/*
* mdadm - manage Linux "md" devices aka RAID arrays.
*
- * Copyright (C) 2001-2009 Neil Brown <neilb@suse.de>
+ * Copyright (C) 2001-2013 Neil Brown <neilb@suse.de>
*
*
* This program is free software; you can redistribute it and/or modify
#include <sys/utsname.h>
#include <sys/wait.h>
#include <sys/un.h>
+#include <sys/resource.h>
+#include <sys/vfs.h>
+#include <linux/magic.h>
+#include <poll.h>
#include <ctype.h>
#include <dirent.h>
#include <signal.h>
+#include <dlfcn.h>
+
/*
* following taken from linux/blkpg.h because they aren't
/* The argument structure */
struct blkpg_ioctl_arg {
- int op;
- int flags;
- int datalen;
- void *data;
+ int op;
+ int flags;
+ int datalen;
+ void *data;
};
/* The subfunctions (for the op field) */
aren't permitted). */
#define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); }))
+static int is_dlm_hooks_ready = 0;
+
+int dlm_funs_ready(void)
+{
+ return is_dlm_hooks_ready ? 1 : 0;
+}
+
+static struct dlm_hooks *dlm_hooks = NULL;
+struct dlm_lock_resource *dlm_lock_res = NULL;
+static int ast_called = 0;
+
+struct dlm_lock_resource {
+ dlm_lshandle_t *ls;
+ struct dlm_lksb lksb;
+};
+
+/* Using poll(2) to wait for and dispatch ASTs */
+static int poll_for_ast(dlm_lshandle_t ls)
+{
+ struct pollfd pfd;
+
+ pfd.fd = dlm_hooks->ls_get_fd(ls);
+ pfd.events = POLLIN;
+
+ while (!ast_called)
+ {
+ if (poll(&pfd, 1, 0) < 0)
+ {
+ perror("poll");
+ return -1;
+ }
+ dlm_hooks->dispatch(dlm_hooks->ls_get_fd(ls));
+ }
+ ast_called = 0;
+
+ return 0;
+}
+
+static void dlm_ast(void *arg)
+{
+ ast_called = 1;
+}
+
+static char *cluster_name = NULL;
+/* Create the lockspace, take bitmapXXX locks on all the bitmaps. */
+int cluster_get_dlmlock(int *lockid)
+{
+ int ret = -1;
+ char str[64];
+ int flags = LKF_NOQUEUE;
+
+ ret = get_cluster_name(&cluster_name);
+ if (ret) {
+ pr_err("The md can't get cluster name\n");
+ return -1;
+ }
+
+ dlm_lock_res = xmalloc(sizeof(struct dlm_lock_resource));
+ dlm_lock_res->ls = dlm_hooks->create_lockspace(cluster_name, O_RDWR);
+ if (!dlm_lock_res->ls) {
+ pr_err("%s failed to create lockspace\n", cluster_name);
+ return -ENOMEM;
+ }
+
+ snprintf(str, 64, "bitmap%s", cluster_name);
+ ret = dlm_hooks->ls_lock(dlm_lock_res->ls, LKM_PWMODE, &dlm_lock_res->lksb,
+ flags, str, strlen(str), 0, dlm_ast,
+ dlm_lock_res, NULL, NULL);
+ if (ret) {
+ pr_err("error %d when get PW mode on lock %s\n", errno, str);
+ dlm_hooks->release_lockspace(cluster_name, dlm_lock_res->ls, 1);
+ return ret;
+ }
+
+ /* Wait for it to complete */
+ poll_for_ast(dlm_lock_res->ls);
+ *lockid = dlm_lock_res->lksb.sb_lkid;
+
+ return dlm_lock_res->lksb.sb_status;
+}
+
+int cluster_release_dlmlock(int lockid)
+{
+ int ret = -1;
+
+ if (!cluster_name)
+ return -1;
+
+ ret = dlm_hooks->ls_unlock(dlm_lock_res->ls, lockid, 0,
+ &dlm_lock_res->lksb, dlm_lock_res);
+ if (ret) {
+ pr_err("error %d happened when unlock\n", errno);
+ /* XXX make sure the lock is unlocked eventually */
+ goto out;
+ }
+
+ /* Wait for it to complete */
+ poll_for_ast(dlm_lock_res->ls);
+
+ errno = dlm_lock_res->lksb.sb_status;
+ if (errno != EUNLOCK) {
+ pr_err("error %d happened in ast when unlock lockspace\n", errno);
+ /* XXX make sure the lockspace is unlocked eventually */
+ goto out;
+ }
+
+ ret = dlm_hooks->release_lockspace(cluster_name, dlm_lock_res->ls, 1);
+ if (ret) {
+ pr_err("error %d happened when release lockspace\n", errno);
+ /* XXX make sure the lockspace is released eventually */
+ goto out;
+ }
+ free(dlm_lock_res);
+
+out:
+ return ret;
+}
+
+int md_array_active(int fd)
+{
+ struct mdinfo *sra;
+ struct mdu_array_info_s array;
+ int ret;
+
+ sra = sysfs_read(fd, NULL, GET_ARRAY_STATE);
+ if (sra) {
+ if (sra->array_state != ARRAY_CLEAR &&
+ sra->array_state != ARRAY_INACTIVE &&
+ sra->array_state != ARRAY_UNKNOWN_STATE)
+ ret = 0;
+ else
+ ret = -ENODEV;
+
+ free(sra);
+ } else {
+ /*
+ * GET_ARRAY_INFO doesn't provide access to the proper state
+ * information, so fallback to a basic check for raid_disks != 0
+ */
+ ret = ioctl(fd, GET_ARRAY_INFO, &array);
+ }
+
+ return !ret;
+}
+
+/*
+ * Get array info from the kernel. Longer term we want to deprecate the
+ * ioctl and get it from sysfs.
+ */
+int md_get_array_info(int fd, struct mdu_array_info_s *array)
+{
+ return ioctl(fd, GET_ARRAY_INFO, array);
+}
+
+/*
+ * Set array info
+ */
+int md_set_array_info(int fd, struct mdu_array_info_s *array)
+{
+ return ioctl(fd, SET_ARRAY_INFO, array);
+}
+
+/*
+ * Get disk info from the kernel.
+ */
+int md_get_disk_info(int fd, struct mdu_disk_info_s *disk)
+{
+ return ioctl(fd, GET_DISK_INFO, disk);
+}
+
/*
* Parse a 128 bit uuid in 4 integers
* format is 32 hexx nibbles with options :.<space> separator
int hit = 0; /* number of Hex digIT */
int i;
char c;
- for (i=0; i<4; i++) uuid[i]=0;
+ for (i = 0; i < 4; i++)
+ uuid[i] = 0;
- while ((c= *str++)) {
+ while ((c = *str++) != 0) {
int n;
- if (c>='0' && c<='9')
+ if (c >= '0' && c <= '9')
n = c-'0';
- else if (c>='a' && c <= 'f')
+ else if (c >= 'a' && c <= 'f')
n = 10 + c - 'a';
- else if (c>='A' && c <= 'F')
+ else if (c >= 'A' && c <= 'F')
n = 10 + c - 'A';
else if (strchr(":. -", c))
continue;
return 0;
}
-
-/*
- * Get the md version number.
- * We use the RAID_VERSION ioctl if it is supported
- * If not, but we have a block device with major '9', we assume
- * 0.36.0
- *
- * Return version number as 24 but number - assume version parts
- * always < 255
- */
-
-int md_get_version(int fd)
-{
- struct stat stb;
- mdu_version_t vers;
-
- if (fstat(fd, &stb)<0)
- return -1;
- if ((S_IFMT&stb.st_mode) != S_IFBLK)
- return -1;
-
- if (ioctl(fd, RAID_VERSION, &vers) == 0)
- return (vers.major*10000) + (vers.minor*100) + vers.patchlevel;
- if (errno == EACCES)
- return -1;
- if (major(stb.st_rdev) == MD_MAJOR)
- return (3600);
- return -1;
-}
-
int get_linux_version()
{
struct utsname name;
return (a*1000000)+(b*1000)+c;
}
-#ifndef MDASSEMBLE
int mdadm_version(char *version)
{
int a, b, c;
return (a*1000000)+(b*1000)+c;
}
-long long parse_size(char *size)
+unsigned long long parse_size(char *size)
{
/* parse 'size' which should be a number optionally
* followed by 'K', 'M', or 'G'.
