summaryrefslogtreecommitdiffstats
path: root/plugins/zns/zns.c
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--plugins/zns/zns.c1274
1 files changed, 1274 insertions, 0 deletions
diff --git a/plugins/zns/zns.c b/plugins/zns/zns.c
new file mode 100644
index 0000000..a7a3766
--- /dev/null
+++ b/plugins/zns/zns.c
@@ -0,0 +1,1274 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+#include <fcntl.h>
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <inttypes.h>
+#include <linux/fs.h>
+#include <sys/stat.h>
+
+#include "common.h"
+#include "nvme.h"
+#include "libnvme.h"
+#include "nvme-print.h"
+#include "util/cleanup.h"
+
+#define CREATE_CMD
+#include "zns.h"
+
+static const char *namespace_id = "Namespace identifier to use";
+static const char dash[100] = { [0 ... 99] = '-' };
+
+static int detect_zns(nvme_ns_t ns, int *out_supported)
+{
+ int err = 0;
+ char *zoned;
+
+ *out_supported = 0;
+
+ zoned = nvme_get_attr(nvme_ns_get_sysfs_dir(ns), "queue/zoned");
+ if (!zoned) {
+ *out_supported = 0;
+ return err;
+ }
+
+ *out_supported = strcmp("host-managed", zoned) == 0;
+ free(zoned);
+
+ return err;
+}
+
+static int print_zns_list_ns(nvme_ns_t ns)
+{
+ int supported;
+ int err = 0;
+
+ err = detect_zns(ns, &supported);
+ if (err) {
+ perror("Failed to enumerate namespace");
+ return err;
+ }
+
+ if (supported)
+ nvme_show_list_item(ns);
+
+ return err;
+}
+
+static int print_zns_list(nvme_root_t nvme_root)
+{
+ int err = 0;
+ nvme_host_t h;
+ nvme_subsystem_t s;
+ nvme_ctrl_t c;
+ nvme_ns_t n;
+
+ nvme_for_each_host(nvme_root, h) {
+ nvme_for_each_subsystem(h, s) {
+ nvme_subsystem_for_each_ns(s, n) {
+ err = print_zns_list_ns(n);
+ if (err)
+ return err;
+ }
+
+ nvme_subsystem_for_each_ctrl(s, c) {
+ nvme_ctrl_for_each_ns(c, n) {
+ err = print_zns_list_ns(n);
+ if (err)
+ return err;
+ }
+ }
+ }
+ }
+
+ return err;
+}
+
+static int list(int argc, char **argv, struct command *cmd,
+ struct plugin *plugin)
+{
+ int err = 0;
+ nvme_root_t nvme_root;
+
+ printf("%-21s %-20s %-40s %-9s %-26s %-16s %-8s\n", "Node", "SN",
+ "Model", "Namespace", "Usage", "Format", "FW Rev");
+ printf("%-.21s %-.20s %-.40s %-.9s %-.26s %-.16s %-.8s\n", dash, dash,
+ dash, dash, dash, dash, dash);
+
+ nvme_root = nvme_scan(NULL);
+ if (nvme_root) {
+ err = print_zns_list(nvme_root);
+ nvme_free_tree(nvme_root);
+ } else {
+ fprintf(stderr, "Failed to scan nvme subsystems\n");
+ err = -errno;
+ }
+
+ return err;
+}
+
+static int id_ctrl(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Send a ZNS specific Identify Controller command to\n"
+ "the given device and report information about the specified\n"
+ "controller in various formats.";
+
+ enum nvme_print_flags flags;
+ struct nvme_zns_id_ctrl ctrl;
+ struct nvme_dev *dev;
+ int err = -1;
+
+ struct config {
+ char *output_format;
+ };
+
+ struct config cfg = {
+ .output_format = "normal",
+ };
+
+ OPT_ARGS(opts) = {
+ OPT_FMT("output-format", 'o', &cfg.output_format, output_format),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ err = validate_output_format(cfg.output_format, &flags);
+ if (err < 0)
+ goto close_dev;
+
+ err = nvme_zns_identify_ctrl(dev_fd(dev), &ctrl);
+ if (!err)
+ nvme_show_zns_id_ctrl(&ctrl, flags);
+ else if (err > 0)
+ nvme_show_status(err);
+ else
+ perror("zns identify controller");
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int id_ns(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Send a ZNS specific Identify Namespace command to\n"
+ "the given device and report information about the specified\n"
+ "namespace in varios formats.";
