xemu/tests/test-blockjob-txn.c

265 lines
6.7 KiB
C
Raw Normal View History

/*
* Blockjob transactions tests
*
* Copyright Red Hat, Inc. 2015
*
* Authors:
* Stefan Hajnoczi <stefanha@redhat.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.
*/
#include "qemu/osdep.h"
2016-03-14 08:01:28 +00:00
#include "qapi/error.h"
#include "qemu/main-loop.h"
#include "block/blockjob_int.h"
#include "sysemu/block-backend.h"
typedef struct {
BlockJob common;
unsigned int iterations;
bool use_timer;
int rc;
int *result;
} TestBlockJob;
static void test_block_job_complete(BlockJob *job, void *opaque)
{
BlockDriverState *bs = blk_bs(job->blk);
int rc = (intptr_t)opaque;
if (job_is_cancelled(&job->job)) {
rc = -ECANCELED;
}
block_job_completed(job, rc);
bdrv_unref(bs);
}
static void coroutine_fn test_block_job_run(void *opaque)
{
TestBlockJob *s = opaque;
BlockJob *job = &s->common;
while (s->iterations--) {
if (s->use_timer) {
block_job_sleep_ns(job, 0);
} else {
block_job_yield(job);
}
if (job_is_cancelled(&job->job)) {
break;
}
}
block_job_defer_to_main_loop(job, test_block_job_complete,
(void *)(intptr_t)s->rc);
}
typedef struct {
TestBlockJob *job;
int *result;
} TestBlockJobCBData;
static void test_block_job_cb(void *opaque, int ret)
{
TestBlockJobCBData *data = opaque;
if (!ret && job_is_cancelled(&data->job->common.job)) {
ret = -ECANCELED;
}
*data->result = ret;
g_free(data);
}
static const BlockJobDriver test_block_job_driver = {
.job_driver = {
.instance_size = sizeof(TestBlockJob),
.free = block_job_free,
},
.start = test_block_job_run,
};
/* Create a block job that completes with a given return code after a given
* number of event loop iterations. The return code is stored in the given
* result pointer.
*
* The event loop iterations can either be handled automatically with a 0 delay
* timer, or they can be stepped manually by entering the coroutine.
*/
static BlockJob *test_block_job_start(unsigned int iterations,
bool use_timer,
int rc, int *result, BlockJobTxn *txn)
{
BlockDriverState *bs;
TestBlockJob *s;
TestBlockJobCBData *data;
static unsigned counter;
char job_id[24];
data = g_new0(TestBlockJobCBData, 1);
bs = bdrv_open("null-co://", NULL, NULL, 0, &error_abort);
g_assert_nonnull(bs);
snprintf(job_id, sizeof(job_id), "job%u", counter++);
s = block_job_create(job_id, &test_block_job_driver, txn, bs,
0, BLK_PERM_ALL, 0, BLOCK_JOB_DEFAULT,
test_block_job_cb, data, &error_abort);
s->iterations = iterations;
s->use_timer = use_timer;
s->rc = rc;
s->result = result;
data->job = s;
data->result = result;
return &s->common;
}
static void test_single_job(int expected)
{
BlockJob *job;
BlockJobTxn *txn;
int result = -EINPROGRESS;
txn = block_job_txn_new();
job = test_block_job_start(1, true, expected, &result, txn);
block_job_start(job);
if (expected == -ECANCELED) {
block/mirror: change the semantic of 'force' of block-job-cancel When doing drive mirror to a low speed shared storage, if there was heavy BLK IO write workload in VM after the 'ready' event, drive mirror block job can't be canceled immediately, it would keep running until the heavy BLK IO workload stopped in the VM. Libvirt depends on the current block-job-cancel semantics, which is that when used without a flag after the 'ready' event, the command blocks until data is in sync. However, these semantics are awkward in other situations, for example, people may use drive mirror for realtime backups while still wanting to use block live migration. Libvirt cannot start a block live migration while another drive mirror is in progress, but the user would rather abandon the backup attempt as broken and proceed with the live migration than be stuck waiting for the current drive mirror backup to finish. The drive-mirror command already includes a 'force' flag, which libvirt does not use, although it documented the flag as only being useful to quit a job which is paused. However, since quitting a paused job has the same effect as abandoning a backup in a non-paused job (namely, the destination file is not in sync, and the command completes immediately), we can just improve the documentation to make the force flag obviously useful. Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Jeff Cody <jcody@redhat.com> Cc: Kevin Wolf <kwolf@redhat.com> Cc: Max Reitz <mreitz@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: John Snow <jsnow@redhat.com> Reported-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Liang Li <liliangleo@didichuxing.com> Signed-off-by: Jeff Cody <jcody@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2018-03-13 12:12:16 +00:00
block_job_cancel(job, false);
}
while (result == -EINPROGRESS) {
aio_poll(qemu_get_aio_context(), true);
}
g_assert_cmpint(result, ==, expected);
block_job_txn_unref(txn);
}
static void test_single_job_success(void)
{
test_single_job(0);
}
static void test_single_job_failure(void)