* Without a suffix, K is assumed.
* Number returned is in sectors (half-K)
+ * INVALID_SECTORS returned on error.
*/
char *c;
long long s = strtoll(size, &c, 10);
c++;
s *= 1024 * 1024 * 2;
break;
+ case 's': /* sectors */
+ c++;
+ break;
}
- }
+ } else
+ s = INVALID_SECTORS;
if (*c)
- s = 0;
+ s = INVALID_SECTORS;
return s;
}
{
/* Parse the layout string for 'faulty' */
int ln = strcspn(layout, "0123456789");
- char *m = strdup(layout);
+ char *m = xstrdup(layout);
int mode;
m[ln] = 0;
mode = map_name(faultylayout, m);
return mode | (atoi(layout+ln)<< ModeShift);
}
-#endif
+
+long parse_num(char *num)
+{
+ /* Either return a valid number, or -1 */
+ char *c;
+ long rv = strtol(num, &c, 10);
+ if (rv < 0 || *c || !num[0])
+ return -1;
+ else
+ return rv;
+}
+
+int parse_cluster_confirm_arg(char *input, char **devname, int *slot)
+{
+ char *dev;
+ *slot = strtoul(input, &dev, 10);
+ if (dev == input || dev[0] != ':')
+ return -1;
+ *devname = dev+1;
+ return 0;
+}
void remove_partitions(int fd)
{
a.datalen = sizeof(p);
a.flags = 0;
memset(a.data, 0, a.datalen);
- for (p.pno=0; p.pno < 16; p.pno++)
+ for (p.pno = 0; p.pno < 16; p.pno++)
ioctl(fd, BLKPG, &a);
#endif
}
if (ioctl(fd, BLKPG, &a) == 0)
/* Very unlikely, but not a partition */
return 0;
- if (errno == ENXIO)
+ if (errno == ENXIO || errno == ENOTTY)
/* not a partition */
return 0;
return rv;
}
-int enough(int level, int raid_disks, int layout, int clean,
- char *avail, int avail_disks)
+int enough(int level, int raid_disks, int layout, int clean, char *avail)
{
int copies, first;
+ int i;
+ int avail_disks = 0;
+
+ for (i = 0; i < raid_disks; i++)
+ avail_disks += !!avail[i];
+
switch (level) {
case 10:
/* This is the tricky one - we need to check
* which actual disks are present.
*/
copies = (layout&255)* ((layout>>8) & 255);
- first=0;
+ first = 0;
do {
/* there must be one of the 'copies' form 'first' */
int n = copies;
- int cnt=0;
+ int cnt = 0;
+ int this = first;
while (n--) {
- if (avail[first])
+ if (avail[this])
cnt++;
- first = (first+1) % raid_disks;
+ this = (this+1) % raid_disks;
}
if (cnt == 0)
return 0;
-
+ first = (first+(layout&255)) % raid_disks;
} while (first != 0);
return 1;
case 1:
return avail_disks >= 1;
case 4:
+ if (avail_disks == raid_disks - 1 &&
+ !avail[raid_disks - 1])
+ /* If just the parity device is missing, then we
+ * have enough, even if not clean
+ */
+ return 1;
+ /* FALL THROUGH */
case 5:
if (clean)
return avail_disks >= raid_disks-1;
}
}
-int enough_fd(int fd)
-{
- struct mdu_array_info_s array;
- struct mdu_disk_info_s disk;
- int avail_disks = 0;
- int i;
- char *avail;
-
- if (ioctl(fd, GET_ARRAY_INFO, &array) != 0 ||
- array.raid_disks <= 0)
- return 0;
- avail = calloc(array.raid_disks, 1);
- for (i=0; i < 1024 && array.nr_disks > 0; i++) {
- disk.number = i;
- if (ioctl(fd, GET_DISK_INFO, &disk) != 0)
- continue;
- if (disk.major == 0 && disk.minor == 0)
- continue;
- array.nr_disks--;
-
- if (! (disk.state & (1<<MD_DISK_SYNC)))
- continue;
- if (disk.raid_disk < 0 || disk.raid_disk >= array.raid_disks)
- continue;
- avail_disks++;
- avail[disk.raid_disk] = 1;
- }
- /* This is used on an active array, so assume it is clean */
- return enough(array.level, array.raid_disks, array.layout,
- 1,
- avail, avail_disks);
-}
-
-
const int uuid_zero[4] = { 0, 0, 0, 0 };
int same_uuid(int a[4], int b[4], int swapuuid)
unsigned char *ac = (unsigned char *)a;
unsigned char *bc = (unsigned char *)b;
int i;
- for (i=0; i<16; i+= 4) {
+ for (i = 0; i < 16; i += 4) {
if (ac[i+0] != bc[i+3] ||
ac[i+1] != bc[i+2] ||
ac[i+2] != bc[i+1] ||
return 0;
}
}
+
void copy_uuid(void *a, int b[4], int swapuuid)
{
if (swapuuid) {
unsigned char *ac = (unsigned char *)a;
unsigned char *bc = (unsigned char *)b;
int i;
- for (i=0; i<16; i+= 4) {
+ for (i = 0; i < 16; i += 4) {
ac[i+0] = bc[i+3];
ac[i+1] = bc[i+2];
ac[i+2] = bc[i+1];
return __fname_from_uuid(info->uuid, (st->ss == &super1) ? 1 : st->ss->swapuuid, buf, sep);
}
-#ifndef MDASSEMBLE
int check_ext2(int fd, char *name)
{
/*
*/
unsigned char sb[1024];
time_t mtime;
- int size, bsize;
+ unsigned long long size;
+ int bsize;
if (lseek(fd, 1024,0)!= 1024)
return 0;
if (read(fd, sb, 1024)!= 1024)
mtime = sb[44]|(sb[45]|(sb[46]|sb[47]<<8)<<8)<<8;
bsize = sb[24]|(sb[25]|(sb[26]|sb[27]<<8)<<8)<<8;
size = sb[4]|(sb[5]|(sb[6]|sb[7]<<8)<<8)<<8;
- fprintf(stderr, Name ": %s appears to contain an ext2fs file system\n",
+ size <<= bsize;
+ pr_err("%s appears to contain an ext2fs file system\n",
name);
- fprintf(stderr," size=%dK mtime=%s",
- size*(1<<bsize), ctime(&mtime));
+ cont_err("size=%lluK mtime=%s", size, ctime(&mtime));
return 1;
}
*
*/
unsigned char sb[1024];
- unsigned long size;
+ unsigned long long size;
if (lseek(fd, 64*1024, 0) != 64*1024)
return 0;
if (read(fd, sb, 1024) != 1024)
return 0;
- if (strncmp((char*)sb+52, "ReIsErFs",8)!=0 &&
- strncmp((char*)sb+52, "ReIsEr2Fs",9)!=0)
+ if (strncmp((char*)sb+52, "ReIsErFs",8) != 0 &&
+ strncmp((char*)sb+52, "ReIsEr2Fs",9) != 0)
return 0;
- fprintf(stderr, Name ": %s appears to contain a reiserfs file system\n",name);
+ pr_err("%s appears to contain a reiserfs file system\n",name);
size = sb[0]|(sb[1]|(sb[2]|sb[3]<<8)<<8)<<8;
- fprintf(stderr, " size = %luK\n", size*4);
+ cont_err("size = %lluK\n", size*4);
return 1;
}
char *level;
struct supertype *st = guess_super(fd);
- if (!st) return 0;
- st->ignore_hw_compat = 1;
- st->ss->load_super(st, fd, name);
- /* Looks like a raid array .. */
- fprintf(stderr, Name ": %s appears to be part of a raid array:\n",
- name);
- st->ss->getinfo_super(st, &info, NULL);
- st->ss->free_super(st);
- crtime = info.array.ctime;
- level = map_num(pers, info.array.level);
- if (!level) level = "-unknown-";
- fprintf(stderr, " level=%s devices=%d ctime=%s",
- level, info.array.raid_disks, ctime(&crtime));
+ if (!st)
+ return 0;
+ if (st->ss->add_to_super != NULL) {
+ st->ss->load_super(st, fd, name);
+ /* Looks like a raid array .. */
+ pr_err("%s appears to be part of a raid array:\n", name);
+ st->ss->getinfo_super(st, &info, NULL);
+ st->ss->free_super(st);
+ crtime = info.array.ctime;
+ level = map_num(pers, info.array.level);
+ if (!level)
+ level = "-unknown-";
+ cont_err("level=%s devices=%d ctime=%s",
+ level, info.array.raid_disks, ctime(&crtime));
+ } else {
+ /* Looks like GPT or MBR */
+ pr_err("partition table exists on %s\n", name);
+ }
return 1;
}
{
char *add = "";
int i;
- for (i=0; i<5; i++) {
+ for (i = 0; i < 5; i++) {
char buf[100];
fprintf(stderr, "%s%s", mesg, add);
fflush(stderr);
return 0;
add = "(y/n) ";
}
- fprintf(stderr, Name ": assuming 'no'\n");
+ pr_err("assuming 'no'\n");
return 0;
}
-#endif /* MDASSEMBLE */
int is_standard(char *dev, int *nump)
{
* 0 if not a standard name.
