Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

# gpg: Signature made Fri Jun 12 15:57:47 2015 BST using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"

* remotes/stefanha/tags/block-pull-request:
  qemu-iotests: expand test 093 to support group throttling
  throttle: Update throttle infrastructure copyright
  throttle: add the name of the ThrottleGroup to BlockDeviceInfo
  throttle: acquire the ThrottleGroup lock in bdrv_swap()
  throttle: Add throttle group support
  throttle: Add throttle group infrastructure tests
  throttle: Add throttle group infrastructure
  throttle: Extract timers from ThrottleState into a separate structure
  raw-posix: Fix .bdrv_co_get_block_status() for unaligned image size
  Revert "iothread: release iothread around aio_poll"

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

3 files changed, 190 insertions, 85 deletions

diff --git a/tests/qemu-iotests/093 b/tests/qemu-iotests/093
index b9096a55d4..c0e9e2b0b5 100755
--- a/tests/qemu-iotests/093
+++ b/tests/qemu-iotests/093
@@ -3,6 +3,7 @@
 # Tests for IO throttling
 #
 # Copyright (C) 2015 Red Hat, Inc.
+# Copyright (C) 2015 Igalia, S.L.
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -22,6 +23,7 @@ import iotests
 
 class ThrottleTestCase(iotests.QMPTestCase):
     test_img = "null-aio://"
+    max_drives = 3
 
     def blockstats(self, device):
         result = self.vm.qmp("query-blockstats")
@@ -32,26 +34,31 @@ class ThrottleTestCase(iotests.QMPTestCase):
         raise Exception("Device not found for blockstats: %s" % device)
 
     def setUp(self):
-        self.vm = iotests.VM().add_drive(self.test_img)
+        self.vm = iotests.VM()
+        for i in range(0, self.max_drives):
+            self.vm.add_drive(self.test_img)
         self.vm.launch()
 
     def tearDown(self):
         self.vm.shutdown()
 
-    def do_test_throttle(self, seconds, params):
+    def do_test_throttle(self, ndrives, seconds, params):
         def check_limit(limit, num):
             # IO throttling algorithm is discrete, allow 10% error so the test
             # is more robust
             return limit == 0 or \
-                   (num < seconds * limit * 1.1
-                   and num > seconds * limit * 0.9)
+                   (num < seconds * limit * 1.1 / ndrives
+                   and num > seconds * limit * 0.9 / ndrives)
 
         nsec_per_sec = 1000000000
 
-        params['device'] = 'drive0'
+        params['group'] = 'test'
 
-        result = self.vm.qmp("block_set_io_throttle", conv_keys=False, **params)
-        self.assert_qmp(result, 'return', {})
+        # Set the I/O throttling parameters to all drives
+        for i in range(0, ndrives):
+            params['device'] = 'drive%d' % i
+            result = self.vm.qmp("block_set_io_throttle", conv_keys=False, **params)
+            self.assert_qmp(result, 'return', {})
 
         # Set vm clock to a known value
         ns = seconds * nsec_per_sec
@@ -66,32 +73,60 @@ class ThrottleTestCase(iotests.QMPTestCase):
                     params['iops'] / 2,
                     params['iops_rd'])
         rd_nr *= seconds * 2
+        rd_nr /= ndrives
         wr_nr = max(params['bps'] / rq_size / 2,
                     params['bps_wr'] / rq_size,
                     params['iops'] / 2,
                     params['iops_wr'])
         wr_nr *= seconds * 2
+        wr_nr /= ndrives
+
+        # Send I/O requests to all drives
         for i in range(rd_nr):
-            self.vm.hmp_qemu_io("drive0", "aio_read %d %d" % (i * rq_size, rq_size))
-        for i in range(wr_nr):
-            self.vm.hmp_qemu_io("drive0", "aio_write %d %d" % (i * rq_size, rq_size))
+            for drive in range(0, ndrives):
+                self.vm.hmp_qemu_io("drive%d" % drive, "aio_read %d %d" %
+                                    (i * rq_size, rq_size))
 
