[SYS_DEV_UNKNOWN] = "Unknown",
[SYS_DEV_SAS] = "SAS",
[SYS_DEV_SATA] = "SATA",
- [SYS_DEV_NVME] = "NVMe"
+ [SYS_DEV_NVME] = "NVMe",
+ [SYS_DEV_VMD] = "VMD"
};
const char *get_sys_dev_type(enum sys_dev_type type)
static struct intel_hba * find_intel_hba(struct intel_hba *hba, struct sys_dev *device)
{
- struct intel_hba *result=NULL;
+ struct intel_hba *result;
+
for (result = hba; result; result = result->next) {
if (result->type == device->type && strcmp(result->path, device->path) == 0)
break;
if (device->type != hba->type)
return 2;
+ /* Always forbid spanning between VMD domains (seen as different controllers by mdadm) */
+ if (device->type == SYS_DEV_VMD && !path_attached_to_hba(device->path, hba->path))
+ return 2;
+
/* Multiple same type HBAs can be used if they share the same OROM */
const struct imsm_orom *device_orom = get_orom_by_device_id(device->dev_id);
* this hba
*/
dir = opendir("/sys/dev/block");
- for (ent = dir ? readdir(dir) : NULL; ent; ent = readdir(dir)) {
+ if (!dir)
+ return 1;
+
+ for (ent = readdir(dir); ent; ent = readdir(dir)) {
int fd;
char model[64];
char vendor[64];
break;
}
sprintf(device, "/sys/dev/block/%d:%d/device/type", major, minor);
- if (load_sys(device, buf) != 0) {
+ if (load_sys(device, buf, sizeof(buf)) != 0) {
if (verbose > 0)
pr_err("failed to read device type for %s\n",
path);
vendor[0] = '\0';
model[0] = '\0';
sprintf(device, "/sys/dev/block/%d:%d/device/vendor", major, minor);
- if (load_sys(device, buf) == 0) {
+ if (load_sys(device, buf, sizeof(buf)) == 0) {
strncpy(vendor, buf, sizeof(vendor));
vendor[sizeof(vendor) - 1] = '\0';
c = (char *) &vendor[sizeof(vendor) - 1];
}
sprintf(device, "/sys/dev/block/%d:%d/device/model", major, minor);
- if (load_sys(device, buf) == 0) {
+ if (load_sys(device, buf, sizeof(buf)) == 0) {
strncpy(model, buf, sizeof(model));
model[sizeof(model) - 1] = '\0';
c = (char *) &model[sizeof(model) - 1];
return err;
}
+static int print_vmd_attached_devs(struct sys_dev *hba)
+{
+ struct dirent *ent;
+ DIR *dir;
+ char path[292];
+ char link[256];
+ char *c, *rp;
+
+ if (hba->type != SYS_DEV_VMD)
+ return 1;
+
+ /* scroll through /sys/dev/block looking for devices attached to
+ * this hba
+ */
+ dir = opendir("/sys/bus/pci/drivers/nvme");
+ if (!dir)
+ return 1;
+
+ for (ent = readdir(dir); ent; ent = readdir(dir)) {
+ int n;
+
+ /* is 'ent' a device? check that the 'subsystem' link exists and
+ * that its target matches 'bus'
+ */
+ sprintf(path, "/sys/bus/pci/drivers/nvme/%s/subsystem",
+ ent->d_name);
+ n = readlink(path, link, sizeof(link));
+ if (n < 0 || n >= (int)sizeof(link))
+ continue;
+ link[n] = '\0';
+ c = strrchr(link, '/');
+ if (!c)
+ continue;
+ if (strncmp("pci", c+1, strlen("pci")) != 0)
+ continue;
+
+ sprintf(path, "/sys/bus/pci/drivers/nvme/%s", ent->d_name);
+ /* if not a intel NVMe - skip it*/
+ if (devpath_to_vendor(path) != 0x8086)
+ continue;
+
+ rp = realpath(path, NULL);
+ if (!rp)
+ continue;
+
+ if (path_attached_to_hba(rp, hba->path)) {
+ printf(" NVMe under VMD : %s\n", rp);
+ }
+ free(rp);
+ }
+
+ closedir(dir);
+ return 0;
+}
+
static void print_found_intel_controllers(struct sys_dev *elem)
{