*/
char *d = strrchr(dev, '/');
- int type=0;
+ int type = 0;
int num;
if (!d)
return 0;
- if (strncmp(d, "/d",2)==0)
- d += 2, type=1; /* /dev/md/dN{pM} */
- else if (strncmp(d, "/md_d", 5)==0)
- d += 5, type=1; /* /dev/md_dN{pM} */
- else if (strncmp(d, "/md", 3)==0)
- d += 3, type=-1; /* /dev/mdN */
- else if (d-dev > 3 && strncmp(d-2, "md/", 3)==0)
- d += 1, type=-1; /* /dev/md/N */
+ if (strncmp(d, "/d",2) == 0)
+ d += 2, type = 1; /* /dev/md/dN{pM} */
+ else if (strncmp(d, "/md_d", 5) == 0)
+ d += 5, type = 1; /* /dev/md_dN{pM} */
+ else if (strncmp(d, "/md", 3) == 0)
+ d += 3, type = -1; /* /dev/mdN */
+ else if (d-dev > 3 && strncmp(d-2, "md/", 3) == 0)
+ d += 1, type = -1; /* /dev/md/N */
else
return 0;
if (!*d)
unsigned int csum;
unsigned int *superc = (unsigned int*) super;
- for(i=0; i<bytes/4; i++)
- newcsum+= superc[i];
+ for(i = 0; i < bytes/4; i++)
+ newcsum += superc[i];
csum = (newcsum& 0xffffffff) + (newcsum>>32);
#ifdef __alpha__
/* The in-kernel checksum calculation is always 16bit on
* the alpha, though it is 32 bit on i386...
- * I wonder what it is elsewhere... (it uses and API in
+ * I wonder what it is elsewhere... (it uses an API in
* a way that it shouldn't).
*/
csum = (csum & 0xffff) + (csum >> 16);
return csum;
}
-#ifndef MDASSEMBLE
char *human_size(long long bytes)
{
- static char buf[30];
+ static char buf[47];
/* We convert bytes to either centi-M{ega,ibi}bytes or
* centi-G{igi,ibi}bytes, with appropriate rounding,
* We allow upto 2048Megabytes before converting to
* gigabytes, as that shows more precision and isn't
* too large a number.
- * Terrabytes are not yet handled.
+ * Terabytes are not yet handled.
*/
if (bytes < 5000*1024)
- buf[0]=0;
+ buf[0] = 0;
else if (bytes < 2*1024LL*1024LL*1024LL) {
- long cMiB = (bytes / ( (1LL<<20) / 200LL ) +1) /2;
+ long cMiB = (bytes * 200LL / (1LL<<20) + 1) / 2;
long cMB = (bytes / ( 1000000LL / 200LL ) +1) /2;
snprintf(buf, sizeof(buf), " (%ld.%02ld MiB %ld.%02ld MB)",
- cMiB/100 , cMiB % 100,
- cMB/100, cMB % 100);
+ cMiB/100, cMiB % 100, cMB/100, cMB % 100);
} else {
- long cGiB = (bytes / ( (1LL<<30) / 200LL ) +1) /2;
+ long cGiB = (bytes * 200LL / (1LL<<30) +1) / 2;
long cGB = (bytes / (1000000000LL/200LL ) +1) /2;
snprintf(buf, sizeof(buf), " (%ld.%02ld GiB %ld.%02ld GB)",
- cGiB/100 , cGiB % 100,
- cGB/100, cGB % 100);
+ cGiB/100, cGiB % 100, cGB/100, cGB % 100);
}
return buf;
}
-char *human_size_brief(long long bytes)
+char *human_size_brief(long long bytes, int prefix)
{
static char buf[30];
+ /* We convert bytes to either centi-M{ega,ibi}bytes or
+ * centi-G{igi,ibi}bytes, with appropriate rounding,
+ * and then print 1/100th of those as a decimal.
+ * We allow upto 2048Megabytes before converting to
+ * gigabytes, as that shows more precision and isn't
+ * too large a number.
+ * Terabytes are not yet handled.
+ *
+ * If prefix == IEC, we mean prefixes like kibi,mebi,gibi etc.
+ * If prefix == JEDEC, we mean prefixes like kilo,mega,giga etc.