-        start_rd_bytes, start_rd_iops, start_wr_bytes, start_wr_iops = self.blockstats('drive0')
+        for i in range(wr_nr):
+            for drive in range(0, ndrives):
+                self.vm.hmp_qemu_io("drive%d" % drive, "aio_write %d %d" %
+                                    (i * rq_size, rq_size))
+
+        # We'll store the I/O stats for each drive in these arrays
+        start_rd_bytes = [0] * ndrives
+        start_rd_iops  = [0] * ndrives
+        start_wr_bytes = [0] * ndrives
+        start_wr_iops  = [0] * ndrives
+        end_rd_bytes   = [0] * ndrives
+        end_rd_iops    = [0] * ndrives
+        end_wr_bytes   = [0] * ndrives
+        end_wr_iops    = [0] * ndrives
+
+        # Read the stats before advancing the clock
+        for i in range(0, ndrives):
+            start_rd_bytes[i], start_rd_iops[i], start_wr_bytes[i], \
+                start_wr_iops[i] = self.blockstats('drive%d' % i)
 
         self.vm.qtest("clock_step %d" % ns)
-        end_rd_bytes, end_rd_iops, end_wr_bytes, end_wr_iops = self.blockstats('drive0')
-
-        rd_bytes = end_rd_bytes - start_rd_bytes
-        rd_iops = end_rd_iops - start_rd_iops
-        wr_bytes = end_wr_bytes - start_wr_bytes
-        wr_iops = end_wr_iops - start_wr_iops
 
-        self.assertTrue(check_limit(params['bps'], rd_bytes + wr_bytes))
-        self.assertTrue(check_limit(params['bps_rd'], rd_bytes))
-        self.assertTrue(check_limit(params['bps_wr'], wr_bytes))
-        self.assertTrue(check_limit(params['iops'], rd_iops + wr_iops))
-        self.assertTrue(check_limit(params['iops_rd'], rd_iops))
-        self.assertTrue(check_limit(params['iops_wr'], wr_iops))
+        # Read the stats after advancing the clock
+        for i in range(0, ndrives):
+            end_rd_bytes[i], end_rd_iops[i], end_wr_bytes[i], \
+                end_wr_iops[i] = self.blockstats('drive%d' % i)
+
+        # Check that the I/O is within the limits and evenly distributed
+        for i in range(0, ndrives):
+            rd_bytes = end_rd_bytes[i] - start_rd_bytes[i]
+            rd_iops = end_rd_iops[i] - start_rd_iops[i]
+            wr_bytes = end_wr_bytes[i] - start_wr_bytes[i]
+            wr_iops = end_wr_iops[i] - start_wr_iops[i]
+
+            self.assertTrue(check_limit(params['bps'], rd_bytes + wr_bytes))
+            self.assertTrue(check_limit(params['bps_rd'], rd_bytes))
+            self.assertTrue(check_limit(params['bps_wr'], wr_bytes))
+            self.assertTrue(check_limit(params['iops'], rd_iops + wr_iops))
+            self.assertTrue(check_limit(params['iops_rd'], rd_iops))
+            self.assertTrue(check_limit(params['iops_wr'], wr_iops))
 
     def test_all(self):
         params = {"bps": 4096,
@@ -101,11 +136,13 @@ class ThrottleTestCase(iotests.QMPTestCase):
                   "iops_rd": 10,
                   "iops_wr": 10,
                  }
-        # Pick each out of all possible params and test
-        for tk in params:
-            limits = dict([(k, 0) for k in params])
-            limits[tk] = params[tk]
-            self.do_test_throttle(5, limits)
+        # Repeat the test with different numbers of drives
+        for ndrives in range(1, self.max_drives + 1):
+            # Pick each out of all possible params and test
+            for tk in params:
+                limits = dict([(k, 0) for k in params])
+                limits[tk] = params[tk] * ndrives
+                self.do_test_throttle(ndrives, 5, limits)
 
 class ThrottleTestCoroutine(ThrottleTestCase):
     test_img = "null-co://"
diff --git a/tests/test-aio.c b/tests/test-aio.c
index 4b0cb45d31..a7cb5c9915 100644
--- a/tests/test-aio.c
+++ b/tests/test-aio.c
@@ -107,7 +107,6 @@ static void test_notify(void)
 
 typedef struct {
     QemuMutex start_lock;
-    EventNotifier notifier;
     bool thread_acquired;
 } AcquireTestData;
 
@@ -119,8 +118,6 @@ static void *test_acquire_thread(void *opaque)
     qemu_mutex_lock(&data->start_lock);
     qemu_mutex_unlock(&data->start_lock);
 
-    g_usleep(500000);
-    event_notifier_set(&data->notifier);
     aio_context_acquire(ctx);
     aio_context_release(ctx);
 
@@ -129,19 +126,20 @@ static void *test_acquire_thread(void *opaque)
     return NULL;
 }
 