+ const char *vendor_specific = "dump binary vendor fields";
+ const char *human_readable = "show identify in readable format";
+
+ enum nvme_print_flags flags;
+ struct nvme_zns_id_ns ns;
+ struct nvme_id_ns id_ns;
+ struct nvme_dev *dev;
+ int err = -1;
+
+ struct config {
+ char *output_format;
+ __u32 namespace_id;
+ bool human_readable;
+ bool vendor_specific;
+ };
+
+ struct config cfg = {
+ .output_format = "normal",
+ };
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_FLAG("vendor-specific", 'v', &cfg.vendor_specific, vendor_specific),
+ OPT_FMT("output-format", 'o', &cfg.output_format, output_format),
+ OPT_FLAG("human-readable", 'H', &cfg.human_readable, human_readable),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ err = validate_output_format(cfg.output_format, &flags);
+ if (err < 0)
+ goto close_dev;
+ if (cfg.vendor_specific)
+ flags |= VS;
+ if (cfg.human_readable)
+ flags |= VERBOSE;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ err = nvme_identify_ns(dev_fd(dev), cfg.namespace_id, &id_ns);
+ if (err) {
+ nvme_show_status(err);
+ goto close_dev;
+ }
+
+ err = nvme_zns_identify_ns(dev_fd(dev), cfg.namespace_id, &ns);
+ if (!err)
+ nvme_show_zns_id_ns(&ns, &id_ns, flags);
+ else if (err > 0)
+ nvme_show_status(err);
+ else
+ perror("zns identify namespace");
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int zns_mgmt_send(int argc, char **argv, struct command *cmd, struct plugin *plugin,
+ const char *desc, enum nvme_zns_send_action zsa)
+{
+ const char *zslba = "starting LBA of the zone for this command";
+ const char *select_all = "send command to all zones";
+ const char *timeout = "timeout value, in milliseconds";
+ struct nvme_dev *dev;
+ int err, zcapc = 0;
+ char *command;
+ __u32 result;
+
+ struct config {
+ __u64 zslba;
+ __u32 namespace_id;
+ bool select_all;
+ __u32 timeout;
+ };
+
+ struct config cfg = {};
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("start-lba", 's', &cfg.zslba, zslba),
+ OPT_FLAG("select-all", 'a', &cfg.select_all, select_all),
+ OPT_UINT("timeout", 't', &cfg.timeout, timeout),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ goto ret;
+
+ err = asprintf(&command, "%s-%s", plugin->name, cmd->name);
+ if (err < 0)
+ goto close_dev;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto free;
+ }
+ }
+
+ struct nvme_zns_mgmt_send_args args = {
+ .args_size = sizeof(args),
+ .fd = dev_fd(dev),
+ .nsid = cfg.namespace_id,
+ .slba = cfg.zslba,
+ .zsa = zsa,
+ .select_all = cfg.select_all,
+ .zsaso = 0,
+ .data_len = 0,
+ .data = NULL,
+ .timeout = cfg.timeout,
+ .result = &result,
+ };
+ err = nvme_zns_mgmt_send(&args);
+ if (!err) {
+ if (zsa == NVME_ZNS_ZSA_RESET)
+ zcapc = result & 0x1;
+
+ printf("%s: Success, action:%d zone:%"PRIx64" all:%d zcapc:%u nsid:%d\n",
+ command, zsa, (uint64_t)cfg.zslba, (int)cfg.select_all,
+ zcapc, cfg.namespace_id);
+ } else if (err > 0) {
+ nvme_show_status(err);
+ } else {
+ perror(desc);
+ }
+free:
+ free(command);
+close_dev:
+ dev_close(dev);
+ret:
+ return err;
+}
+
+static int get_zdes_bytes(int fd, __u32 nsid)
+{
+ struct nvme_zns_id_ns ns;
+ struct nvme_id_ns id_ns;
+ __u8 lbaf;
+ int err;
+
+ err = nvme_identify_ns(fd, nsid, &id_ns);
+ if (err > 0) {
+ nvme_show_status(err);
+ return -1;
+ } else if (err < 0) {
+ perror("identify namespace");
+ return -1;
+ }
+
+ err = nvme_zns_identify_ns(fd, nsid, &ns);
+ if (err > 0) {
+ nvme_show_status(err);
+ return -1;
+ } else if (err < 0) {
+ perror("zns identify namespace");
+ return -1;
+ }
+
+ nvme_id_ns_flbas_to_lbaf_inuse(id_ns.flbas, &lbaf);
+ return ns.lbafe[lbaf].zdes << 6;
+}
+