+ */
+
if (bytes < 5000*1024)
- snprintf(buf, sizeof(buf), "%ld.%02ldKiB",
- (long)(bytes>>10), (long)(((bytes&1023)*100+512)/1024)
- );
- else if (bytes < 2*1024LL*1024LL*1024LL)
- snprintf(buf, sizeof(buf), "%ld.%02ldMiB",
- (long)(bytes>>20),
- (long)((bytes&0xfffff)+0x100000/200)/(0x100000/100)
- );
+ buf[0] = 0;
+ else if (prefix == IEC) {
+ if (bytes < 2*1024LL*1024LL*1024LL) {
+ long cMiB = (bytes * 200LL / (1LL<<20) +1) /2;
+ snprintf(buf, sizeof(buf), "%ld.%02ldMiB",
+ cMiB/100, cMiB % 100);
+ } else {
+ long cGiB = (bytes * 200LL / (1LL<<30) +1) /2;
+ snprintf(buf, sizeof(buf), "%ld.%02ldGiB",
+ cGiB/100, cGiB % 100);
+ }
+ }
+ else if (prefix == JEDEC) {
+ if (bytes < 2*1024LL*1024LL*1024LL) {
+ long cMB = (bytes / ( 1000000LL / 200LL ) +1) /2;
+ snprintf(buf, sizeof(buf), "%ld.%02ldMB",
+ cMB/100, cMB % 100);
+ } else {
+ long cGB = (bytes / (1000000000LL/200LL ) +1) /2;
+ snprintf(buf, sizeof(buf), "%ld.%02ldGB",
+ cGB/100, cGB % 100);
+ }
+ }
else
- snprintf(buf, sizeof(buf), "%ld.%02ldGiB",
- (long)(bytes>>30),
- (long)(((bytes>>10)&0xfffff)+0x100000/200)/(0x100000/100)
- );
+ buf[0] = 0;
+
return buf;
}
if (near*far == 1)
printf("NO REDUNDANCY");
}
-#endif
unsigned long long calc_array_size(int level, int raid_disks, int layout,
int chunksize, unsigned long long devsize)
+{
+ if (level == 1)
+ return devsize;
+ devsize &= ~(unsigned long long)((chunksize>>9)-1);
+ return get_data_disks(level, layout, raid_disks) * devsize;
+}
+
+int get_data_disks(int level, int layout, int raid_disks)
{
int data_disks = 0;
switch (level) {
- case 0: data_disks = raid_disks; break;
- case 1: data_disks = 1; break;
+ case 0: data_disks = raid_disks;
+ break;
+ case 1: data_disks = 1;
+ break;
case 4:
- case 5: data_disks = raid_disks - 1; break;
- case 6: data_disks = raid_disks - 2; break;
+ case 5: data_disks = raid_disks - 1;
+ break;
+ case 6: data_disks = raid_disks - 2;
+ break;
case 10: data_disks = raid_disks / (layout & 255) / ((layout>>8)&255);
break;
}
- devsize &= ~(unsigned long long)((chunksize>>9)-1);
- return data_disks * devsize;
+
+ return data_disks;
+}
+
+dev_t devnm2devid(char *devnm)
+{
+ /* First look in /sys/block/$DEVNM/dev for %d:%d
+ * If that fails, try parsing out a number
+ */
+ char path[100];
+ char *ep;
+ int fd;
+ int mjr,mnr;
+
+ sprintf(path, "/sys/block/%s/dev", devnm);
+ fd = open(path, O_RDONLY);
+ if (fd >= 0) {
+ char buf[20];
+ int n = read(fd, buf, sizeof(buf));
+ close(fd);
+ if (n > 0)
+ buf[n] = 0;
+ if (n > 0 && sscanf(buf, "%d:%d\n", &mjr, &mnr) == 2)
+ return makedev(mjr, mnr);
+ }
+ if (strncmp(devnm, "md_d", 4) == 0 &&
+ isdigit(devnm[4]) &&
+ (mnr = strtoul(devnm+4, &ep, 10)) >= 0 &&
+ ep > devnm && *ep == 0)
+ return makedev(get_mdp_major(), mnr << MdpMinorShift);
+
+ if (strncmp(devnm, "md", 2) == 0 &&
+ isdigit(devnm[2]) &&
+ (mnr = strtoul(devnm+2, &ep, 10)) >= 0 &&
+ ep > devnm && *ep == 0)
+ return makedev(MD_MAJOR, mnr);
+
+ return 0;
}
-#if !defined(MDASSEMBLE) || defined(MDASSEMBLE) && defined(MDASSEMBLE_AUTO)
-char *get_md_name(int dev)
+char *get_md_name(char *devnm)
{
/* find /dev/md%d or /dev/md/%d or make a device /dev/.tmp.md%d */
/* if dev < 0, want /dev/md/d%d or find mdp in /proc/devices ... */
+
static char devname[50];
struct stat stb;
- dev_t rdev;
+ dev_t rdev = devnm2devid(devnm);
char *dn;
- if (dev < 0) {
- int mdp = get_mdp_major();
- if (mdp < 0) return NULL;
- rdev = makedev(mdp, (-1-dev)<<6);
- snprintf(devname, sizeof(devname), "/dev/md/d%d", -1-dev);
- if (stat(devname, &stb) == 0
- && (S_IFMT&stb.st_mode) == S_IFBLK
- && (stb.st_rdev == rdev))
- return devname;
- } else {
- rdev = makedev(MD_MAJOR, dev);
- snprintf(devname, sizeof(devname), "/dev/md%d", dev);
- if (stat(devname, &stb) == 0
- && (S_IFMT&stb.st_mode) == S_IFBLK
- && (stb.st_rdev == rdev))
- return devname;
-
- snprintf(devname, sizeof(devname), "/dev/md/%d", dev);
+ if (rdev == 0)
+ return 0;
+ if (strncmp(devnm, "md_", 3) == 0) {
+ snprintf(devname, sizeof(devname), "/dev/md/%s",
+ devnm + 3);
if (stat(devname, &stb) == 0
&& (S_IFMT&stb.st_mode) == S_IFBLK
&& (stb.st_rdev == rdev))
return devname;
}
+ snprintf(devname, sizeof(devname), "/dev/%s", devnm);
+ if (stat(devname, &stb) == 0
+ && (S_IFMT&stb.st_mode) == S_IFBLK
+ && (stb.st_rdev == rdev))
+ return devname;
+
+ snprintf(devname, sizeof(devname), "/dev/md/%s", devnm+2);
+ if (stat(devname, &stb) == 0
+ && (S_IFMT&stb.st_mode) == S_IFBLK
+ && (stb.st_rdev == rdev))
+ return devname;
+
dn = map_dev(major(rdev), minor(rdev), 0);
if (dn)
return dn;
- snprintf(devname, sizeof(devname), "/dev/.tmp.md%d", dev);
+ snprintf(devname, sizeof(devname), "/dev/.tmp.%s", devnm);
if (mknod(devname, S_IFBLK | 0600, rdev) == -1)
if (errno != EEXIST)
return NULL;
void put_md_name(char *name)
{
- if (strncmp(name, "/dev/.tmp.md", 12)==0)
+ if (strncmp(name, "/dev/.tmp.md", 12) == 0)
unlink(name);
}
-int find_free_devnum(int use_partitions)
+int get_maj_min(char *dev, int *major, int *minor)
{
- int devnum;
- for (devnum = 127; devnum != 128;
- devnum = devnum ? devnum-1 : (1<<20)-1) {
- char *dn;
- int _devnum;
-
- _devnum = use_partitions ? (-1-devnum) : devnum;
- if (mddev_busy(_devnum))
- continue;
- /* make sure it is new to /dev too, at least as a
- * non-standard */
- dn = map_dev(dev2major(_devnum), dev2minor(_devnum), 0);
- if (dn && ! is_standard(dn, NULL))
- continue;
- break;
- }
- if (devnum == 128)
- return NoMdDev;
- return use_partitions ? (-1-devnum) : devnum;
+ char *e;
+ *major = strtoul(dev, &e, 0);
+ return (e > dev && *e == ':' && e[1] &&
+ (*minor = strtoul(e+1, &e, 0)) >= 0 &&
+ *e == 0);
}
-#endif /* !defined(MDASSEMBLE) || defined(MDASSEMBLE) && defined(MDASSEMBLE_AUTO) */
int dev_open(char *dev, int flags)
{
/* like 'open', but if 'dev' matches %d:%d, create a temp
* block device and open that
*/
- char *e;
int fd = -1;