{
test_single_job(-EIO);
}
static void test_single_job_cancel(void)
{
test_single_job(-ECANCELED);
}
static void test_pair_jobs(int expected1, int expected2)
{
BlockJob *job1;
BlockJob *job2;
BlockJobTxn *txn;
int result1 = -EINPROGRESS;
int result2 = -EINPROGRESS;
txn = block_job_txn_new();
job1 = test_block_job_start(1, true, expected1, &result1, txn);
job2 = test_block_job_start(2, true, expected2, &result2, txn);
block_job_start(job1);
block_job_start(job2);
/* Release our reference now to trigger as many nice
* use-after-free bugs as possible.
*/
block_job_txn_unref(txn);
if (expected1 == -ECANCELED) {
block/mirror: change the semantic of 'force' of block-job-cancel When doing drive mirror to a low speed shared storage, if there was heavy BLK IO write workload in VM after the 'ready' event, drive mirror block job can't be canceled immediately, it would keep running until the heavy BLK IO workload stopped in the VM. Libvirt depends on the current block-job-cancel semantics, which is that when used without a flag after the 'ready' event, the command blocks until data is in sync. However, these semantics are awkward in other situations, for example, people may use drive mirror for realtime backups while still wanting to use block live migration. Libvirt cannot start a block live migration while another drive mirror is in progress, but the user would rather abandon the backup attempt as broken and proceed with the live migration than be stuck waiting for the current drive mirror backup to finish. The drive-mirror command already includes a 'force' flag, which libvirt does not use, although it documented the flag as only being useful to quit a job which is paused. However, since quitting a paused job has the same effect as abandoning a backup in a non-paused job (namely, the destination file is not in sync, and the command completes immediately), we can just improve the documentation to make the force flag obviously useful. Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Jeff Cody <jcody@redhat.com> Cc: Kevin Wolf <kwolf@redhat.com> Cc: Max Reitz <mreitz@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: John Snow <jsnow@redhat.com> Reported-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Liang Li <liliangleo@didichuxing.com> Signed-off-by: Jeff Cody <jcody@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2018-03-13 12:12:16 +00:00
block_job_cancel(job1, false);
}
if (expected2 == -ECANCELED) {
block/mirror: change the semantic of 'force' of block-job-cancel When doing drive mirror to a low speed shared storage, if there was heavy BLK IO write workload in VM after the 'ready' event, drive mirror block job can't be canceled immediately, it would keep running until the heavy BLK IO workload stopped in the VM. Libvirt depends on the current block-job-cancel semantics, which is that when used without a flag after the 'ready' event, the command blocks until data is in sync. However, these semantics are awkward in other situations, for example, people may use drive mirror for realtime backups while still wanting to use block live migration. Libvirt cannot start a block live migration while another drive mirror is in progress, but the user would rather abandon the backup attempt as broken and proceed with the live migration than be stuck waiting for the current drive mirror backup to finish. The drive-mirror command already includes a 'force' flag, which libvirt does not use, although it documented the flag as only being useful to quit a job which is paused. However, since quitting a paused job has the same effect as abandoning a backup in a non-paused job (namely, the destination file is not in sync, and the command completes immediately), we can just improve the documentation to make the force flag obviously useful. Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Jeff Cody <jcody@redhat.com> Cc: Kevin Wolf <kwolf@redhat.com> Cc: Max Reitz <mreitz@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: John Snow <jsnow@redhat.com> Reported-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Liang Li <liliangleo@didichuxing.com> Signed-off-by: Jeff Cody <jcody@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2018-03-13 12:12:16 +00:00
block_job_cancel(job2, false);
}
while (result1 == -EINPROGRESS || result2 == -EINPROGRESS) {
aio_poll(qemu_get_aio_context(), true);
}
/* Failure or cancellation of one job cancels the other job */
if (expected1 != 0) {
expected2 = -ECANCELED;
} else if (expected2 != 0) {
expected1 = -ECANCELED;
}
g_assert_cmpint(result1, ==, expected1);
g_assert_cmpint(result2, ==, expected2);
}
static void test_pair_jobs_success(void)
{
test_pair_jobs(0, 0);
}
static void test_pair_jobs_failure(void)
{
/* Test both orderings. The two jobs run for a different number of
* iterations so the code path is different depending on which job fails
* first.
