Tarantool development patches archive
 help / color / mirror / Atom feed
From: Aleksandr Lyapunov <alyapunov@tarantool.org>
To: tarantool-patches@dev.tarantool.org
Subject: [Tarantool-patches] [PATCH v3 07/13] txm: save does_require_old_tuple flag in txn_stmt
Date: Wed, 15 Jul 2020 16:55:30 +0300	[thread overview]
Message-ID: <1594821336-14468-8-git-send-email-alyapunov@tarantool.org> (raw)
In-Reply-To: <1594821336-14468-1-git-send-email-alyapunov@tarantool.org>

That flag is needed for transactional conflict manager - if any
other transaction replaces old_tuple before current one and the flag
is set - the current transaction will be aborted.
For example REPLACE just replaces a key, no matter what tuple
lays in the index and thus does_require_old_tuple = false.
In contrast, UPDATE makes new tuple using old_tuple and thus
the statement will require old_tuple (does_require_old_tuple = true).
INSERT also does_require_old_tuple = true because it requires
old_tuple to be NULL.

Part of #4897
---
 src/box/memtx_space.c | 17 +++++++++++++++++
 src/box/txn.c         |  3 +++
 src/box/txn.h         | 13 +++++++++++++
 3 files changed, 33 insertions(+)

diff --git a/src/box/memtx_space.c b/src/box/memtx_space.c
index 8452ab4..e48ed3a 100644
--- a/src/box/memtx_space.c
+++ b/src/box/memtx_space.c
@@ -316,6 +316,10 @@ memtx_space_execute_replace(struct space *space, struct txn *txn,
 	if (stmt->new_tuple == NULL)
 		return -1;
 	tuple_ref(stmt->new_tuple);
+
+	if (mode == DUP_INSERT)
+		stmt->does_require_old_tuple = true;
+
 	if (memtx_space->replace(space, NULL, stmt->new_tuple,
 				 mode, &stmt->old_tuple) != 0)
 		return -1;
@@ -342,6 +346,13 @@ memtx_space_execute_delete(struct space *space, struct txn *txn,
 	struct tuple *old_tuple;
 	if (index_get(pk, key, part_count, &old_tuple) != 0)
 		return -1;
+
+	/*
+	 * We have to delete exactly old_tuple just because we return it as
+	 * a result.
+	 */
+	stmt->does_require_old_tuple = true;
+
 	if (old_tuple != NULL &&
 	    memtx_space->replace(space, old_tuple, NULL,
 				 DUP_REPLACE_OR_INSERT, &stmt->old_tuple) != 0)
@@ -390,6 +401,9 @@ memtx_space_execute_update(struct space *space, struct txn *txn,
 	if (stmt->new_tuple == NULL)
 		return -1;
 	tuple_ref(stmt->new_tuple);
+
+	stmt->does_require_old_tuple = true;
+
 	if (memtx_space->replace(space, old_tuple, stmt->new_tuple,
 				 DUP_REPLACE, &stmt->old_tuple) != 0)
 		return -1;
@@ -496,6 +510,9 @@ memtx_space_execute_upsert(struct space *space, struct txn *txn,
 			stmt->new_tuple = NULL;
 		}
 	}
+
+	stmt->does_require_old_tuple = true;
+
 	/*
 	 * It's OK to use DUP_REPLACE_OR_INSERT: we don't risk
 	 * inserting a new tuple if the old one exists, since
diff --git a/src/box/txn.c b/src/box/txn.c
index 0372047..d254edb 100644
--- a/src/box/txn.c
+++ b/src/box/txn.c
@@ -116,6 +116,7 @@ txn_stmt_new(struct region *region)
 	stmt->engine_savepoint = NULL;
 	stmt->row = NULL;
 	stmt->has_triggers = false;
+	stmt->does_require_old_tuple = false;
 	return stmt;
 }
 
