Bug 18914 (Calling certain SPs from triggers fail)
Bug 20713 (Functions will not not continue for SQLSTATE VALUE '42S02')
Bug 21825 (Incorrect message error deleting records in a table with a
trigger for inserting)
Bug 22580 (DROP TABLE in nested stored procedure causes strange dependency
error)
Bug 25345 (Cursors from Functions)
This fix resolves a long standing issue originally reported with bug 8407,
which affect the behavior of Stored Procedures, Stored Functions and Trigger
in many different ways, causing symptoms reported by all the bugs listed.
In all cases, the root cause of the problem traces back to 8407 and how the
server locks tables involved with sub statements.
Prior to this fix, the implementation of stored routines would:
- compute the transitive closure of all the tables referenced by a top level
statement
- open and lock all the tables involved
- execute the top level statement
"transitive closure of tables" means collecting:
- all the tables,
- all the stored functions,
- all the views,
- all the table triggers
- all the stored procedures
involved, and recursively inspect these objects definition to find more
references to more objects, until the list of every object referenced does
not grow any more.
This mechanism is known as "pre-locking" tables before execution.
The motivation for locking all the tables (possibly) used at once is to
prevent dead locks.
One problem with this approach is that, if the execution path the code
really takes during runtime does not use a given table, and if the table is
missing, the server would not execute the statement.
This in particular has a major impact on triggers, since a missing table
referenced by an update/delete trigger would prevent an insert trigger to run.
Another problem is that stored routines might define SQL exception handlers
to deal with missing tables, but the server implementation would never give
user code a chance to execute this logic, since the routine is never
executed when a missing table cause the pre-locking code to fail.
With this fix, the internal implementation of the pre-locking code has been
relaxed of some constraints, so that failure to open a table does not
necessarily prevent execution of a stored routine.
In particular, the pre-locking mechanism is now behaving as follows:
1) the first step, to compute the transitive closure of all the tables
possibly referenced by a statement, is unchanged.
2) the next step, which is to open all the tables involved, only attempts
to open the tables added by the pre-locking code, but silently fails without
reporting any error or invoking any exception handler is the table is not
present. This is achieved by trapping internal errors with
Prelock_error_handler
3) the locking step only locks tables that were successfully opened.
4) when executing sub statements, the list of tables used by each statements
is evaluated as before. The tables needed by the sub statement are expected
to be already opened and locked. Statement referencing tables that were not
opened in step 2) will fail to find the table in the open list, and only at
this point will execution of the user code fail.
5) when a runtime exception is raised at 4), the instruction continuation
destination (the next instruction to execute in case of SQL continue
handlers) is evaluated.
This is achieved with sp_instr::exec_open_and_lock_tables()
6) if a user exception handler is present in the stored routine, that
handler is invoked as usual, so that ER_NO_SUCH_TABLE exceptions can be
trapped by stored routines. If no handler exists, then the runtime execution
will fail as expected.
With all these changes, a side effect is that view security is impacted, in
two different ways.
First, a view defined as "select stored_function()", where the stored
function references a table that may not exist, is considered valid.
The rationale is that, because the stored function might trap exceptions
during execution and still return a valid result, there is no way to decide
when the view is created if a missing table really cause the view to be invalid.
Secondly, testing for existence of tables is now done later during
execution. View security, which consist of trapping errors and return a
generic ER_VIEW_INVALID (to prevent disclosing information) was only
implemented at very specific phases covering *opening* tables, but not
covering the runtime execution. Because of this existing limitation,
errors that were previously trapped and converted into ER_VIEW_INVALID are
not trapped, causing table names to be reported to the user.
This change is exposing an existing problem, which is independent and will
be resolved separately.
mysql-test/r/information_schema_db.result:
Revised the pre-locking code implementation, aligned the tests.
mysql-test/r/sp-error.result:
Revised the pre-locking code implementation, aligned the tests.
mysql-test/r/sp.result:
Revised the pre-locking code implementation, aligned the tests.
mysql-test/r/trigger.result:
Revised the pre-locking code implementation, aligned the tests.
mysql-test/r/view.result:
Revised the pre-locking code implementation, aligned the tests.
mysql-test/t/sp-error.test:
Revised the pre-locking code implementation, aligned the tests.
mysql-test/t/sp.test:
Revised the pre-locking code implementation, aligned the tests.
mysql-test/t/trigger.test:
Revised the pre-locking code implementation, aligned the tests.
sql/lock.cc:
table->placeholder now checks for schema_table
sql/mysqld.cc:
my_message_sql(): invoke internal exception handlers
sql/sp_head.cc:
exec_open_and_lock_tables(): open and lock tables, or return the
continuation destination of this instruction
sql/sp_head.h:
exec_open_and_lock_tables(): open and lock tables, or return the
continuation destination of this instruction
sql/sql_base.cc:
Prelock_error_handler: delay open table errors until execution
sql/sql_class.cc:
THD: add internal error handler, as an exception mechanism.
