Document every single live=true use cases on why it should be done in the
live manner. Also document on the parameter so new precopy cases should
always use live=off unless with explicit reasonings.
Cc: Thomas Huth <th...@redhat.com>
Cc: Juan Quintela <quint...@redhat.com>
Cc: Daniel P. Berrangé <berra...@redhat.com>
Signed-off-by: Peter Xu <pet...@redhat.com>
---
tests/qtest/migration-test.c | 37 ++++++++++++++++++++++++++++++------
1 file changed, 31 insertions(+), 6 deletions(-)
diff --git a/tests/qtest/migration-test.c b/tests/qtest/migration-test.c
index 298291f01c..d2cd71e6cf 100644
--- a/tests/qtest/migration-test.c
+++ b/tests/qtest/migration-test.c
@@ -682,7 +682,14 @@ typedef struct {
*/
unsigned int iterations;
- /* Optional: whether the guest CPUs should be running during migration */
+ /*
+ * Optional: whether the guest CPUs should be running during a precopy
+ * migration test. We used to always run with live but it took much
+ * longer so we reduced live tests to only the ones that have solid
+ * reason to be tested live-only. For each of the new test cases for
+ * precopy please provide justifications to use live explicitly (please
+ * refer to existing ones with live=true), or use live=off by default.
+ */
bool live;
/* Postcopy specific fields */
@@ -1592,7 +1599,10 @@ static void test_precopy_unix_plain(void)
MigrateCommon args = {
.listen_uri = uri,
.connect_uri = uri,
-
+ /*
+ * The simplest use case of precopy, covering smoke tests of
+ * get-dirty-log dirty tracking.
+ */
.live = true,
};
@@ -1609,7 +1619,10 @@ static void test_precopy_unix_dirty_ring(void)
},
.listen_uri = uri,
.connect_uri = uri,
-
+ /*
+ * Besides the precopy/unix basic test, cover dirty ring interface
+ * rather than get-dirty-log.
+ */
.live = true,
};
@@ -1721,10 +1734,12 @@ static void test_precopy_unix_xbzrle(void)
MigrateCommon args = {
.connect_uri = uri,
.listen_uri = uri,
-
.start_hook = test_migrate_xbzrle_start,
-
.iterations = 2,
+ /*
+ * XBZRLE needs pages to be modified when doing the 2nd+ round
+ * iteration to have real data pushed to the stream.
+ */
.live = true,
};
@@ -1743,6 +1758,11 @@ static void test_precopy_unix_compress(void)
* the previous iteration.
*/
.iterations = 2,
+ /*
+ * We make sure the compressor can always work well even if guest
+ * memory is changing. See commit 34ab9e9743 where we used to fix
+ * a bug when only trigger-able with guest memory changing.
+ */
.live = true,
};
@@ -1761,6 +1781,7 @@ static void test_precopy_unix_compress_nowait(void)
* the previous iteration.
*/
.iterations = 2,
+ /* Same reason for the wait version of precopy compress test */
.live = true,
};
@@ -2170,7 +2191,11 @@ static void test_multifd_tcp_none(void)
MigrateCommon args = {
.listen_uri = "defer",
.start_hook = test_migrate_precopy_tcp_multifd_start,
-
+ /*
+ * Multifd is more complicated than most of the features, it
+ * directly takes guest page buffers when sending, make sure
+ * everything will work alright even if guest page is changing.
+ */
.live = true,
};
test_precopy_common(&args);
--
2.40.1