@@ -360,6 +361,8 @@ txn_commit_stmt(struct txn *txn, struct request *request)
 	 */
 	if (stmt->space != NULL && !rlist_empty(&stmt->space->on_replace) &&
 	    stmt->space->run_triggers && (stmt->old_tuple || stmt->new_tuple)) {
+		/* Triggers see old_tuple and that tuple must remain the same */
+		stmt->does_require_old_tuple = true;
 		int rc = 0;
 		if(!space_is_temporary(stmt->space)) {
 			rc = trigger_run(&stmt->space->on_replace, txn);
diff --git a/src/box/txn.h b/src/box/txn.h
index e261852..962ada0 100644
--- a/src/box/txn.h
+++ b/src/box/txn.h
@@ -175,6 +175,19 @@ struct txn_stmt {
 	struct xrow_header *row;
 	/** on_commit and/or on_rollback list is not empty. */
 	bool has_triggers;
+	/**
+	 * Whether the stmt requires to replace exactly old_tuple (member).
+	 * That flag is needed for transactional conflict manager - if any
+	 * other transaction replaces old_tuple before current one and the flag
+	 * is set - the current transaction will be aborted.
+	 * For example REPLACE just replaces a key, no matter what tuple
+	 * lays in the index and thus does_require_old_tuple = false.
+	 * In contrast, UPDATE makes new tuple using old_tuple and thus
+	 * the statement will require old_tuple (does_require_old_tuple = true).
+	 * INSERT also does_require_old_tuple = true because it requires
+	 * old_tuple to be NULL.
+	 */
+	bool does_require_old_tuple;
 	/** Commit/rollback triggers associated with this statement. */
 	struct rlist on_commit;
 	struct rlist on_rollback;
-- 
2.7.4

  parent reply	other threads:[~2020-07-15 13:55 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-15 13:55 [Tarantool-patches] [PATCH v3 00/13] Transaction engine for memtx engine Aleksandr Lyapunov
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 01/13] Update license file (2020) Aleksandr Lyapunov
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 02/13] Check data_offset overflow in struct tuple Aleksandr Lyapunov
2020-07-16 14:27   ` Nikita Pettik
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 03/13] vinyl: rename tx_manager -> vy_tx_manager Aleksandr Lyapunov
2020-07-15 16:04   ` Nikita Pettik
2020-07-16  8:17     ` Aleksandr Lyapunov
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 04/13] txm: introduce dirty tuples Aleksandr Lyapunov
2020-07-15 16:22   ` Nikita Pettik
2020-07-16  0:05   ` Vladislav Shpilevoy
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 05/13] txm: save txn in txn_stmt Aleksandr Lyapunov
2020-07-15 16:23   ` Nikita Pettik
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 06/13] txm: add TX status Aleksandr Lyapunov
2020-07-15 16:42   ` Nikita Pettik
2020-07-16  0:08   ` Vladislav Shpilevoy
2020-07-15 13:55 ` Aleksandr Lyapunov [this message]
2020-07-15 16:49   ` [Tarantool-patches] [PATCH v3 07/13] txm: save does_require_old_tuple flag in txn_stmt Nikita Pettik
2020-07-16  0:09   ` Vladislav Shpilevoy
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 08/13] txm: introduce tx manager Aleksandr Lyapunov
2020-07-15 16:51   ` Nikita Pettik
2020-07-15 22:01   ` Vladislav Shpilevoy
2020-07-16  0:10   ` Vladislav Shpilevoy
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 09/13] tmx: introduce prepare sequence number Aleksandr Lyapunov
2020-07-15 17:13   ` Nikita Pettik
2020-07-16  0:11   ` Vladislav Shpilevoy
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 10/13] tmx: introduce conflict tracker Aleksandr Lyapunov
2020-07-16  0:16   ` Vladislav Shpilevoy
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 11/13] txm: introduce txm_story Aleksandr Lyapunov
2020-07-16  0:20   ` Vladislav Shpilevoy
2020-07-17  6:16     ` Aleksandr Lyapunov
2020-07-16 22:25   ` Vladislav Shpilevoy
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 12/13] txm: clarify all fetched tuples Aleksandr Lyapunov
2020-07-15 13:55 ` [Tarantool-patches] [PATCH v3 13/13] tmx: use new tx manager in memtx Aleksandr Lyapunov
2020-07-16 22:26   ` Vladislav Shpilevoy
2020-07-17  5:08     ` Aleksandr Lyapunov
2020-07-23 20:58       ` Vladislav Shpilevoy
2020-07-15 15:47 ` [Tarantool-patches] [PATCH v3 00/13] Transaction engine for memtx engine Aleksandr Lyapunov
2020-07-15 16:38 ` Aleksandr Lyapunov
2020-07-15 16:39 ` Aleksandr Lyapunov
2020-07-15 16:40 ` Aleksandr Lyapunov
2020-07-16  0:05 ` Vladislav Shpilevoy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1594821336-14468-8-git-send-email-alyapunov@tarantool.org \
    --to=alyapunov@tarantool.org \
    --cc=tarantool-patches@dev.tarantool.org \
    --subject='Re: [Tarantool-patches] [PATCH v3 07/13] txm: save does_require_old_tuple flag in txn_stmt' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox