* sysfs - extract md related information from sysfs. Part of:
* mdadm - manage Linux "md" devices aka RAID arrays.
*
- * Copyright (C) 2006 Neil Brown <neilb@suse.de>
+ * Copyright (C) 2006-2009 Neil Brown <neilb@suse.de>
*
*
* This program is free software; you can redistribute it and/or modify
}
if (devnum == NoMdDev)
return;
- if (devnum >= 0)
- sprintf(mdi->sys_name, "md%d", devnum);
- else
- sprintf(mdi->sys_name, "md_d%d",
- -1-devnum);
+ fmt_devname(mdi->sys_name, devnum);
}
struct mdinfo *sysfs_read(int fd, int devnum, unsigned long options)
{
- /* Longest possible name in sysfs, mounted at /sys, is
- * /sys/block/md_dXXX/md/dev-XXXXX/block/dev
- * /sys/block/md_dXXX/md/metadata_version
- * which is about 41 characters. 50 should do for now
- */
- char fname[50];
- char buf[1024];
+ char fname[PATH_MAX];
+ char buf[PATH_MAX];
char *base;
char *dbase;
struct mdinfo *sra;
if (options & GET_SAFEMODE) {
int scale = 1;
int dot = 0;
- int i;
+ unsigned i;
unsigned long msec;
size_t len;
}
}
- dev->next = sra->devs;
- sra->devs = dev;
-
strcpy(dev->sys_name, de->d_name);
dev->disk.raid_disk = strtoul(buf, &ep, 10);
if (*ep) dev->disk.raid_disk = -1;
strcpy(dbase, "block/dev");
- if (load_sys(fname, buf))
- goto abort;
+ if (load_sys(fname, buf)) {
+ /* assume this is a stale reference to a hot
+ * removed device
+ */
+ free(dev);
+ continue;
+ }
sscanf(buf, "%d:%d", &dev->disk.major, &dev->disk.minor);
+ /* special case check for block devices that can go 'offline' */
+ strcpy(dbase, "block/device/state");
+ if (load_sys(fname, buf) == 0 &&
+ strncmp(buf, "offline", 7) == 0) {
+ free(dev);
+ continue;
+ }
+
+ /* finally add this disk to the array */
+ dev->next = sra->devs;
+ sra->devs = dev;
+
if (options & GET_OFFSET) {
strcpy(dbase, "offset");
if (load_sys(fname, buf))
char fname[50];
int n;
if (fstat(fd, &stb)) return 0;
- if (major(stb.st_rdev) != get_mdp_major())
+ if (major(stb.st_rdev) != (unsigned)get_mdp_major())
sprintf(fname, "/sys/block/md%d/md/component_size",
(int)minor(stb.st_rdev));
else
char *name, char *val)
{
char fname[50];
- int n;
+ unsigned int n;
int fd;
sprintf(fname, "/sys/block/%s/md/%s/%s",
return -1;
n = write(fd, event, strlen(event));
close(fd);
+ if (n != (int)strlen(event)) {
+ dprintf(Name ": failed to write '%s' to '%s' (%s)\n",
+ event, fname, strerror(errno));
+ return -1;
+ }
return 0;
}
-int sysfs_get_ll(struct mdinfo *sra, struct mdinfo *dev,
- char *name, unsigned long long *val)
+int sysfs_attribute_available(struct mdinfo *sra, struct mdinfo *dev, char *name)
+{
+ char fname[50];
+ struct stat st;
+
+ sprintf(fname, "/sys/block/%s/md/%s/%s",
+ sra->sys_name, dev?dev->sys_name:"", name);
+
+ return stat(fname, &st) == 0;
+}
+
+int sysfs_get_fd(struct mdinfo *sra, struct mdinfo *dev,
+ char *name)
{
char fname[50];
- char buf[50];
- int n;
int fd;
- char *ep;
+
sprintf(fname, "/sys/block/%s/md/%s/%s",
sra->sys_name, dev?dev->sys_name:"", name);
- fd = open(fname, O_RDONLY);
+ fd = open(fname, O_RDWR);
if (fd < 0)
- return -1;
+ fd = open(fname, O_RDONLY);
+ return fd;
+}
+
