From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:33086) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bVjJp-00080f-3t for qemu-devel@nongnu.org; Fri, 05 Aug 2016 13:54:02 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1bVjJk-00071S-Sb for qemu-devel@nongnu.org; Fri, 05 Aug 2016 13:54:00 -0400 Received: from mx1.redhat.com ([209.132.183.28]:34586) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bVjJk-00071B-KY for qemu-devel@nongnu.org; Fri, 05 Aug 2016 13:53:56 -0400 Date: Fri, 5 Aug 2016 18:53:49 +0100 From: "Dr. David Alan Gilbert" Message-ID: <20160805175348.GJ2022@work-vm> References: <1470227172-13704-1-git-send-email-zhang.zhanghailiang@huawei.com> <1470227172-13704-13-git-send-email-zhang.zhanghailiang@huawei.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1470227172-13704-13-git-send-email-zhang.zhanghailiang@huawei.com> Subject: Re: [Qemu-devel] [PATCH COLO-Frame v18 12/34] COLO: Load VMState into buffer before restore it List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: zhanghailiang , berrange@redhat.com Cc: qemu-devel@nongnu.org, amit.shah@redhat.com, quintela@redhat.com, peter.huangpeng@huawei.com, eddie.dong@intel.com, wency@cn.fujitsu.com, lizhijian@cn.fujitsu.com, zhangchen.fnst@cn.fujitsu.com, xiecl.fnst@cn.fujitsu.com, Gonglei * zhanghailiang (zhang.zhanghailiang@huawei.com) wrote: > We should not destroy the state of SVM (Secondary VM) until we receive > the whole state from the PVM (Primary VM), in case the primary fails in > the middle of sending the state, so, here we cache the device state in > Secondary before restore it. > > Besides, we should call qemu_system_reset() before load VM state, > which can ensure the data is intact. > > Signed-off-by: zhanghailiang > Signed-off-by: Li Zhijian > Signed-off-by: Gonglei > Reviewed-by: Dr. David Alan Gilbert > Cc: Dr. David Alan Gilbert > --- > v17: > - Replace the old buffer API with the new channel buffer API. > v16: > - Rename colo_get_cmd_value() to colo_receive_mesage_value(); > v13: > - Fix the define of colo_get_cmd_value() to use 'Error **errp' instead of > return value. > v12: > - Use the new helper colo_get_cmd_value() instead of colo_ctl_get() > --- > migration/colo.c | 70 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-- > 1 file changed, 68 insertions(+), 2 deletions(-) > > diff --git a/migration/colo.c b/migration/colo.c > index 0401629..8fa2104 100644 > --- a/migration/colo.c > +++ b/migration/colo.c > @@ -115,6 +115,28 @@ static void colo_receive_check_message(QEMUFile *f, COLOMessage expect_msg, > } > } > > +static uint64_t colo_receive_message_value(QEMUFile *f, uint32_t expect_msg, > + Error **errp) > +{ > + Error *local_err = NULL; > + uint64_t value; > + int ret; > + > + colo_receive_check_message(f, expect_msg, &local_err); > + if (local_err) { > + error_propagate(errp, local_err); > + return 0; > + } > + > + value = qemu_get_be64(f); > + ret = qemu_file_get_error(f); > + if (ret < 0) { > + error_setg_errno(errp, -ret, "Failed to get value for COLO message: %s", > + COLOMessage_lookup[expect_msg]); > + } > + return value; > +} > + > static int colo_do_checkpoint_transaction(MigrationState *s, > QIOChannelBuffer *bioc, > QEMUFile *fb) > @@ -286,6 +308,10 @@ static void colo_wait_handle_message(QEMUFile *f, int *checkpoint_request, > void *colo_process_incoming_thread(void *opaque) > { > MigrationIncomingState *mis = opaque; > + QEMUFile *fb = NULL; > + QIOChannelBuffer *bioc = NULL; /* Cache incoming device state */ > + uint64_t total_size; > + uint64_t value; > Error *local_err = NULL; > int ret; > > @@ -310,6 +336,10 @@ void *colo_process_incoming_thread(void *opaque) > goto out; > } > > + bioc = qio_channel_buffer_new(COLO_BUFFER_BASE_SIZE); > + fb = qemu_fopen_channel_input(QIO_CHANNEL(bioc)); > + object_unref(OBJECT(bioc)); > + > colo_send_message(mis->to_src_file, COLO_MESSAGE_CHECKPOINT_READY, > &local_err); > if (local_err) { > @@ -337,7 +367,30 @@ void *colo_process_incoming_thread(void *opaque) > goto out; > } > > - /* TODO: read migration data into colo buffer */ > + value = colo_receive_message_value(mis->from_src_file, > + COLO_MESSAGE_VMSTATE_SIZE, &local_err); > + if (local_err) { > + goto out; > + } > + > + /* > + * Read VM device state data into channel buffer, > + * It's better to re-use the memory allocated. > + * Here we need to handle the channel buffer directly. > + */ > + if (value > bioc->capacity) { > + bioc->capacity = value; > + bioc->data = g_realloc(bioc->data, bioc->capacity); > + } Dan: Are you ok with that, or should there be some io/channel-buffer.c call to do it? Dave > + total_size = qemu_get_buffer(mis->from_src_file, bioc->data, value); > + if (total_size != value) { > + error_report("Got %lu VMState data, less than expected %lu", > + total_size, value); > + ret = -EINVAL; > + goto out; > + } > + bioc->usage = total_size; > + qio_channel_io_seek(QIO_CHANNEL(bioc), 0, 0, NULL); > > colo_send_message(mis->to_src_file, COLO_MESSAGE_VMSTATE_RECEIVED, > &local_err); > @@ -345,7 +398,16 @@ void *colo_process_incoming_thread(void *opaque) > goto out; > } > > - /* TODO: load vm state */ > + qemu_mutex_lock_iothread(); > + qemu_system_reset(VMRESET_SILENT); > + if (qemu_loadvm_state(fb) < 0) { > + error_report("COLO: loadvm failed"); > + qemu_mutex_unlock_iothread(); > + goto out; > + } > + qemu_mutex_unlock_iothread(); > + > + /* TODO: flush vm state */ > > colo_send_message(mis->to_src_file, COLO_MESSAGE_VMSTATE_LOADED, > &local_err); > @@ -360,6 +422,10 @@ out: > error_report_err(local_err); > } > > + if (fb) { > + qemu_fclose(fb); > + } > + > qemu_mutex_lock_iothread(); > colo_release_ram_cache(); > qemu_mutex_unlock_iothread(); > -- > 1.8.3.1 > > -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK