diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2023-04-07 12:43:34 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2023-04-07 12:43:34 +0000 |
commit | 8e1dcd44fcad7826e5f8fa6c5caa6c23989518d9 (patch) | |
tree | 8a5e1e9d02b2562d57d57ef88e26d4778ebeeab3 /plugins/ocp | |
parent | Releasing debian version 2.4+really2.3-2. (diff) | |
download | nvme-cli-8e1dcd44fcad7826e5f8fa6c5caa6c23989518d9.tar.xz nvme-cli-8e1dcd44fcad7826e5f8fa6c5caa6c23989518d9.zip |
Merging upstream version 2.4+really2.4.
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'plugins/ocp')
-rw-r--r-- | plugins/ocp/meson.build | 1 | ||||
-rw-r--r-- | plugins/ocp/ocp-clear-fw-update-history.c | 56 | ||||
-rw-r--r-- | plugins/ocp/ocp-clear-fw-update-history.h | 3 | ||||
-rw-r--r-- | plugins/ocp/ocp-nvme.c | 1276 | ||||
-rw-r--r-- | plugins/ocp/ocp-nvme.h | 12 | ||||
-rw-r--r-- | plugins/ocp/ocp-smart-extended-log.c | 352 | ||||
-rw-r--r-- | plugins/ocp/ocp-smart-extended-log.h | 18 | ||||
-rw-r--r-- | plugins/ocp/ocp-utils.c | 69 | ||||
-rw-r--r-- | plugins/ocp/ocp-utils.h | 2 |
9 files changed, 1025 insertions, 764 deletions
diff --git a/plugins/ocp/meson.build b/plugins/ocp/meson.build index a4e5d20..641239a 100644 --- a/plugins/ocp/meson.build +++ b/plugins/ocp/meson.build @@ -2,5 +2,6 @@ sources += [ 'plugins/ocp/ocp-utils.c', 'plugins/ocp/ocp-nvme.c', 'plugins/ocp/ocp-clear-fw-update-history.c', + 'plugins/ocp/ocp-smart-extended-log.c', ] diff --git a/plugins/ocp/ocp-clear-fw-update-history.c b/plugins/ocp/ocp-clear-fw-update-history.c index fef09cf..b9235b8 100644 --- a/plugins/ocp/ocp-clear-fw-update-history.c +++ b/plugins/ocp/ocp-clear-fw-update-history.c @@ -15,59 +15,7 @@ static const __u8 OCP_FID_CLEAR_FW_ACTIVATION_HISTORY = 0xC1; int ocp_clear_fw_update_history(int argc, char **argv, struct command *cmd, struct plugin *plugin) { const char *desc = "OCP Clear Firmware Update History"; - __u32 result = 0; - __u32 clear_fw_history = 1 << 31; - struct nvme_dev *dev; - int uuid_index = 0; - bool no_uuid = false; - int err; - OPT_ARGS(opts) = { - OPT_FLAG("no-uuid", 'n', &no_uuid, - "Skip UUID index search (UUID index not required for OCP 1.0)"), - OPT_END() - }; - - err = parse_and_open(&dev, argc, argv, desc, opts); - if (err) - return err; - if (no_uuid == false) { - // OCP 2.0 requires UUID index support - err = ocp_get_uuid_index(dev, &uuid_index); - if (err || uuid_index == 0) { - fprintf(stderr, "ERROR: No OCP UUID index found\n"); - goto close_dev; - } - } - - struct nvme_set_features_args args = { - .result = &result, - .data = NULL, - .args_size = sizeof(args), - .fd = dev_fd(dev), - .timeout = NVME_DEFAULT_IOCTL_TIMEOUT, - .nsid = 0, - .cdw11 = clear_fw_history, - .cdw12 = 0, - .cdw13 = 0, - .cdw15 = 0, - .data_len = 0, - .save = 0, - .uuidx = uuid_index, - .fid = OCP_FID_CLEAR_FW_ACTIVATION_HISTORY, - }; - - err = nvme_set_features(&args); - - if (err == 0) - printf("Success : %s\n", desc); - else if (err > 0) - nvme_show_status(err); - else - printf("Fail : %s\n", desc); -close_dev: - /* Redundant close() to make static code analysis happy */ - close(dev->direct.fd); - dev_close(dev); - return err; + return ocp_clear_feature(argc, argv, desc, + OCP_FID_CLEAR_FW_ACTIVATION_HISTORY); } diff --git a/plugins/ocp/ocp-clear-fw-update-history.h b/plugins/ocp/ocp-clear-fw-update-history.h index 25fb6b1..cd0844c 100644 --- a/plugins/ocp/ocp-clear-fw-update-history.h +++ b/plugins/ocp/ocp-clear-fw-update-history.h @@ -6,4 +6,5 @@ * leonardo.da.cunha@solidigm.com */ -int ocp_clear_fw_update_history(int argc, char **argv, struct command *cmd, struct plugin *plugin); +int ocp_clear_fw_update_history(int argc, char **argv, + struct command *cmd, struct plugin *plugin); diff --git a/plugins/ocp/ocp-nvme.c b/plugins/ocp/ocp-nvme.c index 14a5f30..a864363 100644 --- a/plugins/ocp/ocp-nvme.c +++ b/plugins/ocp/ocp-nvme.c @@ -3,7 +3,7 @@ * * Authors: Arthur Shau <arthurshau@fb.com>, * Wei Zhang <wzhang@fb.com>, - * Venkat Ramesh <venkatraghavan@fb.com> + * Venkat Ramesh <venkatraghavan@fb.com> */ #include <stdio.h> #include <string.h> @@ -22,753 +22,627 @@ #include "linux/types.h" #include "util/types.h" #include "nvme-print.h" + +#include "ocp-smart-extended-log.h" #include "ocp-clear-fw-update-history.h" #define CREATE_CMD #include "ocp-nvme.h" +#include "ocp-utils.h" -/* C0 SCAO Log Page */ -#define C0_SMART_CLOUD_ATTR_LEN 0x200 -#define C0_SMART_CLOUD_ATTR_OPCODE 0xC0 -#define C0_GUID_LENGTH 16 -#define C0_ACTIVE_BUCKET_TIMER_INCREMENT 5 -#define C0_ACTIVE_THRESHOLD_INCREMENT 5 -#define C0_MINIMUM_WINDOW_INCREMENT 100 - -static __u8 scao_guid[C0_GUID_LENGTH] = { 0xC5, 0xAF, 0x10, 0x28, 0xEA, 0xBF, - 0xF2, 0xA4, 0x9C, 0x4F, 0x6F, 0x7C, 0xC9, 0x14, 0xD5, 0xAF }; +#define C0_ACTIVE_BUCKET_TIMER_INCREMENT 5 +#define C0_ACTIVE_THRESHOLD_INCREMENT 5 +#define C0_MINIMUM_WINDOW_INCREMENT 100 /* C3 Latency Monitor Log Page */ -#define C3_LATENCY_MON_LOG_BUF_LEN 0x200 -#define C3_LATENCY_MON_OPCODE 0xC3 -#define C3_LATENCY_MON_VERSION 0x0001 -#define C3_GUID_LENGTH 16 -static __u8 lat_mon_guid[C3_GUID_LENGTH] = { 0x92, 0x7a, 0xc0, 0x8c, 0xd0, 0x84, - 0x6c, 0x9c, 0x70, 0x43, 0xe6, 0xd4, 0x58, 0x5e, 0xd4, 0x85 }; - -#define READ 0 -#define WRITE 1 -#define TRIM 2 -#define RESERVED 3 - -typedef enum { - SCAO_PMUW = 0, /* Physical media units written */ - SCAO_PMUR = 16, /* Physical media units read */ - SCAO_BUNBR = 32, /* Bad user nand blocks raw */ - SCAO_BUNBN = 38, /* Bad user nand blocks normalized */ - SCAO_BSNBR = 40, /* Bad system nand blocks raw */ - SCAO_BSNBN = 46, /* Bad system nand blocks normalized */ - SCAO_XRC = 48, /* XOR recovery count */ - SCAO_UREC = 56, /* Uncorrectable read error count */ - SCAO_SEEC = 64, /* Soft ecc error count */ - SCAO_EECE = 72, /* End to end corrected errors */ - SCAO_EEDC = 76, /* End to end detected errors */ - SCAO_SDPU = 80, /* System data percent used */ - SCAO_RFSC = 81, /* Refresh counts */ - SCAO_MXUDEC = 88, /* Max User data erase counts */ - SCAO_MNUDEC = 92, /* Min User data erase counts */ - SCAO_NTTE = 96, /* Number of Thermal throttling events */ - SCAO_CTS = 97, /* Current throttling status */ - SCAO_EVF = 98, /* Errata Version Field */ - SCAO_PVF = 99, /* Point Version Field */ - SCAO_MIVF = 101, /* Minor Version Field */ - SCAO_MAVF = 103, /* Major Version Field */ - SCAO_PCEC = 104, /* PCIe correctable error count */ - SCAO_ICS = 112, /* Incomplete shutdowns */ - SCAO_PFB = 120, /* Percent free blocks */ - SCAO_CPH = 128, /* Capacitor health */ - SCAO_NEV = 130, /* NVMe Errata Version */ - SCAO_UIO = 136, /* Unaligned I/O */ - SCAO_SVN = 144, /* Security Version Number */ - SCAO_NUSE = 152, /* NUSE - Namespace utilization */ - SCAO_PSC = 160, /* PLP start count */ - SCAO_EEST = 176, /* Endurance estimate */ - SCAO_PLRC = 192, /* PCIe Link Retraining Count */ - SCAO_LPV = 494, /* Log page version */ - SCAO_LPG = 496, /* Log page GUID */ -} SMART_CLOUD_ATTRIBUTE_OFFSETS; +#define C3_LATENCY_MON_LOG_BUF_LEN 0x200 +#define C3_LATENCY_MON_OPCODE 0xC3 +#define C3_LATENCY_MON_VERSION 0x0001 +#define C3_GUID_LENGTH 16 +static __u8 lat_mon_guid[C3_GUID_LENGTH] = { + 0x92, 0x7a, 0xc0, 0x8c, + 0xd0, 0x84, 0x6c, 0x9c, + 0x70, 0x43, 0xe6, 0xd4, + 0x58, 0x5e, 0xd4, 0x85 +}; + +#define READ 0 +#define WRITE 1 +#define TRIM 2 +#define RESERVED 3 struct __attribute__((__packed__)) ssd_latency_monitor_log { - __u8 feature_status; /* 0x00 */ - __u8 rsvd1; /* 0x01 */ - __le16 active_bucket_timer; /* 0x02 */ - __le16 active_bucket_timer_threshold; /* 0x04 */ - __u8 active_threshold_a; /* 0x06 */ - __u8 active_threshold_b; /* 0x07 */ - __u8 active_threshold_c; /* 0x08 */ - __u8 active_threshold_d; /* 0x09 */ - __le16 active_latency_config; /* 0x0A */ - __u8 active_latency_min_window; /* 0x0C */ - __u8 rsvd2[0x13]; /* 0x0D */ - - __le32 active_bucket_counter[4][4] ; /* 0x20 - 0x5F */ - __le64 active_latency_timestamp[4][3]; /* 0x60 - 0xBF */ - __le16 active_measured_latency[4][3]; /* 0xC0 - 0xD7 */ - __le16 active_latency_stamp_units; /* 0xD8 */ - __u8 rsvd3[0x16]; /* 0xDA */ - - __le32 static_bucket_counter[4][4] ; /* 0xF0 - 0x12F */ - __le64 static_latency_timestamp[4][3]; /* 0x130 - 0x18F */ - __le16 static_measured_latency[4][3]; /* 0x190 - 0x1A7 */ - __le16 static_latency_stamp_units; /* 0x1A8 */ - __u8 rsvd4[0x16]; /* 0x1AA */ - - __le16 debug_log_trigger_enable; /* 0x1C0 */ - __le16 debug_log_measured_latency; /* 0x1C2 */ - __le64 debug_log_latency_stamp; /* 0x1C4 */ - __le16 debug_log_ptr; /* 0x1CC */ - __le16 debug_log_counter_trigger; /* 0x1CE */ - __u8 debug_log_stamp_units; /* 0x1D0 */ - __u8 rsvd5[0x1D]; /* 0x1D1 */ - - __le16 log_page_version; /* 0x1EE */ - __u8 log_page_guid[0x10]; /* 0x1F0 */ + __u8 feature_status; /* 0x00 */ + __u8 rsvd1; /* 0x01 */ + __le16 active_bucket_timer; /* 0x02 */ + __le16 active_bucket_timer_threshold; /* 0x04 */ + __u8 active_threshold_a; /* 0x06 */ + __u8 active_threshold_b; /* 0x07 */ + __u8 active_threshold_c; /* 0x08 */ + __u8 active_threshold_d; /* 0x09 */ + __le16 active_latency_config; /* 0x0A */ + __u8 active_latency_min_window; /* 0x0C */ + __u8 rsvd2[0x13]; /* 0x0D */ + + __le32 active_bucket_counter[4][4]; /* 0x20 - 0x5F */ + __le64 active_latency_timestamp[4][3]; /* 0x60 - 0xBF */ + __le16 active_measured_latency[4][3]; /* 0xC0 - 0xD7 */ + __le16 active_latency_stamp_units; /* 0xD8 */ + __u8 rsvd3[0x16]; /* 0xDA */ + + __le32 static_bucket_counter[4][4]; /* 0x0F0 - 0x12F */ + __le64 static_latency_timestamp[4][3]; /* 0x130 - 0x18F */ + __le16 static_measured_latency[4][3]; /* 0x190 - 0x1A7 */ + __le16 static_latency_stamp_units; /* 0x1A8 */ + __u8 rsvd4[0x16]; /* 0x1AA */ + + __le16 debug_log_trigger_enable; /* 0x1C0 */ + __le16 debug_log_measured_latency; /* 0x1C2 */ + __le64 debug_log_latency_stamp; /* 0x1C4 */ + __le16 debug_log_ptr; /* 0x1CC */ + __le16 debug_log_counter_trigger; /* 0x1CE */ + __u8 debug_log_stamp_units; /* 0x1D0 */ + __u8 rsvd5[0x1D]; /* 0x1D1 */ + + __le16 log_page_version; /* 0x1EE */ + __u8 log_page_guid[0x10]; /* 0x1F0 */ }; +static const __u8 OCP_FID_CLEAR_PCIE_CORRECTABLE_ERROR_COUNTERS = 0xC3; + static int convert_ts(time_t time, char *ts_buf) { - struct tm gmTimeInfo; - time_t time_Human, time_ms; - char buf[80]; + struct tm gmTimeInfo; + time_t time_Human, time_ms; + char buf[80]; - time_Human = time/1000; - time_ms = time % 1000; + time_Human = time/1000; + time_ms = time % 1000; - gmtime_r((const time_t *)&time_Human, &gmTimeInfo); + gmtime_r((const time_t *)&time_Human, &gmTimeInfo); - strftime(buf, sizeof(buf), "%Y-%m-%d %H:%M:%S", &gmTimeInfo); - sprintf(ts_buf, "%s.%03ld GMT", buf, time_ms); + strftime(buf, sizeof(buf), "%Y-%m-%d %H:%M:%S", &gmTimeInfo); + sprintf(ts_buf, "%s.%03ld GMT", buf, time_ms); - return 0; + return 0; } -static void ocp_print_C0_log_normal(void *data) +static int ocp_print_C3_log_normal(struct nvme_dev *dev, + struct ssd_latency_monitor_log *log_data) { - __u8 *log_data = (__u8*)data; - uint16_t smart_log_ver = 0; - - printf("SMART Cloud Attributes :- \n"); - - printf(" Physical media units written - %"PRIu64" %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW+8] & 0xFFFFFFFFFFFFFFFF), - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW] & 0xFFFFFFFFFFFFFFFF)); - printf(" Physical media units read - %"PRIu64" %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR+8] & 0xFFFFFFFFFFFFFFFF), - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR] & 0xFFFFFFFFFFFFFFFF)); - printf(" Bad user nand blocks - Raw %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BUNBR] & 0x0000FFFFFFFFFFFF)); - printf(" Bad user nand blocks - Normalized %d\n", - (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BUNBN])); - printf(" Bad system nand blocks - Raw %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BSNBR] & 0x0000FFFFFFFFFFFF)); - printf(" Bad system nand blocks - Normalized %d\n", - (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BSNBN])); - printf(" XOR recovery count %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_XRC])); - printf(" Uncorrectable read error count %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UREC])); - printf(" Soft ecc error count %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SEEC])); - printf(" End to end corrected errors %"PRIu32"\n", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EECE])); - printf(" End to end detected errors %"PRIu32"\n", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EEDC])); - printf(" System data percent used %d\n", - (__u8)log_data[SCAO_SDPU]); - printf(" Refresh counts %"PRIu64"\n", - (uint64_t)(le64_to_cpu(*(uint64_t *)&log_data[SCAO_RFSC])& 0x00FFFFFFFFFFFFFF)); - printf(" Max User data erase counts %"PRIu32"\n", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MXUDEC])); - printf(" Min User data erase counts %"PRIu32"\n", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MNUDEC])); - printf(" Number of Thermal throttling events %d\n", - (__u8)log_data[SCAO_NTTE]); - printf(" Current throttling status 0x%x\n", - (__u8)log_data[SCAO_CTS]); - printf(" PCIe correctable error count %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PCEC])); - printf(" Incomplete shutdowns %"PRIu32"\n", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_ICS])); - printf(" Percent free blocks %d\n", - (__u8)log_data[SCAO_PFB]); - printf(" Capacitor health %"PRIu16"\n", - (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_CPH])); - printf(" Unaligned I/O %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UIO])); - printf(" Security Version Number %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SVN])); - printf(" NUSE - Namespace utilization %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_NUSE])); - printf(" PLP start count %s\n", - uint128_t_to_string(le128_to_cpu(&log_data[SCAO_PSC]))); - printf(" Endurance estimate %s\n", - uint128_t_to_string(le128_to_cpu(&log_data[SCAO_EEST]))); - smart_log_ver = (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_LPV]); - printf(" Log page version %"PRIu16"\n",smart_log_ver); - printf(" Log page GUID 0x"); - printf("%"PRIx64"%"PRIx64"\n",(uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG + 8]), - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG])); - if(smart_log_ver > 2) { - printf(" Errata Version Field %d\n", - (__u8)log_data[SCAO_EVF]); - printf(" Point Version Field %"PRIu16"\n", - (uint16_t)log_data[SCAO_PVF]); - printf(" Minor Version Field %"PRIu16"\n", - (uint16_t)log_data[SCAO_MIVF]); - printf(" Major Version Field %d\n", - (__u8)log_data[SCAO_MAVF]); - printf(" NVMe Errata Version %d\n", - (__u8)log_data[SCAO_NEV]); - printf(" PCIe Link Retraining Count %"PRIu64"\n", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PLRC])); - } - printf("\n"); + char ts_buf[128]; + int i, j; + int pos = 0; + + printf("-Latency Monitor/C3 Log Page Data-\n"); + printf(" Controller : %s\n", dev->name); + printf(" Feature Status 0x%x\n", + log_data->feature_status); + printf(" Active Bucket Timer %d min\n", + C0_ACTIVE_BUCKET_TIMER_INCREMENT * + le16_to_cpu(log_data->active_bucket_timer)); + printf(" Active Bucket Timer Threshold %d min\n", + C0_ACTIVE_BUCKET_TIMER_INCREMENT * + le16_to_cpu(log_data->active_bucket_timer_threshold)); + printf(" Active Threshold A %d ms\n", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_a+1)); + printf(" Active Threshold B %d ms\n", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_b+1)); + printf(" Active Threshold C %d ms\n", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_c+1)); + printf(" Active Threshold D %d ms\n", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_d+1)); + printf(" Active Latency Minimum Window %d ms\n", + C0_MINIMUM_WINDOW_INCREMENT * + le16_to_cpu(log_data->active_latency_min_window)); + printf(" Active Latency Stamp Units %d\n", + le16_to_cpu(log_data->active_latency_stamp_units)); + printf(" Static Latency Stamp Units %d\n", + le16_to_cpu(log_data->static_latency_stamp_units)); + printf(" Debug Log Trigger Enable %d\n", + le16_to_cpu(log_data->debug_log_trigger_enable)); + + printf(" Read Write Deallocate/Trim \n"); + for (i = 0; i <= 3; i++) { + printf(" Active Latency Mode: Bucket %d %27d %27d %27d\n", + i, + log_data->active_latency_config & (1 << pos), + log_data->active_latency_config & (1 << pos), + log_data->active_latency_config & (1 << pos)); + } + printf("\n"); + + for (i = 0; i <= 3; i++) { + printf(" Active Bucket Counter: Bucket %d %27d %27d %27d \n", + i, + le32_to_cpu(log_data->active_bucket_counter[i][READ]), + le32_to_cpu(log_data->active_bucket_counter[i][WRITE]), + le32_to_cpu(log_data->active_bucket_counter[i][TRIM])); + } + + for (i = 0; i <= 3; i++) { + printf(" Active Measured Latency: Bucket %d %27d ms %27d ms %27d ms\n", + i, + le16_to_cpu(log_data->active_measured_latency[i][READ]), + le16_to_cpu(log_data->active_measured_latency[i][WRITE]), + le16_to_cpu(log_data->active_measured_latency[i][TRIM])); + } + + for (i = 0; i <= 3; i++) { + printf(" Active Latency Time Stamp: Bucket %d ", i); + for (j = 0; j <= 2; j++) { + if (le64_to_cpu(log_data->active_latency_timestamp[i][j]) == -1) + printf(" N/A "); + else { + convert_ts(le64_to_cpu(log_data->active_latency_timestamp[i][j]), ts_buf); + printf("%s ", ts_buf); + } + } + printf("\n"); + } + + for (i = 0; i <= 3; i++) { + printf(" Static Bucket Counter: Bucket %d %27d %27d %27d\n", + i, + le32_to_cpu(log_data->static_bucket_counter[i][READ]), + le32_to_cpu(log_data->static_bucket_counter[i][WRITE]), + le32_to_cpu(log_data->static_bucket_counter[i][TRIM])); + } + + for (i = 0; i <= 3; i++) { + printf(" Static Measured Latency: Bucket %d %27d ms %27d ms %27d ms \n", + i, + le16_to_cpu(log_data->static_measured_latency[i][READ]), + le16_to_cpu(log_data->static_measured_latency[i][WRITE]), + le16_to_cpu(log_data->static_measured_latency[i][TRIM])); + } + + for (i = 0; i <= 3; i++) { + printf(" Static Latency Time Stamp: Bucket %d ", i); + for (j = 0; j <= 2; j++) { + if (le64_to_cpu(log_data->static_latency_timestamp[i][j]) == -1) + printf(" N/A "); + else { + convert_ts(le64_to_cpu(log_data->static_latency_timestamp[i][j]), ts_buf); + printf("%s ", ts_buf); + } + } + printf("\n"); + } + + return 0; } -static void ocp_print_C0_log_json(void *data) +static void ocp_print_C3_log_json(struct ssd_latency_monitor_log *log_data) { - __u8 *log_data = (__u8*)data; - struct json_object *root; - struct json_object *pmuw; - struct json_object *pmur; - uint16_t smart_log_ver = 0; - - root = json_create_object(); - pmuw = json_create_object(); - pmur = json_create_object(); - - json_object_add_value_uint64(pmuw, "hi", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW+8] & 0xFFFFFFFFFFFFFFFF)); - json_object_add_value_uint64(pmuw, "lo", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW] & 0xFFFFFFFFFFFFFFFF)); - json_object_add_value_object(root, "Physical media units written", pmuw); - json_object_add_value_uint64(pmur, "hi", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR+8] & 0xFFFFFFFFFFFFFFFF)); - json_object_add_value_uint64(pmur, "lo", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR] & 0xFFFFFFFFFFFFFFFF)); - json_object_add_value_object(root, "Physical media units read", pmur); - json_object_add_value_uint64(root, "Bad user nand blocks - Raw", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BUNBR] & 0x0000FFFFFFFFFFFF)); - json_object_add_value_uint(root, "Bad user nand blocks - Normalized", - (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BUNBN])); - json_object_add_value_uint64(root, "Bad system nand blocks - Raw", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BSNBR] & 0x0000FFFFFFFFFFFF)); - json_object_add_value_uint(root, "Bad system nand blocks - Normalized", - (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BSNBN])); - json_object_add_value_uint64(root, "XOR recovery count", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_XRC])); - json_object_add_value_uint64(root, "Uncorrectable read error count", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UREC])); - json_object_add_value_uint64(root, "Soft ecc error count", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SEEC])); - json_object_add_value_uint(root, "End to end corrected errors", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EECE])); - json_object_add_value_uint(root, "End to end detected errors", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EEDC])); - json_object_add_value_uint(root, "System data percent used", - (__u8)log_data[SCAO_SDPU]); - json_object_add_value_uint64(root, "Refresh counts", - (uint64_t)(le64_to_cpu(*(uint64_t *)&log_data[SCAO_RFSC])& 0x00FFFFFFFFFFFFFF)); - json_object_add_value_uint(root, "Max User data erase counts", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MXUDEC])); - json_object_add_value_uint(root, "Min User data erase counts", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MNUDEC])); - json_object_add_value_uint(root, "Number of Thermal throttling events", - (__u8)log_data[SCAO_NTTE]); - json_object_add_value_uint(root, "Current throttling status", - (__u8)log_data[SCAO_CTS]); - json_object_add_value_uint64(root, "PCIe correctable error count", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PCEC])); - json_object_add_value_uint(root, "Incomplete shutdowns", - (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_ICS])); - json_object_add_value_uint(root, "Percent free blocks", - (__u8)log_data[SCAO_PFB]); - json_object_add_value_uint(root, "Capacitor health", - (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_CPH])); - json_object_add_value_uint64(root, "Unaligned I/O", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UIO])); - json_object_add_value_uint64(root, "Security Version Number", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SVN])); - json_object_add_value_uint64(root, "NUSE - Namespace utilization", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_NUSE])); - json_object_add_value_uint128(root, "PLP start count", - le128_to_cpu(&log_data[SCAO_PSC])); - json_object_add_value_uint128(root, "Endurance estimate", - le128_to_cpu(&log_data[SCAO_EEST])); - smart_log_ver = (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_LPV]); - json_object_add_value_uint(root, "Log page version", smart_log_ver); - char guid[40]; - memset((void*)guid, 0, 40); - sprintf((char*)guid, "0x%"PRIx64"%"PRIx64"",(uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG + 8]), - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG])); - json_object_add_value_string(root, "Log page GUID", guid); - if(smart_log_ver > 2){ - json_object_add_value_uint(root, "Errata Version Field", - (__u8)log_data[SCAO_EVF]); - json_object_add_value_uint(root, "Point Version Field", - (uint16_t)log_data[SCAO_PVF]); - json_object_add_value_uint(root, "Minor Version Field", - (uint16_t)log_data[SCAO_MIVF]); - json_object_add_value_uint(root, "Major Version Field", - (__u8)log_data[SCAO_MAVF]); - json_object_add_value_uint(root, "NVMe Errata Version", - (__u8)log_data[SCAO_NEV]); - json_object_add_value_uint(root, "PCIe Link Retraining Count", - (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PLRC])); - } - json_print_object(root, NULL); - printf("\n"); - json_free_object(root); + struct json_object *root; + char ts_buf[128]; + char buf[128]; + int i, j; + int pos = 0; + char *operation[3] = {"Read", "Write", "Trim"}; + + root = json_create_object(); + + json_object_add_value_uint(root, "Feature Status", + log_data->feature_status); + json_object_add_value_uint(root, "Active Bucket Timer", + C0_ACTIVE_BUCKET_TIMER_INCREMENT * + le16_to_cpu(log_data->active_bucket_timer)); + json_object_add_value_uint(root, "Active Bucket Timer Threshold", + C0_ACTIVE_BUCKET_TIMER_INCREMENT * + le16_to_cpu(log_data->active_bucket_timer_threshold)); + json_object_add_value_uint(root, "Active Threshold A", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_a + 1)); + json_object_add_value_uint(root, "Active Threshold B", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_b + 1)); + json_object_add_value_uint(root, "Active Threshold C", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_c + 1)); + json_object_add_value_uint(root, "Active Threshold D", + C0_ACTIVE_THRESHOLD_INCREMENT * + le16_to_cpu(log_data->active_threshold_d + 1)); + json_object_add_value_uint(root, "Active Lantency Minimum Window", + C0_MINIMUM_WINDOW_INCREMENT * + le16_to_cpu(log_data->active_latency_min_window)); + json_object_add_value_uint(root, "Active Latency Stamp Units", + le16_to_cpu(log_data->active_latency_stamp_units)); + json_object_add_value_uint(root, "Static Latency Stamp Units", + le16_to_cpu(log_data->static_latency_stamp_units)); + json_object_add_value_uint(root, "Debug Log Trigger Enable", + le16_to_cpu(log_data->debug_log_trigger_enable)); + + for (i = 0; i <= 3; i++) { + struct json_object *bucket; + + bucket = json_create_object(); + sprintf(buf, "Active Latency Mode: Bucket %d", i); + for (j = 0; j <= 2; j++) { + json_object_add_value_uint(bucket, operation[j], + log_data->active_latency_config & (1 << pos)); + } + json_object_add_value_object(root, buf, bucket); + } + + for (i = 0; i <= 3; i++) { + struct json_object *bucket; + + bucket = json_create_object(); + sprintf(buf, "Active Bucket Counter: Bucket %d", i); + for (j = 0; j <= 2; j++) { + json_object_add_value_uint(bucket, operation[j], + le32_to_cpu(log_data->active_bucket_counter[i][j])); + } + json_object_add_value_object(root, buf, bucket); + } + + for (i = 0; i <= 3; i++) { + struct json_object *bucket; + + bucket = json_create_object(); + sprintf(buf, "Active Measured Latency: Bucket %d", i); + for (j = 0; j <= 2; j++) { + json_object_add_value_uint(bucket, operation[j], + le16_to_cpu(log_data->active_measured_latency[i][j])); + } + json_object_add_value_object(root, buf, bucket); + } + + for (i = 0; i <= 3; i++) { + struct json_object *bucket; + + bucket = json_create_object(); + sprintf(buf, "Active Latency Time Stamp: Bucket %d", i); + for (j = 0; j <= 2; j++) { + if (le64_to_cpu(log_data->active_latency_timestamp[i][j]) == -1) + json_object_add_value_string(bucket, operation[j], "NA"); + else { + convert_ts(le64_to_cpu(log_data->active_latency_timestamp[i][j]), ts_buf); + json_object_add_value_string(bucket, operation[j], ts_buf); + } + } + json_object_add_value_object(root, buf, bucket); + } + + for (i = 0; i <= 3; i++) { + struct json_object *bucket; + + bucket = json_create_object(); + sprintf(buf, "Static Bucket Counter: Bucket %d", i); + for (j = 0; j <= 2; j++) { + json_object_add_value_uint(bucket, operation[j], + le32_to_cpu(log_data->static_bucket_counter[i][j])); + } + json_object_add_value_object(root, buf, bucket); + } + + for (i = 0; i <= 3; i++) { + struct json_object *bucket; + + bucket = json_create_object(); + sprintf(buf, "Static Measured Latency: Bucket %d", i); + for (j = 0; j <= 2; j++) { + json_object_add_value_uint(bucket, operation[j], + le16_to_cpu(log_data->static_measured_latency[i][j])); + } + json_object_add_value_object(root, buf, bucket); + } + + for (i = 0; i <= 3; i++) { + struct json_object *bucket; + + bucket = json_create_object(); + sprintf(buf, "Static Latency Time Stamp: Bucket %d", i); + for (j = 0; j <= 2; j++) { + if (le64_to_cpu(log_data->static_latency_timestamp[i][j]) == -1) + json_object_add_value_string(bucket, operation[j], "NA"); + else { + convert_ts(le64_to_cpu(log_data->static_latency_timestamp[i][j]), ts_buf); + json_object_add_value_string(bucket, operation[j], ts_buf); + } + } + json_object_add_value_object(root, buf, bucket); + } + + json_print_object(root, NULL); + printf("\n"); + + json_free_object(root); } -static int get_c0_log_page(int fd, char *format) +static int get_c3_log_page(struct nvme_dev *dev, char *format) { - int ret = 0; - int fmt = -1; - __u8 *data; - int i; - - fmt = validate_output_format(format); - if (fmt < 0) { - fprintf(stderr, "ERROR : OCP : invalid output format\n"); - return fmt; - } - - if ((data = (__u8 *) malloc(sizeof(__u8) * C0_SMART_CLOUD_ATTR_LEN)) == NULL) { - fprintf(stderr, "ERROR : OCP : malloc : %s\n", strerror(errno)); - return -1; - } - memset(data, 0, sizeof (__u8) * C0_SMART_CLOUD_ATTR_LEN); - - ret = nvme_get_log_simple(fd, C0_SMART_CLOUD_ATTR_OPCODE, - C0_SMART_CLOUD_ATTR_LEN, data); - - if (strcmp(format, "json")) - fprintf(stderr, "NVMe Status:%s(%x)\n", - nvme_status_to_string(ret, false), ret); - - if (ret == 0) { - - /* check log page guid */ - /* Verify GUID matches */ - for (i=0; i<16; i++) { - if (scao_guid[i] != data[SCAO_LPG + i]) { - fprintf(stderr, "ERROR : OCP : Unknown GUID in C0 Log Page data\n"); - int j; - fprintf(stderr, "ERROR : OCP : Expected GUID: 0x"); - for (j = 0; j<16; j++) { - fprintf(stderr, "%x", scao_guid[j]); - } - fprintf(stderr, "\nERROR : OCP : Actual GUID: 0x"); - for (j = 0; j<16; j++) { - fprintf(stderr, "%x", data[SCAO_LPG + j]); - } - fprintf(stderr, "\n"); - - ret = -1; - goto out; - } - } - - /* print the data */ - switch (fmt) { - case NORMAL: - ocp_print_C0_log_normal(data); - break; - case JSON: - ocp_print_C0_log_json(data); - break; - } - } else { - fprintf(stderr, "ERROR : OCP : Unable to read C0 data from buffer\n"); - } + struct ssd_latency_monitor_log *log_data; + int ret = 0; + int fmt = -1; + __u8 *data; + int i; + + fmt = validate_output_format(format); + if (fmt < 0) { + fprintf(stderr, "ERROR : OCP : invalid output format\n"); + return fmt; + } + + data = malloc(sizeof(__u8) * C3_LATENCY_MON_LOG_BUF_LEN); + if (!data) { + fprintf(stderr, "ERROR : OCP : malloc : %s\n", strerror(errno)); + return -1; + } + memset(data, 0, sizeof(__u8) * C3_LATENCY_MON_LOG_BUF_LEN); + + ret = nvme_get_log_simple(dev_fd(dev), C3_LATENCY_MON_OPCODE, + C3_LATENCY_MON_LOG_BUF_LEN, data); + + if (strcmp(format, "json")) + fprintf(stderr, + "NVMe Status:%s(%x)\n", + nvme_status_to_string(ret, false), + ret); + + if (ret == 0) { + log_data = (struct ssd_latency_monitor_log *)data; + + /* check log page version */ + if (log_data->log_page_version != C3_LATENCY_MON_VERSION) { + fprintf(stderr, + "ERROR : OCP : invalid latency monitor version\n"); + ret = -1; + goto out; + } + + /* check log page guid */ + /* Verify GUID matches */ + for (i = 0; i < 16; i++) { + if (lat_mon_guid[i] != log_data->log_page_guid[i]) { + int j; + + fprintf(stderr, "ERROR : OCP : Unknown GUID in C3 Log Page data\n"); + fprintf(stderr, "ERROR : OCP : Expected GUID: 0x"); + for (j = 0; j < 16; j++) { + fprintf(stderr, "%x", lat_mon_guid[j]); + } + + fprintf(stderr, "\nERROR : OCP : Actual GUID: 0x"); + for (j = 0; j < 16; j++) { + fprintf(stderr, "%x", log_data->log_page_guid[j]); + } + fprintf(stderr, "\n"); + + ret = -1; + goto out; + } + } + + switch (fmt) { + case NORMAL: + ocp_print_C3_log_normal(dev, log_data); + break; + case JSON: + ocp_print_C3_log_json(log_data); + break; + } + } else { + fprintf(stderr, + "ERROR : OCP : Unable to read C3 data from buffer\n"); + } out: - free(data); - return ret; + free(data); + return ret; +} + +static int smart_add_log(int argc, char **argv, struct command *cmd, + struct plugin *plugin) +{ + return ocp_smart_add_log(argc, argv, cmd, plugin); } -static int ocp_smart_add_log(int argc, char **argv, struct command *cmd, - struct plugin *plugin) +static int ocp_latency_monitor_log(int argc, char **argv, + struct command *command, + struct plugin *plugin) { - const char *desc = "Retrieve latency monitor log data."; + const char *desc = "Retrieve latency monitor log data."; struct nvme_dev *dev; - int ret = 0; - - struct config { - char *output_format; - }; - - struct config cfg = { - .output_format = "normal", - }; - - OPT_ARGS(opts) = { - OPT_FMT("output-format", 'o', &cfg.output_format, "output Format: normal|json"), - OPT_END() - }; - - ret = parse_and_open(&dev, argc, argv, desc, opts); - if (ret) - return ret; - - ret = get_c0_log_page(dev_fd(dev), cfg.output_format); - if (ret) - fprintf(stderr, "ERROR : OCP : Failure reading the C0 Log Page, ret = %d\n", - ret); - dev_close(dev); - return ret; + int ret = 0; + + struct config { + char *output_format; + }; + + struct config cfg = { + .output_format = "normal", + }; + + OPT_ARGS(opts) = { + OPT_FMT("output-format", 'o', &cfg.output_format, + "output Format: normal|json"), + OPT_END() + }; + + ret = parse_and_open(&dev, argc, argv, desc, opts); + if (ret) + return ret; + + ret = get_c3_log_page(dev, cfg.output_format); + if (ret) + fprintf(stderr, + "ERROR : OCP : Failure reading the C3 Log Page, ret = %d\n", + ret); + + dev_close(dev); + return ret; } -static int ocp_print_C3_log_normal(struct nvme_dev *dev, - struct ssd_latency_monitor_log *log_data) +static int clear_fw_update_history(int argc, char **argv, + struct command *cmd, struct plugin *plugin) { - printf("-Latency Monitor/C3 Log Page Data- \n"); - printf(" Controller : %s\n", dev->name); - int i, j; - int pos = 0; - char ts_buf[128]; - - printf(" Feature Status 0x%x \n", - log_data->feature_status); - printf(" Active Bucket Timer %d min \n", - C0_ACTIVE_BUCKET_TIMER_INCREMENT * - le16_to_cpu(log_data->active_bucket_timer)); - printf(" Active Bucket Timer Threshold %d min \n", - C0_ACTIVE_BUCKET_TIMER_INCREMENT * - le16_to_cpu(log_data->active_bucket_timer_threshold)); - printf(" Active Threshold A %d ms \n", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_a+1)); - printf(" Active Threshold B %d ms \n", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_b+1)); - printf(" Active Threshold C %d ms \n", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_c+1)); - printf(" Active Threshold D %d ms \n", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_d+1)); - printf(" Active Latency Minimum Window %d ms \n", - C0_MINIMUM_WINDOW_INCREMENT * - le16_to_cpu(log_data->active_latency_min_window)); - printf(" Active Latency Stamp Units %d \n", - le16_to_cpu(log_data->active_latency_stamp_units)); - printf(" Static Latency Stamp Units %d \n", - le16_to_cpu(log_data->static_latency_stamp_units)); - printf(" Debug Log Trigger Enable %d \n", - le16_to_cpu(log_data->debug_log_trigger_enable)); - - printf(" Read Write Deallocate/Trim \n"); - for (i = 0; i <= 3; i++) { - printf(" Active Latency Mode: Bucket %d %27d %27d %27d \n", - i, - log_data->active_latency_config & (1 << pos), - log_data->active_latency_config & (1 << pos), - log_data->active_latency_config & (1 << pos)); - } - printf("\n"); - for (i = 0; i <= 3; i++) { - printf(" Active Bucket Counter: Bucket %d %27d %27d %27d \n", - i, - le32_to_cpu(log_data->active_bucket_counter[i][READ]), - le32_to_cpu(log_data->active_bucket_counter[i][WRITE]), - le32_to_cpu(log_data->active_bucket_counter[i][TRIM])); - } - - for (i = 0; i <= 3; i++) { - printf(" Active Measured Latency: Bucket %d %27d ms %27d ms %27d ms \n", - i, - le16_to_cpu(log_data->active_measured_latency[i][READ]), - le16_to_cpu(log_data->active_measured_latency[i][WRITE]), - le16_to_cpu(log_data->active_measured_latency[i][TRIM])); - } - - for (i = 0; i <= 3; i++) { - printf(" Active Latency Time Stamp: Bucket %d ", i); - for (j = 0; j <= 2; j++) { - if (le64_to_cpu(log_data->active_latency_timestamp[i][j]) == -1) - printf(" N/A "); - else { - convert_ts(le64_to_cpu(log_data->active_latency_timestamp[i][j]), ts_buf); - printf("%s ", ts_buf); - } - } - printf("\n"); - } - - for (i = 0; i <= 3; i++) { - printf(" Static Bucket Counter: Bucket %d %27d %27d %27d \n", - i, - le32_to_cpu(log_data->static_bucket_counter[i][READ]), - le32_to_cpu(log_data->static_bucket_counter[i][WRITE]), - le32_to_cpu(log_data->static_bucket_counter[i][TRIM])); - } - - for (i = 0; i <= 3; i++) { - printf(" Static Measured Latency: Bucket %d %27d ms %27d ms %27d ms \n", - i, - le16_to_cpu(log_data->static_measured_latency[i][READ]), - le16_to_cpu(log_data->static_measured_latency[i][WRITE]), - le16_to_cpu(log_data->static_measured_latency[i][TRIM])); - } - - for (i = 0; i <= 3; i++) { - printf(" Static Latency Time Stamp: Bucket %d ", i); - for (j = 0; j <= 2; j++) { - if (le64_to_cpu(log_data->static_latency_timestamp[i][j]) == -1) - printf(" N/A "); - else { - convert_ts(le64_to_cpu(log_data->static_latency_timestamp[i][j]), ts_buf); - printf("%s ", ts_buf); - } - } - printf("\n"); - } - - return 0; + return ocp_clear_fw_update_history(argc, argv, cmd, plugin); } -static void ocp_print_C3_log_json(struct ssd_latency_monitor_log *log_data) +static const char *eol_plp_failure_mode_to_string(__u8 mode) { - int i, j; - int pos = 0; - char buf[128]; - char ts_buf[128]; - char *operation[3] = {"Read", "Write", "Trim"}; - struct json_object *root; - root = json_create_object(); - - json_object_add_value_uint(root, "Feature Status", - log_data->feature_status); - json_object_add_value_uint(root, "Active Bucket Timer", - C0_ACTIVE_BUCKET_TIMER_INCREMENT * - le16_to_cpu(log_data->active_bucket_timer)); - json_object_add_value_uint(root, "Active Bucket Timer Threshold", - C0_ACTIVE_BUCKET_TIMER_INCREMENT * - le16_to_cpu(log_data->active_bucket_timer_threshold)); - json_object_add_value_uint(root, "Active Threshold A", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_a+1)); - json_object_add_value_uint(root, "Active Threshold B", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_b+1)); - json_object_add_value_uint(root, "Active Threshold C", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_c+1)); - json_object_add_value_uint(root, "Active Threshold D", - C0_ACTIVE_THRESHOLD_INCREMENT * - le16_to_cpu(log_data->active_threshold_d+1)); - json_object_add_value_uint(root, "Active Lantency Minimum Window", - C0_MINIMUM_WINDOW_INCREMENT * - le16_to_cpu(log_data->active_latency_min_window)); - json_object_add_value_uint(root, "Active Latency Stamp Units", - le16_to_cpu(log_data->active_latency_stamp_units)); - json_object_add_value_uint(root, "Static Latency Stamp Units", - le16_to_cpu(log_data->static_latency_stamp_units)); - json_object_add_value_uint(root, "Debug Log Trigger Enable", - le16_to_cpu(log_data->debug_log_trigger_enable)); - - for (i = 0; i <= 3; i++) { - struct json_object *bucket; - bucket = json_create_object(); - sprintf(buf, "Active Latency Mode: Bucket %d", i); - for (j = 0; j <= 2; j++) { - json_object_add_value_uint(bucket, operation[j], - log_data->active_latency_config & (1 << pos)); - } - json_object_add_value_object(root, buf, bucket); - } - for (i = 0; i <= 3; i++) { - struct json_object *bucket; - bucket = json_create_object(); - sprintf(buf, "Active Bucket Counter: Bucket %d", i); - for (j = 0; j <= 2; j++) { - json_object_add_value_uint(bucket, operation[j], - le32_to_cpu(log_data->active_bucket_counter[i][j])); - } - json_object_add_value_object(root, buf, bucket); - } - for (i = 0; i <= 3; i++) { - struct json_object *bucket; - bucket = json_create_object(); - sprintf(buf, "Active Measured Latency: Bucket %d", i); - for (j = 0; j <= 2; j++) { - json_object_add_value_uint(bucket, operation[j], - le16_to_cpu(log_data->active_measured_latency[i][j])); - } - json_object_add_value_object(root, buf, bucket); - } - for (i = 0; i <= 3; i++) { - struct json_object *bucket; - bucket = json_create_object(); - sprintf(buf, "Active Latency Time Stamp: Bucket %d", i); - for (j = 0; j <= 2; j++) { - if (le64_to_cpu(log_data->active_latency_timestamp[i][j]) == -1) - json_object_add_value_string(bucket, operation[j], "NA"); - else { - convert_ts(le64_to_cpu(log_data->active_latency_timestamp[i][j]), ts_buf); - json_object_add_value_string(bucket, operation[j], ts_buf); - } - } - json_object_add_value_object(root, buf, bucket); - } - for (i = 0; i <= 3; i++) { - struct json_object *bucket; - bucket = json_create_object(); - sprintf(buf, "Static Bucket Counter: Bucket %d", i); - for (j = 0; j <= 2; j++) { - json_object_add_value_uint(bucket, operation[j], - le32_to_cpu(log_data->static_bucket_counter[i][j])); - } - json_object_add_value_object(root, buf, bucket); - } - for (i = 0; i <= 3; i++) { - struct json_object *bucket; - bucket = json_create_object(); - sprintf(buf, "Static Measured Latency: Bucket %d", i); - for (j = 0; j <= 2; j++) { - json_object_add_value_uint(bucket, operation[j], - le16_to_cpu(log_data->static_measured_latency[i][j])); - } - json_object_add_value_object(root, buf, bucket); - } - for (i = 0; i <= 3; i++) { - struct json_object *bucket; - bucket = json_create_object(); - sprintf(buf, "Static Latency Time Stamp: Bucket %d", i); - for (j = 0; j <= 2; j++) { - if (le64_to_cpu(log_data->static_latency_timestamp[i][j]) == -1) - json_object_add_value_string(bucket, operation[j], "NA"); - else { - convert_ts(le64_to_cpu(log_data->static_latency_timestamp[i][j]), ts_buf); - json_object_add_value_string(bucket, operation[j], ts_buf); - } - } - json_object_add_value_object(root, buf, bucket); - } - - json_print_object(root, NULL); - printf("\n"); - - json_free_object(root); + switch (mode) { + case 1: + return "Read only mode (ROM)"; + case 