/*
* mdadm - manage Linux "md" devices aka RAID arrays.
*
- * Copyright (C) 2001-2009 Neil Brown <neilb@suse.de>
+ * Copyright (C) 2001-2012 Neil Brown <neilb@suse.de>
*
*
* This program is free software; you can redistribute it and/or modify
#include <dirent.h>
#include <signal.h>
+int __offroot;
+
/*
* following taken from linux/blkpg.h because they aren't
* anywhere else and it isn't safe to #include linux/ * stuff.
char volname[BLKPG_VOLNAMELTH]; /* volume label */
};
-/* partition table structures so we can check metadata position
- * against the end of the last partition.
- * Only handle MBR ant GPT partition tables.
- */
-struct MBR_part_record {
- __u8 bootable;
- __u8 first_head;
- __u8 first_sector;
- __u8 first_cyl;
- __u8 part_type;
- __u8 last_head;
- __u8 last_sector;
- __u8 last_cyl;
- __u32 first_sect_lba;
- __u32 blocks_num;
-};
-
-struct GPT_part_entry {
- unsigned char type_guid[16];
- unsigned char partition_guid[16];
- unsigned char starting_lba[8];
- unsigned char ending_lba[8];
- unsigned char attr_bits[8];
- unsigned char name[72];
-};
+#include "part.h"
-/* MBR/GPT magic numbers */
-#define MBR_SIGNATURE_MAGIC __cpu_to_le16(0xAA55)
-#define GPT_SIGNATURE_MAGIC __cpu_to_le64(0x5452415020494645ULL)
+/* Force a compilation error if condition is true */
+#define BUILD_BUG_ON(condition) ((void)BUILD_BUG_ON_ZERO(condition))
-#define MBR_SIGNATURE_OFFSET 510
-#define MBR_PARTITION_TABLE_OFFSET 446
-#define MBR_PARTITIONS 4
-#define MBR_GPT_PARTITION_TYPE 0xEE
-#define GPT_ALL_PARTITIONS_OFFSET 80
-#define GPT_ENTRY_SIZE_OFFSET 84
+/* Force a compilation error if condition is true, but also produce a
+ result (of value 0 and type size_t), so the expression can be used
+ e.g. in a structure initializer (or where-ever else comma expressions
+ aren't permitted). */
+#define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); }))
/*
* Parse a 128 bit uuid in 4 integers
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
{
struct utsname name;
char *cp;
- int a,b,c;
+ int a = 0, b = 0,c = 0;
if (uname(&name) <0)
return -1;
cp = name.release;
a = strtoul(cp, &cp, 10);
- if (*cp != '.') return -1;
- b = strtoul(cp+1, &cp, 10);
- if (*cp != '.') return -1;
- c = strtoul(cp+1, NULL, 10);
+ if (*cp == '.')
+ b = strtoul(cp+1, &cp, 10);
+ if (*cp == '.')
+ c = strtoul(cp+1, &cp, 10);
return (a*1000000)+(b*1000)+c;
}
#ifndef MDASSEMBLE
-long long parse_size(char *size)
+int mdadm_version(char *version)
+{
+ int a, b, c;
+ char *cp;
+
+ if (!version)
+ version = Version;
+
+ cp = strchr(version, '-');
+ if (!cp || *(cp+1) != ' ' || *(cp+2) != 'v')
+ return -1;
+ cp += 3;
+ a = strtoul(cp, &cp, 10);
+ if (*cp != '.')
+ return -1;
+ b = strtoul(cp+1, &cp, 10);
+ if (*cp == '.')
+ c = strtoul(cp+1, &cp, 10);
+ else
+ c = 0;
+ if (*cp != ' ' && *cp != '-')
+ return -1;
+ return (a*1000000)+(b*1000)+c;
+}
+
+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)
+ * 0 returned on error.
*/
char *c;
long long s = strtoll(size, &c, 10);
s *= 1024 * 1024 * 2;
break;
}
- }
+ } else
+ s = 0;
if (*c)
s = 0;
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);
}
+
+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;
+}
#endif
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
}
-int enough(int level, int raid_disks, int layout, int clean,
- char *avail, int avail_disks)
+int test_partition(int fd)
+{
+ /* Check if fd is a whole-disk or a partition.
+ * BLKPG will return EINVAL on a partition, and BLKPG_DEL_PARTITION
+ * will return ENXIO on an invalid partition number.