for (; elem; elem = elem->next) {
fprintf(stderr, "SAS ");
else if (elem->type == SYS_DEV_NVME)
fprintf(stderr, "NVMe ");
- fprintf(stderr, "RAID controller");
+
+ if (elem->type == SYS_DEV_VMD)
+ fprintf(stderr, "VMD domain");
+ else
+ fprintf(stderr, "RAID controller");
+
if (elem->pci_id)
fprintf(stderr, " at %s", elem->pci_id);
fprintf(stderr, ".\n");
if (controller_path && (compare_paths(hba->path, controller_path) != 0))
continue;
if (!find_imsm_capability(hba)) {
+ char buf[PATH_MAX];
pr_err("imsm capabilities not found for controller: %s (type %s)\n",
- hba->path, get_sys_dev_type(hba->type));
+ hba->type == SYS_DEV_VMD ? vmd_domain_to_controller(hba, buf) : hba->path,
+ get_sys_dev_type(hba->type));
continue;
}
result = 0;
const struct orom_entry *entry;
for (entry = orom_entries; entry; entry = entry->next) {
- print_imsm_capability(&entry->orom);
+ if (entry->type == SYS_DEV_VMD) {
+ for (hba = list; hba; hba = hba->next) {
+ if (hba->type == SYS_DEV_VMD) {
+ char buf[PATH_MAX];
+ print_imsm_capability(&entry->orom);
+ printf(" I/O Controller : %s (%s)\n",
+ vmd_domain_to_controller(hba, buf), get_sys_dev_type(hba->type));
+ if (print_vmd_attached_devs(hba)) {
+ if (verbose > 0)
+ pr_err("failed to get devices attached to VMD domain.\n");
+ result |= 2;
+ }
+ printf("\n");
+ }
+ }
+ continue;
+ }
- if (imsm_orom_is_nvme(&entry->orom)) {
+ print_imsm_capability(&entry->orom);
+ if (entry->type == SYS_DEV_NVME) {
for (hba = list; hba; hba = hba->next) {
if (hba->type == SYS_DEV_NVME)
printf(" NVMe Device : %s\n", hba->path);
}
+ printf("\n");
continue;
}
for (hba = list; hba; hba = hba->next) {
if (controller_path && (compare_paths(hba->path,controller_path) != 0))
continue;
- if (!find_imsm_capability(hba) && verbose > 0)
- pr_err("IMSM_DETAIL_PLATFORM_ERROR=NO_IMSM_CAPABLE_DEVICE_UNDER_%s\n", hba->path);
+ if (!find_imsm_capability(hba) && verbose > 0) {
+ char buf[PATH_MAX];
+ pr_err("IMSM_DETAIL_PLATFORM_ERROR=NO_IMSM_CAPABLE_DEVICE_UNDER_%s\n",
+ hba->type == SYS_DEV_VMD ? vmd_domain_to_controller(hba, buf) : hba->path);
+ }
else
result = 0;
}
const struct orom_entry *entry;
- for (entry = orom_entries; entry; entry = entry->next)
+ for (entry = orom_entries; entry; entry = entry->next) {
+ if (entry->type == SYS_DEV_VMD) {
+ for (hba = list; hba; hba = hba->next)
+ print_imsm_capability_export(&entry->orom);
+ continue;
+ }
print_imsm_capability_export(&entry->orom);
+ }
return result;
}
static int load_imsm_migr_rec(struct intel_super *super, struct mdinfo *info)
{
struct mdinfo *sd;
- struct dl *dl = NULL;
+ struct dl *dl;
char nm[30];
int retval = -1;
int fd = -1;
struct imsm_dev *dev;
- struct imsm_map *map = NULL;
+ struct imsm_map *map;
int slot = -1;
/* find map under migration */
int len;
struct imsm_update_general_migration_checkpoint *u;
struct imsm_dev *dev;
- struct imsm_map *map = NULL;
+ struct imsm_map *map;
/* find map under migration */
dev = imsm_get_device_during_migration(super);
* for each disk in array */
struct mdinfo *getinfo_super_disks_imsm(struct supertype *st)
{
- struct mdinfo *mddev = NULL;
+ struct mdinfo *mddev;
struct intel_super *super = st->sb;
struct imsm_disk *disk;
