1
0
mirror of https://github.com/MariaDB/server.git synced 2025-07-27 18:02:13 +03:00

Implementation of WL#2897: Complete definer support in the stored routines.

The idea is to add DEFINER-clause in CREATE PROCEDURE and CREATE FUNCTION
statements. Almost all support of definer in stored routines had been already
done before this patch.

NOTE: this patch changes behaviour of dumping stored routines in mysqldump.
Before this patch, mysqldump did not dump DEFINER-clause for stored routines
and this was documented behaviour. In order to get full information about stored
routines, one should have dumped mysql.proc table. This patch changes this
behaviour, so that DEFINER-clause is dumped.

Since DEFINER-clause is not supported in CREATE PROCEDURE | FUNCTION statements
before this patch, the clause is covered by additional version-specific comments.
This commit is contained in:
anozdrin@mysql.com
2006-03-02 15:18:49 +03:00
parent 302239f388
commit fbb5920399
18 changed files with 521 additions and 158 deletions

View File

@ -4145,6 +4145,90 @@ end_with_restore_list:
}
#endif
/*
If the definer is not specified, this means that CREATE-statement missed
DEFINER-clause. DEFINER-clause can be missed in two cases:
- The user submitted a statement w/o the clause. This is a normal
case, we should assign CURRENT_USER as definer.
- Our slave received an updated from the master, that does not
replicate definer for stored rountines. We should also assign
CURRENT_USER as definer here, but also we should mark this routine
as NON-SUID. This is essential for the sake of backward
compatibility.
The problem is the slave thread is running under "special" user (@),
that actually does not exist. In the older versions we do not fail
execution of a stored routine if its definer does not exist and
continue the execution under the authorization of the invoker
(BUG#13198). And now if we try to switch to slave-current-user (@),
we will fail.
Actually, this leads to the inconsistent state of master and
slave (different definers, different SUID behaviour), but it seems,
this is the best we can do.
*/
if (!lex->definer)
{
bool res= FALSE;
Query_arena original_arena;
Query_arena *ps_arena = thd->activate_stmt_arena_if_needed(&original_arena);
if (!(lex->definer= create_default_definer(thd)))
res= TRUE;
if (ps_arena)
thd->restore_active_arena(ps_arena, &original_arena);
if (res)
{
/* Error has been already reported. */
delete lex->sphead;
lex->sphead= 0;
goto error;
}
if (thd->slave_thread)
lex->sphead->m_chistics->suid= SP_IS_NOT_SUID;
}
/*
If the specified definer differs from the current user, we should check
that the current user has SUPER privilege (in order to create a stored
routine under another user one must have SUPER privilege).
*/
else if (strcmp(lex->definer->user.str, thd->security_ctx->priv_user) ||
my_strcasecmp(system_charset_info,
lex->definer->host.str,
thd->security_ctx->priv_host))
{
if (check_global_access(thd, SUPER_ACL))
{
my_error(ER_SPECIFIC_ACCESS_DENIED_ERROR, MYF(0), "SUPER");
delete lex->sphead;
lex->sphead= 0;
goto error;
}
}
/* Check that the specified definer exists. Emit a warning if not. */
#ifndef NO_EMBEDDED_ACCESS_CHECKS
if (!is_acl_user(lex->definer->host.str,
lex->definer->user.str))
{
push_warning_printf(thd,
MYSQL_ERROR::WARN_LEVEL_NOTE,
ER_NO_SUCH_USER,
ER(ER_NO_SUCH_USER),
lex->definer->user.str,
lex->definer->host.str);
}
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
/*
We need to copy name and db in order to use them for
check_routine_access which is called after lex->sphead has