+ */
+ struct blkpg_ioctl_arg a;
+ struct blkpg_partition p;
+ a.op = BLKPG_DEL_PARTITION;
+ a.data = (void*)&p;
+ a.datalen = sizeof(p);
+ a.flags = 0;
+ memset(a.data, 0, a.datalen);
+ p.pno = 1<<30;
+ if (ioctl(fd, BLKPG, &a) == 0)
+ /* Very unlikely, but not a partition */
+ return 0;
+ if (errno == ENXIO)
+ /* not a partition */
+ return 0;
+
+ return 1;
+}
+
+int test_partition_from_id(dev_t id)
+{
+ char buf[20];
+ int fd, rv;
+
+ sprintf(buf, "%d:%d", major(id), minor(id));
+ fd = dev_open(buf, O_RDONLY);
+ if (fd < 0)
+ return -1;
+ rv = test_partition(fd);
+ close(fd);
+ return rv;
+}
+
+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;
while (n--) {
if (avail[first])
cnt++;
}
}
-const int uuid_match_any[4] = { ~0, ~0, ~0, ~0 };
-int same_uuid(int a[4], int b[4], int swapuuid)
+int enough_fd(int fd)
{
- if (memcmp(a, uuid_match_any, sizeof(int[4])) == 0 ||
- memcmp(b, uuid_match_any, sizeof(int[4])) == 0)
- return 1;
+ struct mdu_array_info_s array;
+ struct mdu_disk_info_s disk;
+ int avail_disks = 0;
+ int i, rv;
+ char *avail;
+
+ if (ioctl(fd, GET_ARRAY_INFO, &array) != 0 ||
+ array.raid_disks <= 0)
+ return 0;
+ avail = xcalloc(array.raid_disks, 1);
+ for (i = 0; i < MAX_DISKS && 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 */
+ rv = enough(array.level, array.raid_disks, array.layout,
+ 1, avail);
+ free(avail);
+ return rv;
+}
+
+const int uuid_zero[4] = { 0, 0, 0, 0 };
+int same_uuid(int a[4], int b[4], int swapuuid)
+{
if (swapuuid) {
/* parse uuids are hostendian.
* uuid's from some superblocks are big-ending
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];
char *fname_from_uuid(struct supertype *st, struct mdinfo *info, char *buf, char sep)
{
- return __fname_from_uuid(info->uuid, st->ss->swapuuid, buf, sep);
+ // dirty hack to work around an issue with super1 superblocks...
+ // super1 superblocks need swapuuid set in order for assembly to
+ // work, but can't have it set if we want this printout to match
+ // all the other uuid printouts in super1.c, so we force swapuuid
+ // to 1 to make our printout match the rest of super1
+ return __fname_from_uuid(info->uuid, (st->ss == &super1) ? 1 : st->ss->swapuuid, buf, sep);
}
#ifndef MDASSEMBLE
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",
+ pr_err("%s appears to contain an ext2fs file system\n",
name);
fprintf(stderr," size=%dK mtime=%s",
size*(1<<bsize), ctime(&mtime));
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);
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",
+ pr_err("%s appears to be part of a raid array:\n",
name);
- st->ss->getinfo_super(st, &info);
+ st->ss->getinfo_super(st, &info, NULL);
st->ss->free_super(st);
crtime = info.array.ctime;
level = map_num(pers, info.array.level);
{
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 */
-char *map_num(mapping_t *map, int num)
-{
- while (map->name) {
- if (map->num == num)
- return map->name;
- map++;
- }
- return NULL;
-}
-
-int map_name(mapping_t *map, char *name)
-{
- while (map->name) {
- if (strcmp(map->name, name)==0)
- return map->num;
- map++;
- }
- return UnSet;
-}
-
-
int is_standard(char *dev, int *nump)
{
/* tests if dev is a "standard" md dev name.
* 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)
return type;
}
-
-/*
- * convert a major/minor pair for a block device into a name in /dev, if possible.
- * On the first call, walk /dev collecting name.
- * Put them in a simple linked listfor now.
