->field.cqe_next = (head)->cqh_first; \
+ (elm)->field.cqe_prev = CIRCLEQ_END(head); \
+ if ((head)->cqh_last == CIRCLEQ_END(head)) \
+ (head)->cqh_last = (elm); \
+ else \
+ (head)->cqh_first->field.cqe_prev = (elm); \
+ (head)->cqh_first = (elm); \
+} while (0)
+
+#define CIRCLEQ_INSERT_TAIL(head, elm, field) do { \
+ (elm)->field.cqe_next = CIRCLEQ_END(head); \
+ (elm)->field.cqe_prev = (head)->cqh_last; \
+ if ((head)->cqh_first == CIRCLEQ_END(head)) \
+ (head)->cqh_first = (elm); \
+ else \
+ (head)->cqh_last->field.cqe_next = (elm); \
+ (head)->cqh_last = (elm); \
+} while (0)
+
+#define CIRCLEQ_REMOVE(head, elm, field) do { \
+ if ((elm)->field.cqe_next == CIRCLEQ_END(head)) \
+ (head)->cqh_last = (elm)->field.cqe_prev; \
+ else \
+ (elm)->field.cqe_next->field.cqe_prev = \
+ (elm)->field.cqe_prev; \
+ if ((elm)->field.cqe_prev == CIRCLEQ_END(head)) \
+ (head)->cqh_first = (elm)->field.cqe_next; \
+ else \
+ (elm)->field.cqe_prev->field.cqe_next = \
+ (elm)->field.cqe_next; \
+} while (0)
+
+#define CIRCLEQ_REPLACE(head, elm, elm2, field) do { \
+ if (((elm2)->field.cqe_next = (elm)->field.cqe_next) == \
+ CIRCLEQ_END(head)) \
+ (head).cqh_last = (elm2); \
+ else \
+ (elm2)->field.cqe_next->field.cqe_prev = (elm2); \
+ if (((elm2)->field.cqe_prev = (elm)->field.cqe_prev) == \
+ CIRCLEQ_END(head)) \
+ (head).cqh_first = (elm2); \
+ else \
+ (elm2)->field.cqe_prev->field.cqe_next = (elm2); \
+} while (0)
+
+#endif /* !_SYS_QUEUE_H_ */
diff -up nfs-utils-1.2.3/utils/spnfsd/spnfsd.conf.orig nfs-utils-1.2.3/utils/spnfsd/spnfsd.conf
--- nfs-utils-1.2.3/utils/spnfsd/spnfsd.conf.orig 2010-10-04 15:48:00.083419957 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/spnfsd.conf 2010-10-04 15:48:00.083419957 -0400
@@ -0,0 +1,21 @@
+[General]
+
+Verbosity = 1
+Stripe-size = 8192
+Dense-striping = 0
+Pipefs-Directory = /var/lib/nfs/rpc_pipefs
+DS-Mount-Directory = /spnfs
+
+[DataServers]
+
+NumDS = 2
+
+DS1_IP = 172.16.28.134
+DS1_PORT = 2049
+DS1_ROOT = /pnfs
+DS1_ID = 1
+
+DS2_IP = 172.16.28.141
+DS2_PORT = 2049
+DS2_ROOT = /pnfs
+DS2_ID = 2
diff -up nfs-utils-1.2.3/utils/spnfsd/spnfsd.c.orig nfs-utils-1.2.3/utils/spnfsd/spnfsd.c
--- nfs-utils-1.2.3/utils/spnfsd/spnfsd.c.orig 2010-10-04 15:48:00.083419957 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/spnfsd.c 2010-10-04 15:48:00.083419957 -0400
@@ -0,0 +1,578 @@
+/*
+ * spnfsd.c
+ * Userland daemon for spNFS.
+ * Based heavily on idmapd.c
+ *
+ */
+/*
+ * idmapd.c
+ *
+ * Userland daemon for idmap.
+ *
+ * Copyright (c) 2002 The Regents of the University of Michigan.
+ * All rights reserved.
+ *
+ * Marius Aamodt Eriksen <marius@umich.edu>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. Neither the name of the University nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
+ * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+ * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#define _LARGEFILE64_SOURCE
+#define _FILE_OFFSET_BITS 64
+#include <sys/types.h>
+#include <sys/time.h>
+#include <sys/poll.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <time.h>
+
+#include <err.h>
+#include <errno.h>
+#include <event.h>
+#include <fcntl.h>
+#include <dirent.h>
+#include <unistd.h>
+#include <netdb.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <stdarg.h>
+#include <syslog.h>
+#include <pwd.h>
+#include <grp.h>
+#include <limits.h>
+#include <ctype.h>
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif /* HAVE_CONFIG_H */
+
+#include "cfg.h"
+#include "queue.h"
+#include "nfslib.h"
+
+#include "nfsd4_spnfs.h"
+#include "spnfsd.h"
+
+#ifndef PIPEFS_DIR
+#define PIPEFS_DIR "/var/lib/nfs/rpc_pipefs/"
+#endif
+
+#ifndef DSMOUNT_DIR
+#define DSMOUNT_DIR "/spnfs"
+#endif
+
+/* From Niels */
+#define CONF_SAVE(w, f) do { \
+ char *p = f; \
+ if (p != NULL) \
+ (w) = p; \
+} while (0)
+
+struct spnfs_client {
+ int sc_fd;
+ char sc_path[PATH_MAX]; /* DM: full path to spnfs pipe */
+ struct event sc_event;
+};
+
+static void spnfscb(int, short, void *);
+static int do_mounts(void);
+
+static void spnfs_msg_handler(struct spnfs_client *, struct spnfs_msg *);
+static void send_invalid_msg(int);
+
+size_t strlcat(char *, const char *, size_t);
+size_t strlcpy(char *, const char *, size_t);
+ssize_t atomicio(ssize_t (*)(), int, void *, size_t);
+void mydaemon(int, int);
+void release_parent();
+static int read_config();
+static void dump_config();
+
+int verbose = 0;
+int stripesize = DEFAULT_STRIPE_SIZE;
+int densestriping = 0;
+int num_dev = 1; /* XXX no multiple device support yet */
+int num_ds;
+struct dserver dataservers[SPNFS_MAX_DATA_SERVERS];
+char dsmountdir[PATH_MAX];
+struct spnfs_config spnfs_config;
+
+static int cache_entry_expiration = 0;
+static char pipefsdir[PATH_MAX];
+static char pipefspath[PATH_MAX];
+
+/* SPNFS */
+struct spnfs_client sc;
+
+/* Used by cfg.c */
+char *conf_path;
+
+static void
+msg_format(char *rtnbuff, int rtnbuffsize, int errval,
+ const char *fmt, va_list args)
+{
+ char buff[1024];
+ int n;
+
+ vsnprintf(buff, sizeof(buff), fmt, args);
+
+ if ((n = strlen(buff)) > 0 && buff[n-1] == '\n')
+ buff[--n] = '\0';
+
+ snprintf(rtnbuff, rtnbuffsize, "%s: %s", buff, strerror(errval));
+}
+
+void
+spnfsd_warn(const char *fmt, ...)
+{
+ int errval = errno; /* save this! */
+ char buff[1024];
+ va_list args;
+
+ va_start(args, fmt);
+ msg_format(buff, sizeof(buff), errval, fmt, args);
+ va_end(args);
+
+ syslog(LOG_WARNING, "%s", buff);
+}
+
+void
+spnfsd_warnx(const char *fmt, ...)
+{
+ va_list args;
+
+ va_start(args, fmt);
+ vsyslog(LOG_WARNING, fmt, args);
+ va_end(args);
+}
+
+void
+spnfsd_err(int eval, const char *fmt, ...)
+{
+ int errval = errno; /* save this! */
+ char buff[1024];
+ va_list args;
+
+ va_start(args, fmt);
+ msg_format(buff, sizeof(buff), errval, fmt, args);
+ va_end(args);
+
+ syslog(LOG_ERR, "%s", buff);
+ exit(eval);
+}
+
+void
+spnfsd_errx(int eval, const char *fmt, ...)
+{
+ va_list args;
+
+ va_start(args, fmt);
+ vsyslog(LOG_ERR, fmt, args);
+ va_end(args);
+ exit(eval);
+}
+
+int
+main(int argc, char **argv)
+{
+ int opt, fg = 0;
+ char *progname;
+ struct stat sb;
+ int rc, fd, cmd = 1;
+ int i;
+
+ fd = open("/proc/fs/spnfs/ctl", O_WRONLY);
+ if (fd < 0)
+ spnfsd_errx(1, "kernel init failed (%s)", strerror(errno));
+ rc = write(fd, &cmd, sizeof(int));
+ if (rc < 0 && errno != EEXIST)
+ spnfsd_errx(1, "kernel init failed (%s)", strerror(errno));
+ close(fd);
+
+ conf_path = _PATH_SPNFSDCONF;
+ strlcpy(pipefsdir, PIPEFS_DIR, sizeof(pipefsdir));
+ strlcpy(dsmountdir, DSMOUNT_DIR, sizeof(dsmountdir));
+ strlcpy(sc.sc_path, PIPEFS_DIR, sizeof(sc.sc_path));
+
+ if ((progname = strrchr(argv[0], '/')))
+ progname++;
+ else
+ progname = argv[0];
+
+ openlog(progname, LOG_PID, LOG_DAEMON);
+
+#define GETOPTSTR "vfd:p:U:G:c:CS"
+ opterr=0; /* Turn off error messages */
+ while ((opt = getopt(argc, argv, GETOPTSTR)) != -1) {
+ if (opt == 'c')
+ conf_path = optarg;
+ if (opt == '?') {