+static int zone_mgmt_send(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Zone Management Send";
+ const char *zslba =
+ "starting LBA of the zone for this command(for flush action, last lba to flush)";
+ const char *zsaso = "Zone Send Action Specific Option";
+ const char *select_all = "send command to all zones";
+ const char *zsa = "zone send action";
+ const char *data_len = "buffer length if data required";
+ const char *data = "optional file for data (default stdin)";
+ const char *timeout = "timeout value, in milliseconds";
+
+ int ffd = STDIN_FILENO, err = -1;
+ struct nvme_dev *dev;
+ void *buf = NULL;
+
+ struct config {
+ __u64 zslba;
+ __u32 namespace_id;
+ bool zsaso;
+ bool select_all;
+ __u8 zsa;
+ int data_len;
+ char *file;
+ __u32 timeout;
+ };
+
+ struct config cfg = {};
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("start-lba", 's', &cfg.zslba, zslba),
+ OPT_FLAG("zsaso", 'o', &cfg.zsaso, zsaso),
+ OPT_FLAG("select-all", 'a', &cfg.select_all, select_all),
+ OPT_BYTE("zsa", 'z', &cfg.zsa, zsa),
+ OPT_UINT("data-len", 'l', &cfg.data_len, data_len),
+ OPT_FILE("data", 'd', &cfg.file, data),
+ OPT_UINT("timeout", 't', &cfg.timeout, timeout),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ if (!cfg.zsa) {
+ fprintf(stderr, "zone send action must be specified\n");
+ err = -EINVAL;
+ goto close_dev;
+ }
+
+ if (cfg.zsa == NVME_ZNS_ZSA_SET_DESC_EXT) {
+ if (!cfg.data_len) {
+ int data_len = get_zdes_bytes(dev_fd(dev),
+ cfg.namespace_id);
+
+ if (data_len == 0) {
+ fprintf(stderr, "Zone Descriptor Extensions are not supported\n");
+ goto close_dev;
+ } else if (data_len < 0) {
+ err = data_len;
+ goto close_dev;
+ }
+ cfg.data_len = data_len;
+ }
+ if (posix_memalign(&buf, getpagesize(), cfg.data_len)) {
+ fprintf(stderr, "can not allocate feature payload\n");
+ goto close_dev;
+ }
+ memset(buf, 0, cfg.data_len);
+
+ if (cfg.file) {
+ ffd = open(cfg.file, O_RDONLY);
+ if (ffd < 0) {
+ perror(cfg.file);
+ goto free;
+ }
+ }
+
+ err = read(ffd, (void *)buf, cfg.data_len);
+ if (err < 0) {
+ perror("read");
+ goto close_ffd;
+ }
+ } else {
+ if (cfg.file || cfg.data_len) {
+ fprintf(stderr, "data, data_len only valid with set extended descriptor\n");
+ err = -EINVAL;
+ goto close_dev;
+ }
+ }
+
+ struct nvme_zns_mgmt_send_args args = {
+ .args_size = sizeof(args),
+ .fd = dev_fd(dev),
+ .nsid = cfg.namespace_id,
+ .slba = cfg.zslba,
+ .zsa = cfg.zsa,
+ .select_all = cfg.select_all,
+ .zsaso = cfg.zsaso,
+ .data_len = cfg.data_len,
+ .data = buf,
+ .timeout = cfg.timeout,
+ .result = NULL,
+ };
+ err = nvme_zns_mgmt_send(&args);
+ if (!err)
+ printf("zone-mgmt-send: Success, action:%d zone:%"PRIx64" all:%d nsid:%d\n",
+ cfg.zsa, (uint64_t)cfg.zslba, (int)cfg.select_all, cfg.namespace_id);
+ else if (err > 0)
+ nvme_show_status(err);
+ else
+ perror("zns zone-mgmt-send");
+
+close_ffd:
+ if (cfg.file)
+ close(ffd);
+free:
+ free(buf);
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int close_zone(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Close zones\n";
+
+ return zns_mgmt_send(argc, argv, cmd, plugin, desc, NVME_ZNS_ZSA_CLOSE);
+}
+
+static int finish_zone(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Finish zones\n";
+
+ return zns_mgmt_send(argc, argv, cmd, plugin, desc, NVME_ZNS_ZSA_FINISH);
+}
+
+static int open_zone(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Open zones\n";
+ const char *zslba = "starting LBA of the zone for this command";
+ const char *zrwaa = "Allocate Zone Random Write Area to zone";
+ const char *select_all = "send command to all zones";
+ const char *timeout = "timeout value, in milliseconds";
