#include <sys/stat.h>
#include <limits.h>
+#define NVME_SUBSYS_PATH "/sys/devices/virtual/nvme-subsystem/"
static int devpath_to_ll(const char *dev_path, const char *entry,
unsigned long long *val);
-static __u16 devpath_to_vendor(const char *dev_path);
-
-void free_sys_dev(struct sys_dev **list)
+static void free_sys_dev(struct sys_dev **list)
{
while (*list) {
struct sys_dev *next = (*list)->next;
struct sys_dev *find_driver_devices(const char *bus, const char *driver)
{
/* search sysfs for devices driven by 'driver' */
- char path[292];
- char link[256];
- char *c;
+ char path[PATH_MAX];
+ char link[PATH_MAX];
+ char *c, *p;
DIR *driver_dir;
struct dirent *de;
struct sys_dev *head = NULL;
struct sys_dev *list = NULL;
+ struct sys_dev *vmd = NULL;
enum sys_dev_type type;
unsigned long long dev_id;
+ unsigned long long class;
if (strcmp(driver, "isci") == 0)
type = SYS_DEV_SAS;
else if (strcmp(driver, "ahci") == 0)
type = SYS_DEV_SATA;
+ else if (strcmp(driver, "nvme") == 0) {
+ /* if looking for nvme devs, first look for vmd */
+ vmd = find_driver_devices("pci", "vmd");
+ type = SYS_DEV_NVME;
+ } else if (strcmp(driver, "vmd") == 0)
+ type = SYS_DEV_VMD;
else
type = SYS_DEV_UNKNOWN;
sprintf(path, "/sys/bus/%s/drivers/%s", bus, driver);
driver_dir = opendir(path);
- if (!driver_dir)
+ if (!driver_dir) {
+ if (vmd)
+ free_sys_dev(&vmd);
return NULL;
+ }
for (de = readdir(driver_dir); de; de = readdir(driver_dir)) {
int n;
+ int skip = 0;
/* is 'de' a device? check that the 'subsystem' link exists and
* that its target matches 'bus'
sprintf(path, "/sys/bus/%s/drivers/%s/%s",
bus, driver, de->d_name);
- /* if it's not Intel device skip it. */
- if (devpath_to_vendor(path) != 0x8086)
+ /* if searching for nvme - skip vmd connected one */
+ if (type == SYS_DEV_NVME) {
+ struct sys_dev *dev;
+ char *rp = realpath(path, NULL);
+ for (dev = vmd; dev; dev = dev->next) {
+ if ((strncmp(dev->path, rp, strlen(dev->path)) == 0))
+ skip = 1;
+ }
+ free(rp);
+ }
+
+ /* if it's not Intel device or mark as VMD connected - skip it. */
+ if (devpath_to_vendor(path) != 0x8086 || skip == 1)
continue;
if (devpath_to_ll(path, "device", &dev_id) != 0)
continue;
+ if (devpath_to_ll(path, "class", &class) != 0)
+ continue;
+
+ /*
+ * Each VMD device (domain) adds separate PCI bus, it is better
+ * to store path as a path to that bus (easier further
+ * determination which NVMe dev is connected to this particular
+ * VMD domain).