- */
-struct devmap {
- int major, minor;
- char *name;
- struct devmap *next;
-} *devlist = NULL;
-int devlist_ready = 0;
-
-int add_dev(const char *name, const struct stat *stb, int flag, struct FTW *s)
-{
- struct stat st;
-
- if (S_ISLNK(stb->st_mode)) {
- if (stat(name, &st) != 0)
- return 0;
- stb = &st;
- }
-
- if ((stb->st_mode&S_IFMT)== S_IFBLK) {
- char *n = strdup(name);
- struct devmap *dm = malloc(sizeof(*dm));
- if (strncmp(n, "/dev/./", 7)==0)
- strcpy(n+4, name+6);
- if (dm) {
- dm->major = major(stb->st_rdev);
- dm->minor = minor(stb->st_rdev);
- dm->name = n;
- dm->next = devlist;
- devlist = dm;
- }
- }
- return 0;
-}
-
-#ifndef HAVE_NFTW
-#ifdef HAVE_FTW
-int add_dev_1(const char *name, const struct stat *stb, int flag)
-{
- return add_dev(name, stb, flag, NULL);
-}
-int nftw(const char *path, int (*han)(const char *name, const struct stat *stb, int flag, struct FTW *s), int nopenfd, int flags)
-{
- return ftw(path, add_dev_1, nopenfd);
-}
-#else
-int nftw(const char *path, int (*han)(const char *name, const struct stat *stb, int flag, struct FTW *s), int nopenfd, int flags)
-{
- return 0;
-}
-#endif /* HAVE_FTW */
-#endif /* HAVE_NFTW */
-
-/*
- * Find a block device with the right major/minor number.
- * If we find multiple names, choose the shortest.
- * If we find a name in /dev/md/, we prefer that.
- * This applies only to names for MD devices.
- */
-char *map_dev(int major, int minor, int create)
-{
- struct devmap *p;
- char *regular = NULL, *preferred=NULL;
- int did_check = 0;
-
- if (major == 0 && minor == 0)
- return NULL;
-
- retry:
- if (!devlist_ready) {
- char *dev = "/dev";
- struct stat stb;
- while(devlist) {
- struct devmap *d = devlist;
- devlist = d->next;
- free(d->name);
- free(d);
- }
- if (lstat(dev, &stb)==0 &&
- S_ISLNK(stb.st_mode))
- dev = "/dev/.";
- nftw(dev, add_dev, 10, FTW_PHYS);
- devlist_ready=1;
- did_check = 1;
- }
-
- for (p=devlist; p; p=p->next)
- if (p->major == major &&
- p->minor == minor) {
- if (strncmp(p->name, "/dev/md/",8) == 0) {
- if (preferred == NULL ||
- strlen(p->name) < strlen(preferred))
- preferred = p->name;
- } else {
- if (regular == NULL ||
- strlen(p->name) < strlen(regular))
- regular = p->name;
- }
- }
- if (!regular && !preferred && !did_check) {
- devlist_ready = 0;
- goto retry;
- }
- if (create && !regular && !preferred) {
- static char buf[30];
- snprintf(buf, sizeof(buf), "%d:%d", major, minor);
- regular = buf;
- }
-
- return preferred ? preferred : regular;
-}
-
unsigned long calc_csum(void *super, int bytes)
{
unsigned long long newcsum = 0;
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);
* 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 cMB = (bytes / ( 1000000LL / 200LL ) +1) /2;
{
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 (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)
- );
- else
- snprintf(buf, sizeof(buf), "%ld.%02ldGiB",
- (long)(bytes>>30),
- (long)(((bytes>>10)&0xfffff)+0x100000/200)/(0x100000/100)
- );
+ buf[0] = 0;
+ else if (bytes < 2*1024LL*1024LL*1024LL) {
+ long cMiB = (bytes / ( (1LL<<20) / 200LL ) +1) /2;
+ snprintf(buf, sizeof(buf), " (%ld.%02ldMiB)",
+ cMiB/100 , cMiB % 100);
+ } else {
+ long cGiB = (bytes / ( (1LL<<30) / 200LL ) +1) /2;
+ snprintf(buf, sizeof(buf), " (%ld.%02ldGiB)",
+ cGiB/100 , cGiB % 100);
+ }
+
return buf;
}
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;
-}
-
