[Qemu-devel] [PULL 31/44] block: Compute minimum, maximum and average I/O latencies

[Stefan Hajnoczi <stefanha@redhat.com>]

From: Alberto Garcia <berto@igalia.com>

This patch keeps track of the minimum, maximum and average latencies
of I/O operations during a certain interval of time.

The values are exposed in the BlockDeviceTimedStats structure.

An option to define the intervals to collect these statistics will be
added in a separate patch.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Message-id: c7382dc89622c64f918d09f32815827772628f8e.1446044837.git.berto@igalia.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/accounting.c         | 43 ++++++++++++++++++++++++++++++++++++++
 block/block-backend.c      |  1 +
 block/qapi.c               | 28 +++++++++++++++++++++++++
 include/block/accounting.h | 14 +++++++++++++
 qapi/block-core.json       | 52 +++++++++++++++++++++++++++++++++++++++++++++-
 qmp-commands.hx            | 31 +++++++++++++++++++++++++++
 6 files changed, 168 insertions(+), 1 deletion(-)

diff --git a/block/accounting.c b/block/accounting.c
index 923aeaf..61de8ce 100644
--- a/block/accounting.c
+++ b/block/accounting.c
@@ -35,6 +35,39 @@ void block_acct_init(BlockAcctStats *stats, bool account_invalid,
     stats->account_failed = account_failed;
 }
 
+void block_acct_cleanup(BlockAcctStats *stats)
+{
+    BlockAcctTimedStats *s, *next;
+    QSLIST_FOREACH_SAFE(s, &stats->intervals, entries, next) {
+        g_free(s);
+    }
+}
+
+void block_acct_add_interval(BlockAcctStats *stats, unsigned interval_length)
+{
+    BlockAcctTimedStats *s;
+    unsigned i;
+
+    s = g_new0(BlockAcctTimedStats, 1);
+    s->interval_length = interval_length;
+    QSLIST_INSERT_HEAD(&stats->intervals, s, entries);
+
+    for (i = 0; i < BLOCK_MAX_IOTYPE; i++) {
+        timed_average_init(&s->latency[i], clock_type,
+                           (uint64_t) interval_length * NANOSECONDS_PER_SECOND);
+    }
+}
+
+BlockAcctTimedStats *block_acct_interval_next(BlockAcctStats *stats,
+                                              BlockAcctTimedStats *s)
+{
+    if (s == NULL) {
+        return QSLIST_FIRST(&stats->intervals);
+    } else {
+        return QSLIST_NEXT(s, entries);
+    }
+}
+
 void block_acct_start(BlockAcctStats *stats, BlockAcctCookie *cookie,
                       int64_t bytes, enum BlockAcctType type)
 {
@@ -47,6 +80,7 @@ void block_acct_start(BlockAcctStats *stats, BlockAcctCookie *cookie,
 
 void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie)
 {
+    BlockAcctTimedStats *s;
     int64_t time_ns = qemu_clock_get_ns(clock_type);
     int64_t latency_ns = time_ns - cookie->start_time_ns;
 
@@ -56,6 +90,10 @@ void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie)
     stats->nr_ops[cookie->type]++;
     stats->total_time_ns[cookie->type] += latency_ns;
     stats->last_access_time_ns = time_ns;
+
+    QSLIST_FOREACH(s, &stats->intervals, entries) {
+        timed_average_account(&s->latency[cookie->type], latency_ns);
+    }
 }
 
 void block_acct_failed(BlockAcctStats *stats, BlockAcctCookie *cookie)
@@ -65,11 +103,16 @@ void block_acct_failed(BlockAcctStats *stats, BlockAcctCookie *cookie)
     stats->failed_ops[cookie->type]++;
 
     if (stats->account_failed) {
+        BlockAcctTimedStats *s;
         int64_t time_ns = qemu_clock_get_ns(clock_type);
         int64_t latency_ns = time_ns - cookie->start_time_ns;
 
         stats->total_time_ns[cookie->type] += latency_ns;
         stats->last_access_time_ns = time_ns;
+
+        QSLIST_FOREACH(s, &stats->intervals, entries) {
+            timed_average_account(&s->latency[cookie->type], latency_ns);
+        }
     }
 }
 
diff --git a/block/block-backend.c b/block/block-backend.c
index 19fdaae..fd4f895 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -176,6 +176,7 @@ static void blk_delete(BlockBackend *blk)
     }
     g_free(blk->name);
     drive_info_del(blk->legacy_dinfo);
+    block_acct_cleanup(&blk->stats);
     g_free(blk);
 }
 