+ */
+ if (type == SYS_DEV_VMD) {
+ sprintf(path, "/sys/bus/%s/drivers/%s/%s/domain/device",
+ bus, driver, de->d_name);
+ }
+ p = realpath(path, NULL);
+ if (p == NULL) {
+ pr_err("Unable to get real path for '%s'\n", path);
+ continue;
+ }
+
/* start / add list entry */
if (!head) {
- head = malloc(sizeof(*head));
+ head = xmalloc(sizeof(*head));
list = head;
} else {
- list->next = malloc(sizeof(*head));
+ list->next = xmalloc(sizeof(*head));
list = list->next;
}
}
list->dev_id = (__u16) dev_id;
+ list->class = (__u32) class;
list->type = type;
- list->path = canonicalize_file_name(path);
list->next = NULL;
+ list->path = p;
+
if ((list->pci_id = strrchr(list->path, '/')) != NULL)
list->pci_id++;
}
closedir(driver_dir);
+
+ if (vmd) {
+ if (list)
+ list->next = vmd;
+ else
+ head = vmd;
+ }
+
return head;
}
-
static struct sys_dev *intel_devices=NULL;
+static time_t valid_time = 0;
-static enum sys_dev_type device_type_by_id(__u16 device_id)
+struct sys_dev *device_by_id(__u16 device_id)
{
struct sys_dev *iter;
- for(iter = intel_devices; iter != NULL; iter = iter->next)
+ for (iter = intel_devices; iter != NULL; iter = iter->next)
if (iter->dev_id == device_id)
- return iter->type;
- return SYS_DEV_UNKNOWN;
+ return iter;
+ return NULL;
+}
+
+struct sys_dev *device_by_id_and_path(__u16 device_id, const char *path)
+{
+ struct sys_dev *iter;
+
+ for (iter = intel_devices; iter != NULL; iter = iter->next)
+ if ((iter->dev_id == device_id) && strstr(iter->path, path))
+ return iter;
+ return NULL;
}
static int devpath_to_ll(const char *dev_path, const char *entry, unsigned long long *val)
return n;
}
-
-static __u16 devpath_to_vendor(const char *dev_path)
+__u16 devpath_to_vendor(const char *dev_path)
{
char path[strlen(dev_path) + strlen("/vendor") + 1];
char vendor[7];
return id;
}
+/* Description: Read text value of dev_path/entry field
+ * Parameters:
+ * dev_path - sysfs path to the device
+ * entry - entry to be read
+ * buf - buffer for read value
+ * len - size of buf
+ * verbose - error logging level
+ */
+int devpath_to_char(const char *dev_path, const char *entry, char *buf, int len,
+ int verbose)
+{
+ char path[PATH_MAX];
+
+ snprintf(path, sizeof(path), "%s/%s", dev_path, entry);
+ if (load_sys(path, buf, len)) {
+ if (verbose)
+ pr_err("Cannot read %s, aborting\n", path);
+ return 1;
+ }
+
+ return 0;
+}
+
struct sys_dev *find_intel_devices(void)
{
- struct sys_dev *ahci, *isci;
+ struct sys_dev *ahci, *isci, *nvme;
+
+ if (valid_time > time(0) - 10)
+ return intel_devices;
+
+ if (intel_devices)
+ free_sys_dev(&intel_devices);
isci = find_driver_devices("pci", "isci");
ahci = find_driver_devices("pci", "ahci");
+ /* Searching for NVMe will return list of NVMe and VMD controllers */
+ nvme = find_driver_devices("pci", "nvme");
- if (!ahci) {
+ if (!isci && !ahci) {
+ ahci = nvme;
+ } else if (!ahci) {
ahci = isci;
+ struct sys_dev *elem = ahci;
+ while (elem->next)
+ elem = elem->next;
+ elem->next = nvme;
} else {
struct sys_dev *elem = ahci;
while (elem->next)
elem = elem->next;
elem->next = isci;
+ while (elem->next)
+ elem = elem->next;
+ elem->next = nvme;
}
- return ahci;
+ intel_devices = ahci;
+ valid_time = time(0);
+ return intel_devices;
}
/*
__u8 ver[4];
__u16 vendorID;
__u16 deviceID;
+ __u16 devListOffset;
+ __u16 pciDataStructLen;
+ __u8 pciDataStructRev;
} __attribute__ ((packed));
