From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp51.i.mail.ru (smtp51.i.mail.ru [94.100.177.111]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by dev.tarantool.org (Postfix) with ESMTPS id 458804696C3 for ; Thu, 9 Apr 2020 14:26:01 +0300 (MSK) Date: Thu, 9 Apr 2020 11:26:00 +0000 From: Nikita Pettik Message-ID: <20200409112559.GC3864@tarantool.org> References: <20200409081822.GA21437@atlas> <20200409105556.GA3864@tarantool.org> <20200409110756.GC29198@atlas> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20200409110756.GC29198@atlas> Subject: Re: [Tarantool-patches] [PATCH 1/2] vinyl: init all vars before cleanup in vy_lsm_split_range() List-Id: Tarantool development patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Konstantin Osipov , tarantool-patches@dev.tarantool.org, v.shpilevoy@tarantool.org On 09 Apr 14:07, Konstantin Osipov wrote: > * Nikita Pettik [20/04/09 14:00]: > > On 09 Apr 11:18, Konstantin Osipov wrote: > > > * Nikita Pettik [20/04/09 00:39]: > > > > > > > + struct errinj *inj = errinj(ERRINJ_VY_MAX_TUPLE_SIZE, ERRINJ_INT); > > > > + if (inj != NULL && inj->iparam >= 0) { > > > > + if (inj->iparam-- == 0) > > > > + total_size = vy_max_tuple_size + 1; > > > > + } > > > > > > shouldn't this be under #ifdef > > > > Looks like not: most of other errinj() calls are not wrapped in > > #ifdef NDEBUG macro (in case of debug build errinj() always > > returns NULL). > > In most cases the compiler can optimize these away, but the rules > are complicated. > Regarding other uses of errinj() one should take into > consideration whether it's a hot path or not, whether the code > in which injection is used is small/simple or not, etc. > > I'd say as a rule of thumb anything using errinj and is multiline > is better off using #ifdef. Ok, as you wish, I don't mind this change. Diff: diff --git a/src/box/vy_stmt.c b/src/box/vy_stmt.c index d2469961f..e822137ec 100644 --- a/src/box/vy_stmt.c +++ b/src/box/vy_stmt.c @@ -134,11 +134,13 @@ vy_stmt_alloc(struct tuple_format *format, uint32_t bsize) { uint32_t total_size = sizeof(struct vy_stmt) + format->field_map_size + bsize; +#ifndef NDEBUG struct errinj *inj = errinj(ERRINJ_VY_MAX_TUPLE_SIZE, ERRINJ_INT); if (inj != NULL && inj->iparam >= 0) { if (inj->iparam-- == 0) total_size = vy_max_tuple_size + 1; } +#endif if (unlikely(total_size > vy_max_tuple_size)) { diag_set(ClientError, ER_VINYL_MAX_TUPLE_SIZE, (unsigned) total_size); > > > The patch is LGTM, please solicit another review. > > -- > Konstantin Osipov, Moscow, Russia > https://scylladb.com