sql/sql_class.h:
THD: add internal error handler, as an exception mechanism.
sql/sql_update.cc:
table->placeholder now checks for schema_table
sql/table.cc:
st_table_list::hide_view_error(): masked more errors for view security
sql/table.h:
table->placeholder now checks for schema_table, and unopened tables
The fix is: if user has privileges to view fields and user has any
(insert,select,delete,update) privileges on underlying view
then 'show fields' and select from I_S.COLUMNS table are sucsessful.
mysql-test/r/information_schema_db.result:
Bug#20543 select on information_schema strange warnings, view, different schemas/users
test result
mysql-test/t/information_schema_db.test:
Bug#20543 select on information_schema strange warnings, view, different schemas/users
test case
sql/sql_acl.cc:
Bug#20543 select on information_schema strange warnings, view, different schemas/users
checked that user has privileges on underlying view and if it's true
set allowed_show to true for top view.
sql/sql_show.cc:
Bug#20543 select on information_schema strange warnings, view, different schemas/users
removed unnecessary rights check.'tables->allowed_show' check is used instead
sql/sql_view.cc:
Bug#20543 select on information_schema strange warnings, view, different schemas/users
skip the check of SHOW_VIEW_ACL privilege on underlying view. It is done later during
execution of find_field_in_table_ref function.
sql/table.h:
Bug#20543 select on information_schema strange warnings, view, different schemas/users
'allowed_show' is set during rights check for view. If true then user has privileges
for 'show create view', etc
Bug#18282 "INFORMATION_SCHEMA.TABLES provides inconsistent info about invalid views"
This bug caused crashes or resulted in wrong data being returned
when one tried to obtain information from I_S tables about views
using stored functions.
It was caused by the fact that we were using LEX representing
statement which were doing select from I_S tables as active LEX
when contents of I_S table were built. So state of this LEX both
affected and was affected by open_tables() calls which happened
during this process. This resulted in wrong behavior and in
violations of some of invariants which caused crashes.
This fix tries to solve this problem by properly saving/resetting
and restoring part of LEX which affects and is affected by the
process of opening tables and views in get_all_tables() routine.
To simplify things we separated this part of LEX in a new class
and made LEX its descendant.
mysql-test/r/information_schema_db.result:
test case
mysql-test/t/information_schema_db.test:
test case
sql/sql_lex.cc:
To simplify saving/resetting and restoring part of LEX which
affects and is affected by the process of opening tables and
views we moved it to new class Query_tables_list and made LEX
descendant of this class. Also introduced two LEX methods
which can be used to save and reset or to restore this state.
sql/sql_lex.h:
To simplify saving/resetting and restoring part of LEX which
affects and is affected by the process of opening tables and
views we moved it to new class Query_tables_list and made LEX
descendant of this class. Also introduced two LEX methods
which can be used to save and reset or to restore this state.
sql/sql_show.cc:
Now in get_all_tables() routine we properly save/reset and
restore part of LEX (statement table list and information
about routines used) which affects and is affected by the
process of opening tables and views.
sql/sql_table.cc:
Now we clean-up LEX after opening table (view) in two stages.
In the first stage we call LEX::cleanup_after_one_table_open()
to clean-up selects lists and derived tables state. In the
second stage which happens after close_thread_tables() is
invoked we call Query_tables_list::reset_query_tables_list(FALSE)
to rollback changes in Query_tables_list.
Crash happened when one selected data from one of INFORMATION_SCHEMA
tables and in order to build its contents server had to open view which
used stored function and table or view on which one had not global or
database-level privileges (e.g. had only table-level or had no
privileges at all).
The crash was caused by usage of check_grant() function, which assumes
that either number of tables to be inspected by it is limited explicitly
or table list used and thd->lex->query_tables_own_last value correspond
to each other (the latter should be either 0 or point to next_global
member of one of elements of this table list), in conditions when
above assumptions were not true. This fix just explicitly limits
number of tables to be inspected. Other negative effects which are
caused by the fact that thd->lex->query_tables_own_last might not
be set properly during processing of I_S tables are less disastrous
and will be reported and fixed separetely.
mysql-test/r/information_schema_db.result:
Fix for bug #18113 "SELECT * FROM information_schema.xxx crashes server"
test case
mysql-test/t/information_schema_db.test:
Fix for bug #18113 "SELECT * FROM information_schema.xxx crashes server"
test case
sql/sql_acl.cc:
added note
Fix test result.
table.h, sql_show.cc:
Put I_S tables in lexical order.
sql/sql_show.cc:
Put I_S tables in lexical order.
sql/table.h:
Put I_S tables in lexical order.