int count = 0;
if (devname) {
struct intel_hba *hba = super->hba;
- pr_err("%s is attached to Intel(R) %s RAID controller (%s),\n"
- " but the container is assigned to Intel(R) %s RAID controller (",
+ pr_err("%s is attached to Intel(R) %s %s (%s),\n"
+ " but the container is assigned to Intel(R) %s %s (",
devname,
get_sys_dev_type(hba_name->type),
+ hba_name->type == SYS_DEV_VMD ? "domain" : "RAID controller",
hba_name->pci_id ? : "Err!",
- get_sys_dev_type(super->hba->type));
+ get_sys_dev_type(super->hba->type),
+ hba->type == SYS_DEV_VMD ? "domain" : "RAID controller");
while (hba) {
fprintf(stderr, "%s", hba->pci_id ? : "Err!");
hba = hba->next;
}
fprintf(stderr, ").\n"
- " Mixing devices attached to different controllers is not allowed.\n");
+ " Mixing devices attached to different %s is not allowed.\n",
+ hba_name->type == SYS_DEV_VMD ? "VMD domains" : "controllers");
}
return 2;
}
static int get_super_block(struct intel_super **super_list, char *devnm, char *devname,
int major, int minor, int keep_fd)
{
- struct intel_super*s = NULL;
+ struct intel_super *s;
char nm[32];
int dfd = -1;
int err = 0;
/* retry the load if we might have raced against mdmon */
if (rv == 3) {
- struct mdstat_ent *mdstat = mdstat_by_component(fd2devnm(fd));
+ struct mdstat_ent *mdstat = NULL;
+ char *name = fd2kname(fd);
+
+ if (name)
+ mdstat = mdstat_by_component(name);
if (mdstat && mdmon_running(mdstat->devnm) && getpid() != mdmon_pid(mdstat->devnm)) {
for (retry = 0; retry < 3; retry++) {
{
int fd;
unsigned long long ldsize;
- struct intel_super *super=NULL;
+ struct intel_super *super;
int rv = 0;
if (level != LEVEL_CONTAINER)
int dpa, int verbose)
{
struct mdstat_ent *mdstat = mdstat_read(0, 0);
- struct mdstat_ent *memb = NULL;
+ struct mdstat_ent *memb;
int count = 0;
int num = 0;
- struct md_list *dv = NULL;
+ struct md_list *dv;
int found;
for (memb = mdstat ; memb ; memb = memb->next) {
if (num > 0)
fd = open(path, O_RDONLY, 0);
if ((num <= 0) || (fd < 0)) {
- pr_vrb(": Cannot open %s: %s\n",
+ pr_vrb("Cannot open %s: %s\n",
dev->name, strerror(errno));
}
free(path);
for (vol = mdstat ; vol ; vol = vol->next) {
if ((vol->active > 0) &&
vol->metadata_version &&
- is_container_member(vol, memb->dev)) {
+ is_container_member(vol, memb->devnm)) {
found++;
count++;
}
}
if (*devlist && (found < dpa)) {
dv = xcalloc(1, sizeof(*dv));
- dv->devname = xmalloc(strlen(memb->dev) + strlen("/dev/") + 1);
- sprintf(dv->devname, "%s%s", "/dev/", memb->dev);
+ dv->devname = xmalloc(strlen(memb->devnm) + strlen("/dev/") + 1);
+ sprintf(dv->devname, "%s%s", "/dev/", memb->devnm);
dv->found = found;
dv->used = 0;
dv->next = *devlist;
{
int i;
struct md_list *devlist = NULL;
- struct md_list *dv = NULL;
+ struct md_list *dv;
for(i = 0; i < 12; i++) {
dv = xcalloc(1, sizeof(*dv));
get_devices(const char *hba_path)
{
struct md_list *devlist = NULL;
- struct md_list *dv = NULL;
+ struct md_list *dv;
struct dirent *ent;
DIR *dir;
int err = 0;
{
struct md_list *tmpdev;
int count = 0;
- struct supertype *st = NULL;
+ struct supertype *st;
/* first walk the list of devices to find a consistent set
* that match the criterea, if that is possible.
