[RFC PATCH 1/4] powerpc/papr_scm: Fetch nvdimm performance stats from PHYP
Vaibhav Jain
vaibhav at linux.ibm.com
Mon May 18 21:08:11 AEST 2020
Update papr_scm.c to query dimm performance statistics from PHYP via
H_SCM_PERFORMANCE_STATS hcall and export them to userspace as PAPR
specific NVDIMM attribute 'perf_stats' in sysfs. The patch also
provide a sysfs ABI documentation for the stats being reported and
their meanings.
During NVDIMM probe time in papr_scm_nvdimm_init() a special variant
of H_SCM_PERFORMANCE_STATS hcall is issued to check if collection of
performance statistics is supported or not. If yes then a per-nvdimm
performance stats buffer is of size as returned by PHYP is allocated
and stored along with its length in two newly introduced NVDIMM
private struct members 'perf_stats' and 'len_stat_buffer'.
Signed-off-by: Vaibhav Jain <vaibhav at linux.ibm.com>
---
Documentation/ABI/testing/sysfs-bus-papr-scm | 27 ++++
arch/powerpc/platforms/pseries/papr_scm.c | 156 +++++++++++++++++++
2 files changed, 183 insertions(+)
diff --git a/Documentation/ABI/testing/sysfs-bus-papr-scm b/Documentation/ABI/testing/sysfs-bus-papr-scm
index 6143d06072f1..ad06b3e9c315 100644
--- a/Documentation/ABI/testing/sysfs-bus-papr-scm
+++ b/Documentation/ABI/testing/sysfs-bus-papr-scm
@@ -25,3 +25,30 @@ Description:
NVDIMM have been scrubbed.
* "locked" : Indicating that NVDIMM contents cant
be modified until next power cycle.
+
+What: /sys/bus/nd/devices/nmemX/papr/perf_stats
+Date: May, 2020
+KernelVersion: v5.8
+Contact: linuxppc-dev <linuxppc-dev at lists.ozlabs.org>, linux-nvdimm at lists.01.org,
+Description:
+ (RO) Report various performance stats related to papr-scm NVDIMM
+ device. Each stat is reported on a new line with each line
+ composed of a stat-identifier followed by it value. Below are
+ currently known dimm performance stats which are reported:
+
+ * "CtlResCt" : Controller Reset Count
+ * "CtlResTm" : Controller Reset Elapsed Time
+ * "PonSecs " : Power-on Seconds
+ * "MemLife " : Life Remaining
+ * "CritRscU" : Critical Resource Utilization
+ * "HostLCnt" : Host Load Count
+ * "HostSCnt" : Host Store Count
+ * "HostSDur" : Host Store Duration
+ * "HostLDur" : Host Load Duration
+ * "MedRCnt " : Media Read Count
+ * "MedWCnt " : Media Write Count
+ * "MedRDur " : Media Read Duration
+ * "MedWDur " : Media Write Duration
+ * "CchRHCnt" : Cache Read Hit Count
+ * "CchWHCnt" : Cache Write Hit Count
+ * "FastWCnt" : Fast Write Count
\ No newline at end of file
diff --git a/arch/powerpc/platforms/pseries/papr_scm.c b/arch/powerpc/platforms/pseries/papr_scm.c
index c59bf17ad054..fd9a12275315 100644
--- a/arch/powerpc/platforms/pseries/papr_scm.c
+++ b/arch/powerpc/platforms/pseries/papr_scm.c
@@ -62,6 +62,24 @@
PAPR_SCM_DIMM_HEALTH_FATAL | \
PAPR_SCM_DIMM_HEALTH_UNHEALTHY)
+#define PAPR_SCM_PERF_STATS_EYECATCHER __stringify(SCMSTATS)
+#define PAPR_SCM_PERF_STATS_VERSION 0x1
+
+/* Struct holding a single performance metric */
+struct papr_scm_perf_stat {
+ u8 statistic_id[8];
+ u64 statistic_value;
+};
+
+/* Struct exchanged between kernel and PHYP for fetching drc perf stats */
+struct papr_scm_perf_stats {
+ u8 eye_catcher[8];
+ u32 stats_version; /* Should be 0x01 */
+ u32 num_statistics; /* Number of stats following */
+ /* zero or more performance matrics */
+ struct papr_scm_perf_stat scm_statistics[];
+} __packed;
+
/* private struct associated with each region */
struct papr_scm_priv {
struct platform_device *pdev;
@@ -89,6 +107,12 @@ struct papr_scm_priv {
/* Health information for the dimm */
struct nd_papr_pdsm_health health;
+
+ /* length of the stat buffer as expected by phyp */
+ size_t len_stat_buffer;
+
+ /* Cached version of all performance state */
+ struct papr_scm_perf_stats *perf_stats;
};
static int drc_pmem_bind(struct papr_scm_priv *p)
@@ -194,6 +218,75 @@ static int drc_pmem_query_n_bind(struct papr_scm_priv *p)
return drc_pmem_bind(p);
}
+/*
+ * Query the Dimm performance stats from PHYP and copy them (if returned) to
+ * provided struct papr_scm_perf_stats instance 'stats' of 'size' in bytes.
+ * The value of R4 is copied to 'out' if the pointer is provided.
