From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:33407) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fBLXf-0000he-Fz for qemu-devel@nongnu.org; Wed, 25 Apr 2018 10:37:08 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fBLXe-0008BG-Ii for qemu-devel@nongnu.org; Wed, 25 Apr 2018 10:37:07 -0400 Received: from mail-pf0-x241.google.com ([2607:f8b0:400e:c00::241]:43569) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1fBLXe-0008Al-DT for qemu-devel@nongnu.org; Wed, 25 Apr 2018 10:37:06 -0400 Received: by mail-pf0-x241.google.com with SMTP id j11so15393712pff.10 for ; Wed, 25 Apr 2018 07:37:06 -0700 (PDT) From: Lidong Chen Date: Wed, 25 Apr 2018 22:35:34 +0800 Message-Id: <1524666934-8064-6-git-send-email-lidongchen@tencent.com> In-Reply-To: <1524666934-8064-1-git-send-email-lidongchen@tencent.com> References: <1524666934-8064-1-git-send-email-lidongchen@tencent.com> Subject: [Qemu-devel] [PATCH v2 5/5] migration: Stop rdma yielding during incoming postcopy List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: quintela@redhat.com, dgilbert@redhat.com Cc: qemu-devel@nongnu.org, galsha@mellanox.com, aviadye@mellanox.com, licq@mellanox.com, adido@mellanox.com, Lidong Chen During incoming postcopy, the destination qemu will invoke qemu_rdma_wait_comp_channel in a seprate thread. So does not use rdma yield, and poll the completion channel fd instead. Signed-off-by: Lidong Chen --- migration/rdma.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/migration/rdma.c b/migration/rdma.c index 0652224..4ba9fe2 100644 --- a/migration/rdma.c +++ b/migration/rdma.c @@ -1490,11 +1490,13 @@ static int qemu_rdma_wait_comp_channel(RDMAContext *rdma) * Coroutine doesn't start until migration_fd_process_incoming() * so don't yield unless we know we're running inside of a coroutine. */ - if (rdma->migration_started_on_destination) { + if (rdma->migration_started_on_destination && + migration_incoming_get_current()->state == MIGRATION_STATUS_ACTIVE) { yield_until_fd_readable(rdma->comp_channel->fd); } else { /* This is the source side, we're in a separate thread * or destination prior to migration_fd_process_incoming() + * after postcopy, the destination also in a seprate thread. * we can't yield; so we have to poll the fd. * But we need to be able to handle 'cancel' or an error * without hanging forever. -- 1.8.3.1