*found = 0;
st = match_metadata_desc_imsm("imsm");
if (st == NULL) {
- pr_vrb(": cannot allocate memory for imsm supertype\n");
+ pr_vrb("cannot allocate memory for imsm supertype\n");
return 0;
}
continue;
tst = dup_super(st);
if (tst == NULL) {
- pr_vrb(": cannot allocate memory for imsm supertype\n");
+ pr_vrb("cannot allocate memory for imsm supertype\n");
goto err_1;
}
tmpdev->container = 0;
devid_list = entry->devid_list;
for (dv = devid_list; dv; dv = dv->next) {
-
- struct md_list *devlist = NULL;
+ struct md_list *devlist;
struct sys_dev *device = device_by_id(dv->devid);
char *hba_path;
int found = 0;
else
return 0;
+ /* VMD has one orom entry for all domain, but spanning is not allowed.
+ * VMD arrays should be counted per domain (controller), so skip
+ * domains that are not the given one.
+ */
+ if ((hba->type == SYS_DEV_VMD) &&
+ (strncmp(device->path, hba->path, strlen(device->path)) != 0))
+ continue;
+
devlist = get_devices(hba_path);
/* if no intel devices return zero volumes */
if (devlist == NULL)
{
/* check/set platform and metadata limits/defaults */
if (super->orom && raiddisks > super->orom->dpa) {
- pr_vrb(": platform supports a maximum of %d disks per array\n",
+ pr_vrb("platform supports a maximum of %d disks per array\n",
super->orom->dpa);
return 0;
}
/* capabilities of OROM tested - copied from validate_geometry_imsm_volume */
if (!is_raid_level_supported(super->orom, level, raiddisks)) {
- pr_vrb(": platform does not support raid%d with %d disk%s\n",
+ pr_vrb("platform does not support raid%d with %d disk%s\n",
level, raiddisks, raiddisks > 1 ? "s" : "");
return 0;
}
*chunk = imsm_default_chunk(super->orom);
if (super->orom && !imsm_orom_has_chunk(super->orom, *chunk)) {
- pr_vrb(": platform does not support a chunk size of: %d\n", *chunk);
+ pr_vrb("platform does not support a chunk size of: %d\n", *chunk);
return 0;
}
if (layout != imsm_level_to_layout(level)) {
if (level == 5)
- pr_vrb(": imsm raid 5 only supports the left-asymmetric layout\n");
+ pr_vrb("imsm raid 5 only supports the left-asymmetric layout\n");
else if (level == 10)
- pr_vrb(": imsm raid 10 only supports the n2 layout\n");
+ pr_vrb("imsm raid 10 only supports the n2 layout\n");
else
- pr_vrb(": imsm unknown layout %#x for this raid level %d\n",
+ pr_vrb("imsm unknown layout %#x for this raid level %d\n",
layout, level);
return 0;
}
if (super->orom && (super->orom->attr & IMSM_OROM_ATTR_2TB) == 0 &&
(calc_array_size(level, raiddisks, layout, *chunk, size) >> 32) > 0) {
- pr_vrb(": platform does not support a volume size over 2TB\n");
+ pr_vrb("platform does not support a volume size over 2TB\n");
return 0;
}
int count = count_volumes(super->hba,
super->orom->dpa, verbose);
if (super->orom->vphba <= count) {
- pr_vrb(": platform does not support more than %d raid volumes.\n",
+ pr_vrb("platform does not support more than %d raid volumes.\n",
super->orom->vphba);
return 0;
}
count = count_volumes(super->hba,
super->orom->dpa, verbose);
if (super->orom->vphba <= count) {
- pr_vrb(": platform does not support more than %d raid volumes.\n",
+ pr_vrb("platform does not support more than %d raid volumes.\n",
super->orom->vphba);
return 0;
}
IMSM_T_STATE_DEGRADED)
return NULL;
+ if (get_imsm_map(dev, MAP_0)->map_state == IMSM_T_STATE_UNINITIALIZED) {
+ dprintf("imsm: No spare activation allowed. Volume is not initialized.\n");
+ return NULL;
+ }
+
/*
* If there are any failed disks check state of the other volume.