+ if (strchr(GETOPTSTR, optopt))
+ errx(1, "'-%c' option requires an argument.", optopt);
+ else
+ errx(1, "'-%c' is an invalid argument.", optopt);
+ }
+ }
+ optind = 1;
+
+ if (stat(conf_path, &sb) == -1 &&(errno == ENOENT || errno == EACCES)) {
+ warn("Skipping configuration file \"%s\"", conf_path);
+ conf_path = NULL;
+ } else {
+ conf_init();
+ if (read_config() != 0)
+ spnfsd_err(1, "Invalid config file\n");
+ if (verbose)
+ dump_config();
+ }
+
+ while ((opt = getopt(argc, argv, GETOPTSTR)) != -1)
+ switch (opt) {
+ case 'v':
+ verbose++;
+ break;
+ case 'f':
+ fg = 1;
+ break;
+ case 'p':
+ strlcpy(pipefsdir, optarg, sizeof(pipefsdir));
+ strlcpy(sc.sc_path, optarg, sizeof(sc.sc_path));
+ break;
+ default:
+ break;
+ }
+
+ strncat(pipefsdir, "/nfs", sizeof(pipefsdir));
+ strncat(sc.sc_path, "/nfs/spnfs", sizeof(sc.sc_path));
+ memcpy(pipefspath, sc.sc_path, sizeof(pipefspath));
+
+ spnfs_config.stripe_size = stripesize;
+ spnfs_config.dense_striping = densestriping;
+ spnfs_config.num_ds = num_ds;
+
+ for (i = 0; i < num_ds; i++)
+ sprintf(spnfs_config.ds_dir[i], "%s/%s",
+ dsmountdir, dataservers[i].ds_ip);
+
+ fd = open("/proc/fs/spnfs/config", O_WRONLY);
+ if (fd < 0)
+ spnfsd_errx(1, "error creating config (%s)", strerror(errno));
+ rc = write(fd, &spnfs_config, sizeof(struct spnfs_config));
+ if (rc < 0 && errno != EEXIST)
+ spnfsd_errx(1, "error writing config (%s)", strerror(errno));
+ close(fd);
+
+ signal(SIGHUP, send_invalid_msg);
+
+ if (do_mounts() != 0)
+ spnfsd_err(1, "Mounting DSs failed\n");
+
+
+/* DMXXX in case I forget -f while testing... */
+fg = 1;
+ if (!fg)
+ mydaemon(0, 0);
+
+ event_init();
+
+ if (verbose > 0)
+ spnfsd_warnx("Expiration time is %d seconds.",
+ cache_entry_expiration);
+
+ if ((sc.sc_fd = open(sc.sc_path, O_RDWR, 0)) == -1) {
+ perror("spnfsd open file");
+ } else {
+ event_set(&sc.sc_event, sc.sc_fd, EV_READ, spnfscb, &sc);
+ event_add(&sc.sc_event, NULL);
+ }
+
+ release_parent();
+
+ if (event_dispatch() < 0)
+ spnfsd_errx(1, "main: event_dispatch returns errno %d (%s)",
+ errno, strerror(errno));
+ /* NOTREACHED */
+ return 1;
+}
+
+static void
+spnfs_msg_handler(struct spnfs_client *scp, struct spnfs_msg *im)
+{
+ int err;
+
+ switch (im->im_type) {
+ case SPNFS_TYPE_LAYOUTGET:
+ err = spnfsd_layoutget(im);
+ break;
+ case SPNFS_TYPE_LAYOUTCOMMIT:
+ err = spnfsd_layoutcommit(im);
+ break;
+ case SPNFS_TYPE_LAYOUTRETURN:
+ err = spnfsd_layoutreturn(im);
+ break;
+ case SPNFS_TYPE_GETDEVICEITER:
+ err = spnfsd_getdeviceiter(im);
+ break;
+ case SPNFS_TYPE_GETDEVICEINFO:
+ err = spnfsd_getdeviceinfo(im);
+ break;
+ case SPNFS_TYPE_SETATTR:
+ err = spnfsd_setattr(im);
+ break;
+ case SPNFS_TYPE_OPEN:
+ err = spnfsd_open(im);
+ break;
+ case SPNFS_TYPE_CLOSE:
+ err = spnfsd_close(im);
+ break;
+ case SPNFS_TYPE_CREATE:
+ err = spnfsd_create(im);
+ break;
+ case SPNFS_TYPE_REMOVE:
+ err = spnfsd_remove(im);
+ break;
+ case SPNFS_TYPE_COMMIT:
+ err = spnfsd_commit(im);
+ break;
+ case SPNFS_TYPE_READ:
+ err = spnfsd_read(im);
+ break;
+ case SPNFS_TYPE_WRITE:
+ err = spnfsd_write(im);
+ break;
+ default:
+ spnfsd_warnx("spnfs_msg_handler: Invalid msg type (%d) in message",
+ im->im_type);
+ im->im_status |= SPNFS_STATUS_INVALIDMSG;
+ err = -EINVAL;
+ break;
+ }
+}
+
+static void
+spnfscb(int fd, short which, void *data)
+{
+ struct spnfs_client *scp = data;
+ struct spnfs_msg im;
+ int rval;
+
+ if (which != EV_READ)
+ goto out;
+
+ if (atomicio(read, scp->sc_fd, &im, sizeof(im)) != sizeof(im)) {
+ if (verbose > 0)
+ spnfsd_warn("spnfscb: read(%s)", scp->sc_path);
+ if (errno == EPIPE)
+ return;
+ goto out;
+ }
+
+ spnfs_msg_handler(scp, &im);
+
+ /* XXX: I don't like ignoring this error in the id->name case,
+ * but we've never returned it, and I need to check that the client
+ * can handle it gracefully before starting to return it now. */
+
+ if (im.im_status == SPNFS_STATUS_FAIL)
+ im.im_status = SPNFS_STATUS_SUCCESS;
+
+ if ((rval=atomicio(write, scp->sc_fd, &im, sizeof(im))) != sizeof(im)) {
+ spnfsd_warn("spnfscb: write(%s)", scp->sc_path);
+ }
+
+out:
+ event_add(&scp->sc_event, NULL);
+}
+
+/*
+ * mydaemon creates a pipe between the partent and child
+ * process. The parent process will wait until the
+ * child dies or writes a '1' on the pipe signaling
+ * that it started successfully.