-static struct imsm_orom imsm_orom[SYS_DEV_MAX];
-static int populated_orom[SYS_DEV_MAX];
+struct orom_entry *orom_entries;
+
+const struct orom_entry *get_orom_entry_by_device_id(__u16 dev_id)
+{
+ struct orom_entry *entry;
+ struct devid_list *devid;
+
+ for (entry = orom_entries; entry; entry = entry->next) {
+ for (devid = entry->devid_list; devid; devid = devid->next) {
+ if (devid->devid == dev_id)
+ return entry;
+ }
+ }
+
+ return NULL;
+}
+
+const struct imsm_orom *get_orom_by_device_id(__u16 dev_id)
+{
+ const struct orom_entry *entry = get_orom_entry_by_device_id(dev_id);
+
+ if (entry)
+ return &entry->orom;
+
+ return NULL;
+}
+
+static struct orom_entry *add_orom(const struct imsm_orom *orom)
+{
+ struct orom_entry *list;
+ struct orom_entry *prev = NULL;
+
+ for (list = orom_entries; list; prev = list, list = list->next)
+ ;
+
+ list = xmalloc(sizeof(struct orom_entry));
+ list->orom = *orom;
+ list->devid_list = NULL;
+ list->next = NULL;
+
+ if (prev == NULL)
+ orom_entries = list;
+ else
+ prev->next = list;
+
+ return list;
+}
+
+static void add_orom_device_id(struct orom_entry *entry, __u16 dev_id)
+{
+ struct devid_list *list;
+ struct devid_list *prev = NULL;
+
+ for (list = entry->devid_list; list; prev = list, list = list->next) {
+ if (list->devid == dev_id)
+ return;
+ }
+ list = xmalloc(sizeof(struct devid_list));
+ list->devid = dev_id;
+ list->next = NULL;
+
+ if (prev == NULL)
+ entry->devid_list = list;
+ else
+ prev->next = list;
+}
static int scan(const void *start, const void *end, const void *data)
{
int offset;
- const struct imsm_orom *imsm_mem;
- int dev;
+ const struct imsm_orom *imsm_mem = NULL;
int len = (end - start);
struct pciExpDataStructFormat *ptr= (struct pciExpDataStructFormat *)data;
(ulong) __le16_to_cpu(ptr->vendorID),
(ulong) __le16_to_cpu(ptr->deviceID));
- if (__le16_to_cpu(ptr->vendorID) == 0x8086) {
- /* serach attached intel devices by device id from OROM */
- dev = device_type_by_id(__le16_to_cpu(ptr->deviceID));
- if (dev == SYS_DEV_UNKNOWN)
- return 0;
- }
- else
+ if (__le16_to_cpu(ptr->vendorID) != 0x8086)
+ return 0;
+
+ if (get_orom_by_device_id(ptr->deviceID))
return 0;
for (offset = 0; offset < len; offset += 4) {
- imsm_mem = start + offset;
- if ((memcmp(imsm_mem->signature, "$VER", 4) == 0)) {
- imsm_orom[dev] = *imsm_mem;
- populated_orom[dev] = 1;
- return populated_orom[SYS_DEV_SATA] && populated_orom[SYS_DEV_SAS];
+ const void *mem = start + offset;
+
+ if ((memcmp(mem, IMSM_OROM_SIGNATURE, 4) == 0)) {
+ imsm_mem = mem;
+ break;
}
}
+
+ if (!imsm_mem)
+ return 0;
+
+ struct orom_entry *orom = add_orom(imsm_mem);
+
+ /* only PciDataStructure with revision 3 and above supports devices list. */
+ if (ptr->pciDataStructRev >= 3 && ptr->devListOffset) {
+ const __u16 *dev_list = (void *)ptr + ptr->devListOffset;
+ int i;
+
+ for (i = 0; dev_list[i] != 0; i++)
+ add_orom_device_id(orom, dev_list[i]);
+ } else {
+ add_orom_device_id(orom, __le16_to_cpu(ptr->deviceID));
+ }
+
return 0;
}
-
-const struct imsm_orom *imsm_platform_test(enum sys_dev_type hba_id, int *populated,
- struct imsm_orom *imsm_orom)
+const struct imsm_orom *imsm_platform_test(struct sys_dev *hba)
{
- memset(imsm_orom, 0, sizeof(*imsm_orom));
- imsm_orom->rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
- IMSM_OROM_RLC_RAID10 | IMSM_OROM_RLC_RAID5;
- imsm_orom->sss = IMSM_OROM_SSS_4kB | IMSM_OROM_SSS_8kB |