-int get_mdp_major(void)
-{
-static int mdp_major = -1;
- FILE *fl;
- char *w;
- int have_block = 0;
- int have_devices = 0;
- int last_num = -1;
- if (mdp_major != -1)
- return mdp_major;
- fl = fopen("/proc/devices", "r");
- if (!fl)
- return -1;
- while ((w = conf_word(fl, 1))) {
- if (have_block && strcmp(w, "devices:")==0)
- have_devices = 1;
- have_block = (strcmp(w, "Block")==0);
- if (isdigit(w[0]))
- last_num = atoi(w);
- if (have_devices && strcmp(w, "mdp")==0)
- mdp_major = last_num;
- free(w);
- }
- fclose(fl);
- return mdp_major;
+ return data_disks;
}
#if !defined(MDASSEMBLE) || defined(MDASSEMBLE) && defined(MDASSEMBLE_AUTO)
void put_md_name(char *name)
{
- if (strncmp(name, "/dev/.tmp.md", 12)==0)
+ if (strncmp(name, "/dev/.tmp.md", 12) == 0)
unlink(name);
}
devnum = devnum ? devnum-1 : (1<<20)-1) {
char *dn;
int _devnum;
+ char nbuf[50];
_devnum = use_partitions ? (-1-devnum) : devnum;
if (mddev_busy(_devnum))
continue;
+ sprintf(nbuf, "%s%d", use_partitions?"mdp":"md", devnum);
+ if (!conf_name_is_free(nbuf))
+ continue;
/* make sure it is new to /dev too, at least as a
* non-standard */
dn = map_dev(dev2major(_devnum), dev2minor(_devnum), 0);
int minor;
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) {
- 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|O_DIRECT);
- unlink(devname);
+ 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 (fd < 0) {
+ snprintf(devname, sizeof(devname), "/tmp/.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);
+ }
}
} else
- fd = open(dev, flags|O_DIRECT);
+ fd = open(dev, flags);
return fd;
}
-int open_dev(int devnum)
+int open_dev_flags(int devnum, int flags)
{
char buf[20];
sprintf(buf, "%d:%d", dev2major(devnum), dev2minor(devnum));
- return dev_open(buf, O_RDWR);
+ return dev_open(buf, flags);
+}
+
+int open_dev(int devnum)
+{
+ return open_dev_flags(devnum, O_RDONLY);
}
int open_dev_excl(int devnum)
{
char buf[20];
int i;
+ int flags = O_RDWR;
sprintf(buf, "%d:%d", dev2major(devnum), dev2minor(devnum));
- for (i=0 ; i<25 ; i++) {
- int fd = dev_open(buf, O_RDWR|O_EXCL);
+ for (i = 0 ; i < 25 ; i++) {
+ int fd = dev_open(buf, flags|O_EXCL);
if (fd >= 0)
return fd;
+ if (errno == EACCES && flags == O_RDWR) {
+ flags = O_RDONLY;
+ continue;
+ }
if (errno != EBUSY)
return fd;
usleep(200000);
(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 &&
dprintf("%s: timeout waiting for %s\n", __func__, dev);
}
-struct superswitch *superlist[] = { &super0, &super1, &super_ddf, &super_imsm, NULL };
+struct superswitch *superlist[] =
+{
+ &super0, &super1,
+ &super_ddf, &super_imsm,
+ &mbr, &gpt,
+ NULL };
#if !defined(MDASSEMBLE) || defined(MDASSEMBLE) && defined(MDASSEMBLE_AUTO)
-struct supertype *super_by_fd(int fd)
+struct supertype *super_by_fd(int fd, char **subarrayp)
{
mdu_array_info_t array;
int vers;
char version[20];
int i;
char *subarray = NULL;
+ int container = NoMdDev;
sra = sysfs_read(fd, 0, GET_VERSION);
}
if (minor == -2 && is_subarray(verstr)) {
char *dev = verstr+1;
+
subarray = strchr(dev, '/');
- int devnum;
- if (subarray)
+ if (subarray) {
*subarray++ = '\0';
- devnum = devname2devnum(dev);
- subarray = strdup(subarray);
+ subarray = xstrdup(subarray);
+ }
+ container = devname2devnum(dev);
if (sra)
sysfs_free(sra);
- sra = sysfs_read(-1, devnum, GET_VERSION);
+ sra = sysfs_read(-1, container, GET_VERSION);
if (sra && sra->text_version[0])