+ struct nvme_dev *dev;
+ int err;
+
+ struct config {
+ __u64 zslba;
+ __u32 namespace_id;
+ bool zrwaa;
+ bool select_all;
+ __u32 timeout;
+ };
+
+ struct config cfg = {
+ };
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("start-lba", 's', &cfg.zslba, zslba),
+ OPT_FLAG("zrwaa", 'r', &cfg.zrwaa, zrwaa),
+ OPT_FLAG("select-all", 'a', &cfg.select_all, select_all),
+ OPT_UINT("timeout", 't', &cfg.timeout, timeout),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ struct nvme_zns_mgmt_send_args args = {
+ .args_size = sizeof(args),
+ .fd = dev_fd(dev),
+ .nsid = cfg.namespace_id,
+ .slba = cfg.zslba,
+ .zsa = NVME_ZNS_ZSA_OPEN,
+ .select_all = cfg.select_all,
+ .zsaso = cfg.zrwaa,
+ .data_len = 0,
+ .data = NULL,
+ .timeout = cfg.timeout,
+ .result = NULL,
+ };
+ err = nvme_zns_mgmt_send(&args);
+ if (!err)
+ printf("zns-open-zone: Success zone slba:%"PRIx64" nsid:%d\n",
+ (uint64_t)cfg.zslba, cfg.namespace_id);
+ else
+ nvme_show_status(err);
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int reset_zone(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Reset zones\n";
+
+ return zns_mgmt_send(argc, argv, cmd, plugin, desc, NVME_ZNS_ZSA_RESET);
+}
+
+static int offline_zone(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Offline zones\n";
+
+ return zns_mgmt_send(argc, argv, cmd, plugin, desc, NVME_ZNS_ZSA_OFFLINE);
+}
+
+static int set_zone_desc(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Set Zone Descriptor Extension\n";
+ const char *zslba = "starting LBA of the zone for this command";
+ const char *zrwaa = "Allocate Zone Random Write Area to zone";
+ const char *data = "optional file for zone extension data (default stdin)";
+ const char *timeout = "timeout value, in milliseconds";
+
+ int ffd = STDIN_FILENO, err;
+ struct nvme_dev *dev;
+ void *buf = NULL;
+ int data_len;
+
+ struct config {
+ __u64 zslba;
+ bool zrwaa;
+ __u32 namespace_id;
+ char *file;
+ __u32 timeout;
+ };
+
+ struct config cfg = {};
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("start-lba", 's', &cfg.zslba, zslba),
+ OPT_FLAG("zrwaa", 'r', &cfg.zrwaa, zrwaa),
+ OPT_FILE("data", 'd', &cfg.file, data),
+ OPT_UINT("timeout", 't', &cfg.timeout, timeout),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ data_len = get_zdes_bytes(dev_fd(dev), cfg.namespace_id);
+
+ if (!data_len || data_len < 0) {
+ fprintf(stderr,
+ "zone format does not provide descriptor extension\n");
+ errno = EINVAL;
+ err = -1;
+ goto close_dev;
+ }
+
+ buf = calloc(1, data_len);
+ if (!buf) {
+ perror("could not alloc memory for zone desc");
+ err = -ENOMEM;
+ goto close_dev;
+ }
+
+ if (cfg.file) {
+ ffd = open(cfg.file, O_RDONLY);
+ if (ffd < 0) {
+ perror(cfg.file);
+ err = -1;
+ goto free;
+ }
+ }
+
+ err = read(ffd, (void *)buf, data_len);
+ if (err < 0) {
+ perror("read");
+ goto close_ffd;
+ }
+
+ struct nvme_zns_mgmt_send_args args = {
+ .args_size = sizeof(args),
+ .fd = dev_fd(dev),
+ .nsid = cfg.namespace_id,
+ .slba = cfg.zslba,
+ .zsa = NVME_ZNS_ZSA_SET_DESC_EXT,
+ .select_all = 0,
+ .zsaso = cfg.zrwaa,
+ .data_len = data_len,
+ .data = buf,
+ .timeout = cfg.timeout,
+ .result = NULL,
+ };
+ err = nvme_zns_mgmt_send(&args);
+ if (!err)
+ printf("set-zone-desc: Success, zone:%"PRIx64" nsid:%d\n",
+ (uint64_t)cfg.zslba, cfg.namespace_id);
+ else if (err > 0)
+ nvme_show_status(err);
+ else
+ perror("zns set-zone-desc");
+close_ffd:
+ if (cfg.file)
+ close(ffd);
+free:
+ free(buf);
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+
+static int zrwa_flush_zone(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Flush Explicit ZRWA Range";