diff --git a/block/qapi.c b/block/qapi.c
index 56c8139..4baf6e1 100644
--- a/block/qapi.c
+++ b/block/qapi.c
@@ -346,6 +346,7 @@ static BlockStats *bdrv_query_stats(const BlockDriverState *bs,
     s->stats = g_malloc0(sizeof(*s->stats));
     if (bs->blk) {
         BlockAcctStats *stats = blk_get_stats(bs->blk);
+        BlockAcctTimedStats *ts = NULL;
 
         s->stats->rd_bytes = stats->nr_bytes[BLOCK_ACCT_READ];
         s->stats->wr_bytes = stats->nr_bytes[BLOCK_ACCT_WRITE];
@@ -375,6 +376,33 @@ static BlockStats *bdrv_query_stats(const BlockDriverState *bs,
 
         s->stats->account_invalid = stats->account_invalid;
         s->stats->account_failed = stats->account_failed;
+
+        while ((ts = block_acct_interval_next(stats, ts))) {
+            BlockDeviceTimedStatsList *timed_stats =
+                g_malloc0(sizeof(*timed_stats));
+            BlockDeviceTimedStats *dev_stats = g_malloc0(sizeof(*dev_stats));
+            timed_stats->next = s->stats->timed_stats;
+            timed_stats->value = dev_stats;
+            s->stats->timed_stats = timed_stats;
+
+            TimedAverage *rd = &ts->latency[BLOCK_ACCT_READ];
+            TimedAverage *wr = &ts->latency[BLOCK_ACCT_WRITE];
+            TimedAverage *fl = &ts->latency[BLOCK_ACCT_FLUSH];
+
+            dev_stats->interval_length = ts->interval_length;
+
+            dev_stats->min_rd_latency_ns = timed_average_min(rd);
+            dev_stats->max_rd_latency_ns = timed_average_max(rd);
+            dev_stats->avg_rd_latency_ns = timed_average_avg(rd);
+
+            dev_stats->min_wr_latency_ns = timed_average_min(wr);
+            dev_stats->max_wr_latency_ns = timed_average_max(wr);
+            dev_stats->avg_wr_latency_ns = timed_average_avg(wr);
+
+            dev_stats->min_flush_latency_ns = timed_average_min(fl);
+            dev_stats->max_flush_latency_ns = timed_average_max(fl);
+            dev_stats->avg_flush_latency_ns = timed_average_avg(fl);
+        }
     }
 
     s->stats->wr_highest_offset = bs->wr_highest_offset;
diff --git a/include/block/accounting.h b/include/block/accounting.h
index 0d9b076..1dd582a 100644
--- a/include/block/accounting.h
+++ b/include/block/accounting.h
@@ -28,6 +28,9 @@
 #include <stdbool.h>
 
 #include "qemu/typedefs.h"
+#include "qemu/timed-average.h"
+
+typedef struct BlockAcctTimedStats BlockAcctTimedStats;
 
 enum BlockAcctType {
     BLOCK_ACCT_READ,
@@ -36,6 +39,12 @@ enum BlockAcctType {
     BLOCK_MAX_IOTYPE,
 };
 
+struct BlockAcctTimedStats {
+    TimedAverage latency[BLOCK_MAX_IOTYPE];
+    unsigned interval_length; /* in seconds */
+    QSLIST_ENTRY(BlockAcctTimedStats) entries;
+};
+
 typedef struct BlockAcctStats {
     uint64_t nr_bytes[BLOCK_MAX_IOTYPE];
     uint64_t nr_ops[BLOCK_MAX_IOTYPE];
@@ -44,6 +53,7 @@ typedef struct BlockAcctStats {
     uint64_t total_time_ns[BLOCK_MAX_IOTYPE];
     uint64_t merged[BLOCK_MAX_IOTYPE];
     int64_t last_access_time_ns;
+    QSLIST_HEAD(, BlockAcctTimedStats) intervals;
     bool account_invalid;
     bool account_failed;
 } BlockAcctStats;
@@ -56,6 +66,10 @@ typedef struct BlockAcctCookie {
 
 void block_acct_init(BlockAcctStats *stats, bool account_invalid,
                      bool account_failed);
+void block_acct_cleanup(BlockAcctStats *stats);
+void block_acct_add_interval(BlockAcctStats *stats, unsigned interval_length);
+BlockAcctTimedStats *block_acct_interval_next(BlockAcctStats *stats,
+                                              BlockAcctTimedStats *s);
 void block_acct_start(BlockAcctStats *stats, BlockAcctCookie *cookie,
                       int64_t bytes, enum BlockAcctType type);
 void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie);
diff --git a/qapi/block-core.json b/qapi/block-core.json
index b33663b..e2cfac6 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -414,6 +414,52 @@
 ##
 { 'command': 'query-block', 'returns': ['BlockInfo'] }
 