- IMSM_OROM_SSS_16kB | IMSM_OROM_SSS_32kB |
- IMSM_OROM_SSS_64kB | IMSM_OROM_SSS_128kB |
- IMSM_OROM_SSS_256kB | IMSM_OROM_SSS_512kB |
- IMSM_OROM_SSS_1MB | IMSM_OROM_SSS_2MB;
- imsm_orom->dpa = IMSM_OROM_DISKS_PER_ARRAY;
- imsm_orom->tds = IMSM_OROM_TOTAL_DISKS;
- imsm_orom->vpa = IMSM_OROM_VOLUMES_PER_ARRAY;
- imsm_orom->vphba = IMSM_OROM_VOLUMES_PER_HBA;
- imsm_orom->attr = imsm_orom->rlc | IMSM_OROM_ATTR_ChecksumVerify;
- *populated = 1;
+ struct imsm_orom orom = {
+ .signature = IMSM_OROM_SIGNATURE,
+ .rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
+ IMSM_OROM_RLC_RAID10 | IMSM_OROM_RLC_RAID5,
+ .sss = IMSM_OROM_SSS_4kB | IMSM_OROM_SSS_8kB |
+ IMSM_OROM_SSS_16kB | IMSM_OROM_SSS_32kB |
+ IMSM_OROM_SSS_64kB | IMSM_OROM_SSS_128kB |
+ IMSM_OROM_SSS_256kB | IMSM_OROM_SSS_512kB |
+ IMSM_OROM_SSS_1MB | IMSM_OROM_SSS_2MB,
+ .dpa = IMSM_OROM_DISKS_PER_ARRAY,
+ .tds = IMSM_OROM_TOTAL_DISKS,
+ .vpa = IMSM_OROM_VOLUMES_PER_ARRAY,
+ .vphba = IMSM_OROM_VOLUMES_PER_HBA
+ };
+ orom.attr = orom.rlc | IMSM_OROM_ATTR_ChecksumVerify;
if (check_env("IMSM_TEST_OROM_NORAID5")) {
- imsm_orom->rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
+ orom.rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
IMSM_OROM_RLC_RAID10;
}
- if (check_env("IMSM_TEST_AHCI_EFI_NORAID5") && (hba_id == SYS_DEV_SAS)) {
- imsm_orom->rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
+ if (check_env("IMSM_TEST_AHCI_EFI_NORAID5") && (hba->type == SYS_DEV_SAS)) {
+ orom.rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
IMSM_OROM_RLC_RAID10;
}
- if (check_env("IMSM_TEST_SCU_EFI_NORAID5") && (hba_id == SYS_DEV_SATA)) {
- imsm_orom->rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
+ if (check_env("IMSM_TEST_SCU_EFI_NORAID5") && (hba->type == SYS_DEV_SATA)) {
+ orom.rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
IMSM_OROM_RLC_RAID10;
}
- return imsm_orom;
-}
+ struct orom_entry *ret = add_orom(&orom);
+ add_orom_device_id(ret, hba->dev_id);
+ return &ret->orom;
+}
-static const struct imsm_orom *find_imsm_hba_orom(enum sys_dev_type hba_id)
+static const struct imsm_orom *find_imsm_hba_orom(struct sys_dev *hba)
{
unsigned long align;
- if (hba_id >= SYS_DEV_MAX)
- return NULL;
+ if (check_env("IMSM_TEST_OROM"))
+ return imsm_platform_test(hba);
- /* it's static data so we only need to read it once */
- if (populated_orom[hba_id]) {
- dprintf("OROM CAP: %p, pid: %d pop: %d\n",
- &imsm_orom[hba_id], (int) getpid(), populated_orom[hba_id]);
- return &imsm_orom[hba_id];
- }
- if (check_env("IMSM_TEST_OROM")) {
- dprintf("OROM CAP: %p, pid: %d pop: %d\n",
- &imsm_orom[hba_id], (int) getpid(), populated_orom[hba_id]);
- return imsm_platform_test(hba_id, &populated_orom[hba_id], &imsm_orom[hba_id]);
- }
/* return empty OROM capabilities in EFI test mode */
- if (check_env("IMSM_TEST_AHCI_EFI") ||
- check_env("IMSM_TEST_SCU_EFI"))
+ if (check_env("IMSM_TEST_AHCI_EFI") || check_env("IMSM_TEST_SCU_EFI"))
return NULL;
-
- if (intel_devices != NULL)
- free_sys_dev(&intel_devices);
-
- intel_devices = find_intel_devices();
+ find_intel_devices();
if (intel_devices == NULL)
return NULL;
scan_adapter_roms(scan);
probe_roms_exit();
- if (intel_devices != NULL)
- free_sys_dev(&intel_devices);
- intel_devices = NULL;
-
- if (populated_orom[hba_id])
- return &imsm_orom[hba_id];
- return NULL;
+ return get_orom_by_device_id(hba->dev_id);