* Block rebuild if the another one is failed until failed disks
static struct dl *get_disk_super(struct intel_super *super, int major, int minor)
{
- struct dl *dl = NULL;
+ struct dl *dl;
+
for (dl = super->disks; dl; dl = dl->next)
if ((dl->major == major) && (dl->minor == minor))
return dl;
static int remove_disk_super(struct intel_super *super, int major, int minor)
{
- struct dl *prev = NULL;
+ struct dl *prev;
struct dl *dl;
prev = NULL;
static int add_remove_disk_update(struct intel_super *super)
{
int check_degraded = 0;
- struct dl *disk = NULL;
+ struct dl *disk;
+
/* add/remove some spares to/from the metadata/contrainer */
while (super->disk_mgmt_list) {
struct dl *disk_cfg;
return 1;
}
- if (orom != orom2) {
+ if ((orom != orom2) || ((hba->type == SYS_DEV_VMD) && (hba != hba2))) {
pr_err("WARNING - IMSM container assembled with disks under different HBAs!\n"
" This operation is not supported and can lead to data loss.\n");
return 1;
{
int rv = -1;
struct intel_super *super = st->sb;
- unsigned long long *target_offsets = NULL;
- int *targets = NULL;
+ unsigned long long *target_offsets;
+ int *targets;
int i;
struct imsm_map *map_dest = get_imsm_map(dev, MAP_0);
int new_disks = map_dest->num_members;
{
struct intel_super *super = st->sb;
struct migr_record *migr_rec = super->migr_rec;
- struct imsm_map *map_dest = NULL;
+ struct imsm_map *map_dest;
struct intel_dev *id = NULL;
unsigned long long read_offset;
unsigned long long write_offset;
strncat(disk_path, path, PATH_MAX - strlen(disk_path) - 1);
if (stat(disk_path, &st) == 0) {
struct sys_dev* hba;
- char *path=NULL;
+ char *path;
path = devt_to_devpath(st.st_rdev);
if (path == NULL)
{
struct intel_super *super = st->sb;
struct imsm_super *mpb = super->anchor;
- int update_memory_size = 0;
- struct imsm_update_reshape *u = NULL;
- struct mdinfo *spares = NULL;
+ int update_memory_size;
+ struct imsm_update_reshape *u;
+ struct mdinfo *spares;
int i;
- int delta_disks = 0;
+ int delta_disks;
struct mdinfo *dev;
dprintf("(enter) raid_disks = %i\n", geo->raid_disks);
struct imsm_update_size_change **updatep)
{
struct intel_super *super = st->sb;
- int update_memory_size = 0;
- struct imsm_update_size_change *u = NULL;
+ int update_memory_size;
+ struct imsm_update_size_change *u;
dprintf("(enter) New size = %llu\n", geo->size);
struct imsm_update_reshape_migration **updatep)
{
struct intel_super *super = st->sb;
- int update_memory_size = 0;
- struct imsm_update_reshape_migration *u = NULL;
+ int update_memory_size;
+ struct imsm_update_reshape_migration *u;
struct imsm_dev *dev;
int previous_level = -1;
return ret_val;
}
+#define COMPLETED_OK 0
+#define COMPLETED_NONE 1
+#define COMPLETED_DELAYED 2
+
+static int read_completed(int fd, unsigned long long *val)
+{
+ int ret;
+ char buf[50];
+
+ ret = sysfs_fd_get_str(fd, buf, 50);
+ if (ret < 0)
+ return ret;
+
+ ret = COMPLETED_OK;
+ if (strncmp(buf, "none", 4) == 0) {
+ ret = COMPLETED_NONE;
+ } else if (strncmp(buf, "delayed", 7) == 0) {
+ ret = COMPLETED_DELAYED;
+ } else {
+ char *ep;
+ *val = strtoull(buf, &ep, 0);
+ if (ep == buf || (*ep != 0 && *ep != '\n' && *ep != ' '))
+ ret = -1;
+ }
+ return ret;
+}
+
/*******************************************************************************
* Function: wait_for_reshape_imsm
* Description: Function writes new sync_max value and waits until