+ const char *slba = "LBA to flush up to";
+ const char *timeout = "timeout value, in milliseconds";
+ struct nvme_dev *dev;
+ int err;
+
+ struct config {
+ __u64 lba;
+ __u32 namespace_id;
+ __u32 timeout;
+ };
+
+ struct config cfg = {};
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("lba", 'l', &cfg.lba, slba),
+ OPT_UINT("timeout", 't', &cfg.timeout, timeout),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ struct nvme_zns_mgmt_send_args args = {
+ .args_size = sizeof(args),
+ .fd = dev_fd(dev),
+ .nsid = cfg.namespace_id,
+ .slba = cfg.lba,
+ .zsa = NVME_ZNS_ZSA_ZRWA_FLUSH,
+ .select_all = 0,
+ .zsaso = 0,
+ .data_len = 0,
+ .data = NULL,
+ .timeout = cfg.timeout,
+ .result = NULL,
+ };
+ err = nvme_zns_mgmt_send(&args);
+ if (!err)
+ printf("zrwa-flush-zone: Success, lba:%"PRIx64" nsid:%d\n",
+ (uint64_t)cfg.lba, cfg.namespace_id);
+ else
+ nvme_show_status(err);
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int zone_mgmt_recv(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Zone Management Receive";
+ const char *zslba = "starting LBA of the zone";
+ const char *zra = "Zone Receive Action";
+ const char *zrasf = "Zone Receive Action Specific Field(Reporting Options)";
+ const char *partial = "Zone Receive Action Specific Features(Partial Report)";
+ const char *data_len = "length of data in bytes";
+
+ enum nvme_print_flags flags;
+ struct nvme_dev *dev;
+ void *data = NULL;
+ int err = -1;
+
+ struct config {
+ char *output_format;
+ __u64 zslba;
+ __u32 namespace_id;
+ __u8 zra;
+ __u8 zrasf;
+ bool partial;
+ __u32 data_len;
+ };
+
+ struct config cfg = {
+ .output_format = "normal",
+ };
+
+ OPT_ARGS(opts) = {
+ OPT_FMT("output-format", 'o', &cfg.output_format, output_format),
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("start-lba", 's', &cfg.zslba, zslba),
+ OPT_BYTE("zra", 'z', &cfg.zra, zra),
+ OPT_BYTE("zrasf", 'S', &cfg.zrasf, zrasf),
+ OPT_FLAG("partial", 'p', &cfg.partial, partial),
+ OPT_UINT("data-len", 'l', &cfg.data_len, data_len),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ err = validate_output_format(cfg.output_format, &flags);
+ if (err < 0)
+ goto close_dev;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ if (cfg.zra == NVME_ZNS_ZRA_REPORT_ZONES && !cfg.data_len) {
+ fprintf(stderr, "error: data len is needed for NVME_ZRA_ZONE_REPORT\n");
+ err = -EINVAL;
+ goto close_dev;
+ }
+ if (cfg.data_len) {
+ data = calloc(1, cfg.data_len);
+ if (!data) {
+ perror("could not alloc memory for zone mgmt receive data");
+ err = -ENOMEM;
+ goto close_dev;
+ }
+ }
+
+ struct nvme_zns_mgmt_recv_args args = {
+ .args_size = sizeof(args),
+ .fd = dev_fd(dev),
+ .nsid = cfg.namespace_id,
+ .slba = cfg.zslba,
+ .zra = cfg.zra,
+ .zrasf = cfg.zrasf,
+ .zras_feat = cfg.partial,
+ .data_len = cfg.data_len,
+ .data = data,
+ .timeout = NVME_DEFAULT_IOCTL_TIMEOUT,
+ .result = NULL,
+ };
+ err = nvme_zns_mgmt_recv(&args);
+ if (!err)
+ printf("zone-mgmt-recv: Success, action:%d zone:%"PRIx64" nsid:%d\n",
+ cfg.zra, (uint64_t)cfg.zslba, cfg.namespace_id);
+ else if (err > 0)
+ nvme_show_status(err);
+ else
+ perror("zns zone-mgmt-recv");
+
+ free(data);
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int report_zones(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Retrieve the Report Zones data structure";
+ const char *zslba = "starting LBA of the zone";
+ const char *num_descs = "number of descriptors to retrieve (default: all of them)";
+ const char *state = "state of zones to list";
+ const char *ext = "set to use the extended report zones";