char devname[32];
int major;
int minor;
- if (!dev) return -1;
+ if (!dev)
+ return -1;
flags |= O_DIRECT;
- major = strtoul(dev, &e, 0);
- if (e > dev && *e == ':' && e[1] &&
- (minor = strtoul(e+1, &e, 0)) >= 0 &&
- *e == 0) {
- char *path = map_dev(major, minor, 0);
- if (path)
- fd = open(path, flags);
- if (fd < 0) {
- snprintf(devname, sizeof(devname), "/dev/.tmp.md.%d:%d:%d",
- (int)getpid(), major, minor);
- if (mknod(devname, S_IFBLK|0600, makedev(major, minor))==0) {
- fd = open(devname, flags);
- unlink(devname);
- }
+ if (get_maj_min(dev, &major, &minor)) {
+ snprintf(devname, sizeof(devname), "/dev/.tmp.md.%d:%d:%d",
+ (int)getpid(), major, minor);
+ if (mknod(devname, S_IFBLK|0600, makedev(major, minor)) == 0) {
+ fd = open(devname, flags);
+ unlink(devname);
}
if (fd < 0) {
+ /* Try /tmp as /dev appear to be read-only */
snprintf(devname, sizeof(devname), "/tmp/.tmp.md.%d:%d:%d",
(int)getpid(), major, minor);
- if (mknod(devname, S_IFBLK|0600, makedev(major, minor))==0) {
+ if (mknod(devname, S_IFBLK|0600, makedev(major, minor)) == 0) {
fd = open(devname, flags);
unlink(devname);
}
return fd;
}
-int open_dev_flags(int devnum, int flags)
+int open_dev_flags(char *devnm, int flags)
{
+ dev_t devid;
char buf[20];
- sprintf(buf, "%d:%d", dev2major(devnum), dev2minor(devnum));
+ devid = devnm2devid(devnm);
+ sprintf(buf, "%d:%d", major(devid), minor(devid));
return dev_open(buf, flags);
}
-int open_dev(int devnum)
+int open_dev(char *devnm)
{
- return open_dev_flags(devnum, O_RDONLY);
+ return open_dev_flags(devnm, O_RDONLY);
}
-int open_dev_excl(int devnum)
+int open_dev_excl(char *devnm)
{
char buf[20];
int i;
int flags = O_RDWR;
+ dev_t devid = devnm2devid(devnm);
+ long delay = 1000;
- sprintf(buf, "%d:%d", dev2major(devnum), dev2minor(devnum));
- for (i=0 ; i<25 ; i++) {
+ sprintf(buf, "%d:%d", major(devid), minor(devid));
+ for (i = 0; i < 25; i++) {
int fd = dev_open(buf, flags|O_EXCL);
if (fd >= 0)
return fd;
}
if (errno != EBUSY)
return fd;
- usleep(200000);
+ usleep(delay);
+ if (delay < 200000)
+ delay *= 2;
}
return -1;
}
{
int i;
struct stat stb_want;
+ long delay = 1000;
if (fstat(fd, &stb_want) != 0 ||
(stb_want.st_mode & S_IFMT) != S_IFBLK)
return;
- for (i=0 ; i<25 ; i++) {
+ for (i = 0; i < 25; i++) {
struct stat stb;
if (stat(dev, &stb) == 0 &&
(stb.st_mode & S_IFMT) == S_IFBLK &&
(stb.st_rdev == stb_want.st_rdev))
return;
- usleep(200000);
+ usleep(delay);
+ if (delay < 200000)
+ delay *= 2;
}
if (i == 25)
- dprintf("%s: timeout waiting for %s\n", __func__, dev);
+ dprintf("timeout waiting for %s\n", dev);
}
struct superswitch *superlist[] =
&super0, &super1,
&super_ddf, &super_imsm,
&mbr, &gpt,
- NULL };
-
-#if !defined(MDASSEMBLE) || defined(MDASSEMBLE) && defined(MDASSEMBLE_AUTO)
+ NULL
+};
struct supertype *super_by_fd(int fd, char **subarrayp)
{
char version[20];
int i;
char *subarray = NULL;
- int container = NoMdDev;
+ char container[32] = "";
- sra = sysfs_read(fd, 0, GET_VERSION);
+ sra = sysfs_read(fd, NULL, GET_VERSION);
if (sra) {
vers = sra->array.major_version;
minor = sra->array.minor_version;
verstr = sra->text_version;
} else {
- if (ioctl(fd, GET_ARRAY_INFO, &array))
+ if (md_get_array_info(fd, &array))
array.major_version = array.minor_version = 0;
vers = array.major_version;
minor = array.minor_version;
char *dev = verstr+1;
subarray = strchr(dev, '/');
- if (subarray)
+ if (subarray) {
*subarray++ = '\0';
- subarray = strdup(subarray);
- container = devname2devnum(dev);
- if (sra)
- sysfs_free(sra);
+ subarray = xstrdup(subarray);
+ }
+ strcpy(container, dev);
+ sysfs_free(sra);
sra = sysfs_read(-1, container, GET_VERSION);
if (sra && sra->text_version[0])
verstr = sra->text_version;
verstr = "-no-metadata-";
}
- for (i = 0; st == NULL && superlist[i] ; i++)
+ for (i = 0; st == NULL && superlist[i]; i++)
st = superlist[i]->match_metadata_desc(verstr);
- if (sra)
- sysfs_free(sra);
+ sysfs_free(sra);
if (st) {
st->sb = NULL;
if (subarrayp)
*subarrayp = subarray;
- st->container_dev = container;
- st->devnum = fd2devnum(fd);
+ strcpy(st->container_devnm, container);
+ strcpy(st->devnm, fd2devnm(fd));
} else
free(subarray);
return st;
}
-#endif /* !defined(MDASSEMBLE) || defined(MDASSEMBLE) && defined(MDASSEMBLE_AUTO) */
int dev_size_from_id(dev_t id, unsigned long long *size)
{
if (!orig)
return orig;
- st = malloc(sizeof(*st));
- if (!st)
- return st;
- memset(st, 0, sizeof(*st));
+ st = xcalloc(1, sizeof(*st));
st->ss = orig->ss;
st->max_devs = orig->max_devs;
st->minor_version = orig->minor_version;
+ st->ignore_hw_compat = orig->ignore_hw_compat;
+ st->data_offset = orig->data_offset;
st->sb = NULL;
st->info = NULL;
return st;
*/
struct superswitch *ss;
struct supertype *st;
- time_t besttime = 0;
+ unsigned int besttime = 0;
int bestsuper = -1;
int i;
- st = malloc(sizeof(*st));
- memset(st, 0, sizeof(*st));
- st->container_dev = NoMdDev;
+ st = xcalloc(1, sizeof(*st));
+ st->container_devnm[0] = 0;
- for (i=0 ; superlist[i]; i++) {
+ for (i = 0; superlist[i]; i++) {
int rv;
ss = superlist[i];
if (guess_type == guess_array && ss->add_to_super == NULL)
rv = superlist[bestsuper]->load_super(st, fd, NULL);
if (rv == 0) {
superlist[bestsuper]->free_super(st);
- st->ignore_hw_compat = 0;
return st;
}
}
ldsize <<= 9;
} else {
if (dname)
- fprintf(stderr, Name ": Cannot get size of %s: %s\b",
+ pr_err("Cannot get size of %s: %s\n",
dname, strerror(errno));
return 0;
}
return 1;
}
+/* Return sector size of device in bytes */
+int get_dev_sector_size(int fd, char *dname, unsigned int *sectsizep)
+{
+ unsigned int sectsize;
+
+ if (ioctl(fd, BLKSSZGET, §size) != 0) {
+ if (dname)
+ pr_err("Cannot get sector size of %s: %s\n",
+ dname, strerror(errno));
+ return 0;
+ }
+
+ *sectsizep = sectsize;
+ return 1;
+}
+
/* Return true if this can only be a container, not a member device.