}
#define GUID_STR_MAX 37 /* according to GUID format:
(c) & 0xff, ((c) >> 8) & 0xff, \
(d0), (d1), (d2), (d3), (d4), (d5), (d6), (d7) }})
-
#define SYS_EFI_VAR_PATH "/sys/firmware/efi/vars"
+#define SYS_EFIVARS_PATH "/sys/firmware/efi/efivars"
#define SCU_PROP "RstScuV"
#define AHCI_PROP "RstSataV"
+#define AHCI_SSATA_PROP "RstsSatV"
+#define AHCI_TSATA_PROP "RsttSatV"
+#define VMD_PROP "RstUefiV"
#define VENDOR_GUID \
EFI_GUID(0x193dfefa, 0xa445, 0x4302, 0x99, 0xd8, 0xef, 0x3a, 0xad, 0x1a, 0x04, 0xc6)
-int populated_efi[SYS_DEV_MAX] = { 0, 0 };
-
-static struct imsm_orom imsm_efi[SYS_DEV_MAX];
+#define PCI_CLASS_RAID_CNTRL 0x010400
-const struct imsm_orom *find_imsm_efi(enum sys_dev_type hba_id)
+static int read_efi_var(void *buffer, ssize_t buf_size,
+ const char *variable_name, struct efi_guid guid)
{
- int dfd=-1;
char path[PATH_MAX];
char buf[GUID_STR_MAX];
- int n;
+ int fd;
+ ssize_t n;
- if (hba_id >= SYS_DEV_MAX)
- return NULL;
+ snprintf(path, PATH_MAX, "%s/%s-%s", SYS_EFIVARS_PATH, variable_name, guid_str(buf, guid));
- dprintf("EFI CAP: %p, pid: %d pop: %d\n",
- &imsm_efi[hba_id], (int) getpid(), populated_efi[hba_id]);
+ fd = open(path, O_RDONLY);
+ if (fd < 0)
+ return 1;
- /* it's static data so we only need to read it once */
- if (populated_efi[hba_id]) {
- dprintf("EFI CAP: %p, pid: %d pop: %d\n",
- &imsm_efi[hba_id], (int) getpid(), populated_efi[hba_id]);
- return &imsm_efi[hba_id];
+ /* read the variable attributes and ignore it */
+ n = read(fd, buf, sizeof(__u32));
+ if (n < 0) {
+ close(fd);
+ return 1;
}
- if (check_env("IMSM_TEST_AHCI_EFI") ||
- check_env("IMSM_TEST_SCU_EFI")) {
- dprintf("OROM CAP: %p, pid: %d pop: %d\n",
- &imsm_efi[hba_id], (int) getpid(), populated_efi[hba_id]);
- return imsm_platform_test(hba_id, &populated_efi[hba_id], &imsm_efi[hba_id]);
+
+ /* read the variable data */
+ n = read(fd, buffer, buf_size);
+ close(fd);
+ if (n < buf_size)
+ return 1;
+
+ return 0;
+}
+
+static int read_efi_variable(void *buffer, ssize_t buf_size,
+ const char *variable_name, struct efi_guid guid)
+{
+ char path[PATH_MAX];
+ char buf[GUID_STR_MAX];
+ int dfd;
+ ssize_t n, var_data_len;
+
+ /* Try to read the variable using the new efivarfs interface first.
+ * If that fails, fall back to the old sysfs-efivars interface. */
+ if (!read_efi_var(buffer, buf_size, variable_name, guid))
+ return 0;
+
+ snprintf(path, PATH_MAX, "%s/%s-%s/size", SYS_EFI_VAR_PATH, variable_name, guid_str(buf, guid));
+
+ dprintf("EFI VAR: path=%s\n", path);
+ /* get size of variable data */
+ dfd = open(path, O_RDONLY);
+ if (dfd < 0)
+ return 1;
+
+ n = read(dfd, &buf, sizeof(buf));
+ close(dfd);
+ if (n < 0)
+ return 1;
+ buf[n] = '\0';
+
+ errno = 0;
+ var_data_len = strtoul(buf, NULL, 16);
+ if ((errno == ERANGE && (var_data_len == LONG_MAX)) ||
+ (errno != 0 && var_data_len == 0))
+ return 1;
+
+ /* get data */
+ snprintf(path, PATH_MAX, "%s/%s-%s/data", SYS_EFI_VAR_PATH, variable_name, guid_str(buf, guid));
+
+ dprintf("EFI VAR: path=%s\n", path);
+ dfd = open(path, O_RDONLY);
+ if (dfd < 0)
+ return 1;
+
+ n = read(dfd, buffer, buf_size);
+ close(dfd);
+ if (n != var_data_len || n < buf_size) {
+ return 1;
}
+
+ return 0;
+}
+
+const struct imsm_orom *find_imsm_efi(struct sys_dev *hba)
+{
+ struct imsm_orom orom;
+ struct orom_entry *ret;
+ static const char * const sata_efivars[] = {AHCI_PROP, AHCI_SSATA_PROP,