verstr = sra->text_version;
else
sysfs_free(sra);
if (st) {
st->sb = NULL;
- if (subarray) {
- strncpy(st->subarray, subarray, 32);
- st->subarray[31] = 0;
- free(subarray);
- } else
- st->subarray[0] = 0;
- }
+ if (subarrayp)
+ *subarrayp = subarray;
+ st->container_dev = container;
+ st->devnum = fd2devnum(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)
+{
+ char buf[20];
+ int fd;
+
+ sprintf(buf, "%d:%d", major(id), minor(id));
+ fd = dev_open(buf, O_RDONLY);
+ if (fd < 0)
+ return 0;
+ if (get_dev_size(fd, NULL, size)) {
+ close(fd);
+ return 1;
+ }
+ close(fd);
+ return 0;
+}
struct supertype *dup_super(struct supertype *orig)
{
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;
- strcpy(st->subarray, orig->subarray);
st->sb = NULL;
st->info = NULL;
return st;
}
-struct supertype *guess_super(int fd)
+struct supertype *guess_super_type(int fd, enum guess_types guess_type)
{
/* try each load_super to find the best match,
* and return the best superswitch
*/
struct superswitch *ss;
struct supertype *st;
- unsigned long besttime = 0;
+ time_t besttime = 0;
int bestsuper = -1;
int i;
- st = malloc(sizeof(*st));
- for (i=0 ; superlist[i]; i++) {
+ st = xcalloc(1, sizeof(*st));
+ st->container_dev = NoMdDev;
+
+ for (i = 0 ; superlist[i]; i++) {
int rv;
ss = superlist[i];
+ if (guess_type == guess_array && ss->add_to_super == NULL)
+ continue;
+ if (guess_type == guess_partitions && ss->add_to_super != NULL)
+ continue;
memset(st, 0, sizeof(*st));
+ st->ignore_hw_compat = 1;
rv = ss->load_super(st, fd, NULL);
if (rv == 0) {
struct mdinfo info;
- st->ss->getinfo_super(st, &info);
+ st->ss->getinfo_super(st, &info, NULL);
if (bestsuper == -1 ||
besttime < info.array.ctime) {
bestsuper = i;
if (bestsuper != -1) {
int rv;
memset(st, 0, sizeof(*st));
+ st->ignore_hw_compat = 1;
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\b",
dname, strerror(errno));
return 0;
}
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)
+{
+ unsigned long long size;
+ if (md_get_version(fd) < 0)
+ return 0;
+ if (get_dev_size(fd, NULL, &size) == 0)
+ return 1;
+ if (size == 0)
+ return 1;
+ return 0;
+}
/* Sets endofpart parameter to the last block used by the last GPT partition on the device.
* Returns: 1 if successful
*/
static int get_gpt_last_partition_end(int fd, unsigned long long *endofpart)
{
- unsigned char buf[512];
- __u64 *buf64;
+ struct GPT gpt;
unsigned char empty_gpt_entry[16]= {0};
struct GPT_part_entry *part;
+ char buf[512];
unsigned long long curr_part_end;
unsigned all_partitions, entry_size;
- int part_nr;
+ unsigned part_nr;
*endofpart = 0;
- /* read GPT header */
+ BUILD_BUG_ON(sizeof(gpt) != 512);
+ /* skip protective MBR */
lseek(fd, 512, SEEK_SET);
- if (read(fd, buf, 512) != 512)
+ /* read GPT header */
+ if (read(fd, &gpt, 512) != 512)
return 0;
/* get the number of partition entries and the entry size */
- all_partitions = __le32_to_cpu(buf[GPT_ALL_PARTITIONS_OFFSET]);
- entry_size = __le32_to_cpu(buf[GPT_ENTRY_SIZE_OFFSET]);
+ all_partitions = __le32_to_cpu(gpt.part_cnt);
+ entry_size = __le32_to_cpu(gpt.part_size);
/* Check GPT signature*/
- buf64 = (__u64*)buf;
- if (buf64[0] != GPT_SIGNATURE_MAGIC)
+ if (gpt.magic != GPT_SIGNATURE_MAGIC)
return -1;
/* sanity checks */
if (all_partitions > 1024 ||
- entry_size > 512)
+ entry_size > sizeof(buf))
return -1;
- /* read first GPT partition entries */
- if (read(fd, buf, 512) != 512)