+int sysfs_fd_get_ll(int fd, unsigned long long *val)
+{
+ char buf[50];
+ int n;
+ char *ep;
+
+ lseek(fd, 0, 0);
n = read(fd, buf, sizeof(buf));
- close(fd);
if (n <= 0)
- return -1;
+ return -2;
buf[n] = 0;
*val = strtoull(buf, &ep, 0);
if (ep == buf || (*ep != 0 && *ep != '\n' && *ep != ' '))
return 0;
}
+int sysfs_get_ll(struct mdinfo *sra, struct mdinfo *dev,
+ char *name, unsigned long long *val)
+{
+ int n;
+ int fd;
+
+ fd = sysfs_get_fd(sra, dev, name);
+ if (fd < 0)
+ return -1;
+ n = sysfs_fd_get_ll(fd, val);
+ close(fd);
+ return n;
+}
+
+int sysfs_fd_get_str(int fd, char *val, int size)
+{
+ int n;
+
+ lseek(fd, 0, 0);
+ n = read(fd, val, size);
+ if (n <= 0)
+ return -1;
+ val[n] = 0;
+ return n;
+}
+
+int sysfs_get_str(struct mdinfo *sra, struct mdinfo *dev,
+ char *name, char *val, int size)
+{
+ int n;
+ int fd;
+
+ fd = sysfs_get_fd(sra, dev, name);
+ if (fd < 0)
+ return -1;
+ n = sysfs_fd_get_str(fd, val, size);
+ close(fd);
+ return n;
+}
+
int sysfs_set_safemode(struct mdinfo *sra, unsigned long ms)
{
unsigned long sec;
{
int rv = 0;
char ver[100];
+ int raid_disks = info->array.raid_disks;
ver[0] = 0;
if (info->array.major_version == -1 &&
info->array.minor_version == -2) {
+ char buf[1024];
+
strcat(strcpy(ver, "external:"), info->text_version);
+ /* meta version might already be set if we are setting
+ * new geometry for a reshape. In that case we don't
+ * want to over-write the 'readonly' flag that is
+ * stored in the metadata version. So read the current
+ * version first, and preserve the flag
+ */
+ if (sysfs_get_str(info, NULL, "metadata_version",
+ buf, 1024) > 0)
+ if (strlen(buf) >= 9 && buf[9] == '-')
+ ver[9] = '-';
+
if ((vers % 100) < 2 ||
sysfs_set_str(info, NULL, "metadata_version",
ver) < 0) {
return 0; /* FIXME */
rv |= sysfs_set_str(info, NULL, "level",
map_num(pers, info->array.level));
- rv |= sysfs_set_num(info, NULL, "raid_disks", info->array.raid_disks);
+ if (info->reshape_active && info->delta_disks != UnSet)
+ raid_disks -= info->delta_disks;
+ rv |= sysfs_set_num(info, NULL, "raid_disks", raid_disks);
rv |= sysfs_set_num(info, NULL, "chunk_size", info->array.chunk_size);
rv |= sysfs_set_num(info, NULL, "layout", info->array.layout);
rv |= sysfs_set_num(info, NULL, "component_size", info->component_size/2);
+ if (info->custom_array_size) {
+ int rc;
+
+ rc = sysfs_set_num(info, NULL, "array_size",
+ info->custom_array_size/2);
+ if (rc && errno == ENOENT) {
+ fprintf(stderr, Name ": This kernel does not "
+ "have the md/array_size attribute, "
+ "the array may be larger than expected\n");
+ rc = 0;
+ }
+ rv |= rc;
+ }
+
if (info->array.level > 0)
rv |= sysfs_set_num(info, NULL, "resync_start", info->resync_start);
+
+ if (info->reshape_active) {
+ rv |= sysfs_set_num(info, NULL, "reshape_position",
+ info->reshape_progress);
+ rv |= sysfs_set_num(info, NULL, "chunk_size", info->new_chunk);
+ rv |= sysfs_set_num(info, NULL, "layout", info->new_layout);
+ rv |= sysfs_set_num(info, NULL, "raid_disks",
+ info->array.raid_disks);
+ /* We don't set 'new_level' here. That can only happen
+ * once the reshape completes.
+ */
+ }
return rv;
}
-int sysfs_add_disk(struct mdinfo *sra, struct mdinfo *sd)
+int sysfs_add_disk(struct mdinfo *sra, struct mdinfo *sd, int resume)
{
- char dv[100];
- char nm[100];
+ char dv[PATH_MAX];
+ char nm[PATH_MAX];
char *dname;
int rv;
memset(nm, 0, sizeof(nm));
sprintf(dv, "/sys/dev/block/%d:%d", sd->disk.major, sd->disk.minor);
- rv = readlink(dv, nm, sizeof(nm));
+ rv = readlink(dv, nm, sizeof(nm)-1);
if (rv <= 0)
return -1;
nm[rv] = '\0';
strcpy(sd->sys_name, "dev-");
strcpy(sd->sys_name+4, dname);
+ /* test write to see if 'recovery_start' is available */
+ if (resume && sd->recovery_start < MaxSector &&
+ sysfs_set_num(sra, sd, "recovery_start", 0)) {
+ sysfs_set_str(sra, sd, "state", "remove");
+ return -1;
+ }
+
rv = sysfs_set_num(sra, sd, "offset", sd->data_offset);
rv |= sysfs_set_num(sra, sd, "size", (sd->component_size+1) / 2);
if (sra->array.level != LEVEL_CONTAINER) {
- rv |= sysfs_set_num(sra, sd, "slot", sd->disk.raid_disk);
-// rv |= sysfs_set_str(sra, sd, "state", "in_sync");
+ if (sd->recovery_start == MaxSector)
+ /* This can correctly fail if array isn't started,
+ * yet, so just ignore status for now.
+ */
+ sysfs_set_str(sra, sd, "state", "insync");
+ if (sd->disk.raid_disk >= 0)
+ rv |= sysfs_set_num(sra, sd, "slot", sd->disk.raid_disk);
+ if (resume)
+ sysfs_set_num(sra, sd, "recovery_start", sd->recovery_start);
}
return rv;
}
/* from an open block device, try to retrieve it scsi_id */
struct stat st;
char path[256];
- char *c1, *c2;
DIR *dir;
struct dirent *de;
+ int host, bus, target, lun;
if (fstat(fd, &st))
return 1;
- snprintf(path, sizeof(path), "/sys/dev/block/%d:%d/device",
+ snprintf(path, sizeof(path), "/sys/dev/block/%d:%d/device/scsi_device",
major(st.st_rdev), minor(st.st_rdev));
dir = opendir(path);
if (!dir)
return 1;
- de = readdir(dir);
- while (de) {
- if (strncmp("scsi_disk:", de->d_name,
- strlen("scsi_disk:")) == 0)
+ for (de = readdir(dir); de; de = readdir(dir)) {
+ int count;
+
+ if (de->d_type != DT_DIR)
+ continue;
+
+ count = sscanf(de->d_name, "%d:%d:%d:%d", &host, &bus, &target, &lun);
+ if (count == 4)
break;
- de = readdir(dir);
}
closedir(dir);
if (!de)
return 1;
- c1 = strchr(de->d_name, ':');
- c1++;
- c2 = strchr(c1, ':');
- *c2 = '\0';
- *id = strtol(c1, NULL, 10) << 24; /* host */
- c1 = c2 + 1;
- c2 = strchr(c1, ':');
- *c2 = '\0';
- *id |= strtol(c1, NULL, 10) << 16; /* channel */
- c1 = c2 + 1;
- c2 = strchr(c1, ':');
- *c2 = '\0';
- *id |= strtol(c1, NULL, 10) << 8; /* lun */
- c1 = c2 + 1;
- *id |= strtol(c1, NULL, 10); /* id */
-
+ *id = (host << 24) | (bus << 16) | (target << 8) | (lun << 0);
return 0;
}
}
n = read(fd, buf, sizeof(buf)-1);
close(fd);
+ if (n < 0)
+ continue;
buf[n] = 0;
if (sscanf(buf, "%d:%d%c", &mj, &mn, &c) != 3 ||
c != '\n') {
else
return found;
}
+
+int sysfs_freeze_array(struct mdinfo *sra)
+{
+ /* Try to freeze resync/rebuild on this array/container.
+ * Return -1 if the array is busy,
+ * return -2 container cannot be frozen,
+ * return 0 if this kernel doesn't support 'frozen'
+ * return 1 if it worked.
+ */
+ char buf[20];
+
+ if (!sysfs_attribute_available(sra, NULL, "sync_action"))
+ return 1; /* no sync_action == frozen */
+ if (sysfs_get_str(sra, NULL, "sync_action", buf, 20) <= 0)
+ return 0;
+ if (strcmp(buf, "idle\n") != 0 &&
+ strcmp(buf, "frozen\n") != 0)
+ return -1;
+ if (sysfs_set_str(sra, NULL, "sync_action", "frozen") < 0)
+ return 0;
+ return 1;
+}