-static void dummy_notifier_read(EventNotifier *n)
+static void dummy_notifier_read(EventNotifier *unused)
 {
-    event_notifier_test_and_clear(n);
+    g_assert(false); /* should never be invoked */
 }
 
 static void test_acquire(void)
 {
     QemuThread thread;
+    EventNotifier notifier;
     AcquireTestData data;
 
     /* Dummy event notifier ensures aio_poll() will block */
-    event_notifier_init(&data.notifier, false);
-    aio_set_event_notifier(ctx, &data.notifier, dummy_notifier_read);
+    event_notifier_init(&notifier, false);
+    aio_set_event_notifier(ctx, &notifier, dummy_notifier_read);
     g_assert(!aio_poll(ctx, false)); /* consume aio_notify() */
 
     qemu_mutex_init(&data.start_lock);
@@ -155,13 +153,12 @@ static void test_acquire(void)
     /* Block in aio_poll(), let other thread kick us and acquire context */
     aio_context_acquire(ctx);
     qemu_mutex_unlock(&data.start_lock); /* let the thread run */
-    g_assert(aio_poll(ctx, true));
-    g_assert(!data.thread_acquired);
+    g_assert(!aio_poll(ctx, true));
     aio_context_release(ctx);
 
     qemu_thread_join(&thread);
-    aio_set_event_notifier(ctx, &data.notifier, NULL);
-    event_notifier_cleanup(&data.notifier);
+    aio_set_event_notifier(ctx, &notifier, NULL);
+    event_notifier_cleanup(&notifier);
 
     g_assert(data.thread_acquired);
 }
diff --git a/tests/test-throttle.c b/tests/test-throttle.c
index d8ba415e43..016844546a 100644
--- a/tests/test-throttle.c
+++ b/tests/test-throttle.c
@@ -1,10 +1,12 @@
 /*
  * Throttle infrastructure tests
  *
- * Copyright Nodalink, SARL. 2013
+ * Copyright Nodalink, EURL. 2013-2014
+ * Copyright Igalia, S.L. 2015
  *
  * Authors:
- *  Benoît Canet     <benoit.canet@irqsave.net>
+ *  Benoît Canet     <benoit.canet@nodalink.com>
+ *  Alberto Garcia   <berto@igalia.com>
  *
  * This work is licensed under the terms of the GNU LGPL, version 2 or later.
  * See the COPYING.LIB file in the top-level directory.
@@ -15,11 +17,13 @@
 #include "block/aio.h"
 #include "qemu/throttle.h"
 #include "qemu/error-report.h"
+#include "block/throttle-groups.h"
 
 static AioContext     *ctx;
 static LeakyBucket    bkt;
 static ThrottleConfig cfg;
 static ThrottleState  ts;
+static ThrottleTimers tt;
 
 /* useful function */
 static bool double_cmp(double x, double y)
@@ -103,17 +107,19 @@ static void test_init(void)
 {
     int i;
 
-    /* fill the structure with crap */
+    /* fill the structures with crap */
     memset(&ts, 1, sizeof(ts));
+    memset(&tt, 1, sizeof(tt));
 
-    /* init the structure */
-    throttle_init(&ts, ctx, QEMU_CLOCK_VIRTUAL,
-                  read_timer_cb, write_timer_cb, &ts);
+    /* init structures */
+    throttle_init(&ts);
+    throttle_timers_init(&tt, ctx, QEMU_CLOCK_VIRTUAL,
+                         read_timer_cb, write_timer_cb, &ts);
 
     /* check initialized fields */
-    g_assert(ts.clock_type == QEMU_CLOCK_VIRTUAL);
-    g_assert(ts.timers[0]);
-    g_assert(ts.timers[1]);
+    g_assert(tt.clock_type == QEMU_CLOCK_VIRTUAL);
+    g_assert(tt.timers[0]);
+    g_assert(tt.timers[1]);
 
     /* check other fields where cleared */
     g_assert(!ts.previous_leak);
@@ -124,17 +130,18 @@ static void test_init(void)
         g_assert(!ts.cfg.buckets[i].level);
     }
 
-    throttle_destroy(&ts);
+    throttle_timers_destroy(&tt);
 }
 
 static void test_destroy(void)
 {
     int i;
-    throttle_init(&ts, ctx, QEMU_CLOCK_VIRTUAL,
-                  read_timer_cb, write_timer_cb, &ts);
-    throttle_destroy(&ts);
+    throttle_init(&ts);
+    throttle_timers_init(&tt, ctx, QEMU_CLOCK_VIRTUAL,
+                         read_timer_cb, write_timer_cb, &ts);
+    throttle_timers_destroy(&tt);
     for (i = 0; i < 2; i++) {
-        g_assert(!ts.timers[i]);
+        g_assert(!tt.timers[i]);
     }
 }
 