- return 0;
+ part = (struct GPT_part_entry *)buf;
- part = (struct GPT_part_entry*)buf;
+ for (part_nr = 0; part_nr < all_partitions; part_nr++) {
+ /* read partition entry */
+ if (read(fd, buf, entry_size) != (ssize_t)entry_size)
+ return 0;
- for (part_nr=0; part_nr < all_partitions; part_nr++) {
/* is this valid partition? */
if (memcmp(part->type_guid, empty_gpt_entry, 16) != 0) {
/* check the last lba for the current partition */
- buf64 = (__u64*)part->ending_lba;
- curr_part_end = __le64_to_cpu(buf64[0]);
+ curr_part_end = __le64_to_cpu(part->ending_lba);
if (curr_part_end > *endofpart)
*endofpart = curr_part_end;
}
- part = (struct GPT_part_entry*)((unsigned char*)part + entry_size);
-
- if ((unsigned char *)part >= buf + 512) {
- if (read(fd, buf, 512) != 512)
- return 0;
- part = (struct GPT_part_entry*)buf;
- }
}
return 1;
}
*/
static int get_last_partition_end(int fd, unsigned long long *endofpart)
{
- unsigned char boot_sect[512];
+ struct MBR boot_sect;
struct MBR_part_record *part;
unsigned long long curr_part_end;
- int part_nr;
+ unsigned part_nr;
int retval = 0;
*endofpart = 0;
+ BUILD_BUG_ON(sizeof(boot_sect) != 512);
/* read MBR */
lseek(fd, 0, 0);
- if (read(fd, boot_sect, 512) != 512)
+ if (read(fd, &boot_sect, 512) != 512)
goto abort;
/* check MBP signature */
- if (*((__u16*)(boot_sect + MBR_SIGNATURE_OFFSET))
- == MBR_SIGNATURE_MAGIC) {
+ if (boot_sect.magic == MBR_SIGNATURE_MAGIC) {
retval = 1;
/* found the correct signature */
- part = (struct MBR_part_record*)
- (boot_sect + MBR_PARTITION_TABLE_OFFSET);
+ part = boot_sect.parts;
- for (part_nr=0; part_nr < MBR_PARTITIONS; part_nr++) {
+ for (part_nr = 0; part_nr < MBR_PARTITIONS; part_nr++) {
/* check for GPT type */
if (part->part_type == MBR_GPT_PARTITION_TYPE) {
retval = get_gpt_last_partition_end(fd, endofpart);
return retval;
}
-int check_partitions(int fd, char *dname, unsigned long long freesize)
+int check_partitions(int fd, char *dname, unsigned long long freesize,
+ unsigned long long size)
{
/*
* Check where the last partition ends
/* 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 */
+ pr_err("array size is too small to cover all partitions on %s.\n",
+ dname);
return 1;
}
}
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 < ainf->raid_disks + ainf->nr_disks ; d++)
- if (ioctl(mdfd, GET_DISK_INFO, disk) == 0)
+ for (d = 0 ; d < MAX_DISKS ; d++) {
+ if (ioctl(mdfd, GET_DISK_INFO, disk) == 0 &&
+ (disk->major || disk->minor))
return;
+ }
}
int open_container(int fd)
continue;
n = read(dfd, buf, sizeof(buf));
close(dfd);
- if (n <= 0 || n >= sizeof(buf))
+ if (n <= 0 || (unsigned)n >= sizeof(buf))
continue;
buf[n] = 0;
if (sscanf(buf, "%d:%d", &major, &minor) != 2)
return -1;
}
+struct superswitch *version_to_superswitch(char *vers)
+{
+ int i;
+
+ for (i = 0; superlist[i]; i++) {
+ struct superswitch *ss = superlist[i];
+
+ if (strcmp(vers, ss->name) == 0)
+ return ss;
+ }
+
+ return NULL;
+}
+
+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)] != '/')
+ return 0;
+
+ return 1;
+}
+
+int is_subarray_active(char *subarray, char *container)
+{
+ struct mdstat_ent *mdstat = mdstat_read(0, 0);
+ struct mdstat_ent *ent;
+
+ for (ent = mdstat; ent; ent = ent->next)
+ if (is_container_member(ent, container))
+ if (strcmp(to_subarray(ent, container), subarray) == 0)
+ break;
+
+ free_mdstat(mdstat);
+
+ return ent != NULL;
+}
+
+/* open_subarray - opens a subarray in a container
+ * @dev: container device name
+ * @st: empty supertype
+ * @quiet: block reporting errors flag
+ *
+ * On success returns an fd to a container and fills in *st
+ */
+int open_subarray(char *dev, char *subarray, struct supertype *st, int quiet)
+{
+ struct mdinfo *mdi;
+ struct mdinfo *info;
+ int fd, err = 1;
+
+ fd = open(dev, O_RDWR|O_EXCL);
+ if (fd < 0) {
+ if (!quiet)
+ pr_err("Couldn't open %s, aborting\n",
+ dev);
+ return -1;
+ }
+
+ st->devnum = fd2devnum(fd);
+ if (st->devnum == NoMdDev) {
+ if (!quiet)
+ pr_err("Failed to determine device number for %s\n",
+ dev);
+ goto close_fd;
+ }
+
+ mdi = sysfs_read(fd, st->devnum, GET_VERSION|GET_LEVEL);
+ if (!mdi) {
+ if (!quiet)
+ pr_err("Failed to read sysfs for %s\n",
+ dev);
+ goto close_fd;
+ }
+
+ if (mdi->array.level != UnSet) {
+ if (!quiet)
+ 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)
+ pr_err("Operation not supported for %s metadata\n",
+ mdi->text_version);
+ goto free_sysfs;
+ }
+
+ st->devname = devnum2devname(st->devnum);
+ if (!st->devname) {
+ if (!quiet)
+ pr_err("Failed to allocate device name\n");
+ goto free_sysfs;
+ }
+
+ if (!st->ss->load_container) {
+ if (!quiet)
+ pr_err("%s is not a container\n", dev);
+ goto free_name;
+ }
+
+ if (st->ss->load_container(st, fd, NULL)) {
+ if (!quiet)
+ pr_err("Failed to load metadata for %s\n",
+ dev);
+ goto free_name;
+ }
+
+ info = st->ss->container_content(st, subarray);
+ if (!info) {
+ if (!quiet)
+ pr_err("Failed to find subarray-%s in %s\n",
+ subarray, dev);
+ goto free_super;
+ }
+ free(info);
+
+ err = 0;
+
+ free_super:
+ if (err)
+ st->ss->free_super(st);
+ free_name:
+ if (err)
+ free(st->devname);
+ free_sysfs:
+ sysfs_free(mdi);
+ close_fd:
+ if (err)
+ close(fd);
+
+ if (err)
+ return -1;
+ else
+ return fd;
+}
+
int add_disk(int mdfd, struct supertype *st,
struct mdinfo *sra, struct mdinfo *info)
{
if (sd2 == info)
break;
if (sd2 == NULL) {
- sd2 = malloc(sizeof(*sd2));
+ sd2 = xmalloc(sizeof(*sd2));
*sd2 = *info;
sd2->next = sra->devs;
sra->devs = sd2;
return rv;
}
+int remove_disk(int mdfd, struct supertype *st,
+ 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 set_array_info(int mdfd, struct supertype *st, struct mdinfo *info)
{
/* Initialise kernel's knowledge of array.
return recovery_start;
}
-char *devnum2devname(int num)
-{
- char name[100];
- if (num >= 0)
- sprintf(name, "md%d", num);
- else
- sprintf(name, "md_d%d", -1-num);
- return strdup(name);
-}
-
-int devname2devnum(char *name)
-{
- char *ep;
- int num;
- if (strncmp(name, "md_d", 4)==0)
- num = -1-strtoul(name+4, &ep, 10);
- else
- num = strtoul(name+2, &ep, 10);
- return num;
-}
-
-int stat2devnum(struct stat *st)
-{
- char path[30];
- char link[200];
- char *cp;
- int n;
-
- if ((S_IFMT & st->st_mode) == S_IFBLK) {
- if (major(st->st_rdev) == MD_MAJOR)
- return minor(st->st_rdev);
- else if (major(st->st_rdev) == get_mdp_major())
- return -1- (minor(st->st_rdev)>>MdpMinorShift);
-
- /* must be an extended-minor partition. Look at the
- * /sys/dev/block/%d:%d link which must look like
- * ../../block/mdXXX/mdXXXpYY
- */
- sprintf(path, "/sys/dev/block/%d:%d", major(st->st_rdev),
- minor(st->st_rdev));
- n = readlink(path, link, sizeof(link)-1);
- if (n <= 0)
- return NoMdDev;
- link[n] = 0;
- cp = strrchr(link, '/');
- if (cp) *cp = 0;
- cp = strchr(link, '/');