+ const char *part = "set to use the partial report";
+ const char *verbose = "show report zones verbosity";
+
+ enum nvme_print_flags flags;
+ int zdes = 0, err = -1;
+ struct nvme_dev *dev;
+ __u32 report_size;
+ struct nvme_zone_report *report, *buff;
+ _cleanup_huge_ struct nvme_mem_huge mh = { 0, };
+
+ unsigned int nr_zones_chunks = 1024, /* 1024 entries * 64 bytes per entry = 64k byte transfer */
+ nr_zones_retrieved = 0,
+ nr_zones,
+ log_len;
+ __u64 offset;
+ int total_nr_zones = 0;
+ struct nvme_zns_id_ns id_zns;
+ struct nvme_id_ns id_ns;
+ uint8_t lbaf;
+ __le64 zsze;
+ struct json_object *zone_list = NULL;
+
+ struct config {
+ char *output_format;
+ __u64 zslba;
+ __u32 namespace_id;
+ int num_descs;
+ int state;
+ bool verbose;
+ bool extended;
+ bool partial;
+ };
+
+ struct config cfg = {
+ .output_format = "normal",
+ .num_descs = -1,
+ };
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("start-lba", 's', &cfg.zslba, zslba),
+ OPT_UINT("descs", 'd', &cfg.num_descs, num_descs),
+ OPT_UINT("state", 'S', &cfg.state, state),
+ OPT_FMT("output-format", 'o', &cfg.output_format, output_format),
+ OPT_FLAG("verbose", 'v', &cfg.verbose, verbose),
+ OPT_FLAG("extended", 'e', &cfg.extended, ext),
+ OPT_FLAG("partial", 'p', &cfg.partial, part),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ err = validate_output_format(cfg.output_format, &flags);
+ if (err < 0)
+ goto close_dev;
+ if (cfg.verbose)
+ flags |= VERBOSE;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ if (cfg.extended) {
+ zdes = get_zdes_bytes(dev_fd(dev), cfg.namespace_id);
+ if (zdes < 0) {
+ err = zdes;
+ goto close_dev;
+ }
+ }
+
+ err = nvme_identify_ns(dev_fd(dev), cfg.namespace_id, &id_ns);
+ if (err) {
+ nvme_show_status(err);
+ goto close_dev;
+ }
+
+ err = nvme_zns_identify_ns(dev_fd(dev), cfg.namespace_id, &id_zns);
+ if (!err) {
+ /* get zsze field from zns id ns data - needed for offset calculation */
+ nvme_id_ns_flbas_to_lbaf_inuse(id_ns.flbas, &lbaf);
+ zsze = le64_to_cpu(id_zns.lbafe[lbaf].zsze);
+ } else {
+ nvme_show_status(err);
+ goto close_dev;
+ }
+
+ log_len = sizeof(struct nvme_zone_report);
+ buff = calloc(1, log_len);
+ if (!buff) {
+ err = -ENOMEM;
+ goto close_dev;
+ }
+
+ err = nvme_zns_report_zones(dev_fd(dev), cfg.namespace_id, 0,
+ cfg.state, false, false,
+ log_len, buff,
+ NVME_DEFAULT_IOCTL_TIMEOUT, NULL);
+ if (err > 0) {
+ nvme_show_status(err);
+ goto free_buff;
+ } else if (err < 0) {
+ perror("zns report-zones");
+ goto free_buff;
+ }
+
+ total_nr_zones = le64_to_cpu(buff->nr_zones);
+
+ if (cfg.num_descs == -1)
+ cfg.num_descs = total_nr_zones;
+
+ nr_zones = cfg.num_descs;
+ if (nr_zones < nr_zones_chunks)
+ nr_zones_chunks = nr_zones;
+
+ log_len = sizeof(struct nvme_zone_report) + ((sizeof(struct nvme_zns_desc) * nr_zones_chunks) + (nr_zones_chunks * zdes));
+ report_size = log_len;
+
+ report = nvme_alloc_huge(report_size, &mh);
+ if (!report) {
+ perror("alloc");
+ err = -ENOMEM;
+ goto close_dev;
+ }
+
+ offset = cfg.zslba;
+
+ nvme_zns_start_zone_list(total_nr_zones, &zone_list, flags);
+
+ while (nr_zones_retrieved < nr_zones) {
+ if (nr_zones_retrieved >= nr_zones)
+ break;
+
+ if (nr_zones_retrieved + nr_zones_chunks > nr_zones) {
+ nr_zones_chunks = nr_zones - nr_zones_retrieved;
+ log_len = sizeof(struct nvme_zone_report) + ((sizeof(struct nvme_zns_desc) * nr_zones_chunks) + (nr_zones_chunks * zdes));
+ }
+
+ err = nvme_zns_report_zones(dev_fd(dev), cfg.namespace_id,
+ offset,
+ cfg.state, cfg.extended,
+ cfg.partial, log_len, report,
+ NVME_DEFAULT_IOCTL_TIMEOUT, NULL);
+ if (err > 0) {
+ nvme_show_status(err);
