blk-throttle: add interface for per-cgroup target latency
Here we introduce per-cgroup latency target. The target determines how a
cgroup can afford latency increasement. We will use the target latency
to calculate a threshold and use it to schedule IO for cgroups. If a
cgroup's bandwidth is below its low limit but its average latency is
below the threshold, other cgroups can safely dispatch more IO even
their bandwidth is higher than their low limits. On the other hand, if
the first cgroup's latency is higher than the threshold, other cgroups
are throttled to their low limits. So the target latency determines how
we efficiently utilize free disk resource without sacifice of worload's
IO latency.
For example, assume 4k IO average latency is 50us when disk isn't
congested. A cgroup sets the target latency to 30us. Then the cgroup can
accept 50+30=80us IO latency. If the cgroupt's average IO latency is
90us and its bandwidth is below low limit, other cgroups are throttled
to their low limit. If the cgroup's average IO latency is 60us, other
cgroups are allowed to dispatch more IO. When other cgroups dispatch
more IO, the first cgroup's IO latency will increase. If it increases to
81us, we then throttle other cgroups.
User will configure the interface in this way:
echo "8:16 rbps=2097152 wbps=max latency=100 idle=200" > io.low
latency is in microsecond unit
By default, latency target is 0, which means to guarantee IO latency.
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 0ea8698..6e1c298 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -25,6 +25,8 @@
#define DFL_IDLE_THRESHOLD_SSD (1000L) /* 1 ms */
#define DFL_IDLE_THRESHOLD_HD (100L * 1000) /* 100 ms */
#define MAX_IDLE_TIME (5L * 1000 * 1000) /* 5 s */
+/* default latency target is 0, eg, guarantee IO latency by default */
+#define DFL_LATENCY_TARGET (0)
static struct blkcg_policy blkcg_policy_throtl;
@@ -152,6 +154,7 @@
unsigned long last_check_time;
+ unsigned long latency_target; /* us */
/* When did we start a new slice */
unsigned long slice_start[2];
unsigned long slice_end[2];
@@ -449,6 +452,8 @@
tg->iops_conf[WRITE][LIMIT_MAX] = UINT_MAX;
/* LIMIT_LOW will have default value 0 */
+ tg->latency_target = DFL_LATENCY_TARGET;
+
return &tg->pd;
}
@@ -1445,6 +1450,7 @@
u64 bps_dft;
unsigned int iops_dft;
char idle_time[26] = "";
+ char latency_time[26] = "";
if (!dname)
return 0;
@@ -1461,8 +1467,9 @@
tg->bps_conf[WRITE][off] == bps_dft &&
tg->iops_conf[READ][off] == iops_dft &&
tg->iops_conf[WRITE][off] == iops_dft &&
- (off != LIMIT_LOW || tg->idletime_threshold ==
- tg->td->dft_idletime_threshold))
+ (off != LIMIT_LOW ||
+ (tg->idletime_threshold == tg->td->dft_idletime_threshold &&
+ tg->latency_target == DFL_LATENCY_TARGET)))
return 0;
if (tg->bps_conf[READ][off] != bps_dft)
@@ -1483,10 +1490,17 @@
else
snprintf(idle_time, sizeof(idle_time), " idle=%lu",
tg->idletime_threshold);
+
+ if (tg->latency_target == ULONG_MAX)
+ strcpy(latency_time, " latency=max");
+ else
+ snprintf(latency_time, sizeof(latency_time),
+ " latency=%lu", tg->latency_target);
}
- seq_printf(sf, "%s rbps=%s wbps=%s riops=%s wiops=%s%s\n",
- dname, bufs[0], bufs[1], bufs[2], bufs[3], idle_time);
+ seq_printf(sf, "%s rbps=%s wbps=%s riops=%s wiops=%s%s%s\n",
+ dname, bufs[0], bufs[1], bufs[2], bufs[3], idle_time,
+ latency_time);
return 0;
}
@@ -1505,6 +1519,7 @@
struct throtl_grp *tg;
u64 v[4];
unsigned long idle_time;
+ unsigned long latency_time;
int ret;
int index = of_cft(of)->private;
@@ -1520,6 +1535,7 @@
v[3] = tg->iops_conf[WRITE][index];
idle_time = tg->idletime_threshold;
+ latency_time = tg->latency_target;
while (true) {
char tok[27]; /* wiops=18446744073709551616 */
char *p;
@@ -1553,6 +1569,8 @@
v[3] = min_t(u64, val, UINT_MAX);
else if (off == LIMIT_LOW && !strcmp(tok, "idle"))
idle_time = val;
+ else if (off == LIMIT_LOW && !strcmp(tok, "latency"))
+ latency_time = val;
else
goto out_finish;
}
@@ -1583,6 +1601,8 @@
tg->td->limit_index = LIMIT_LOW;
tg->idletime_threshold = (idle_time == ULONG_MAX) ?
ULONG_MAX : idle_time;
+ tg->latency_target = (latency_time == ULONG_MAX) ?
+ ULONG_MAX : latency_time;
}
tg_conf_updated(tg);
ret = 0;