2: + return "Write through mode (WTM)"; + case 3: + return "Normal mode"; + default: + break; + } + + return "Reserved"; } -static int get_c3_log_page(struct nvme_dev *dev, char *format) +static int eol_plp_failure_mode_get(struct nvme_dev *dev, const __u32 nsid, + const __u8 fid, __u8 sel) { - int ret = 0; - int fmt = -1; - __u8 *data; - int i; - struct ssd_latency_monitor_log *log_data; - - fmt = validate_output_format(format); - if (fmt < 0) { - fprintf(stderr, "ERROR : OCP : invalid output format\n"); - return fmt; - } - - if ((data = (__u8 *) malloc(sizeof(__u8) * C3_LATENCY_MON_LOG_BUF_LEN)) == NULL) { - fprintf(stderr, "ERROR : OCP : malloc : %s\n", strerror(errno)); - return -1; - } - memset(data, 0, sizeof (__u8) * C3_LATENCY_MON_LOG_BUF_LEN); - - ret = nvme_get_log_simple(dev_fd(dev), C3_LATENCY_MON_OPCODE, - C3_LATENCY_MON_LOG_BUF_LEN, data); - - if (strcmp(format, "json")) - fprintf(stderr, - "NVMe Status:%s(%x)\n", - nvme_status_to_string(ret, false), - ret); - - if (ret == 0) { - log_data = (struct ssd_latency_monitor_log*)data; - - /* check log page version */ - if (log_data->log_page_version != C3_LATENCY_MON_VERSION) { - fprintf(stderr, - "ERROR : OCP : invalid latency monitor version\n"); - ret = -1; - goto out; - } - - /* check log page guid */ - /* Verify GUID matches */ - for (i=0; i<16; i++) { - if (lat_mon_guid[i] != log_data->log_page_guid[i]) { - fprintf(stderr,"ERROR : OCP : Unknown GUID in C3 Log Page data\n"); - int j; - fprintf(stderr, "ERROR : OCP : Expected GUID: 0x"); - for (j = 0; j<16; j++) { - fprintf(stderr, "%x", lat_mon_guid[j]); - } - fprintf(stderr, "\nERROR : OCP : Actual GUID: 0x"); - for (j = 0; j<16; j++) { - fprintf(stderr, "%x", log_data->log_page_guid[j]); - } - fprintf(stderr, "\n"); - - ret = -1; - goto out; - } - } - - switch (fmt) { - case NORMAL: - ocp_print_C3_log_normal(dev, log_data); - break; - case JSON: - ocp_print_C3_log_json(log_data); - break; - } - } else { - fprintf(stderr, - "ERROR : OCP : Unable to read C3 data from buffer\n"); - } + __u32 result; + int err; + + struct nvme_get_features_args args = { + .args_size = sizeof(args), + .fd = dev_fd(dev), + .fid = fid, + .nsid = nsid, + .sel = sel, + .cdw11 = 0, + .uuidx = 0, + .data_len = 0, + .data = NULL, + .timeout = NVME_DEFAULT_IOCTL_TIMEOUT, + .result = &result, + }; + + err = nvme_get_features(&args); + if (!err) { + printf("End of Life Behavior (feature: %#0*x): %#0*x (%s: %s)\n", + fid ? 4 : 2, fid, result ? 10 : 8, result, + nvme_select_to_string(sel), + eol_plp_failure_mode_to_string(result)); + if (sel == NVME_GET_FEATURES_SEL_SUPPORTED) + nvme_show_select_result(result); + } else { + printf("Could not get feature: %#0*x.\n", fid ? 4 : 2, fid); + } + + return err; +} -out: - free(data); - return ret; +static int eol_plp_failure_mode_set(struct nvme_dev *dev, const __u32 nsid, + const __u8 fid, __u8 mode, bool save, + bool uuid) +{ + __u32 result; + int err; + int uuid_index = 0; + + if (uuid) { + /* OCP 2.0 requires UUID index support */ + err = ocp_get_uuid_index(dev, &uuid_index); + if (err || !uuid_index) { + fprintf(stderr, "ERROR: No OCP UUID index found\n"); + return err; + } + } + + + struct nvme_set_features_args args = { + .args_size = sizeof(args), + .fd = dev_fd(dev), + .fid = fid, + .nsid = nsid, + .cdw11 = mode << 30, + .cdw12 = 0, + .save = save, + .uuidx = uuid_index, + .cdw15 = 0, + .data_len = 0, + .data = NULL, + .timeout = NVME_DEFAULT_IOCTL_TIMEOUT, + .result = &result, + }; + + err = nvme_set_features(&args); + if (err > 0) { + nvme_show_status(err); + } else if (err < 0) { + perror("Define EOL/PLP failure mode"); + fprintf(stderr, "Command failed while parsing.\n"); + } else { + printf("Successfully set mode (feature: %#0*x): %#0*x (%s: %s).\n", + fid ? 4 : 2, fid, mode ? 10 : 8, mode, + save ? "Save" : "Not save", + eol_plp_failure_mode_to_string(mode)); + } + + return err; } -static int ocp_latency_monitor_log(int argc, char **argv, struct command *command, - struct plugin *plugin) +static int eol_plp_failure_mode(int argc, char **argv, struct command *cmd, + struct plugin *plugin) { - const char *desc = "Retrieve latency monitor log data."; + const char *desc = "Define EOL or PLP circuitry failure mode.\n"\ + "No argument prints current mode."; + const char *mode = "[0-3]: default/rom/wtm/normal"; + const char *save = "Specifies that the controller shall save the attribute"; + const char *sel = "[0-3,8]: current/default/saved/supported/changed"; + const __u32 nsid = 0; + const __u8 fid = 0xc2; struct nvme_dev *dev; - int ret = 0; - - struct config { - char *output_format; - }; - - struct config cfg = { - .output_format = "normal", - }; - - OPT_ARGS(opts) = { - OPT_FMT("output-format", 'o', &cfg.output_format, - "output Format: normal|json"), - OPT_END() - }; - - ret = parse_and_open(&dev, argc, argv, desc, opts); - if (ret) - return ret; - - ret = get_c3_log_page(dev, cfg.output_format); - if (ret) - fprintf(stderr, - "ERROR : OCP : Failure reading the C3 Log Page, ret = %d\n", - ret); - dev_close(dev); - return ret; + int err; + + struct config { + __u8 mode; + bool save; + __u8 sel; + }; + + struct config cfg = { + .mode = 0, + .save = false, + .sel = 0, + }; + + OPT_ARGS(opts) = { + OPT_BYTE("mode", 'm', &cfg.mode, mode), + OPT_FLAG("save", 's', &cfg.save, save), + OPT_BYTE("sel", 'S', &cfg.sel, sel), + OPT_FLAG("no-uuid", 'n', NULL, + "Skip UUID index search (UUID index not required for OCP 1.0)"), + OPT_END() + }; + + err = parse_and_open(&dev, argc, argv, desc, opts); + if (err) + return err; + + if (argconfig_parse_seen(opts, "mode")) + err = eol_plp_failure_mode_set(dev, nsid, fid, cfg.mode, + cfg.save, + !argconfig_parse_seen(opts, "no-uuid")); + else + err = eol_plp_failure_mode_get(dev, nsid, fid, cfg.sel); + + dev_close(dev); + + return err; } -static int clear_fw_update_history(int argc, char **argv, struct command *cmd, - struct plugin *plugin) +static int clear_pcie_corectable_error_counters(int argc, char **argv, + struct command *cmd, + struct plugin *plugin) { - return ocp_clear_fw_update_history(argc, argv, cmd, plugin); + const char *desc = "OCP Clear PCIe Correctable Error Counters"; + + return ocp_clear_feature(argc, argv, desc, + OCP_FID_CLEAR_PCIE_CORRECTABLE_ERROR_COUNTERS); } diff --git a/plugins/ocp/ocp-nvme.h b/plugins/ocp/ocp-nvme.h index c20646a..dc9e154 100644 --- a/plugins/ocp/ocp-nvme.h +++ b/plugins/ocp/ocp-nvme.h @@ -3,7 +3,7 @@ * * Authors: Arthur Shau <arthurshau@fb.com>, * Wei Zhang <wzhang@fb.com>, - * Venkat Ramesh <venkatraghavan@fb.com> + * Venkat Ramesh <venkatraghavan@fb.com> */ #undef CMD_INC_FILE #define CMD_INC_FILE plugins/ocp/ocp-nvme @@ -15,11 +15,11 @@ PLUGIN(NAME("ocp", "OCP cloud SSD extensions", NVME_VERSION), COMMAND_LIST( - ENTRY("smart-add-log", "Retrieve extended SMART Information", ocp_smart_add_log) - ENTRY("latency-monitor-log", "Get Latency Monitor Log Page", - ocp_latency_monitor_log) - ENTRY("clear-fw-activate-history", "Clear firmware update history log", - clear_fw_update_history) + ENTRY("smart-add-log", "Retrieve extended SMART Information", smart_add_log) + ENTRY("latency-monitor-log", "Get Latency Monitor Log Page", ocp_latency_monitor_log) + ENTRY("clear-fw-activate-history", "Clear firmware update history log", clear_fw_update_history) + ENTRY("eol-plp-failure-mode", "Define EOL or PLP circuitry failure mode.", eol_plp_failure_mode) + ENTRY("clear-pcie-correctable-error-counters", "Clear PCIe correctable error counters", clear_pcie_corectable_error_counters) ) ); diff --git a/plugins/ocp/ocp-smart-extended-log.c b/plugins/ocp/ocp-smart-extended-log.c new file mode 100644 index 0000000..37b62e9 --- /dev/null +++ b/plugins/ocp/ocp-smart-extended-log.c @@ -0,0 +1,352 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* Copyright (c) 2022 Meta Platforms, Inc. + * + * Authors: Arthur Shau <arthurshau@fb.com>, + * Wei Zhang <wzhang@fb.com>, + * Venkat Ramesh <venkatraghavan@fb.com> + */ + +#include "ocp-smart-extended-log.h" + +#include <errno.h> +#include <stdio.h> + +#include "common.h" +#include "nvme-print.h" + +/* C0 SCAO Log Page */ +#define C0_SMART_CLOUD_ATTR_LEN 0x200 +#define