+ AHCI_TSATA_PROP};
+ unsigned long i;
+
+ if (check_env("IMSM_TEST_AHCI_EFI") || check_env("IMSM_TEST_SCU_EFI"))
+ return imsm_platform_test(hba);
+
/* OROM test is set, return that there is no EFI capabilities */
- if (check_env("IMSM_TEST_OROM")) {
+ if (check_env("IMSM_TEST_OROM"))
return NULL;
- }
- if (hba_id == SYS_DEV_SAS)
- snprintf(path, PATH_MAX, "%s/%s-%s", SYS_EFI_VAR_PATH, SCU_PROP, guid_str(buf, VENDOR_GUID));
- else
- snprintf(path, PATH_MAX, "%s/%s-%s", SYS_EFI_VAR_PATH, AHCI_PROP, guid_str(buf, VENDOR_GUID));
- dprintf("EFI VAR: path=%s\n", path);
- dfd = open(path, O_RDONLY);
- if (dfd < 0) {
- populated_efi[hba_id] = 0;
+ switch (hba->type) {
+ case SYS_DEV_SAS:
+ if (!read_efi_variable(&orom, sizeof(orom), SCU_PROP,
+ VENDOR_GUID))
+ break;
+
return NULL;
- }
- n = read(dfd, &imsm_efi[hba_id], sizeof(imsm_efi[0]));
- close(dfd);
- if (n < (int) (sizeof(imsm_efi[0]))) {
+ case SYS_DEV_SATA:
+ if (hba->class != PCI_CLASS_RAID_CNTRL)
+ return NULL;
+
+ for (i = 0; i < ARRAY_SIZE(sata_efivars); i++) {
+ if (!read_efi_variable(&orom, sizeof(orom),
+ sata_efivars[i], VENDOR_GUID))
+ break;
+
+ }
+ if (i == ARRAY_SIZE(sata_efivars))
+ return NULL;
+
+ break;
+ case SYS_DEV_VMD:
+ if (!read_efi_variable(&orom, sizeof(orom), VMD_PROP,
+ VENDOR_GUID))
+ break;
+ return NULL;
+ default:
return NULL;
}
- populated_efi[hba_id] = 1;
- return &imsm_efi[hba_id];
+
+ ret = add_orom(&orom);
+ add_orom_device_id(ret, hba->dev_id);
+ ret->type = hba->type;
+
+ return &ret->orom;
}
-/*
- * backward interface compatibility
- */
-const struct imsm_orom *find_imsm_orom(void)
+const struct imsm_orom *find_imsm_nvme(struct sys_dev *hba)
{
- return find_imsm_hba_orom(SYS_DEV_SATA);
+ static struct orom_entry *nvme_orom;
+
+ if (hba->type != SYS_DEV_NVME)
+ return NULL;
+
+ if (!nvme_orom) {
+ struct imsm_orom nvme_orom_compat = {
+ .signature = IMSM_NVME_OROM_COMPAT_SIGNATURE,
+ .rlc = IMSM_OROM_RLC_RAID0 | IMSM_OROM_RLC_RAID1 |
+ IMSM_OROM_RLC_RAID10 | IMSM_OROM_RLC_RAID5,
+ .sss = IMSM_OROM_SSS_4kB | IMSM_OROM_SSS_8kB |
+ IMSM_OROM_SSS_16kB | IMSM_OROM_SSS_32kB |
+ IMSM_OROM_SSS_64kB | IMSM_OROM_SSS_128kB,
+ .dpa = IMSM_OROM_DISKS_PER_ARRAY_NVME,
+ .tds = IMSM_OROM_TOTAL_DISKS_NVME,
+ .vpa = IMSM_OROM_VOLUMES_PER_ARRAY,
+ .vphba = IMSM_OROM_TOTAL_DISKS_NVME / 2 * IMSM_OROM_VOLUMES_PER_ARRAY,
+ .attr = IMSM_OROM_ATTR_2TB | IMSM_OROM_ATTR_2TB_DISK,
+ .driver_features = IMSM_OROM_CAPABILITIES_EnterpriseSystem
+ };
+ nvme_orom = add_orom(&nvme_orom_compat);
+ }
+ add_orom_device_id(nvme_orom, hba->dev_id);
+ nvme_orom->type = SYS_DEV_NVME;
+ return &nvme_orom->orom;
}
-const struct imsm_orom *find_imsm_capability(enum sys_dev_type hba_id)
+const struct imsm_orom *find_imsm_capability(struct sys_dev *hba)
{
- const struct imsm_orom *cap=NULL;
+ const struct imsm_orom *cap = get_orom_by_device_id(hba->dev_id);
+ if (cap)
+ return cap;
- if ((cap = find_imsm_efi(hba_id)) != NULL)
+ if (hba->type == SYS_DEV_NVME)
+ return find_imsm_nvme(hba);
+ if ((cap = find_imsm_efi(hba)) != NULL)
return cap;
- if ((cap = find_imsm_hba_orom(hba_id)) != NULL)
+ if ((cap = find_imsm_hba_orom(hba)) != NULL)
return cap;
+
return NULL;
}
-char *devt_to_devpath(dev_t dev)
+/* Check whether the nvme device is represented by nvme subsytem,
+ * if yes virtual path should be changed to hardware device path,
+ * to allow IMSM capabilities detection.