+ */
+static int drc_pmem_query_stats(struct papr_scm_priv *p,
+ struct papr_scm_perf_stats *buff_stats,
+ size_t size, unsigned int num_stats,
+ uint64_t *out)
+{
+ unsigned long ret[PLPAR_HCALL_BUFSIZE];
+ struct papr_scm_perf_stat *stats;
+ s64 rc, i;
+
+ /* Setup the out buffer */
+ if (buff_stats) {
+ memcpy(buff_stats->eye_catcher,
+ PAPR_SCM_PERF_STATS_EYECATCHER, 8);
+ buff_stats->stats_version =
+ cpu_to_be32(PAPR_SCM_PERF_STATS_VERSION);
+ buff_stats->num_statistics =
+ cpu_to_be32(num_stats);
+ } else {
+ /* In case of no out buffer ignore the size */
+ size = 0;
+ }
+
+ /*
+ * Do the HCALL asking PHYP for info and if R4 was requested
+ * return its value in 'out' variable.
+ */
+ rc = plpar_hcall(H_SCM_PERFORMANCE_STATS, ret, p->drc_index,
+ __pa(buff_stats), size);
+ if (out)
+ *out = ret[0];
+
+ if (rc == H_PARTIAL) {
+ dev_err(&p->pdev->dev,
+ "Unknown performance stats, Err:0x%016lX\n", ret[0]);
+ return -ENOENT;
+ } else if (rc != H_SUCCESS) {
+ dev_err(&p->pdev->dev,
+ "Failed to query performance stats, Err:%lld\n", rc);
+ return -ENXIO;
+ }
+
+ /* Successfully fetched the requested stats from phyp */
+ if (size != 0) {
+ buff_stats->num_statistics =
+ be32_to_cpu(buff_stats->num_statistics);
+
+ /* Transform the stats buffer values from BE to cpu native */
+ for (i = 0, stats = buff_stats->scm_statistics;
+ i < buff_stats->num_statistics; ++i) {
+ stats[i].statistic_value =
+ be64_to_cpu(stats[i].statistic_value);
+ }
+ dev_dbg(&p->pdev->dev,
+ "Performance stats returned %d stats\n",
+ buff_stats->num_statistics);
+ } else {
+ /* Handle case where stat buffer size was requested */
+ dev_dbg(&p->pdev->dev,
+ "Performance stats size %ld\n", ret[0]);
+ }
+
+ return 0;
+}
+
/*
* Issue hcall to retrieve dimm health info and populate papr_scm_priv with the
* health information.
@@ -563,6 +656,47 @@ static int papr_scm_ndctl(struct nvdimm_bus_descriptor *nd_desc,
return *cmd_rc;
}
+static ssize_t perf_stats_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ int index, rc;
+ struct seq_buf s;
+ struct nvdimm *dimm = to_nvdimm(dev);
+ struct papr_scm_priv *p = nvdimm_provider_data(dimm);
+ struct papr_scm_perf_stat *stats = p->perf_stats->scm_statistics;
+
+ if (!p->len_stat_buffer)
+ return -ENOENT;
+
+ seq_buf_init(&s, buf, PAGE_SIZE);
+
+ /* Protect concurrent modifications to papr_scm_priv */
+ rc = mutex_lock_interruptible(&p->health_mutex);
+ if (rc)
+ return rc;
+
+ /* Ask phyp to return all dimm perf stats */
+ rc = drc_pmem_query_stats(p, p->perf_stats, p->len_stat_buffer, 0,
+ NULL);
+ if (!rc) {
+ /*
+ * Go through the returned output buffer and print stats and
+ * values. Since statistic_id is essentially a char string of
+ * 8 bytes, simply use the string format specifier to print it.
+ */
+ for (index = 0; index < p->perf_stats->num_statistics;
+ ++index) {
+ seq_buf_printf(&s, "%.8s = 0x%016llX\n",
+ stats[index].statistic_id,
+ stats[index].statistic_value);
+ }
+ }
+
+ mutex_unlock(&p->health_mutex);
+ return rc ? rc : seq_buf_used(&s);
+}
+DEVICE_ATTR_RO(perf_stats);
+
static ssize_t flags_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
@@ -615,6 +749,7 @@ DEVICE_ATTR_RO(flags);
/* papr_scm specific dimm attributes */
static struct attribute *papr_scm_nd_attributes[] = {
&dev_attr_flags.attr,
+ &dev_attr_perf_stats.attr,
NULL,
};
@@ -635,6 +770,7 @@ static int papr_scm_nvdimm_init(struct papr_scm_priv *p)
struct nd_region_desc ndr_desc;
unsigned long dimm_flags;
int target_nid, online_nid;
+ u64 stat_size;
p->bus_desc.ndctl = papr_scm_ndctl;
p->bus_desc.module = THIS_MODULE;
@@ -698,6 +834,26 @@ static int papr_scm_nvdimm_init(struct papr_scm_priv *p)
dev_info(dev, "Region registered with target node %d and online node %d",
target_nid, online_nid);
+ /* Try retriving the stat buffer and see if its supported */
+ if (!drc_pmem_query_stats(p, NULL, 0, 0, &stat_size)) {
+ /* Allocate the buffer for phyp where stats are written */
+ p->perf_stats = kzalloc(stat_size, GFP_KERNEL);
+
+ /* Failed allocation is non fatal and results in limited data */
+ if (!p->perf_stats)
+ dev_dbg(&p->pdev->dev,
+ "Unable to allocate %llu bytes for perf-state",
+ stat_size);
+ else
+ p->len_stat_buffer = (size_t)stat_size;
+ } else {
+ dev_dbg(&p->pdev->dev, "Unable to retrieve performace stats\n");
+ }
+
+ /* Check if perf-stats buffer was allocated */
+ if (!p->len_stat_buffer)
+ dev_info(&p->pdev->dev, "Limited dimm info available\n");
+
return 0;
err: nvdimm_bus_unregister(p->bus);
--
2.26.2
More information about the Linuxppc-dev
mailing list