+ break;
+ }
+
+ if (!err)
+ nvme_show_zns_report_zones(report, nr_zones_chunks,
+ zdes, log_len, zone_list, flags);
+
+ nr_zones_retrieved += nr_zones_chunks;
+ offset = le64_to_cpu(report->entries[nr_zones_chunks-1].zslba) + zsze;
+ }
+
+ nvme_zns_finish_zone_list(total_nr_zones, zone_list, flags);
+
+free_buff:
+ free(buff);
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int zone_append(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "The zone append command is used to write to a zone\n"
+ "using the slba of the zone, and the write will be appended from the\n"
+ "write pointer of the zone";
+ const char *zslba = "starting LBA of the zone";
+ const char *data = "file containing data to write";
+ const char *metadata = "file with metadata to be written";
+ const char *limited_retry = "limit media access attempts";
+ const char *fua = "force unit access";
+ const char *prinfo = "protection information action and checks field";
+ const char *piremap = "protection information remap (for type 1 PI)";
+ const char *ref_tag = "reference tag for end-to-end PI";
+ const char *lbat = "logical block application tag for end-to-end PI";
+ const char *lbatm = "logical block application tag mask for end-to-end PI";
+ const char *metadata_size = "size of metadata in bytes";
+ const char *data_size = "size of data in bytes";
+ const char *latency = "output latency statistics";
+
+ int err = -1, dfd = STDIN_FILENO, mfd = STDIN_FILENO;
+ unsigned int lba_size, meta_size;
+ void *buf = NULL, *mbuf = NULL;
+ __u16 nblocks, control = 0;
+ struct nvme_dev *dev;
+ __u64 result;
+ __u8 lba_index;
+ struct timeval start_time, end_time;
+
+ struct nvme_id_ns ns;
+
+ struct config {
+ char *data;
+ char *metadata;
+ __u64 zslba;
+ __u64 data_size;
+ __u64 metadata_size;
+ bool limited_retry;
+ bool fua;
+ __u32 namespace_id;
+ __u64 ref_tag;
+ __u16 lbat;
+ __u16 lbatm;
+ __u8 prinfo;
+ bool piremap;
+ bool latency;
+ };
+
+ struct config cfg = {};
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_SUFFIX("zslba", 's', &cfg.zslba, zslba),
+ OPT_SUFFIX("data-size", 'z', &cfg.data_size, data_size),
+ OPT_SUFFIX("metadata-size", 'y', &cfg.metadata_size, metadata_size),
+ OPT_FILE("data", 'd', &cfg.data, data),
+ OPT_FILE("metadata", 'M', &cfg.metadata, metadata),
+ OPT_FLAG("limited-retry", 'l', &cfg.limited_retry, limited_retry),
+ OPT_FLAG("force-unit-access", 'f', &cfg.fua, fua),
+ OPT_SUFFIX("ref-tag", 'r', &cfg.ref_tag, ref_tag),
+ OPT_SHRT("app-tag-mask", 'm', &cfg.lbatm, lbatm),
+ OPT_SHRT("app-tag", 'a', &cfg.lbat, lbat),
+ OPT_BYTE("prinfo", 'p', &cfg.prinfo, prinfo),
+ OPT_FLAG("piremap", 'P', &cfg.piremap, piremap),
+ OPT_FLAG("latency", 't', &cfg.latency, latency),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ if (!cfg.data_size) {
+ fprintf(stderr, "Append size not provided\n");
+ errno = EINVAL;
+ goto close_dev;
+ }
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ err = nvme_identify_ns(dev_fd(dev), cfg.namespace_id, &ns);
+ if (err) {
+ nvme_show_status(err);
+ goto close_dev;
+ }
+
+ nvme_id_ns_flbas_to_lbaf_inuse(ns.flbas, &lba_index);
+ lba_size = 1 << ns.lbaf[lba_index].ds;
+ if (cfg.data_size & (lba_size - 1)) {
+ fprintf(stderr,
+ "Data size:%#"PRIx64" not aligned to lba size:%#x\n",
+ (uint64_t)cfg.data_size, lba_size);
+ errno = EINVAL;
+ goto close_dev;
+ }
+
+ meta_size = ns.lbaf[lba_index].ms;
+ if (meta_size && !(meta_size == 8 && (cfg.prinfo & 0x8)) &&
+ (!cfg.metadata_size || cfg.metadata_size % meta_size)) {
+ fprintf(stderr,
+ "Metadata size:%#"PRIx64" not aligned to metadata size:%#x\n",
+ (uint64_t)cfg.metadata_size, meta_size);