int wait_for_reshape_imsm(struct mdinfo *sra, int ndata)
{
int fd = sysfs_get_fd(sra, NULL, "sync_completed");
+ int retry = 3;
unsigned long long completed;
/* to_complete : new sync_max position */
unsigned long long to_complete = sra->reshape_progress;
return 1;
}
- if (sysfs_fd_get_ll(fd, &completed) < 0) {
- dprintf("cannot read reshape_position (no reshape in progres)\n");
- close(fd);
- return 0;
- }
+ do {
+ if (sysfs_fd_get_ll(fd, &completed) < 0) {
+ if (!retry) {
+ dprintf("cannot read reshape_position (no reshape in progres)\n");
+ close(fd);
+ return 1;
+ }
+ usleep(30000);
+ } else
+ break;
+ } while (retry--);
if (completed > position_to_set) {
dprintf("wrong next position to set %llu (%llu)\n",
}
do {
+ int rc;
char action[20];
- sysfs_wait(fd, NULL);
+ int timeout = 3000;
+
+ sysfs_wait(fd, &timeout);
if (sysfs_get_str(sra, NULL, "sync_action",
action, 20) > 0 &&
- strncmp(action, "reshape", 7) != 0)
- break;
- if (sysfs_fd_get_ll(fd, &completed) < 0) {
+ strncmp(action, "reshape", 7) != 0) {
+ if (strncmp(action, "idle", 4) == 0)
+ break;
+ close(fd);
+ return -1;
+ }
+
+ rc = read_completed(fd, &completed);
+ if (rc < 0) {
dprintf("cannot read reshape_position (in loop)\n");
close(fd);
return 1;
- }
+ } else if (rc == COMPLETED_NONE)
+ break;
} while (completed < position_to_set);
+
close(fd);
return 0;
-
}
/*******************************************************************************
* Function: imsm_manage_reshape
* Description: Function finds array under reshape and it manages reshape
* process. It creates stripes backups (if required) and sets
- * checheckpoits.
+ * checkpoints.
* Parameters:
* afd : Backup handle (nattive) - not used
* sra : general array info
{
int ret_val = 0;
struct intel_super *super = st->sb;
- struct intel_dev *dv = NULL;
+ struct intel_dev *dv;
struct imsm_dev *dev = NULL;
struct imsm_map *map_src;
int migr_vol_qan = 0;
int degraded = 0;
int source_layout = 0;
- if (!fds || !offsets || !sra)
+ if (!sra)
+ return ret_val;
+
+ if (!fds || !offsets)
goto abort;
/* Find volume during the reshape */
}
/* Only one volume can migrate at the same time */
if (migr_vol_qan != 1) {
- pr_err(": %s", migr_vol_qan ?
+ pr_err("%s", migr_vol_qan ?
"Number of migrating volumes greater than 1\n" :
"There is no volume during migrationg\n");
goto abort;
start = current_position * 512;
- /* allign reading start to old geometry */
+ /* align reading start to old geometry */
start_buf_shift = start % old_data_stripe_length;
start_src = start - start_buf_shift;
* to backup alligned to source array
* [bytes]
*/
- unsigned long long next_step_filler = 0;
+ unsigned long long next_step_filler;
unsigned long long copy_length = next_step * 512;
/* allign copy area length to stripe in old geometry */
sra->reshape_progress = next_step;
/* wait until reshape finish */
- if (wait_for_reshape_imsm(sra, ndata) < 0) {
+ if (wait_for_reshape_imsm(sra, ndata)) {
dprintf("wait_for_reshape_imsm returned error!\n");
goto abort;
}
ret_val = 1;
abort:
free(buf);
- abort_reshape(sra);
+ /* See Grow.c: abort_reshape() for further explanation */
+ sysfs_set_num(sra, NULL, "suspend_lo", 0x7FFFFFFFFFFFFFFFULL);
+ sysfs_set_num(sra, NULL, "suspend_hi", 0);
+ sysfs_set_num(sra, NULL, "suspend_lo", 0);
return ret_val;
}
projects / thirdparty / mdadm.git / blobdiff