@@ -170,11 +177,12 @@ static void test_config_functions(void)
 
     orig_cfg.op_size = 1;
 
-    throttle_init(&ts, ctx, QEMU_CLOCK_VIRTUAL,
-                  read_timer_cb, write_timer_cb, &ts);
+    throttle_init(&ts);
+    throttle_timers_init(&tt, ctx, QEMU_CLOCK_VIRTUAL,
+                         read_timer_cb, write_timer_cb, &ts);
     /* structure reset by throttle_init previous_leak should be null */
     g_assert(!ts.previous_leak);
-    throttle_config(&ts, &orig_cfg);
+    throttle_config(&ts, &tt, &orig_cfg);
 
     /* has previous leak been initialized by throttle_config ? */
     g_assert(ts.previous_leak);
@@ -182,7 +190,7 @@ static void test_config_functions(void)
     /* get back the fixed configuration */
     throttle_get_config(&ts, &final_cfg);
 
-    throttle_destroy(&ts);
+    throttle_timers_destroy(&tt);
 
     g_assert(final_cfg.buckets[THROTTLE_BPS_TOTAL].avg == 153);
     g_assert(final_cfg.buckets[THROTTLE_BPS_READ].avg  == 56);
@@ -323,43 +331,47 @@ static void test_is_valid(void)
 
 static void test_have_timer(void)
 {
-    /* zero the structure */
+    /* zero structures */
     memset(&ts, 0, sizeof(ts));
+    memset(&tt, 0, sizeof(tt));
 
     /* no timer set should return false */
-    g_assert(!throttle_have_timer(&ts));
+    g_assert(!throttle_timers_are_initialized(&tt));
 
-    /* init the structure */
-    throttle_init(&ts, ctx, QEMU_CLOCK_VIRTUAL,
-                  read_timer_cb, write_timer_cb, &ts);
+    /* init structures */
+    throttle_init(&ts);
+    throttle_timers_init(&tt, ctx, QEMU_CLOCK_VIRTUAL,
+                         read_timer_cb, write_timer_cb, &ts);
 
     /* timer set by init should return true */
-    g_assert(throttle_have_timer(&ts));
+    g_assert(throttle_timers_are_initialized(&tt));
 
-    throttle_destroy(&ts);
+    throttle_timers_destroy(&tt);
 }
 
 static void test_detach_attach(void)
 {
-    /* zero the structure */
+    /* zero structures */
     memset(&ts, 0, sizeof(ts));
+    memset(&tt, 0, sizeof(tt));
 
     /* init the structure */
-    throttle_init(&ts, ctx, QEMU_CLOCK_VIRTUAL,
-                  read_timer_cb, write_timer_cb, &ts);
+    throttle_init(&ts);
+    throttle_timers_init(&tt, ctx, QEMU_CLOCK_VIRTUAL,
+                         read_timer_cb, write_timer_cb, &ts);
 
     /* timer set by init should return true */
-    g_assert(throttle_have_timer(&ts));
+    g_assert(throttle_timers_are_initialized(&tt));
 
     /* timer should no longer exist after detaching */
-    throttle_detach_aio_context(&ts);
-    g_assert(!throttle_have_timer(&ts));
+    throttle_timers_detach_aio_context(&tt);
+    g_assert(!throttle_timers_are_initialized(&tt));
 
     /* timer should exist again after attaching */
-    throttle_attach_aio_context(&ts, ctx);
-    g_assert(throttle_have_timer(&ts));
+    throttle_timers_attach_aio_context(&tt, ctx);
+    g_assert(throttle_timers_are_initialized(&tt));
 
-    throttle_destroy(&ts);
+    throttle_timers_destroy(&tt);
 }
 
 static bool do_test_accounting(bool is_ops, /* are we testing bps or ops */
@@ -387,9 +399,10 @@ static bool do_test_accounting(bool is_ops, /* are we testing bps or ops */
 
     cfg.op_size = op_size;
 
-    throttle_init(&ts, ctx, QEMU_CLOCK_VIRTUAL,
-                  read_timer_cb, write_timer_cb, &ts);
-    throttle_config(&ts, &cfg);
+    throttle_init(&ts);
+    throttle_timers_init(&tt, ctx, QEMU_CLOCK_VIRTUAL,
+                         read_timer_cb, write_timer_cb, &ts);
+    throttle_config(&ts, &tt, &cfg);
 