+ */
+int pipefds[2] = { -1, -1};
+
+void
+mydaemon(int nochdir, int noclose)
+{
+ int pid, status, tempfd;
+
+ if (pipe(pipefds) < 0)
+ err(1, "mydaemon: pipe() failed: errno %d", errno);
+
+ if ((pid = fork ()) < 0)
+ err(1, "mydaemon: fork() failed: errno %d", errno);
+
+ if (pid != 0) {
+ /*
+ * Parent. Wait for status from child.
+ */
+ close(pipefds[1]);
+ if (read(pipefds[0], &status, 1) != 1)
+ exit(1);
+ exit (0);
+ }
+ /* Child. */
+ close(pipefds[0]);
+ setsid ();
+ if (nochdir == 0) {
+ if (chdir ("/") == -1)
+ err(1, "mydaemon: chdir() failed: errno %d", errno);
+ }
+
+ while (pipefds[1] <= 2) {
+ pipefds[1] = dup(pipefds[1]);
+ if (pipefds[1] < 0)
+ err(1, "mydaemon: dup() failed: errno %d", errno);
+ }
+
+ if (noclose == 0) {
+ tempfd = open("/dev/null", O_RDWR);
+ if (tempfd < 0)
+ tempfd = open("/", O_RDONLY);
+ if (tempfd >= 0) {
+ dup2(tempfd, 0);
+ dup2(tempfd, 1);
+ dup2(tempfd, 2);
+ closeall(3);
+ } else
+ closeall(0);
+ }
+
+ return;
+}
+
+void
+release_parent()
+{
+ int status;
+
+ if (pipefds[1] > 0) {
+ write(pipefds[1], &status, 1);
+ close(pipefds[1]);
+ pipefds[1] = -1;
+ }
+}
+
+static int
+read_config()
+{
+ char *xpipefsdir = NULL;
+ char *xdsmountdir = NULL;
+ int ds;
+ char ipstr[20], portstr[20], rootstr[20], dsidstr[20];
+
+ verbose = conf_get_num("General", "Verbosity", 0);
+ stripesize = conf_get_num("General", "Stripe-size",DEFAULT_STRIPE_SIZE);
+ densestriping = conf_get_num("General", "Dense-striping", 0);
+ CONF_SAVE(xpipefsdir, conf_get_str("General", "Pipefs-Directory"));
+ if (xpipefsdir != NULL)
+ strlcpy(pipefsdir, xpipefsdir, sizeof(pipefsdir));
+ CONF_SAVE(xdsmountdir, conf_get_str("General", "DS-Mount-Directory"));
+ if (xdsmountdir != NULL)
+ strlcpy(dsmountdir, xdsmountdir, sizeof(dsmountdir));
+ num_ds = conf_get_num("DataServers", "NumDS", 0);
+ if (num_ds < 1 || num_ds > SPNFS_MAX_DATA_SERVERS)
+ spnfsd_err(1, "Invalid number of data servers in config: %d\n",
+ num_ds);
+ for (ds = 1; ds <= num_ds ; ds++) {
+ sprintf(ipstr, "DS%d_IP", ds);
+ sprintf(portstr, "DS%d_PORT", ds);
+ sprintf(rootstr, "DS%d_ROOT", ds);
+ sprintf(dsidstr, "DS%d_ID", ds);
+ CONF_SAVE(dataservers[ds-1].ds_ip,
+ conf_get_str("DataServers", ipstr));
+ if (dataservers[ds-1].ds_ip == NULL)
+ spnfsd_err(1, "Missing IP for DS%d\n", ds);
+ dataservers[ds-1].ds_port =
+ conf_get_num("DataServers", portstr, DEFAULT_DS_PORT);
+ CONF_SAVE(dataservers[ds-1].ds_path,
+ conf_get_str("DataServers", rootstr));
+ if (dataservers[ds-1].ds_ip == NULL)
+ spnfsd_err(1, "Missing IP for DS%d\n", ds);
+ dataservers[ds-1].ds_id =
+ conf_get_num("DataServers", dsidstr, -1);
+ if (dataservers[ds-1].ds_id < 0)
+ spnfsd_err(1, "Missing or invalid ID for DS%d\n", ds);
+ }
+
+ return 0;
+}
+
+static void
+dump_config()
+{
+ int ds;
+
+ printf("Verbosity: %d\n", verbose);
+ printf("Stripe size: %d\n", stripesize);
+ printf("Dense striping: %d\n", densestriping);
+ printf("Number of data servers: %d\n", num_ds);
+
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ printf("DS%d IP: %s\n", ds+1, dataservers[ds].ds_ip);
+ printf("DS%d PORT: %d\n", ds+1, dataservers[ds].ds_port);
+ printf("DS%d ROOT: %s\n", ds+1, dataservers[ds].ds_path);
+ printf("DS%d DSID: %d\n", ds+1, dataservers[ds].ds_id);
+ }
+}
+
+static int
+do_mounts()
+{
+ int ds;
+ char cmd[1024];
+
+ return 0;
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ sprintf(cmd, "mkdir -p %s/%s", dsmountdir,
+ dataservers[ds].ds_ip);
+ system(cmd);
+ sprintf(cmd, "mount -t nfs4 %s:%s %s/%s",
+ dataservers[ds].ds_ip, dataservers[ds].ds_path,
+ dsmountdir, dataservers[ds].ds_ip);
+ system(cmd);
+ }
+}
+
+static void
+send_invalid_msg(int signum)
+{
+ struct spnfs_msg im;
+ int fd, rval;
+
+ im.im_status = SPNFS_STATUS_FAIL;
+
+ if ((fd = open(pipefspath, O_RDWR, 0)) == -1) {
+ perror("spnfsd open pipe");
+ } else {
+ if ((rval=atomicio(write, fd, &im, sizeof(im))) != sizeof(im)) {
+ spnfsd_warn("send_invalid_msg: write(%s)", pipefspath);
+ }
+ }
+}
diff -up nfs-utils-1.2.3/utils/spnfsd/spnfsd.h.orig nfs-utils-1.2.3/utils/spnfsd/spnfsd.h
--- nfs-utils-1.2.3/utils/spnfsd/spnfsd.h.orig 2010-10-04 15:48:00.084495938 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/spnfsd.h 2010-10-04 15:48:00.084495938 -0400
@@ -0,0 +1,70 @@
+/*
+ * spnfsd.h
+ *
+ * spNFS - simple pNFS implementation with userspace daemon
+ *
+ */
+/******************************************************************************
+
+(c) 2007 Network Appliance, Inc. All Rights Reserved.
+
+Network Appliance provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#ifndef _PATH_SPNFSDCONF
+#define _PATH_SPNFSDCONF "/etc/spnfsd.conf"
+#endif
+
+#ifndef DEFAULT_STRIPE_SIZE
+#define DEFAULT_STRIPE_SIZE 4096
+#endif
+
+#ifndef DEFAULT_DS_PORT
+#define DEFAULT_DS_PORT 2049
+#endif
+
+struct dserver {
+ char *ds_ip;
+ int ds_port;
+ char *ds_path;
+ int ds_id;
+};
+
+/* DMXXX future struct for whole config */
+struct spnfsd_config {
+ int verbose;
+ int stripesize;
+ int densestriping;
+ int num_ds;
+ struct dserver dataservers[SPNFS_MAX_DATA_SERVERS];
+};
+
+int spnfsd_layoutget(struct spnfs_msg *);
+int spnfsd_layoutcommit(struct spnfs_msg *);
+int spnfsd_layoutreturn(struct spnfs_msg *);
+int spnfsd_getdeviceiter(struct spnfs_msg *);
+int spnfsd_getdeviceinfo(struct spnfs_msg *);
+int spnfsd_setattr(struct spnfs_msg *);
+int spnfsd_open(struct spnfs_msg *);
+int spnfsd_close(struct spnfs_msg *);
+int spnfsd_create(struct spnfs_msg *);
+int spnfsd_remove(struct spnfs_msg *);
+int spnfsd_commit(struct spnfs_msg *);
+int spnfsd_read(struct spnfs_msg *);
+int spnfsd_write(struct spnfs_msg *);
+int spnfsd_getfh(char *, unsigned char *, unsigned int *);
diff -up nfs-utils-1.2.3/utils/spnfsd/spnfsd_ops.c.orig nfs-utils-1.2.3/utils/spnfsd/spnfsd_ops.c
--- nfs-utils-1.2.3/utils/spnfsd/spnfsd_ops.c.orig 2010-10-04 15:48:00.084495938 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/spnfsd_ops.c 2010-10-04 15:48:00.085512992 -0400
@@ -0,0 +1,439 @@
+/******************************************************************************
+
+(c) 2007 Network Appliance, Inc. All Rights Reserved.
+
+Network Appliance provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#define _LARGEFILE64_SOURCE
+#define _FILE_OFFSET_BITS 64
+#include "nfsd4_spnfs.h"
+#include "spnfsd.h"
+#include "nfs/nfs.h"
+#include <string.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <errno.h>
+
+
+extern int stripesize;
+extern int densestriping;
+extern int num_ds;
+extern int num_dev;
+extern char dsmountdir[];
+extern struct dserver dataservers[SPNFS_MAX_DATA_SERVERS];
+size_t strlcat(char *, const char *, size_t);
+size_t strlcpy(char *, const char *, size_t);
+
+int
+spnfsd_layoutget(struct spnfs_msg *im)
+{
+ int ds;
+ int rc;
+ char fullpath[1024]; /* MSXXX */
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.layoutget_res.status = 0;
+ im->im_res.layoutget_res.devid = 1; /* XXX */
+ im->im_res.layoutget_res.stripe_size = stripesize;
+ if (densestriping)
+ im->im_res.layoutget_res.stripe_type = 1; /* DMXXX enum */
+ else
+ im->im_res.layoutget_res.stripe_type = 0; /* DMXXX ^^^^ */
+ im->im_res.layoutget_res.stripe_count = num_ds;
+
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ memset(im->im_res.layoutget_res.flist[ds].fh_val, 0, 128); /*DMXXX*/
+ sprintf(fullpath, "%s/%s/%lu.%lu",
+ dsmountdir, dataservers[ds].ds_ip,
+ im->im_args.layoutget_args.inode,
+ im->im_args.layoutget_args.generation);
+ rc = spnfsd_getfh(fullpath,
+ im->im_res.layoutget_res.flist[ds].fh_val,
+ &im->im_res.layoutget_res.flist[ds].fh_len);
+ if (rc < 0) {
+ /* MSXXX needs error msg/handling */
+ im->im_res.layoutget_res.status = ENOENT;
+ return -1;
+ }
+
+ /*
+ * MSXXX another hack...fix the hardcoding.