* i.e. is and md device and size is zero
*/
int must_be_container(int fd)
{
+ struct mdinfo *mdi;
unsigned long long size;
- if (md_get_version(fd) < 0)
+
+ mdi = sysfs_read(fd, NULL, GET_VERSION);
+ if (!mdi)
return 0;
+ sysfs_free(mdi);
+
if (get_dev_size(fd, NULL, &size) == 0)
return 1;
if (size == 0)
{
struct GPT gpt;
unsigned char empty_gpt_entry[16]= {0};
- struct GPT_part_entry part;
+ struct GPT_part_entry *part;
+ char buf[512];
unsigned long long curr_part_end;
unsigned all_partitions, entry_size;
unsigned part_nr;
+ unsigned int sector_size = 0;
*endofpart = 0;
BUILD_BUG_ON(sizeof(gpt) != 512);
/* skip protective MBR */
- lseek(fd, 512, SEEK_SET);
+ if (!get_dev_sector_size(fd, NULL, §or_size))
+ return 0;
+ lseek(fd, sector_size, SEEK_SET);
/* read GPT header */
if (read(fd, &gpt, 512) != 512)
return 0;
/* sanity checks */
if (all_partitions > 1024 ||
- entry_size > 512)
+ entry_size > sizeof(buf))
return -1;
- for (part_nr=0; part_nr < all_partitions; part_nr++) {
+ part = (struct GPT_part_entry *)buf;
+
+ /* set offset to third block (GPT entries) */
+ lseek(fd, sector_size*2, SEEK_SET);
+ for (part_nr = 0; part_nr < all_partitions; part_nr++) {
/* read partition entry */
- if (read(fd, &part, entry_size) != (ssize_t)entry_size)
+ if (read(fd, buf, entry_size) != (ssize_t)entry_size)
return 0;
/* is this valid partition? */
- if (memcmp(part.type_guid, empty_gpt_entry, 16) != 0) {
+ if (memcmp(part->type_guid, empty_gpt_entry, 16) != 0) {
/* check the last lba for the current partition */
- curr_part_end = __le64_to_cpu(part.ending_lba);
+ curr_part_end = __le64_to_cpu(part->ending_lba);
if (curr_part_end > *endofpart)
*endofpart = curr_part_end;
}
static int get_last_partition_end(int fd, unsigned long long *endofpart)
{
struct MBR boot_sect;
- struct MBR_part_record *part;
unsigned long long curr_part_end;
unsigned part_nr;
+ unsigned int sector_size;
int retval = 0;
*endofpart = 0;
if (boot_sect.magic == MBR_SIGNATURE_MAGIC) {
retval = 1;
/* found the correct signature */
- part = boot_sect.parts;
- for (part_nr=0; part_nr < MBR_PARTITIONS; part_nr++) {
+ for (part_nr = 0; part_nr < MBR_PARTITIONS; part_nr++) {
+ /*
+ * Have to make every access through boot_sect rather
+ * than using a pointer to the partition table (or an
+ * entry), since the entries are not properly aligned.
+ */
+
/* check for GPT type */
- if (part->part_type == MBR_GPT_PARTITION_TYPE) {
+ if (boot_sect.parts[part_nr].part_type ==
+ MBR_GPT_PARTITION_TYPE) {
retval = get_gpt_last_partition_end(fd, endofpart);
break;
}
/* check the last used lba for the current partition */
- curr_part_end = __le32_to_cpu(part->first_sect_lba) +
- __le32_to_cpu(part->blocks_num);
+ curr_part_end =
+ __le32_to_cpu(boot_sect.parts[part_nr].first_sect_lba) +
+ __le32_to_cpu(boot_sect.parts[part_nr].blocks_num);
if (curr_part_end > *endofpart)
*endofpart = curr_part_end;
-
- part++;
}
} else {
/* Unknown partition table */
retval = -1;
}
+ /* calculate number of 512-byte blocks */
+ if (get_dev_sector_size(fd, NULL, §or_size))
+ *endofpart *= (sector_size / 512);
abort:
return retval;
}
* Check where the last partition ends
*/
unsigned long long endofpart;
- int ret;
- if ((ret = get_last_partition_end(fd, &endofpart)) > 0) {
+ if (get_last_partition_end(fd, &endofpart) > 0) {
/* There appears to be a partition table here */
if (freesize == 0) {
/* partitions will not be visible in new device */
- fprintf(stderr,
- Name ": partition table exists on %s but will be lost or\n"
- " meaningless after creating array\n",
- dname);
+ pr_err("partition table exists on %s but will be lost or\n"
+ " meaningless after creating array\n",
+ dname);
return 1;
} else if (endofpart > freesize) {
/* last partition overlaps metadata */
- fprintf(stderr,
- Name ": metadata will over-write last partition on %s.\n",
- dname);
+ pr_err("metadata will over-write last partition on %s.\n",
+ dname);
return 1;
} else if (size && endofpart > size) {
/* partitions will be truncated in new device */
- fprintf(stderr,
- Name ": array size is too small to cover all partitions on %s.\n",
- dname);
+ pr_err("array size is too small to cover all partitions on %s.\n",
+ dname);
return 1;
}
}
return 0;
}
-void get_one_disk(int mdfd, mdu_array_info_t *ainf, mdu_disk_info_t *disk)
-{
- int d;
-
- ioctl(mdfd, GET_ARRAY_INFO, ainf);
- for (d = 0 ; d < 1024 ; d++) {
- if (ioctl(mdfd, GET_DISK_INFO, disk) == 0 &&
- (disk->major || disk->minor))
- return;
- }
-}
-
int open_container(int fd)
{
/* 'fd' is a block device. Find out if it is in use
continue;
if (de->d_name[0] == '.')
continue;
+ /* Need to make sure it is a container and not a volume */
+ sprintf(e, "/%s/md/metadata_version", de->d_name);
+ dfd = open(path, O_RDONLY);
+ if (dfd < 0)
+ continue;
+ n = read(dfd, buf, sizeof(buf));
+ close(dfd);
+ if (n <= 0 || (unsigned)n >= sizeof(buf))
+ continue;
+ buf[n] = 0;
+ if (strncmp(buf, "external", 8) != 0 ||
+ n < 10 ||
+ buf[9] == '/')
+ continue;
sprintf(e, "/%s/dev", de->d_name);
dfd = open(path, O_RDONLY);
if (dfd < 0)
return NULL;
}
+int metadata_container_matches(char *metadata, char *devnm)
+{
+ /* Check if 'devnm' is the container named in 'metadata'
+ * which is
+ * /containername/componentname or
+ * -containername/componentname
+ */
+ int l;
+ if (*metadata != '/' && *metadata != '-')
+ return 0;
+ l = strlen(devnm);
+ if (strncmp(metadata+1, devnm, l) != 0)
+ return 0;
+ if (metadata[l+1] != '/')
+ return 0;
+ return 1;
+}
+
+int metadata_subdev_matches(char *metadata, char *devnm)
+{
+ /* Check if 'devnm' is the subdev named in 'metadata'
+ * which is
+ * /containername/subdev or
+ * -containername/subdev
+ */
+ char *sl;
+ if (*metadata != '/' && *metadata != '-')
+ return 0;
+ sl = strchr(metadata+1, '/');
+ if (!sl)
+ return 0;
+ if (strcmp(sl+1, devnm) == 0)
+ return 1;
+ return 0;
+}
+
int is_container_member(struct mdstat_ent *mdstat, char *container)
{
if (mdstat->metadata_version == NULL ||
strncmp(mdstat->metadata_version, "external:", 9) != 0 ||
- !is_subarray(mdstat->metadata_version+9) ||
