[lustre-devel] [PATCH 8/8] staging: lustre: selftest: fix time reporting
James Simmons
jsimmons at infradead.org
Sun Jun 24 14:53:52 PDT 2018
Migrate crp_stamp to nanoseconds and both timestamps nd_stamp,
sn_start to ktime. The move away from jiffies, which can vary on
platforms, to something that is the consistent on every node.
Currently jiffies data was being transmitted over the wire and
the other node could have a different HZ value configured which
lead to incorrect results. This will ensure that the reported
results to the user will always be correct.
Signed-off-by: James Simmons <uja.ornl at yahoo.com>
WC-bug-id: https://jira.whamcloud.com/browse/LU-9019
Reviewed-on: https://review.whamcloud.com/31041
Reviewed-by: Doug Oucharek <dougso at me.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin at intel.com>
Reviewed-by: Oleg Drokin <green at whamcloud.com>
Signed-off-by: James Simmons <jsimmons at infradead.org>
---
.../lustre/include/uapi/linux/lnet/lnetst.h | 2 +-
drivers/staging/lustre/lnet/selftest/conrpc.c | 39 +++++++++++-----------
drivers/staging/lustre/lnet/selftest/conrpc.h | 2 +-
drivers/staging/lustre/lnet/selftest/console.c | 4 +--
drivers/staging/lustre/lnet/selftest/console.h | 2 +-
drivers/staging/lustre/lnet/selftest/framework.c | 4 +--
drivers/staging/lustre/lnet/selftest/selftest.h | 2 +-
drivers/staging/lustre/lnet/selftest/timer.c | 8 ++---
8 files changed, 32 insertions(+), 31 deletions(-)
diff --git a/drivers/staging/lustre/include/uapi/linux/lnet/lnetst.h b/drivers/staging/lustre/include/uapi/linux/lnet/lnetst.h
index a4f9ff0..7cfc241 100644
--- a/drivers/staging/lustre/include/uapi/linux/lnet/lnetst.h
+++ b/drivers/staging/lustre/include/uapi/linux/lnet/lnetst.h
@@ -71,7 +71,7 @@
struct lst_sid {
lnet_nid_t ses_nid; /* nid of console node */
- __u64 ses_stamp; /* time stamp */
+ __s64 ses_stamp; /* time stamp */
}; /*** session id */
extern struct lst_sid LST_INVALID_SID;
diff --git a/drivers/staging/lustre/lnet/selftest/conrpc.c b/drivers/staging/lustre/lnet/selftest/conrpc.c
index 0dabade..95cbd1a 100644
--- a/drivers/staging/lustre/lnet/selftest/conrpc.c
+++ b/drivers/staging/lustre/lnet/selftest/conrpc.c
@@ -70,11 +70,11 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
/* not an orphan RPC */
crpc->crp_finished = 1;
- if (!crpc->crp_stamp) {
+ if (!crpc->crp_stamp_ns) {
/* not aborted */
LASSERT(!crpc->crp_status);
- crpc->crp_stamp = jiffies;
+ crpc->crp_stamp_ns = ktime_get_ns();
crpc->crp_status = rpc->crpc_status;
}
@@ -102,7 +102,7 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
crpc->crp_finished = 0;
crpc->crp_unpacked = 0;
crpc->crp_status = 0;
- crpc->crp_stamp = 0;
+ crpc->crp_stamp_ns = 0;
crpc->crp_embedded = embedded;
INIT_LIST_HEAD(&crpc->crp_link);
@@ -294,16 +294,16 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
spin_lock(&rpc->crpc_lock);
if (!crpc->crp_posted || /* not posted */
- crpc->crp_stamp) { /* rpc done or aborted already */
- if (!crpc->crp_stamp) {
- crpc->crp_stamp = jiffies;
+ crpc->crp_stamp_ns) { /* rpc done or aborted already */
+ if (!crpc->crp_stamp_ns) {
+ crpc->crp_stamp_ns = ktime_get_ns();
crpc->crp_status = -EINTR;
}
spin_unlock(&rpc->crpc_lock);
continue;
}
- crpc->crp_stamp = jiffies;
+ crpc->crp_stamp_ns = ktime_get_ns();