+ * The fh's fsid_type is incremented by 8 to get
+ * around stateid checking.
+ */
+ im->im_res.layoutget_res.flist[ds].fh_val[2] += 8;
+ }
+
+ return 0;
+}
+
+int
+spnfsd_layoutcommit(struct spnfs_msg *im)
+{
+ char basename[1024]; /* DMXXX */
+ char fullpath[1024]; /* DMXXX */
+ int ds;
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.layoutcommit_res.status = 0;
+ sprintf(basename, "%lu.%lu", im->im_args.layoutcommit_args.inode,
+ im->im_args.layoutcommit_args.generation);
+
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ sprintf(fullpath, "%s/%s/%s", dsmountdir,
+ dataservers[ds].ds_ip, basename);
+ truncate(fullpath, im->im_args.layoutcommit_args.file_size);
+ }
+
+ return 0;
+}
+
+int
+spnfsd_layoutreturn(struct spnfs_msg *im)
+{
+ return 0;
+}
+
+int
+spnfsd_getdeviceiter(struct spnfs_msg *im)
+{
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.getdeviceiter_res.status = 0;
+
+ /* verifier ignored for now */
+ if (im->im_args.getdeviceiter_args.cookie >= num_dev)
+ im->im_res.getdeviceiter_res.eof = 1;
+ else {
+ /* XXX just hardcoded for now...fix this */
+ im->im_res.getdeviceiter_res.devid = 1;
+ im->im_res.getdeviceiter_res.cookie = im->im_args.getdeviceiter_args.cookie + 1;
+ im->im_res.getdeviceiter_res.verf = 0;
+ im->im_res.getdeviceiter_res.eof = 0;
+ }
+
+ return 0;
+}
+
+int
+spnfsd_getdeviceinfo(struct spnfs_msg *im)
+{
+ struct spnfs_device *devp;
+ struct spnfs_data_server *dsp;
+ u_int32_t devid;
+ int ds;
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.getdeviceinfo_res.status = 0;
+
+ devid = im->im_args.getdeviceinfo_args.devid;
+
+ /* XXX fix this if/when we support multiple devices */
+ if (devid != 1) {
+ im->im_res.getdeviceinfo_res.status = -ENOENT;
+ return -1;
+ }
+
+ devp = &im->im_res.getdeviceinfo_res.devinfo;
+ devp->dscount = num_ds;
+
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ dsp = &devp->dslist[ds];
+ dsp->dsid = dataservers[ds].ds_id;
+ memset(dsp->netid, 0, 5);
+ strlcpy(dsp->netid, "tcp", 4);
+ sprintf(dsp->addr, "%s.%d.%d",
+ dataservers[ds].ds_ip,
+ dataservers[ds].ds_port >> 8,
+ dataservers[ds].ds_port & 0xff);
+ }
+
+ return 0;
+}
+
+int
+spnfsd_setattr(struct spnfs_msg *im)
+{
+ char basename[1024]; /* DMXXX */
+ char fullpath[1024]; /* DMXXX */
+ int ds;
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.setattr_res.status = 0;
+ sprintf(basename, "%lu.%lu", im->im_args.setattr_args.inode,
+ im->im_args.setattr_args.generation);
+
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ sprintf(fullpath, "%s/%s/%s", dsmountdir,
+ dataservers[ds].ds_ip, basename);
+ truncate(fullpath, im->im_args.setattr_args.file_size);
+ }
+
+ return 0;
+}
+
+int
+spnfsd_open(struct spnfs_msg *im)
+{
+ char basename[1024]; /* DMXXX */
+ char fullpath[1024]; /* DMXXX */
+ char dirpath[1024]; /* DMXXX */
+ int ds;
+ int fd, status;
+ struct stat buf;
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.open_res.status = 0;
+ sprintf(basename, "%lu.%lu", im->im_args.open_args.inode,
+ im->im_args.open_args.generation);
+
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ sprintf(dirpath, "%s/%s", dsmountdir,
+ dataservers[ds].ds_ip);
+ sprintf(fullpath, "%s/%s", dirpath, basename);
+ status = stat(dirpath, &buf);
+ im->im_res.open_res.status = errno;
+ if (status != 0) {
+ perror(dirpath);
+ break;
+ }
+ if ((fd = open(fullpath, O_WRONLY|O_CREAT, 0777)) < 0) {
+ perror(fullpath);
+ im->im_res.open_res.status = errno;
+ break;
+ } else {
+ if (fchmod(fd, 0777) != 0) {
+ /* DM: we'll consider this non-fatal for now */
+ perror("chmod stripe");
+ }
+ im->im_res.open_res.status = 0;
+ close(fd);
+ }
+ }
+
+ return im->im_res.open_res.status;
+}
+
+int
+spnfsd_close(struct spnfs_msg *im)
+{
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ printf("spnfsd_close received: %d\n", im->im_args.close_args.x);
+ im->im_res.close_res.y = 7331;
+ return 0;
+}
+
+int
+spnfsd_create(struct spnfs_msg *im)
+{
+ return 0;
+}
+
+int
+spnfsd_remove(struct spnfs_msg *im)
+{
+ char basename[1024]; /* DMXXX */
+ char fullpath[1024]; /* DMXXX */
+ int ds;
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.remove_res.status = 0;
+ sprintf(basename, "%lu.%lu", im->im_args.remove_args.inode,
+ im->im_args.remove_args.generation);