C0_SMART_CLOUD_ATTR_OPCODE 0xC0 +#define C0_GUID_LENGTH 16 + +static __u8 scao_guid[C0_GUID_LENGTH] = { + 0xC5, 0xAF, 0x10, 0x28, + 0xEA, 0xBF, 0xF2, 0xA4, + 0x9C, 0x4F, 0x6F, 0x7C, + 0xC9, 0x14, 0xD5, 0xAF +}; + +typedef enum { + SCAO_PMUW = 0, /* Physical media units written */ + SCAO_PMUR = 16, /* Physical media units read */ + SCAO_BUNBR = 32, /* Bad user nand blocks raw */ + SCAO_BUNBN = 38, /* Bad user nand blocks normalized */ + SCAO_BSNBR = 40, /* Bad system nand blocks raw */ + SCAO_BSNBN = 46, /* Bad system nand blocks normalized */ + SCAO_XRC = 48, /* XOR recovery count */ + SCAO_UREC = 56, /* Uncorrectable read error count */ + SCAO_SEEC = 64, /* Soft ecc error count */ + SCAO_EEDC = 72, /* End to end detected errors */ + SCAO_EECE = 76, /* End to end corrected errors */ + SCAO_SDPU = 80, /* System data percent used */ + SCAO_RFSC = 81, /* Refresh counts */ + SCAO_MXUDEC = 88, /* Max User data erase counts */ + SCAO_MNUDEC = 92, /* Min User data erase counts */ + SCAO_NTTE = 96, /* Number of Thermal throttling events */ + SCAO_CTS = 97, /* Current throttling status */ + SCAO_EVF = 98, /* Errata Version Field */ + SCAO_PVF = 99, /* Point Version Field */ + SCAO_MIVF = 101, /* Minor Version Field */ + SCAO_MAVF = 103, /* Major Version Field */ + SCAO_PCEC = 104, /* PCIe correctable error count */ + SCAO_ICS = 112, /* Incomplete shutdowns */ + SCAO_PFB = 120, /* Percent free blocks */ + SCAO_CPH = 128, /* Capacitor health */ + SCAO_NEV = 130, /* NVMe Errata Version */ + SCAO_UIO = 136, /* Unaligned I/O */ + SCAO_SVN = 144, /* Security Version Number */ + SCAO_NUSE = 152, /* NUSE - Namespace utilization */ + SCAO_PSC = 160, /* PLP start count */ + SCAO_EEST = 176, /* Endurance estimate */ + SCAO_PLRC = 192, /* PCIe Link Retraining Count */ + SCAO_PSCC = 200, /* Power State Change Count */ + SCAO_LPV = 494, /* Log page version */ + SCAO_LPG = 496, /* Log page GUID */ +} SMART_CLOUD_ATTRIBUTE_OFFSETS; + +static void ocp_print_C0_log_normal(void *data) +{ + uint16_t smart_log_ver = 0; + __u8 *log_data = data; + + printf("SMART Cloud Attributes :-\n"); + + printf(" Physical media units written - %"PRIu64" %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW + 8] & 0xFFFFFFFFFFFFFFFF), + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW] & 0xFFFFFFFFFFFFFFFF)); + printf(" Physical media units read - %"PRIu64" %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR + 8] & 0xFFFFFFFFFFFFFFFF), + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR] & 0xFFFFFFFFFFFFFFFF)); + printf(" Bad user nand blocks - Raw %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BUNBR] & 0x0000FFFFFFFFFFFF)); + printf(" Bad user nand blocks - Normalized %d\n", + (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BUNBN])); + printf(" Bad system nand blocks - Raw %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BSNBR] & 0x0000FFFFFFFFFFFF)); + printf(" Bad system nand blocks - Normalized %d\n", + (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BSNBN])); + printf(" XOR recovery count %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_XRC])); + printf(" Uncorrectable read error count %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UREC])); + printf(" Soft ecc error count %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SEEC])); + printf(" End to end detected errors %"PRIu32"\n", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EEDC])); + printf(" End to end corrected errors %"PRIu32"\n", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EECE])); + printf(" System data percent used %d\n", + (__u8)log_data[SCAO_SDPU]); + printf(" Refresh counts %"PRIu64"\n", + (uint64_t)(le64_to_cpu(*(uint64_t *)&log_data[SCAO_RFSC]) & 0x00FFFFFFFFFFFFFF)); + printf(" Max User data erase counts %"PRIu32"\n", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MXUDEC])); + printf(" Min User data erase counts %"PRIu32"\n", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MNUDEC])); + printf(" Number of Thermal throttling events %d\n", + (__u8)log_data[SCAO_NTTE]); + printf(" Current throttling status 0x%x\n", + (__u8)log_data[SCAO_CTS]); + printf(" PCIe correctable error count %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PCEC])); + printf(" Incomplete shutdowns %"PRIu32"\n", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_ICS])); + printf(" Percent free blocks %d\n", + (__u8)log_data[SCAO_PFB]); + printf(" Capacitor health %"PRIu16"\n", + (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_CPH])); + printf(" Unaligned I/O %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UIO])); + printf(" Security Version Number %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SVN])); + printf(" NUSE - Namespace utilization %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_NUSE])); + printf(" PLP start count %s\n", + uint128_t_to_string(le128_to_cpu(&log_data[SCAO_PSC]))); + printf(" Endurance estimate %s\n", + uint128_t_to_string(le128_to_cpu(&log_data[SCAO_EEST]))); + smart_log_ver = (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_LPV]); + printf(" Log page version %"PRIu16"\n", smart_log_ver); + printf(" Log page GUID 0x"); + printf("%"PRIx64"%"PRIx64"\n", (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG + 8]), + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG])); + if (smart_log_ver > 2) { + printf(" Errata Version Field %d\n", + (__u8)log_data[SCAO_EVF]); + printf(" Point Version Field %"PRIu16"\n", + le16_to_cpu(*(uint16_t *)&log_data[SCAO_PVF])); + printf(" Minor Version Field %"PRIu16"\n", + le16_to_cpu(*(uint16_t *)&log_data[SCAO_MIVF])); + printf(" Major Version Field %d\n", + (__u8)log_data[SCAO_MAVF]); + printf(" NVMe Errata Version %d\n", + (__u8)log_data[SCAO_NEV]); + printf(" PCIe Link Retraining Count %"PRIu64"\n", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PLRC])); + printf(" Power State Change Count %"PRIu64"\n", + le64_to_cpu(*(uint64_t *)&log_data[SCAO_PSCC])); + } + printf("\n"); +} + +static void ocp_print_C0_log_json(void *data) +{ + struct json_object *root; + struct json_object *pmuw; + struct json_object *pmur; + uint16_t smart_log_ver = 0; + __u8 *log_data = data; + char guid[40]; + + root = json_create_object(); + pmuw = json_create_object(); + pmur = json_create_object(); + + json_object_add_value_uint64(pmuw, "hi", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW + 8] & 0xFFFFFFFFFFFFFFFF)); + json_object_add_value_uint64(pmuw, "lo", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUW] & 0xFFFFFFFFFFFFFFFF)); + json_object_add_value_object(root, "Physical media units written", pmuw); + json_object_add_value_uint64(pmur, "hi", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR + 8] & 0xFFFFFFFFFFFFFFFF)); + json_object_add_value_uint64(pmur, "lo", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PMUR] & 0xFFFFFFFFFFFFFFFF)); + json_object_add_value_object(root, "Physical media units read", pmur); + json_object_add_value_uint64(root, "Bad user nand blocks - Raw", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BUNBR] & 0x0000FFFFFFFFFFFF)); + json_object_add_value_uint(root, "Bad user nand blocks - Normalized", + (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BUNBN])); + json_object_add_value_uint64(root, "Bad system nand blocks - Raw", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_BSNBR] & 0x0000FFFFFFFFFFFF)); + json_object_add_value_uint(root, "Bad system nand blocks - Normalized", + (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_BSNBN])); + json_object_add_value_uint64(root, "XOR recovery count", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_XRC])); + json_object_add_value_uint64(root, "Uncorrectable read error count", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UREC])); + json_object_add_value_uint64(root, "Soft ecc error count", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SEEC])); + json_object_add_value_uint(root, "End to end detected errors", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EEDC])); + json_object_add_value_uint(root, "End to end corrected errors", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_EECE])); + json_object_add_value_uint(root, "System