crpc->crp_status = error;
spin_unlock(&rpc->crpc_lock);
@@ -314,10 +314,10 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
continue;
nd = crpc->crp_node;
- if (time_after(nd->nd_stamp, crpc->crp_stamp))
+ if (ktime_to_ns(nd->nd_stamp) > crpc->crp_stamp_ns)
continue;
- nd->nd_stamp = crpc->crp_stamp;
+ nd->nd_stamp = ktime_set(0, crpc->crp_stamp_ns);
nd->nd_state = LST_NODE_DOWN;
}
}
@@ -390,7 +390,7 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
struct srpc_generic_reply *rep;
LASSERT(nd && rpc);
- LASSERT(crpc->crp_stamp);
+ LASSERT(crpc->crp_stamp_ns);
if (crpc->crp_status) {
*msgpp = NULL;
@@ -403,10 +403,10 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
crpc->crp_unpacked = 1;
}
- if (time_after(nd->nd_stamp, crpc->crp_stamp))
+ if (ktime_to_ns(nd->nd_stamp) > crpc->crp_stamp_ns)
return 0;
- nd->nd_stamp = crpc->crp_stamp;
+ nd->nd_stamp = ktime_set(0, crpc->crp_stamp_ns);
rep = &(*msgpp)->msg_body.reply;
if (rep->sid.ses_nid == LNET_NID_ANY)
@@ -433,7 +433,7 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
list_for_each_entry(crpc, &trans->tas_rpcs_list, crp_link) {
lstcon_rpc_stat_total(stat, 1);
- LASSERT(crpc->crp_stamp);
+ LASSERT(crpc->crp_stamp_ns);
error = lstcon_rpc_get_reply(crpc, &rep);
if (error) {
@@ -474,9 +474,9 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
struct lstcon_rpc *crpc;
struct srpc_msg *msg;
struct lstcon_node *nd;
- long dur;
struct timeval tv;
int error;
+ s64 dur;
LASSERT(head_up);
@@ -493,15 +493,15 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
ent = list_entry(next, struct lstcon_rpc_ent, rpe_link);
- LASSERT(crpc->crp_stamp);
+ LASSERT(crpc->crp_stamp_ns);
error = lstcon_rpc_get_reply(crpc, &msg);
nd = crpc->crp_node;
- dur = (long)(crpc->crp_stamp -
- (unsigned long)console_session.ses_id.ses_stamp);
- jiffies_to_timeval(dur, &tv);
+ dur = crpc->crp_stamp_ns -
+ console_session.ses_id.ses_stamp * NSEC_PER_MSEC;
+ tv = ns_to_timeval(dur);
if (copy_to_user(&ent->rpe_peer, &nd->nd_id,
sizeof(struct lnet_process_id)) ||
@@ -1246,7 +1246,8 @@ void lstcon_rpc_stat_reply(struct lstcon_rpc_trans *, struct srpc_msg *,
if (nd->nd_state != LST_NODE_ACTIVE)
continue;
- intv = (jiffies - nd->nd_stamp) / msecs_to_jiffies(MSEC_PER_SEC);
+ intv = div_u64(ktime_ms_delta(ktime_get(), nd->nd_stamp),
+ MSEC_PER_SEC);
if (intv < nd->nd_timeout / 2)
continue;
diff --git a/drivers/staging/lustre/lnet/selftest/conrpc.h b/drivers/staging/lustre/lnet/selftest/conrpc.h
index ce2f92d..dfc31b8 100644
--- a/drivers/staging/lustre/lnet/selftest/conrpc.h
+++ b/drivers/staging/lustre/lnet/selftest/conrpc.h
@@ -70,7 +70,7 @@ struct lstcon_rpc {
/** RPC is embedded in other structure and can't free it */
unsigned int crp_embedded:1;
int crp_status; /* console rpc errors */
- unsigned long crp_stamp; /* replied time stamp */
+ s64 crp_stamp_ns; /* replied time stamp */
};
struct lstcon_rpc_trans {
diff --git a/drivers/staging/lustre/lnet/selftest/console.c b/drivers/staging/lustre/lnet/selftest/console.c
index 3c1c1b5..8e6f875 100644
--- a/drivers/staging/lustre/lnet/selftest/console.c
+++ b/drivers/staging/lustre/lnet/selftest/console.c
@@ -97,7 +97,7 @@
ndl->ndl_node->nd_ref = 1;
ndl->ndl_node->nd_id = id;
- ndl->ndl_node->nd_stamp = jiffies;
+ ndl->ndl_node->nd_stamp = ktime_get();