+
+ for (ds = 0 ; ds < num_ds ; ds++) {
+ sprintf(fullpath, "%s/%s/%s", dsmountdir,
+ dataservers[ds].ds_ip, basename);
+ unlink(fullpath);
+ }
+
+ return 0;
+}
+
+int
+spnfsd_commit(struct spnfs_msg *im)
+{
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ return 0;
+}
+
+int
+min (unsigned int x, unsigned int y)
+{
+ if (x<y) return x; else return y;
+}
+
+/* DMXXX: for read and write, there's an issue with partially completed i/o */
+int
+spnfsd_read(struct spnfs_msg *im)
+{
+ unsigned long inode = im->im_args.read_args.inode;
+ unsigned long generation = im->im_args.read_args.generation;
+ loff_t offset = im->im_args.read_args.offset;
+ unsigned long len = im->im_args.read_args.len;
+ int ds, iolen;
+ loff_t soffset;
+ int bufoffset = 0;
+ char fullpath[1024]; /* DMXXX */
+ int fd, err;
+ int completed = 0;
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.read_res.status = 0;
+ if (len > SPNFS_MAX_IO) {
+ im->im_res.read_res.status = -1;
+ return 0;
+ }
+ while (len > 0) {
+ ds = (offset / stripesize) % num_ds;
+ if (densestriping == 0)
+ soffset = offset;
+ else
+ soffset = (offset / num_ds) + (offset % stripesize);
+ iolen = min(len, stripesize - (offset % stripesize));
+
+ sprintf(fullpath, "%s/%s/%lu.%lu", dsmountdir,
+ dataservers[ds].ds_ip, inode, generation);
+ fd = open(fullpath, O_RDONLY);
+ if (fd < 0) {
+ perror(fullpath);
+ im->im_res.read_res.status = -errno;
+ return 0; /* DMXXX */
+ }
+ /* DM: add some error checking */
+ lseek64(fd, offset, SEEK_SET);
+ err = read(fd,
+ (void *)(im->im_res.read_res.data+bufoffset), iolen);
+ close(fd);
+ if (err < 0) {
+ perror("read");
+ im->im_res.read_res.status = -errno;
+ return 0; /* DMXXX */
+ }
+
+ if (err == 0)
+ break;
+ iolen = err; /* number of bytes read */
+ completed += iolen;
+ len -= iolen;
+ offset += iolen;
+ bufoffset += iolen;
+ }
+ im->im_res.read_res.status = completed;
+
+ return 0;
+}
+
+int
+spnfsd_write(struct spnfs_msg *im)
+{
+ unsigned long inode = im->im_args.write_args.inode;
+ unsigned long generation = im->im_args.write_args.generation;
+ loff_t offset = im->im_args.write_args.offset;
+ size_t len = im->im_args.write_args.len;
+ char *wbuf = im->im_args.write_args.data;
+ int ds, iolen;
+ loff_t soffset;
+ int bufoffset = 0;
+ char fullpath[1024]; /* DMXXX */
+ int fd, err;
+ int completed = 0;
+
+ im->im_status = SPNFS_STATUS_SUCCESS;
+ im->im_res.write_res.status = 0;
+ if (len > SPNFS_MAX_IO) {
+ printf("write length > SPNFS_MAX_IO\n");
+ im->im_res.write_res.status = -1;
+ return 0;
+ }
+ while (len > 0) {
+ ds = (offset / stripesize) % num_ds;
+ if (densestriping == 0)
+ soffset = offset;
+ else
+ soffset = (offset / num_ds) + (offset % stripesize);
+ iolen = min(len, stripesize - (offset % stripesize));
+
+ sprintf(fullpath, "%s/%s/%lu.%lu", dsmountdir,
+ dataservers[ds].ds_ip, inode, generation);
+ fd = open(fullpath, O_WRONLY);
+ if (fd < 0) {
+ perror(fullpath);
+ im->im_res.write_res.status = -errno;
+ return 0; /* DMXXX */
+ }
+ /* DM: add some error checking */
+ lseek64(fd, offset, SEEK_SET);
+ err = write(fd, (void *)(wbuf+bufoffset), iolen);
+ close(fd);
+ if (err < 0) {
+ perror("write");
+ im->im_res.write_res.status = -errno;
+ return 0; /* DMXXX */
+ }
+
+ iolen = err; /* number of bytes read */
+ completed += iolen;
+ len -= iolen;
+ offset += iolen;
+ bufoffset += iolen;
+ }
+ im->im_res.write_res.status = completed;
+ return 0;
+}
+
+int
+spnfsd_getfh(char *path, unsigned char *fh_val, unsigned int *fh_len)
+{
+ int fd, proc_fd;
+ unsigned char res[130]; /* XXX align this to proper structure */
+ extern void spnfsd_errx(int, const char *, ...);
+
+ /*
+ * hack to make sure there's an active data structure for this
+ * inode in the nfs client kernel space
+ */
+ if ((fd = open(path, O_RDONLY)) < 0) {
+ perror(path);
+ return -1;
+ }
+
+ /* fd to fh conversion */
+ proc_fd = open("/proc/fs/spnfs/getfh", O_RDWR);
+ if (proc_fd < 0) {
+ spnfsd_errx(1, "open getfh failed (%s)", strerror(errno));
+ }
+ if (write(proc_fd, &fd, sizeof(int)) < 0) {
+ spnfsd_errx(1, "write getfh failed (%s)", strerror(errno));
+ }
+ if (read(proc_fd, res, 130) <= 0) {
+ spnfsd_errx(1, "read getfh failed (%s)", strerror(errno));
+ }
+ close(proc_fd);
+
+ /* XXX use proper structure */