     /* account a read */
     throttle_account(&ts, false, size);
@@ -414,7 +427,7 @@ static bool do_test_accounting(bool is_ops, /* are we testing bps or ops */
         return false;
     }
 
-    throttle_destroy(&ts);
+    throttle_timers_destroy(&tt);
 
     return true;
 }
@@ -490,23 +503,80 @@ static void test_accounting(void)
                                 (64.0 / 13)));
 }
 
+static void test_groups(void)
+{
+    ThrottleConfig cfg1, cfg2;
+    BlockDriverState *bdrv1, *bdrv2, *bdrv3;
+
+    bdrv1 = bdrv_new();
+    bdrv2 = bdrv_new();
+    bdrv3 = bdrv_new();
+
+    g_assert(bdrv1->throttle_state == NULL);
+    g_assert(bdrv2->throttle_state == NULL);
+    g_assert(bdrv3->throttle_state == NULL);
+
+    throttle_group_register_bs(bdrv1, "bar");
+    throttle_group_register_bs(bdrv2, "foo");
+    throttle_group_register_bs(bdrv3, "bar");
+
+    g_assert(bdrv1->throttle_state != NULL);
+    g_assert(bdrv2->throttle_state != NULL);
+    g_assert(bdrv3->throttle_state != NULL);
+
+    g_assert(!strcmp(throttle_group_get_name(bdrv1), "bar"));
+    g_assert(!strcmp(throttle_group_get_name(bdrv2), "foo"));
+    g_assert(bdrv1->throttle_state == bdrv3->throttle_state);
+
+    /* Setting the config of a group member affects the whole group */
+    memset(&cfg1, 0, sizeof(cfg1));
+    cfg1.buckets[THROTTLE_BPS_READ].avg  = 500000;
+    cfg1.buckets[THROTTLE_BPS_WRITE].avg = 285000;
+    cfg1.buckets[THROTTLE_OPS_READ].avg  = 20000;
+    cfg1.buckets[THROTTLE_OPS_WRITE].avg = 12000;
+    throttle_group_config(bdrv1, &cfg1);
+
+    throttle_group_get_config(bdrv1, &cfg1);
+    throttle_group_get_config(bdrv3, &cfg2);
+    g_assert(!memcmp(&cfg1, &cfg2, sizeof(cfg1)));
+
+    cfg2.buckets[THROTTLE_BPS_READ].avg  = 4547;
+    cfg2.buckets[THROTTLE_BPS_WRITE].avg = 1349;
+    cfg2.buckets[THROTTLE_OPS_READ].avg  = 123;
+    cfg2.buckets[THROTTLE_OPS_WRITE].avg = 86;
+    throttle_group_config(bdrv3, &cfg1);
+
+    throttle_group_get_config(bdrv1, &cfg1);
+    throttle_group_get_config(bdrv3, &cfg2);
+    g_assert(!memcmp(&cfg1, &cfg2, sizeof(cfg1)));
+
+    throttle_group_unregister_bs(bdrv1);
+    throttle_group_unregister_bs(bdrv2);
+    throttle_group_unregister_bs(bdrv3);
+
+    g_assert(bdrv1->throttle_state == NULL);
+    g_assert(bdrv2->throttle_state == NULL);
+    g_assert(bdrv3->throttle_state == NULL);
+}
+
 int main(int argc, char **argv)
 {
-    GSource *src;
     Error *local_error = NULL;
 
-    init_clocks();
+    qemu_init_main_loop(&local_error);
+    ctx = qemu_get_aio_context();
 
-    ctx = aio_context_new(&local_error);
     if (!ctx) {
         error_report("Failed to create AIO Context: '%s'",
-                     error_get_pretty(local_error));
-        error_free(local_error);
+                     local_error ? error_get_pretty(local_error) :
+                     "Failed to initialize the QEMU main loop");
+        if (local_error) {
+            error_free(local_error);
+        }
         exit(1);
     }
-    src = aio_get_g_source(ctx);
-    g_source_attach(src, NULL);
-    g_source_unref(src);
+
+    bdrv_init();
 
     do {} while (g_main_context_iteration(NULL, false));
 
@@ -523,6 +593,7 @@ int main(int argc, char **argv)
     g_test_add_func("/throttle/config/is_valid",    test_is_valid);
     g_test_add_func("/throttle/config_functions",   test_config_functions);
     g_test_add_func("/throttle/accounting",         test_accounting);
+    g_test_add_func("/throttle/groups",             test_groups);
     return g_test_run();
 }