*/
test_pair_jobs(-EIO, 0);
test_pair_jobs(0, -EIO);
}
static void test_pair_jobs_cancel(void)
{
test_pair_jobs(-ECANCELED, 0);
test_pair_jobs(0, -ECANCELED);
}
static void test_pair_jobs_fail_cancel_race(void)
{
BlockJob *job1;
BlockJob *job2;
BlockJobTxn *txn;
int result1 = -EINPROGRESS;
int result2 = -EINPROGRESS;
txn = block_job_txn_new();
job1 = test_block_job_start(1, true, -ECANCELED, &result1, txn);
job2 = test_block_job_start(2, false, 0, &result2, txn);
block_job_start(job1);
block_job_start(job2);
block/mirror: change the semantic of 'force' of block-job-cancel When doing drive mirror to a low speed shared storage, if there was heavy BLK IO write workload in VM after the 'ready' event, drive mirror block job can't be canceled immediately, it would keep running until the heavy BLK IO workload stopped in the VM. Libvirt depends on the current block-job-cancel semantics, which is that when used without a flag after the 'ready' event, the command blocks until data is in sync. However, these semantics are awkward in other situations, for example, people may use drive mirror for realtime backups while still wanting to use block live migration. Libvirt cannot start a block live migration while another drive mirror is in progress, but the user would rather abandon the backup attempt as broken and proceed with the live migration than be stuck waiting for the current drive mirror backup to finish. The drive-mirror command already includes a 'force' flag, which libvirt does not use, although it documented the flag as only being useful to quit a job which is paused. However, since quitting a paused job has the same effect as abandoning a backup in a non-paused job (namely, the destination file is not in sync, and the command completes immediately), we can just improve the documentation to make the force flag obviously useful. Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Jeff Cody <jcody@redhat.com> Cc: Kevin Wolf <kwolf@redhat.com> Cc: Max Reitz <mreitz@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: John Snow <jsnow@redhat.com> Reported-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Huaitong Han <huanhuaitong@didichuxing.com> Signed-off-by: Liang Li <liliangleo@didichuxing.com> Signed-off-by: Jeff Cody <jcody@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2018-03-13 12:12:16 +00:00
block_job_cancel(job1, false);
/* Now make job2 finish before the main loop kicks jobs. This simulates
* the race between a pending kick and another job completing.
*/
block_job_enter(job2);
block_job_enter(job2);
while (result1 == -EINPROGRESS || result2 == -EINPROGRESS) {
aio_poll(qemu_get_aio_context(), true);
}
g_assert_cmpint(result1, ==, -ECANCELED);
g_assert_cmpint(result2, ==, -ECANCELED);
block_job_txn_unref(txn);
}
int main(int argc, char **argv)
{
qemu_init_main_loop(&error_abort);
bdrv_init();
g_test_init(&argc, &argv, NULL);
g_test_add_func("/single/success", test_single_job_success);
g_test_add_func("/single/failure", test_single_job_failure);
g_test_add_func("/single/cancel", test_single_job_cancel);
g_test_add_func("/pair/success", test_pair_jobs_success);
g_test_add_func("/pair/failure", test_pair_jobs_failure);
g_test_add_func("/pair/cancel", test_pair_jobs_cancel);
g_test_add_func("/pair/fail-cancel-race", test_pair_jobs_fail_cancel_race);
return g_test_run();
}