+ *fh_len = (short)res[0];
+ memcpy(fh_val, &res[2], *fh_len);
+
+ close(fd);
+ return 0;
+}
diff -up nfs-utils-1.2.3/utils/spnfsd/strlcat.c.orig nfs-utils-1.2.3/utils/spnfsd/strlcat.c
--- nfs-utils-1.2.3/utils/spnfsd/strlcat.c.orig 2010-10-04 15:48:00.085512992 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/strlcat.c 2010-10-04 15:48:00.085512992 -0400
@@ -0,0 +1,77 @@
+/* $OpenBSD: strlcat.c,v 1.8 2001/05/13 15:40:15 deraadt Exp $ */
+
+/*
+ * Copyright (c) 1998 Todd C. Miller <Todd.Miller@courtesan.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. The name of the author may not be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
+ * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
+ * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+ * THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+ * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+ * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
+ * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
+ * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
+ * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#if defined(LIBC_SCCS) && !defined(lint)
+static char *rcsid = "$OpenBSD: strlcat.c,v 1.8 2001/05/13 15:40:15 deraadt Exp $";
+#endif /* LIBC_SCCS and not lint */
+
+#include <sys/types.h>
+#include <string.h>
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif /* HAVE_CONFIG_H */
+
+/*
+ * Appends src to string dst of size siz (unlike strncat, siz is the
+ * full size of dst, not space left). At most siz-1 characters
+ * will be copied. Always NUL terminates (unless siz <= strlen(dst)).
+ * Returns strlen(src) + MIN(siz, strlen(initial dst)).
+ * If retval >= siz, truncation occurred.
+ */
+size_t
+strlcat(dst, src, siz)
+ char *dst;
+ const char *src;
+ size_t siz;
+{
+ register char *d = dst;
+ register const char *s = src;
+ register size_t n = siz;
+ size_t dlen;
+
+ /* Find the end of dst and adjust bytes left but don't go past end */
+ while (n-- != 0 && *d != '\0')
+ d++;
+ dlen = d - dst;
+ n = siz - dlen;
+
+ if (n == 0)
+ return(dlen + strlen(s));
+ while (*s != '\0') {
+ if (n != 1) {
+ *d++ = *s;
+ n--;
+ }
+ s++;
+ }
+ *d = '\0';
+
+ return(dlen + (s - src)); /* count does not include NUL */
+}
diff -up nfs-utils-1.2.3/utils/spnfsd/strlcpy.c.orig nfs-utils-1.2.3/utils/spnfsd/strlcpy.c
--- nfs-utils-1.2.3/utils/spnfsd/strlcpy.c.orig 2010-10-04 15:48:00.086338923 -0400
+++ nfs-utils-1.2.3/utils/spnfsd/strlcpy.c 2010-10-04 15:48:00.086338923 -0400
@@ -0,0 +1,73 @@
+/* $OpenBSD: strlcpy.c,v 1.5 2001/05/13 15:40:16 deraadt Exp $ */
+
+/*
+ * Copyright (c) 1998 Todd C. Miller <Todd.Miller@courtesan.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. The name of the author may not be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
+ * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
+ * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+ * THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+ * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+ * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
+ * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
+ * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
+ * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#if defined(LIBC_SCCS) && !defined(lint)
+static char *rcsid = "$OpenBSD: strlcpy.c,v 1.5 2001/05/13 15:40:16 deraadt Exp $";
+#endif /* LIBC_SCCS and not lint */
+
+#include <sys/types.h>
+#include <string.h>
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif /* HAVE_CONFIG_H */
+
+/*
+ * Copy src to string dst of size siz. At most siz-1 characters
+ * will be copied. Always NUL terminates (unless siz == 0).
+ * Returns strlen(src); if retval >= siz, truncation occurred.
+ */
+size_t
+strlcpy(dst, src, siz)
+ char *dst;
+ const char *src;
+ size_t siz;
+{
+ register char *d = dst;
+ register const char *s = src;
+ register size_t n = siz;
+
+ /* Copy as many bytes as will fit */
+ if (n != 0 && --n != 0) {
+ do {
+ if ((*d++ = *s++) == 0)
+ break;
+ } while (--n != 0);
+ }
+
+ /* Not enough room in dst, add NUL and traverse rest of src */
+ if (n == 0) {
+ if (siz != 0)
+ *d = '\0'; /* NUL-terminate dst */
+ while (*s++)
+ ;
+ }
+
+ return(s - src - 1); /* count does not include NUL */
+}