ndl->ndl_node->nd_state = LST_NODE_UNKNOWN;
ndl->ndl_node->nd_timeout = 0;
memset(&ndl->ndl_node->nd_ping, 0, sizeof(struct lstcon_rpc));
@@ -1700,7 +1700,7 @@ static void lstcon_group_ndlink_release(struct lstcon_group *,
LNetGetId(1, &id);
sid->ses_nid = id.nid;
- sid->ses_stamp = jiffies;
+ sid->ses_stamp = div_u64(ktime_get_ns(), NSEC_PER_MSEC);
}
int
diff --git a/drivers/staging/lustre/lnet/selftest/console.h b/drivers/staging/lustre/lnet/selftest/console.h
index 2826205..eaad07c 100644
--- a/drivers/staging/lustre/lnet/selftest/console.h
+++ b/drivers/staging/lustre/lnet/selftest/console.h
@@ -51,7 +51,7 @@ struct lstcon_node {
int nd_ref; /* reference count */
int nd_state; /* state of the node */
int nd_timeout; /* session timeout */
- unsigned long nd_stamp; /* timestamp of last replied RPC */
+ ktime_t nd_stamp; /* timestamp of last replied RPC */
struct lstcon_rpc nd_ping; /* ping rpc */
};
diff --git a/drivers/staging/lustre/lnet/selftest/framework.c b/drivers/staging/lustre/lnet/selftest/framework.c
index 741af10..6768aef 100644
--- a/drivers/staging/lustre/lnet/selftest/framework.c
+++ b/drivers/staging/lustre/lnet/selftest/framework.c
@@ -272,7 +272,7 @@
sn->sn_id = sid;
sn->sn_features = features;
sn->sn_timeout = session_timeout;
- sn->sn_started = jiffies;
+ sn->sn_started = ktime_get();
timer->stt_data = sn;
timer->stt_func = sfw_session_expired;
@@ -384,7 +384,7 @@
* send over the msecs since the session was started
* with 32 bits to send, this is ~49 days
*/
- cnt->running_ms = jiffies_to_msecs(jiffies - sn->sn_started);
+ cnt->running_ms = ktime_ms_delta(ktime_get(), sn->sn_started);
cnt->brw_errors = atomic_read(&sn->sn_brw_errors);
cnt->ping_errors = atomic_read(&sn->sn_ping_errors);
cnt->zombie_sessions = atomic_read(&sfw_data.fw_nzombies);
diff --git a/drivers/staging/lustre/lnet/selftest/selftest.h b/drivers/staging/lustre/lnet/selftest/selftest.h
index 8737fa9..76e02cd 100644
--- a/drivers/staging/lustre/lnet/selftest/selftest.h
+++ b/drivers/staging/lustre/lnet/selftest/selftest.h
@@ -331,7 +331,7 @@ struct sfw_session {
atomic_t sn_refcount;
atomic_t sn_brw_errors;
atomic_t sn_ping_errors;
- unsigned long sn_started;
+ ktime_t sn_started;
};
#define sfw_sid_equal(sid0, sid1) ((sid0).ses_nid == (sid1).ses_nid && \
diff --git a/drivers/staging/lustre/lnet/selftest/timer.c b/drivers/staging/lustre/lnet/selftest/timer.c
index 582f252..c83473e 100644
--- a/drivers/staging/lustre/lnet/selftest/timer.c
+++ b/drivers/staging/lustre/lnet/selftest/timer.c
@@ -55,7 +55,7 @@
static struct st_timer_data {
spinlock_t stt_lock;
- unsigned long stt_prev_slot; /* start time of the slot processed
+ time64_t stt_prev_slot; /* start time of the slot processed
* previously
*/
struct list_head stt_hash[STTIMER_NSLOTS];
@@ -144,18 +144,18 @@
}
static int
-stt_check_timers(unsigned long *last)
+stt_check_timers(time64_t *last)
{
int expired = 0;
time64_t now;
- unsigned long this_slot;
+ time64_t this_slot;
now = ktime_get_real_seconds();
this_slot = now & STTIMER_SLOTTIMEMASK;
spin_lock(&stt_data.stt_lock);
- while (time_after_eq(this_slot, *last)) {
+ while (this_slot >= *last) {
expired += stt_expire_list(STTIMER_SLOT(this_slot), now);
this_slot = this_slot - STTIMER_SLOTTIME;
}
--
1.8.3.1
More information about the lustre-devel
mailing list