- if (cp && strncmp(cp, "/md", 3) == 0)
- return devname2devnum(cp+1);
- }
- return NoMdDev;
-
-}
-
-int fd2devnum(int fd)
-{
- struct stat stb;
- if (fstat(fd, &stb) == 0)
- return stat2devnum(&stb);
- return NoMdDev;
-}
-
-char *pid_dir = VAR_RUN;
-
int mdmon_pid(int devnum)
{
char path[100];
char pid[10];
int fd;
int n;
- sprintf(path, "%s/%s.pid", pid_dir, devnum2devname(devnum));
+ char *devname = devnum2devname(devnum);
+
+ sprintf(path, "%s/%s.pid", MDMON_DIR, devname);
+ free(devname);
+
fd = open(path, O_RDONLY | O_NOATIME, 0);
if (fd < 0)
int start_mdmon(int devnum)
{
- int i;
+ int i, skipped;
int len;
- pid_t pid;
+ pid_t pid;
int status;
char pathbuf[1024];
char *paths[4] = {
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;
switch(fork()) {
case 0:
/* FIXME yuk. CLOSE_EXEC?? */
- for (i=3; i < 100; i++)
- close(i);
- for (i=0; paths[i]; i++)
- if (paths[i][0])
- execl(paths[i], "mdmon",
- devnum2devname(devnum),
- NULL);
+ skipped = 0;
+ for (i = 3; skipped < 20; i++)
+ if (close(i) < 0)
+ skipped++;
+ else
+ skipped = 0;
+
+ for (i = 0; paths[i]; i++)
+ if (paths[i][0]) {
+ if (__offroot) {
+ execl(paths[i], "mdmon", "--offroot",
+ devnum2devname(devnum),
+ NULL);
+ } else {
+ execl(paths[i], "mdmon",
+ devnum2devname(devnum),
+ NULL);
+ }
+ }
exit(1);
- case -1: fprintf(stderr, Name ": cannot run mdmon. "
+ case -1: pr_err("cannot run mdmon. "
"Array remains readonly\n");
return -1;
default: /* parent - good */
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;
mu->space = NULL;
+ mu->space_list = NULL;
mu->next = NULL;
*st->update_tail = mu;
st->update_tail = &mu->next;
unsigned int __invalid_size_argument_for_IOC = 0;
#endif
+int experimental(void)
+{
+ if (check_env("MDADM_EXPERIMENTAL"))
+ return 1;
+ else {
+ pr_err("To use this feature MDADM_EXPERIMENTAL"
+ " environment variable has to be defined.\n");
+ return 0;
+ }
+}
+
+/* Pick all spares matching given criteria from a container
+ * if min_size == 0 do not check size
+ * if domlist == NULL do not check domains
+ * if spare_group given add it to domains of each spare
+ * metadata allows to test domains using metadata of destination array */
+struct mdinfo *container_choose_spares(struct supertype *st,
+ unsigned long long min_size,
+ struct domainlist *domlist,
+ char *spare_group,
+ const char *metadata, int get_one)
+{
+ struct mdinfo *d, **dp, *disks = NULL;
+
+ /* get list of all disks in container */
+ if (st->ss->getinfo_super_disks)
+ disks = st->ss->getinfo_super_disks(st);
+
+ if (!disks)
+ return disks;
+ /* find spare devices on the list */
+ dp = &disks->devs;
+ disks->array.spare_disks = 0;
+ while (*dp) {
+ int found = 0;
+ d = *dp;
+ if (d->disk.state == 0) {
+ /* check if size is acceptable */
+ unsigned long long dev_size;
+ dev_t dev = makedev(d->disk.major,d->disk.minor);
+
+ if (!min_size ||
+ (dev_size_from_id(dev, &dev_size) &&
+ dev_size >= min_size))
+ found = 1;
+ /* check if domain matches */
+ if (found && domlist) {
+ struct dev_policy *pol = devnum_policy(dev);
+ if (spare_group)
+ pol_add(&pol, pol_domain,
+ spare_group, NULL);
+ if (domain_test(domlist, pol, metadata) != 1)
+ found = 0;
+ dev_policy_free(pol);
+ }
+ }
+ if (found) {
+ dp = &d->next;
+ disks->array.spare_disks++;
+ if (get_one) {
+ sysfs_free(*dp);
+ d->next = NULL;
+ }
+ } else {
+ *dp = d->next;
+ d->next = NULL;
+ sysfs_free(d);
+ }
+ }
+ return disks;
+}