haraldwolff/qemu-patch-raspberry4
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity

migration: defer migration_end & blk_mig_cleanup

Browse source 
Because of the patch 3ea3b7fa9af067982f34b of kvm, which introduces a
lazy collapsing of small sptes into large sptes mechanism, now
migration_end() is a time consuming operation because it calls
memroy_global_dirty_log_stop(), which will trigger the dropping of small
sptes operation and takes about dozens of milliseconds, so call
migration_end() before all the vmsate data has already been transferred
to the destination will prolong VM downtime. This operation should be
deferred after all the data has been transferred to the destination.

blk_mig_cleanup() can be deferred too.

For a VM with 8G RAM, this patch can reduce the VM downtime about 30 ms.

Signed-off-by: Liang Li <liang.z.li@intel.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>al3
Reviewed-by: Amit Shah <amit.shah@redhat.com>al3
Signed-off-by: Juan Quintela <quintela@redhat.com>al3
This commit is contained in:
Liang Li 2015-11-02 15:37:00 +08:00 committed by Juan Quintela
parent 79cf9fad34
commit 94f5a43704
3 changed files with 6 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
migration/block.c
View file

@ -750,7 +750,6 @@ static int block_save_complete(QEMUFile *f, void *opaque)
qemu_put_be64(f, BLK_MIG_FLAG_EOS);
blk_mig_cleanup();
return 0;
}

13
migration/migration.c
View file

@ -613,12 +613,9 @@ static void migrate_fd_cleanup(void *opaque)
assert(s->state != MIGRATION_STATUS_ACTIVE);
if (s->state != MIGRATION_STATUS_COMPLETED) {
qemu_savevm_state_cancel();
if (s->state == MIGRATION_STATUS_CANCELLING) {
migrate_set_state(s, MIGRATION_STATUS_CANCELLING,
MIGRATION_STATUS_CANCELLED);
}
if (s->state == MIGRATION_STATUS_CANCELLING) {
migrate_set_state(s, MIGRATION_STATUS_CANCELLING,
MIGRATION_STATUS_CANCELLED);
}
notifier_list_notify(&migration_state_notifiers, s);
@ -1028,6 +1025,7 @@ static void *migration_thread(void *opaque)
int64_t initial_bytes = 0;
int64_t max_size = 0;
int64_t start_time = initial_time;
int64_t end_time;
bool old_vm_running = false;
rcu_register_thread();
@ -1089,10 +1087,11 @@ static void *migration_thread(void *opaque)
/* If we enabled cpu throttling for auto-converge, turn it off. */
cpu_throttle_stop();
end_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
qemu_mutex_lock_iothread();
qemu_savevm_state_cancel();
if (s->state == MIGRATION_STATUS_COMPLETED) {
int64_t end_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
uint64_t transferred_bytes = qemu_ftell(s->file);
s->total_time = end_time - s->total_time;
s->downtime = end_time - start_time;

1
migration/ram.c
View file

@ -1344,7 +1344,6 @@ static int ram_save_complete(QEMUFile *f, void *opaque)
rcu_read_unlock();
migration_end();
qemu_put_be64(f, RAM_SAVE_FLAG_EOS);
return 0;