+ errno = EINVAL;
+ goto close_dev;
+ }
+
+ if (cfg.prinfo > 0xf) {
+ fprintf(stderr, "Invalid value for prinfo:%#x\n", cfg.prinfo);
+ errno = EINVAL;
+ goto close_dev;
+ }
+
+ if (cfg.data) {
+ dfd = open(cfg.data, O_RDONLY);
+ if (dfd < 0) {
+ perror(cfg.data);
+ goto close_dev;
+ }
+ }
+
+ if (posix_memalign(&buf, getpagesize(), cfg.data_size)) {
+ fprintf(stderr, "No memory for data size:%"PRIx64"\n",
+ (uint64_t)cfg.data_size);
+ goto close_dfd;
+ }
+
+ memset(buf, 0, cfg.data_size);
+ err = read(dfd, buf, cfg.data_size);
+ if (err < 0) {
+ perror("read-data");
+ goto free_data;
+ }
+
+ if (cfg.metadata) {
+ mfd = open(cfg.metadata, O_RDONLY);
+ if (mfd < 0) {
+ perror(cfg.metadata);
+ err = -1;
+ goto free_data;
+ }
+ }
+
+ if (cfg.metadata_size) {
+ if (posix_memalign(&mbuf, getpagesize(), meta_size)) {
+ fprintf(stderr, "No memory for metadata size:%d\n",
+ meta_size);
+ err = -1;
+ goto close_mfd;
+ }
+
+ memset(mbuf, 0, cfg.metadata_size);
+ err = read(mfd, mbuf, cfg.metadata_size);
+ if (err < 0) {
+ perror("read-metadata");
+ goto free_meta;
+ }
+ }
+
+ nblocks = (cfg.data_size / lba_size) - 1;
+ control |= (cfg.prinfo << 10);
+ if (cfg.limited_retry)
+ control |= NVME_IO_LR;
+ if (cfg.fua)
+ control |= NVME_IO_FUA;
+ if (cfg.piremap)
+ control |= NVME_IO_ZNS_APPEND_PIREMAP;
+
+ struct nvme_zns_append_args args = {
+ .args_size = sizeof(args),
+ .fd = dev_fd(dev),
+ .nsid = cfg.namespace_id,
+ .zslba = cfg.zslba,
+ .nlb = nblocks,
+ .control = control,
+ .ilbrt_u64 = cfg.ref_tag,
+ .lbat = cfg.lbat,
+ .lbatm = cfg.lbatm,
+ .data_len = cfg.data_size,
+ .data = buf,
+ .metadata_len = cfg.metadata_size,
+ .metadata = mbuf,
+ .timeout = NVME_DEFAULT_IOCTL_TIMEOUT,
+ .result = &result,
+ };
+
+ gettimeofday(&start_time, NULL);
+ err = nvme_zns_append(&args);
+ gettimeofday(&end_time, NULL);
+ if (cfg.latency)
+ printf(" latency: zone append: %llu us\n",
+ elapsed_utime(start_time, end_time));
+
+ if (!err)
+ printf("Success appended data to LBA %"PRIx64"\n", (uint64_t)result);
+ else if (err > 0)
+ nvme_show_status(err);
+ else
+ perror("zns zone-append");
+
+free_meta:
+ free(mbuf);
+close_mfd:
+ if (cfg.metadata)
+ close(mfd);
+free_data:
+ free(buf);
+close_dfd:
+ if (cfg.data)
+ close(dfd);
+close_dev:
+ dev_close(dev);
+ return err;
+}
+
+static int changed_zone_list(int argc, char **argv, struct command *cmd, struct plugin *plugin)
+{
+ const char *desc = "Retrieve Changed Zone log for the given device";
+ const char *rae = "retain an asynchronous event";
+
+ struct nvme_zns_changed_zone_log log;
+ enum nvme_print_flags flags;
+ struct nvme_dev *dev;
+ int err = -1;
+
+ struct config {
+ char *output_format;
+ __u32 namespace_id;
+ bool rae;
+ };
+
+ struct config cfg = {
+ .output_format = "normal",
+ };
+
+ OPT_ARGS(opts) = {
+ OPT_UINT("namespace-id", 'n', &cfg.namespace_id, namespace_id),
+ OPT_FMT("output-format", 'o', &cfg.output_format, output_format),
+ OPT_FLAG("rae", 'r', &cfg.rae, rae),
+ OPT_END()
+ };
+
+ err = parse_and_open(&dev, argc, argv, desc, opts);
+ if (err)
+ return errno;
+
+ err = validate_output_format(cfg.output_format, &flags);
+ if (err < 0)
+ goto close_dev;
+
+ if (!cfg.namespace_id) {
+ err = nvme_get_nsid(dev_fd(dev), &cfg.namespace_id);
+ if (err < 0) {
+ perror("get-namespace-id");
+ goto close_dev;
+ }
+ }
+
+ err = nvme_get_log_zns_changed_zones(dev_fd(dev), cfg.namespace_id,
+ cfg.rae, &log);
+ if (!err)
+ nvme_show_zns_changed(&log, flags);
+ else if (err > 0)
+ nvme_show_status(err);
+ else
+ perror("zns changed-zone-list");
+
+close_dev:
+ dev_close(dev);
+ return err;
+}