- strncmp(mdstat->metadata_version+10, container, strlen(container)) != 0 ||
- mdstat->metadata_version[10+strlen(container)] != '/')
+ !metadata_container_matches(mdstat->metadata_version+9, container))
return 0;
return 1;
struct mdinfo *mdi;
struct mdinfo *info;
int fd, err = 1;
+ char *_devnm;
fd = open(dev, O_RDWR|O_EXCL);
if (fd < 0) {
if (!quiet)
- fprintf(stderr, Name ": Couldn't open %s, aborting\n",
+ pr_err("Couldn't open %s, aborting\n",
dev);
- return 2;
+ return -1;
}
- st->devnum = fd2devnum(fd);
- if (st->devnum == NoMdDev) {
+ _devnm = fd2devnm(fd);
+ if (_devnm == NULL) {
if (!quiet)
- fprintf(stderr,
- Name ": Failed to determine device number for %s\n",
- dev);
+ pr_err("Failed to determine device number for %s\n",
+ dev);
goto close_fd;
}
+ strcpy(st->devnm, _devnm);
- mdi = sysfs_read(fd, st->devnum, GET_VERSION|GET_LEVEL);
+ mdi = sysfs_read(fd, st->devnm, GET_VERSION|GET_LEVEL);
if (!mdi) {
if (!quiet)
- fprintf(stderr, Name ": Failed to read sysfs for %s\n",
+ pr_err("Failed to read sysfs for %s\n",
dev);
goto close_fd;
}
if (mdi->array.level != UnSet) {
if (!quiet)
- fprintf(stderr, Name ": %s is not a container\n", dev);
+ pr_err("%s is not a container\n", dev);
goto free_sysfs;
}
st->ss = version_to_superswitch(mdi->text_version);
if (!st->ss) {
if (!quiet)
- fprintf(stderr,
- Name ": Operation not supported for %s metadata\n",
- mdi->text_version);
+ pr_err("Operation not supported for %s metadata\n",
+ mdi->text_version);
goto free_sysfs;
}
- st->devname = devnum2devname(st->devnum);
- if (!st->devname) {
+ if (st->devnm[0] == 0) {
if (!quiet)
- fprintf(stderr, Name ": Failed to allocate device name\n");
+ pr_err("Failed to allocate device name\n");
goto free_sysfs;
}
if (!st->ss->load_container) {
if (!quiet)
- fprintf(stderr, Name ": %s is not a container\n", dev);
- goto free_name;
+ pr_err("%s is not a container\n", dev);
+ goto free_sysfs;
}
if (st->ss->load_container(st, fd, NULL)) {
if (!quiet)
- fprintf(stderr, Name ": Failed to load metadata for %s\n",
+ pr_err("Failed to load metadata for %s\n",
dev);
- goto free_name;
+ goto free_sysfs;
}
info = st->ss->container_content(st, subarray);
if (!info) {
if (!quiet)
- fprintf(stderr, Name ": Failed to find subarray-%s in %s\n",
+ pr_err("Failed to find subarray-%s in %s\n",
subarray, dev);
goto free_super;
}
free_super:
if (err)
st->ss->free_super(st);
- free_name:
- if (err)
- free(st->devname);
free_sysfs:
sysfs_free(mdi);
close_fd:
{
/* Add a device to an array, in one of 2 ways. */
int rv;
-#ifndef MDASSEMBLE
+
if (st->ss->external) {
if (info->disk.state & (1<<MD_DISK_SYNC))
info->recovery_start = MaxSector;
if (sd2 == info)
break;
if (sd2 == NULL) {
- sd2 = malloc(sizeof(*sd2));
+ sd2 = xmalloc(sizeof(*sd2));
*sd2 = *info;
sd2->next = sra->devs;
sra->devs = sd2;
}
}
} else
-#endif
rv = ioctl(mdfd, ADD_NEW_DISK, &info->disk);
return rv;
}
struct mdinfo *sra, struct mdinfo *info)
{
int rv;
+
/* Remove the disk given by 'info' from the array */
-#ifndef MDASSEMBLE
if (st->ss->external)
rv = sysfs_set_str(sra, info, "slot", "none");
else
-#endif
rv = ioctl(mdfd, HOT_REMOVE_DISK, makedev(info->disk.major,
info->disk.minor));
return rv;
}
+int hot_remove_disk(int mdfd, unsigned long dev, int force)
+{
+ int cnt = force ? 500 : 5;
+ int ret;
+
+ /* HOT_REMOVE_DISK can fail with EBUSY if there are
+ * outstanding IO requests to the device.
+ * In this case, it can be helpful to wait a little while,
+ * up to 5 seconds if 'force' is set, or 50 msec if not.
+ */
+ while ((ret = ioctl(mdfd, HOT_REMOVE_DISK, dev)) == -1 &&
+ errno == EBUSY &&
+ cnt-- > 0)
+ usleep(10000);
+
+ return ret;
+}
+
+int sys_hot_remove_disk(int statefd, int force)
+{
+ int cnt = force ? 500 : 5;
+ int ret;
+
+ while ((ret = write(statefd, "remove", 6)) == -1 &&
+ errno == EBUSY &&
+ cnt-- > 0)
+ usleep(10000);
+ return ret == 6 ? 0 : -1;
+}
+
int set_array_info(int mdfd, struct supertype *st, struct mdinfo *info)
{
/* Initialise kernel's knowledge of array.
* This varies between externally managed arrays
* and older kernels
*/
- int vers = md_get_version(mdfd);
+ mdu_array_info_t inf;
int rv;
-#ifndef MDASSEMBLE
if (st->ss->external)
- rv = sysfs_set_array(info, vers);
- else
-#endif
- if ((vers % 100) >= 1) { /* can use different versions */
- mdu_array_info_t inf;
- memset(&inf, 0, sizeof(inf));
- inf.major_version = info->array.major_version;
- inf.minor_version = info->array.minor_version;
- rv = ioctl(mdfd, SET_ARRAY_INFO, &inf);
- } else
- rv = ioctl(mdfd, SET_ARRAY_INFO, NULL);
+ return sysfs_set_array(info, 9003);
+
+ memset(&inf, 0, sizeof(inf));
+ inf.major_version = info->array.major_version;
+ inf.minor_version = info->array.minor_version;
+ rv = md_set_array_info(mdfd, &inf);
+
return rv;
}
return recovery_start;
}
-int mdmon_pid(int devnum)
+int mdmon_pid(char *devnm)
{
char path[100];
char pid[10];
int fd;
int n;
- char *devname = devnum2devname(devnum);
- sprintf(path, "%s/%s.pid", MDMON_DIR, devname);
- free(devname);
+ sprintf(path, "%s/%s.pid", MDMON_DIR, devnm);
fd = open(path, O_RDONLY | O_NOATIME, 0);
return atoi(pid);
}
-int mdmon_running(int devnum)
+int mdmon_running(char *devnm)
{
- int pid = mdmon_pid(devnum);
+ int pid = mdmon_pid(devnm);
if (pid <= 0)
return 0;
if (kill(pid, 0) == 0)
return 0;
}
-int start_mdmon(int devnum)
+int start_mdmon(char *devnm)
{
int i, skipped;
int len;
- pid_t pid;
+ pid_t pid;
int status;
char pathbuf[1024];
char *paths[4] = {
pathbuf,
- "/sbin/mdmon",
- "mdmon",
+ BINDIR "/mdmon",
+ "./mdmon",
NULL
};
if (check_env("MDADM_NO_MDMON"))
return 0;
- len = readlink("/proc/self/exe", pathbuf, sizeof(pathbuf));
+ len = readlink("/proc/self/exe", pathbuf, sizeof(pathbuf)-1);
if (len > 0) {
char *sl;
pathbuf[len] = 0;
} else
pathbuf[0] = '\0';
+ /* First try to run systemctl */
+ if (!check_env("MDADM_NO_SYSTEMCTL"))
+ switch(fork()) {
+ case 0:
+ /* FIXME yuk. CLOSE_EXEC?? */
+ skipped = 0;
+ for (i = 3; skipped < 20; i++)
+ if (close(i) < 0)
+ skipped++;
+ else
+ skipped = 0;
+
+ /* Don't want to see error messages from
+ * systemctl. If the service doesn't exist,
+ * we start mdmon ourselves.