+
+##
+# @BlockDeviceTimedStats:
+#
+# Statistics of a block device during a given interval of time.
+#
+# @interval_length: Interval used for calculating the statistics,
+#                   in seconds.
+#
+# @min_rd_latency_ns: Minimum latency of read operations in the
+#                     defined interval, in nanoseconds.
+#
+# @min_wr_latency_ns: Minimum latency of write operations in the
+#                     defined interval, in nanoseconds.
+#
+# @min_flush_latency_ns: Minimum latency of flush operations in the
+#                        defined interval, in nanoseconds.
+#
+# @max_rd_latency_ns: Maximum latency of read operations in the
+#                     defined interval, in nanoseconds.
+#
+# @max_wr_latency_ns: Maximum latency of write operations in the
+#                     defined interval, in nanoseconds.
+#
+# @max_flush_latency_ns: Maximum latency of flush operations in the
+#                        defined interval, in nanoseconds.
+#
+# @avg_rd_latency_ns: Average latency of read operations in the
+#                     defined interval, in nanoseconds.
+#
+# @avg_wr_latency_ns: Average latency of write operations in the
+#                     defined interval, in nanoseconds.
+#
+# @avg_flush_latency_ns: Average latency of flush operations in the
+#                        defined interval, in nanoseconds.
+#
+# Since: 2.5
+##
+
+{ 'struct': 'BlockDeviceTimedStats',
+  'data': { 'interval_length': 'int', 'min_rd_latency_ns': 'int',
+            'max_rd_latency_ns': 'int', 'avg_rd_latency_ns': 'int',
+            'min_wr_latency_ns': 'int', 'max_wr_latency_ns': 'int',
+            'avg_wr_latency_ns': 'int', 'min_flush_latency_ns': 'int',
+            'max_flush_latency_ns': 'int', 'avg_flush_latency_ns': 'int' } }
+
 ##
 # @BlockDeviceStats:
 #
@@ -476,6 +522,9 @@
 # @account_failed: Whether failed operations are included in the
 #                  latency and last access statistics (Since 2.5)
 #
+# @timed_stats: Statistics specific to the set of previously defined
+#               intervals of time (Since 2.5)
+#
 # Since: 0.14.0
 ##
 { 'struct': 'BlockDeviceStats',
@@ -487,7 +536,8 @@
            'failed_rd_operations': 'int', 'failed_wr_operations': 'int',
            'failed_flush_operations': 'int', 'invalid_rd_operations': 'int',
            'invalid_wr_operations': 'int', 'invalid_flush_operations': 'int',
-           'account_invalid': 'bool', 'account_failed': 'bool' } }
+           'account_invalid': 'bool', 'account_failed': 'bool',
+           'timed_stats': ['BlockDeviceTimedStats'] } }
 
 ##
 # @BlockStats:
diff --git a/qmp-commands.hx b/qmp-commands.hx
index 70cfea5..e9b3a76 100644
--- a/qmp-commands.hx
+++ b/qmp-commands.hx
@@ -2547,6 +2547,37 @@ Each json-object contain the following:
     - "account_failed": whether failed operations are included in the
                          latency and last access statistics
                          (json-bool)
+    - "timed_stats": A json-array containing statistics collected in
+                     specific intervals, with the following members:
+        - "interval_length": interval used for calculating the
+                             statistics, in seconds (json-int)
+        - "min_rd_latency_ns": minimum latency of read operations in
+                               the defined interval, in nanoseconds
+                               (json-int)
+        - "min_wr_latency_ns": minimum latency of write operations in
+                               the defined interval, in nanoseconds
+                               (json-int)
+        - "min_flush_latency_ns": minimum latency of flush operations
+                                  in the defined interval, in
+                                  nanoseconds (json-int)
+        - "max_rd_latency_ns": maximum latency of read operations in
+                               the defined interval, in nanoseconds
+                               (json-int)
+        - "max_wr_latency_ns": maximum latency of write operations in
+                               the defined interval, in nanoseconds
+                               (json-int)
+        - "max_flush_latency_ns": maximum latency of flush operations
+                                  in the defined interval, in
+                                  nanoseconds (json-int)
+        - "avg_rd_latency_ns": average latency of read operations in
+                               the defined interval, in nanoseconds
+                               (json-int)
+        - "avg_wr_latency_ns": average latency of write operations in
+                               the defined interval, in nanoseconds
+                               (json-int)
+        - "avg_flush_latency_ns": average latency of flush operations
+                                  in the defined interval, in
+                                  nanoseconds (json-int)
 - "parent": Contains recursively the statistics of the underlying
             protocol (e.g. the host file for a qcow2 image). If there is
             no underlying protocol, this field is omitted
-- 
2.5.0