From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-lj1-f195.google.com (mail-lj1-f195.google.com [209.85.208.195]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by dev.tarantool.org (Postfix) with ESMTPS id A3ADE430407 for ; Tue, 18 Aug 2020 00:54:03 +0300 (MSK) Received: by mail-lj1-f195.google.com with SMTP id t6so19140114ljk.9 for ; Mon, 17 Aug 2020 14:54:03 -0700 (PDT) Date: Tue, 18 Aug 2020 00:54:00 +0300 From: Cyrill Gorcunov Message-ID: <20200817215400.GK2074@grain> References: <20200817133918.875558-1-gorcunov@gmail.com> <78aa1ff6-34c1-767c-7f50-21c93236d8e7@tarantool.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <78aa1ff6-34c1-767c-7f50-21c93236d8e7@tarantool.org> Subject: Re: [Tarantool-patches] [PATCH v8 0/9] qsync: write CONFIRM/ROLLBACK without txn engine List-Id: Tarantool development patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Vladislav Shpilevoy Cc: tml On Mon, Aug 17, 2020 at 11:24:40PM +0200, Vladislav Shpilevoy wrote: > process_synchro_row() is still present on the branch. Sigh. Overlooked, thanks! I've force-pushed the branch, just in case here is a commit on top. We can simply drop this helper it harmless. --- diff --git a/src/box/applier.cc b/src/box/applier.cc index 83f6da461..5d7c35ca6 100644 --- a/src/box/applier.cc +++ b/src/box/applier.cc @@ -270,45 +270,11 @@ process_nop(struct request *request) return txn_commit_stmt(txn, request); } -/* - * CONFIRM/ROLLBACK rows aren't dml requests and require special - * handling: instead of performing some operations on spaces, - * processing these requests requires txn_limbo to either confirm - * or rollback some of its entries. - */ -static int -process_synchro_row(struct request *request) -{ - assert(iproto_type_is_synchro_request(request->header->type)); - struct txn *txn = in_txn(); - - struct synchro_request syn_req; - if (xrow_decode_synchro(request->header, &syn_req) != 0) - return -1; - assert(txn->n_applier_rows == 0); - /* - * This is not really a transaction. It just uses txn API - * to put the data into WAL. And obviously it should not - * go to the limbo and block on the very same sync - * transaction which it tries to confirm now. - */ - txn_set_flag(txn, TXN_FORCE_ASYNC); - - if (txn_begin_stmt(txn, NULL) != 0) - return -1; - if (txn_commit_stmt(txn, request) != 0) - return -1; - return txn_limbo_process(&txn_limbo, &syn_req); -} - static int apply_row(struct xrow_header *row) { struct request request; - if (iproto_type_is_synchro_request(row->type)) { - request.header = row; - return process_synchro_row(&request); - } + assert(!iproto_type_is_synchro_request(row->type)); if (xrow_decode_dml(row, &request, dml_request_key_map(row->type)) != 0) return -1; if (request.type == IPROTO_NOP)