+ * Returns:
+ * hardware path to device - if the device is represented via
+ * nvme virtual subsytem
+ * NULL - if the device is not represented via nvme virtual subsytem
+ */
+char *get_nvme_multipath_dev_hw_path(const char *dev_path)
+{
+ DIR *dir;
+ struct dirent *ent;
+ char *rp = NULL;
+
+ if (strncmp(dev_path, NVME_SUBSYS_PATH, strlen(NVME_SUBSYS_PATH)) != 0)
+ return NULL;
+
+ dir = opendir(dev_path);
+ if (!dir)
+ return NULL;
+
+ for (ent = readdir(dir); ent; ent = readdir(dir)) {
+ char buf[strlen(dev_path) + strlen(ent->d_name) + 1];
+
+ /* Check if dir is a controller, ignore namespaces*/
+ if (!(strncmp(ent->d_name, "nvme", 4) == 0) ||
+ (strrchr(ent->d_name, 'n') != &ent->d_name[0]))
+ continue;
+
+ sprintf(buf, "%s/%s", dev_path, ent->d_name);
+ rp = realpath(buf, NULL);
+ break;
+ }
+
+ closedir(dir);
+ return rp;
+}
+
+/* Description: Return part or whole realpath for the dev
+ * Parameters:
+ * dev - the device to be quered
+ * dev_level - level of "/device" entries. It allows to caller to access
+ * virtual or physical devices which are on "path" to quered
+ * one.
+ * buf - optional, must be PATH_MAX size. If set, then will be used.
+ */
+char *devt_to_devpath(dev_t dev, int dev_level, char *buf)
{
- char device[46];
+ char device[PATH_MAX];
+ char *hw_path;
+ int i;
+ unsigned long device_free_len = sizeof(device) - 1;
+ char dev_str[] = "/device";
+ unsigned long dev_str_len = strlen(dev_str);
+
+ snprintf(device, sizeof(device), "/sys/dev/block/%d:%d", major(dev),
+ minor(dev));
+
+ /* If caller wants block device, return path to it even if it is exposed
+ * via virtual layer.
+ */
+ if (dev_level == 0)
+ return realpath(device, buf);
+
+ device_free_len -= strlen(device);
+ for (i = 0; i < dev_level; i++) {
+ if (device_free_len < dev_str_len)
+ return NULL;
+
+ strncat(device, dev_str, device_free_len);
- sprintf(device, "/sys/dev/block/%d:%d/device", major(dev), minor(dev));
- return canonicalize_file_name(device);
+ /* Resolve nvme-subsystem abstraction if needed
+ */
+ device_free_len -= dev_str_len;
+ if (i == 0) {
+ char rp[PATH_MAX];
+
+ if (!realpath(device, rp))
+ return NULL;
+ hw_path = get_nvme_multipath_dev_hw_path(rp);
+ if (hw_path) {
+ strcpy(device, hw_path);
+ device_free_len = sizeof(device) -
+ strlen(device) - 1;
+ free(hw_path);
+ }
+ }
+ }
+
+ return realpath(device, buf);
}
-char *diskfd_to_devpath(int fd)
+char *diskfd_to_devpath(int fd, int dev_level, char *buf)
{
/* return the device path for a disk, return NULL on error or fd
* refers to a partition
if (!S_ISBLK(st.st_mode))
return NULL;
- return devt_to_devpath(st.st_rdev);
+ return devt_to_devpath(st.st_rdev, dev_level, buf);
}
int path_attached_to_hba(const char *disk_path, const char *hba_path)
int devt_attached_to_hba(dev_t dev, const char *hba_path)
{
- char *disk_path = devt_to_devpath(dev);
+ char *disk_path = devt_to_devpath(dev, 1, NULL);
int rc = path_attached_to_hba(disk_path, hba_path);
if (disk_path)
int disk_attached_to_hba(int fd, const char *hba_path)
{
- char *disk_path = diskfd_to_devpath(fd);
+ char *disk_path = diskfd_to_devpath(fd, 1, NULL);
int rc = path_attached_to_hba(disk_path, hba_path);
if (disk_path)
return rc;
}
+
+char *vmd_domain_to_controller(struct sys_dev *hba, char *buf)
+{
+ struct dirent *ent;
+ DIR *dir;
+ char path[PATH_MAX];
+
+ if (!hba)
+ return NULL;
+
+ if (hba->type != SYS_DEV_VMD)
+ return NULL;
+
+ dir = opendir("/sys/bus/pci/drivers/vmd");
+ if (!dir)
+ return NULL;
+
+ for (ent = readdir(dir); ent; ent = readdir(dir)) {
+ sprintf(path, "/sys/bus/pci/drivers/vmd/%s/domain/device",
+ ent->d_name);
+
+ if (!realpath(path, buf))
+ continue;
+
+ if (strncmp(buf, hba->path, strlen(buf)) == 0) {
+ sprintf(path, "/sys/bus/pci/drivers/vmd/%s", ent->d_name);
+ closedir(dir);
+ return realpath(path, buf);
+ }
+ }
+
+ closedir(dir);
+ return NULL;
+}
+
+/* Scan over all controller's namespaces and compare nsid value to verify if
+ * current one is supported. The routine doesn't check IMSM capabilities for
+ * namespace. Only one nvme namespace is supported by IMSM.
+ * Paramteres:
+ * fd - open descriptor to the nvme namespace
+ * verbose - error logging level
+ * Returns:
+ * 1 - if namespace is supported
+ * 0 - otherwise
+ */
+int imsm_is_nvme_namespace_supported(int fd, int verbose)
+{
+ DIR *dir = NULL;
+ struct dirent *ent;
+ char cntrl_path[PATH_MAX];
+ char ns_path[PATH_MAX];
+ unsigned long long lowest_nsid = ULLONG_MAX;
+ unsigned long long this_nsid;
+ int rv = 0;
+
+
+ if (!diskfd_to_devpath(fd, 1, cntrl_path) ||
+ !diskfd_to_devpath(fd, 0, ns_path)) {
+ if (verbose)
+ pr_err("Cannot get device paths\n");
+ goto abort;
+ }
+
+
+ if (devpath_to_ll(ns_path, "nsid", &this_nsid)) {
+ if (verbose)
+ pr_err("Cannot read nsid value for %s",
+ basename(ns_path));
+ goto abort;
+ }
+
+ dir = opendir(cntrl_path);
+ if (!dir)
+ goto abort;
+
+ /* The lowest nvme namespace is supported */
+ for (ent = readdir(dir); ent; ent = readdir(dir)) {
+ unsigned long long curr_nsid;
+ char curr_ns_path[PATH_MAX + 256];
+
+ if (!strstr(ent->d_name, "nvme"))
+ continue;
+
+ snprintf(curr_ns_path, sizeof(curr_ns_path), "%s/%s",
+ cntrl_path, ent->d_name);
+
+ if (devpath_to_ll(curr_ns_path, "nsid", &curr_nsid))
+ goto abort;
+
+ if (lowest_nsid > curr_nsid)
+ lowest_nsid = curr_nsid;
+ }
+
+ if (this_nsid == lowest_nsid)
+ rv = 1;
+ else if (verbose)
+ pr_err("IMSM is supported on the lowest NVMe namespace\n");
+
+abort:
+ if (dir)
+ closedir(dir);
+
+ return rv;
+}
+
+/* Verify if multipath is supported by NVMe controller
+ * Returns:
+ * 0 - not supported
+ * 1 - supported
+ */
+int is_multipath_nvme(int disk_fd)
+{
+ char ns_path[PATH_MAX];
+
+ if (!diskfd_to_devpath(disk_fd, 0, ns_path))
+ return 0;
+
+ if (strncmp(ns_path, NVME_SUBSYS_PATH, strlen(NVME_SUBSYS_PATH)) == 0)
+ return 1;
+
+ return 0;
+}
projects / thirdparty / mdadm.git / blobdiff