+ */
+ close(2);
+ open("/dev/null", O_WRONLY);
+ snprintf(pathbuf, sizeof(pathbuf), "mdmon@%s.service",
+ devnm);
+ status = execl("/usr/bin/systemctl", "systemctl",
+ "start",
+ pathbuf, NULL);
+ status = execl("/bin/systemctl", "systemctl", "start",
+ pathbuf, NULL);
+ exit(1);
+ case -1: pr_err("cannot run mdmon. Array remains readonly\n");
+ return -1;
+ default: /* parent - good */
+ pid = wait(&status);
+ if (pid >= 0 && status == 0)
+ return 0;
+ }
+
+ /* That failed, try running mdmon directly */
switch(fork()) {
case 0:
/* FIXME yuk. CLOSE_EXEC?? */
skipped = 0;
- for (i=3; skipped < 20; i++)
+ for (i = 3; skipped < 20; i++)
if (close(i) < 0)
skipped++;
else
skipped = 0;
- for (i=0; paths[i]; i++)
- if (paths[i][0])
- execl(paths[i], "mdmon",
- devnum2devname(devnum),
- NULL);
+ for (i = 0; paths[i]; i++)
+ if (paths[i][0]) {
+ execl(paths[i], paths[i],
+ devnm, NULL);
+ }
exit(1);
- case -1: fprintf(stderr, Name ": cannot run mdmon. "
- "Array remains readonly\n");
+ case -1: pr_err("cannot run mdmon. Array remains readonly\n");
return -1;
default: /* parent - good */
pid = wait(&status);
- if (pid < 0 || status != 0)
+ if (pid < 0 || status != 0) {
+ pr_err("failed to launch mdmon. Array remains readonly\n");
return -1;
+ }
}
return 0;
}
-int check_env(char *name)
-{
- char *val = getenv(name);
-
- if (val && atoi(val) == 1)
- return 1;
-
- return 0;
-}
-
__u32 random32(void)
{
__u32 rv;
return rv;
}
-#ifndef MDASSEMBLE
+void random_uuid(__u8 *buf)
+{
+ int fd, i, len;
+ __u32 r[4];
+
+ fd = open("/dev/urandom", O_RDONLY);
+ if (fd < 0)
+ goto use_random;
+ len = read(fd, buf, 16);
+ close(fd);
+ if (len != 16)
+ goto use_random;
+
+ return;
+
+use_random:
+ for (i = 0; i < 4; i++)
+ r[i] = random();
+ memcpy(buf, r, 16);
+}
+
int flush_metadata_updates(struct supertype *st)
{
int sfd;
return -1;
}
- sfd = connect_monitor(devnum2devname(st->container_dev));
+ sfd = connect_monitor(st->container_devnm);
if (sfd < 0)
return -1;
void append_metadata_update(struct supertype *st, void *buf, int len)
{
- struct metadata_update *mu = malloc(sizeof(*mu));
+ struct metadata_update *mu = xmalloc(sizeof(*mu));
mu->buf = buf;
mu->len = len;
*st->update_tail = mu;
st->update_tail = &mu->next;
}
-#endif /* MDASSEMBLE */
#ifdef __TINYC__
/* tinyc doesn't optimize this check in ioctl.h out ... */
if (check_env("MDADM_EXPERIMENTAL"))
return 1;
else {
- fprintf(stderr, Name ": To use this feature MDADM_EXPERIMENTAL enviroment variable has to defined.\n");
+ pr_err("To use this feature MDADM_EXPERIMENTAL environment variable has to be defined.\n");
return 0;
}
}
found = 1;
/* check if domain matches */
if (found && domlist) {
- struct dev_policy *pol = devnum_policy(dev);
+ struct dev_policy *pol = devid_policy(dev);
if (spare_group)
pol_add(&pol, pol_domain,
spare_group, NULL);
}
return disks;
}
+
+/* Checks if paths point to the same device
+ * Returns 0 if they do.
+ * Returns 1 if they don't.
+ * Returns -1 if something went wrong,
+ * e.g. paths are empty or the files
+ * they point to don't exist */
+int compare_paths (char* path1, char* path2)
+{
+ struct stat st1,st2;
+
+ if (path1 == NULL || path2 == NULL)
+ return -1;
+ if (stat(path1,&st1) != 0)
+ return -1;
+ if (stat(path2,&st2) != 0)
+ return -1;
+ if ((st1.st_ino == st2.st_ino) && (st1.st_dev == st2.st_dev))
+ return 0;
+ return 1;
+}
+
+/* Make sure we can open as many devices as needed */
+void enable_fds(int devices)
+{
+ unsigned int fds = 20 + devices;
+ struct rlimit lim;
+ if (getrlimit(RLIMIT_NOFILE, &lim) != 0
+ || lim.rlim_cur >= fds)
+ return;
+ if (lim.rlim_max < fds)
+ lim.rlim_max = fds;
+ lim.rlim_cur = fds;
+ setrlimit(RLIMIT_NOFILE, &lim);
+}
+
+int in_initrd(void)
+{
+ /* This is based on similar function in systemd. */
+ struct statfs s;
+ /* statfs.f_type is signed long on s390x and MIPS, causing all
+ sorts of sign extension problems with RAMFS_MAGIC being
+ defined as 0x858458f6 */
+ return statfs("/", &s) >= 0 &&
+ ((unsigned long)s.f_type == TMPFS_MAGIC ||
+ ((unsigned long)s.f_type & 0xFFFFFFFFUL) ==
+ ((unsigned long)RAMFS_MAGIC & 0xFFFFFFFFUL));
+}
+
+void reopen_mddev(int mdfd)
+{
+ /* Re-open without any O_EXCL, but keep
+ * the same fd
+ */
+ char *devnm;
+ int fd;
+ devnm = fd2devnm(mdfd);
+ close(mdfd);
+ fd = open_dev(devnm);
+ if (fd >= 0 && fd != mdfd)
+ dup2(fd, mdfd);
+}
+
+static struct cmap_hooks *cmap_hooks = NULL;
+static int is_cmap_hooks_ready = 0;
+
+void set_cmap_hooks(void)
+{
+ cmap_hooks = xmalloc(sizeof(struct cmap_hooks));
+ cmap_hooks->cmap_handle = dlopen("libcmap.so.4", RTLD_NOW | RTLD_LOCAL);
+ if (!cmap_hooks->cmap_handle)
+ return;
+
+ cmap_hooks->initialize = dlsym(cmap_hooks->cmap_handle, "cmap_initialize");
+ cmap_hooks->get_string = dlsym(cmap_hooks->cmap_handle, "cmap_get_string");
+ cmap_hooks->finalize = dlsym(cmap_hooks->cmap_handle, "cmap_finalize");
+
+ if (!cmap_hooks->initialize || !cmap_hooks->get_string ||
+ !cmap_hooks->finalize)
+ dlclose(cmap_hooks->cmap_handle);
+ else
+ is_cmap_hooks_ready = 1;
+}
+
+int get_cluster_name(char **cluster_name)
+{
+ int rv = -1;
+ cmap_handle_t handle;
+
+ if (!is_cmap_hooks_ready)
+ return rv;
+
+ rv = cmap_hooks->initialize(&handle);
+ if (rv != CS_OK)
+ goto out;
+
+ rv = cmap_hooks->get_string(handle, "totem.cluster_name", cluster_name);
+ if (rv != CS_OK) {
+ free(*cluster_name);
+ rv = -1;
+ goto name_err;
+ }
+
+ rv = 0;
+name_err:
+ cmap_hooks->finalize(handle);
+out:
+ return rv;
+}
+
+void set_dlm_hooks(void)
+{
+ dlm_hooks = xmalloc(sizeof(struct dlm_hooks));
+ dlm_hooks->dlm_handle = dlopen("libdlm_lt.so.3", RTLD_NOW | RTLD_LOCAL);
+ if (!dlm_hooks->dlm_handle)
+ return;
+
+ dlm_hooks->create_lockspace = dlsym(dlm_hooks->dlm_handle, "dlm_create_lockspace");
+ dlm_hooks->release_lockspace = dlsym(dlm_hooks->dlm_handle, "dlm_release_lockspace");
+ dlm_hooks->ls_lock = dlsym(dlm_hooks->dlm_handle, "dlm_ls_lock");
+ dlm_hooks->ls_unlock = dlsym(dlm_hooks->dlm_handle, "dlm_ls_unlock");
+ dlm_hooks->ls_get_fd = dlsym(dlm_hooks->dlm_handle, "dlm_ls_get_fd");
+ dlm_hooks->dispatch = dlsym(dlm_hooks->dlm_handle, "dlm_dispatch");
+
+ if (!dlm_hooks->create_lockspace || !dlm_hooks->ls_lock ||
+ !dlm_hooks->ls_unlock || !dlm_hooks->release_lockspace ||
+ !dlm_hooks->ls_get_fd || !dlm_hooks->dispatch)
+ dlclose(dlm_hooks->dlm_handle);
+ else
+ is_dlm_hooks_ready = 1;
+}
+
+void set_hooks(void)
+{
+ set_dlm_hooks();
+ set_cmap_hooks();
+}