data percent used", + (__u8)log_data[SCAO_SDPU]); + json_object_add_value_uint64(root, "Refresh counts", + (uint64_t)(le64_to_cpu(*(uint64_t *)&log_data[SCAO_RFSC]) & 0x00FFFFFFFFFFFFFF)); + json_object_add_value_uint(root, "Max User data erase counts", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MXUDEC])); + json_object_add_value_uint(root, "Min User data erase counts", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_MNUDEC])); + json_object_add_value_uint(root, "Number of Thermal throttling events", + (__u8)log_data[SCAO_NTTE]); + json_object_add_value_uint(root, "Current throttling status", + (__u8)log_data[SCAO_CTS]); + json_object_add_value_uint64(root, "PCIe correctable error count", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PCEC])); + json_object_add_value_uint(root, "Incomplete shutdowns", + (uint32_t)le32_to_cpu(*(uint32_t *)&log_data[SCAO_ICS])); + json_object_add_value_uint(root, "Percent free blocks", + (__u8)log_data[SCAO_PFB]); + json_object_add_value_uint(root, "Capacitor health", + (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_CPH])); + json_object_add_value_uint64(root, "Unaligned I/O", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_UIO])); + json_object_add_value_uint64(root, "Security Version Number", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_SVN])); + json_object_add_value_uint64(root, "NUSE - Namespace utilization", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_NUSE])); + json_object_add_value_uint128(root, "PLP start count", + le128_to_cpu(&log_data[SCAO_PSC])); + json_object_add_value_uint128(root, "Endurance estimate", + le128_to_cpu(&log_data[SCAO_EEST])); + smart_log_ver = (uint16_t)le16_to_cpu(*(uint16_t *)&log_data[SCAO_LPV]); + + json_object_add_value_uint(root, "Log page version", smart_log_ver); + + memset((void *)guid, 0, 40); + sprintf((char *)guid, "0x%"PRIx64"%"PRIx64"", (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG + 8]), + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_LPG])); + json_object_add_value_string(root, "Log page GUID", guid); + + if (smart_log_ver > 2) { + json_object_add_value_uint(root, "Errata Version Field", + (__u8)log_data[SCAO_EVF]); + json_object_add_value_uint(root, "Point Version Field", + le16_to_cpu(*(uint16_t *)&log_data[SCAO_PVF])); + json_object_add_value_uint(root, "Minor Version Field", + le16_to_cpu(*(uint16_t *)&log_data[SCAO_MIVF])); + json_object_add_value_uint(root, "Major Version Field", + (__u8)log_data[SCAO_MAVF]); + json_object_add_value_uint(root, "NVMe Errata Version", + (__u8)log_data[SCAO_NEV]); + json_object_add_value_uint(root, "PCIe Link Retraining Count", + (uint64_t)le64_to_cpu(*(uint64_t *)&log_data[SCAO_PLRC])); + json_object_add_value_uint(root, "Power State Change Count", + le64_to_cpu(*(uint64_t *)&log_data[SCAO_PSCC])); + } + json_print_object(root, NULL); + printf("\n"); + json_free_object(root); +} + +static int get_c0_log_page(int fd, char *format) +{ + __u8 *data; + int i; + int ret = 0; + int fmt = -1; + + fmt = validate_output_format(format); + if (fmt < 0) { + fprintf(stderr, "ERROR : OCP : invalid output format\n"); + return fmt; + } + + data = malloc(sizeof(__u8) * C0_SMART_CLOUD_ATTR_LEN); + if (!data) { + fprintf(stderr, "ERROR : OCP : malloc : %s\n", strerror(errno)); + return -1; + } + memset(data, 0, sizeof(__u8) * C0_SMART_CLOUD_ATTR_LEN); + + ret = nvme_get_log_simple(fd, C0_SMART_CLOUD_ATTR_OPCODE, + C0_SMART_CLOUD_ATTR_LEN, data); + + if (strcmp(format, "json")) + fprintf(stderr, "NVMe Status:%s(%x)\n", + nvme_status_to_string(ret, false), ret); + + if (ret == 0) { + /* check log page guid */ + /* Verify GUID matches */ + for (i = 0; i < 16; i++) { + if (scao_guid[i] != data[SCAO_LPG + i]) { + int j; + + fprintf(stderr, "ERROR : OCP : Unknown GUID in C0 Log Page data\n"); + fprintf(stderr, "ERROR : OCP : Expected GUID: 0x"); + for (j = 0; j < 16; j++) { + fprintf(stderr, "%x", scao_guid[j]); + } + + fprintf(stderr, "\nERROR : OCP : Actual GUID: 0x"); + for (j = 0; j < 16; j++) { + fprintf(stderr, "%x", data[SCAO_LPG + j]); + } + fprintf(stderr, "\n"); + + ret = -1; + goto out; + } + } + + /* print the data */ + switch (fmt) { + case NORMAL: + ocp_print_C0_log_normal(data); + break; + case JSON: + ocp_print_C0_log_json(data); + break; + } + } else { + fprintf(stderr, "ERROR : OCP : Unable to read C0 data from buffer\n"); + } + +out: + free(data); + return ret; +} + +int ocp_smart_add_log(int argc, char **argv, struct command *cmd, + struct plugin *plugin) +{ + const char *desc = "Retrieve the extended SMART health data."; + struct nvme_dev *dev; + int ret = 0; + + struct config { + char *output_format; + }; + + struct config cfg = { + .output_format = "normal", + }; + + OPT_ARGS(opts) = { + OPT_FMT("output-format", 'o', &cfg.output_format, "output Format: normal|json"), + OPT_END() + }; + + ret = parse_and_open(&dev, argc, argv, desc, opts); + if (ret) + return ret; + + ret = get_c0_log_page(dev_fd(dev), cfg.output_format); + if (ret) + fprintf(stderr, "ERROR : OCP : Failure reading the C0 Log Page, ret = %d\n", + ret); + dev_close(dev); + return ret; +} diff --git a/plugins/ocp/ocp-smart-extended-log.h b/plugins/ocp/ocp-smart-extended-log.h new file mode 100644 index 0000000..42c1f98 --- /dev/null +++ b/plugins/ocp/ocp-smart-extended-log.h @@ -0,0 +1,18 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* Copyright (c) 2022 Meta Platforms, Inc. + * + * Authors: Arthur Shau <arthurshau@fb.com>, + * Wei Zhang <wzhang@fb.com>, + * Venkat Ramesh <venkatraghavan@fb.com> + */ + +#ifndef OCP_SMART_EXTENDED_LOG_H +#define OCP_SMART_EXTENDED_LOG_H + +struct command; +struct plugin; + +int ocp_smart_add_log(int argc, char **argv, struct command *cmd, + struct plugin *plugin); + +#endif diff --git a/plugins/ocp/ocp-utils.c b/plugins/ocp/ocp-utils.c index 9294c05..a37a58c 100644 --- a/plugins/ocp/ocp-utils.c +++ b/plugins/ocp/ocp-utils.c @@ -5,11 +5,13 @@ * Author: leonardo.da.cunha@solidigm.com */ +#include <unistd.h> #include "ocp-utils.h" +#include "nvme-print.h" const unsigned char ocp_uuid[NVME_UUID_LEN] = { - 0x6f, 0xbe, 0x56, 0x8f, 0x99, 0x29, 0x1d, 0xa2, 0x94, 0x47, - 0x94, 0xe0, 0x5b, 0xd5, 0x94, 0xc1 }; + 0xc1, 0x94, 0xd5, 0x5b, 0xe0, 0x94, 0x47, 0x94, 0xa2, 0x1d, + 0x29, 0x99, 0x8f, 0x56, 0xbe, 0x6f }; int ocp_get_uuid_index(struct nvme_dev *dev, int *index) { @@ -28,3 +30,66 @@ int ocp_get_uuid_index(struct nvme_dev *dev, int *index) } return err; } + +int ocp_clear_feature(int argc, char **argv, const char *desc, const __u8 fid) +{ + __u32 result = 0; + __u32 clear = 1 << 31; + struct nvme_dev *dev; + int uuid_index = 0; + bool uuid = true; + int err; + + OPT_ARGS(opts) = { + OPT_FLAG("no-uuid", 'n', NULL, + "Skip UUID index search (UUID index not required for OCP 1.0)"), + OPT_END() + }; + + err = parse_and_open(&dev, argc, argv, desc, opts); + if (err) + return err; + + if (opts[0].seen) + uuid = false; + + if (uuid) { + /* OCP 2.0 requires UUID index support */ + err = ocp_get_uuid_index(dev, &uuid_index); + if (err || !uuid_index) { + fprintf(stderr, "ERROR: No OCP UUID index found\n"); + goto close_dev; + } + } + + struct nvme_set_features_args args = { + .result = &result, + .data = NULL, + .args_size = sizeof(args), + .fd = dev_fd(dev), + .timeout = NVME_DEFAULT_IOCTL_TIMEOUT, + .nsid = 0, + .cdw11 = clear, + .cdw12 = 0, + .cdw13 = 0, + .cdw15 = 0, + .data_len = 0, + .save = 0, + .uuidx = uuid_index, + .fid = fid, + }; + + err = nvme_set_features(&args); + + if (err == 0) + printf("Success : %s\n", desc); + else if (err > 0) + nvme_show_status(err); + else + printf("Fail : %s\n", desc); +close_dev: + /* Redundant close() to make static code analysis happy */ + close(dev->direct.fd); + dev_close(dev); + return err; +} diff --git a/plugins/ocp/ocp-utils.h b/plugins/ocp/ocp-utils.h index 44d0af4..a962169 100644 --- a/plugins/ocp/ocp-utils.h +++ b/plugins/ocp/ocp-utils.h @@ -16,3 +16,5 @@ * Return: Zero if nvme device has UUID list log page, or result of get uuid list otherwise. */ int ocp_get_uuid_index(struct nvme_dev *dev, int *index); + +int ocp_clear_feature(int argc, char **argv, const char *desc, const __u8 fid); |