mysql-test/r/information_schema.result:
Fix test result.
mysql-test/r/information_schema_db.result:
Fix test result.
mysql-test/r/information_schema.result:
fix the test case for fix for bug #12315 - SHOW TABLE STATUS FROM `information_schema`; does not sort tablenames
mysql-test/r/information_schema_db.result:
fix the test case for fix for bug #12315 - SHOW TABLE STATUS FROM `information_schema`; does not sort tablenames
sql/sql_show.cc:
reorder the structure
fix for bug #12315 - SHOW TABLE STATUS FROM `information_schema`; does not sort tablenames
sql/table.h:
reorder the enum
fix for bug #12315 - SHOW TABLE STATUS FROM `information_schema`; does not sort tablenames
"Triggers have the wrong namespace"
"Triggers: duplicate names allowed"
"Triggers: CREATE TRIGGER does not accept fully qualified names"
"SHOW TRIGGERS"
mysql-test/r/information_schema.result:
Added tests for new INFORMATION_SCHEMA.TRIGGERS view and SHOW TRIGGERS command.
mysql-test/r/information_schema_db.result:
INFORMATION_SCHEMA.TRIGGERS view was added.
mysql-test/r/rpl_sp.result:
Now DROP TRIGGER interprets first part of trigger identifier as database
name and not as table name. Adjusted tests properly.
mysql-test/r/trigger.result:
Now DROP TRIGGER interprets first part of trigger identifier as database
name and not as table name. Adjusted tests properly.
Added test checking that triggers have database wide namespace.
Added test for bug #8791 "Triggers: Allowed to create triggers on a subject
table in a different DB".
mysql-test/r/view.result:
Now DROP TRIGGER interprets first part of trigger identifier as database
name and not as table name. Adjusted tests properly.
mysql-test/t/information_schema.test:
Added tests for new INFORMATION_SCHEMA.TRIGGERS view and SHOW TRIGGERS command.
mysql-test/t/rpl_sp.test:
Now DROP TRIGGER interprets first part of trigger identifier as database
name and not as table name. Adjusted tests properly.
mysql-test/t/trigger.test:
Now DROP TRIGGER interprets first part of trigger identifier as database
name and not as table name. Adjusted tests properly.
Added test checking that triggers have database wide namespace.
Added test for bug #8791 "Triggers: Allowed to create triggers on a subject
table in a different DB".
mysql-test/t/view.test:
Now DROP TRIGGER interprets first part of trigger identifier as database
name and not as table name. Adjusted tests properly.
sql/handler.cc:
Added .TRN tho the list of known file extensions assoicated with tables.
sql/item.h:
trg_action_time_type/trg_event_type enums:
Added TRG_ACTION_MAX/TRG_EVENT_MAX elements which should be used instead of
magical values in various loops where we iterate through all types of trigger
action times or/and trigger event types.
sql/lex.h:
Added new symbol "TRIGGERS".
sql/mysql_priv.h:
Added declaration of constant holding extension for trigger name (.TRN) files.
sql/mysqld.cc:
Added statistical variable for SHOW TRIGGERS command.
sql/share/errmsg.txt:
Added error message saying that one attempts to create trigger in wrong schema.
sql/sp.cc:
Replaced magical values with TRG_EVENT_MAX/TRG_ACTION_MAX constants.
sql/sql_base.cc:
open_unireg_entry():
Now Table_triggers_list::check_n_load() has one more argument which
controls whether we should prepare Table_triggers_list with fully functional
triggers or load only their names.
sql/sql_lex.h:
Added element for new SHOW TRIGGERS command to enum_sql_command enum.
sql/sql_parse.cc:
prepare_schema_table():
Added support for SHOW TRIGGERS statement.
sql/sql_show.cc:
Added new INFORMATION_SCHEMA.TRIGGERS view and SHOW TRIGGERS command.
sql/sql_table.cc:
mysql_rm_table_part2():
Replaced simple deletion of .TRG file with call to
Table_triggers_list::drop_all_triggers which will also delete .TRN files
for all triggers associated with table.
sql/sql_trigger.cc:
Now triggers have database wide namespace. To support it we create special .TRN
file with same name as trigger for each trigger. This file contains name of
trigger's table so one does not need to specify it explicitly in DROP TRIGGER.
Moreover DROP TRIGGER treats first part of trigger identifier as database name
now. Updated mysql_create_or_drop_trigger() routine and
Table_triggers_list::create_trigger()/drop_trigger()/check_n_load() methods
accordingly. Added add_table_for_trigger() routine and
Table_triggers_list::drop_all_triggers() method.
Added Table_triggers_list::get_trigger_info() for obtaining trigger metadata.
sql/sql_trigger.h:
Table_triggers_list:
Use TRG_EVENT_MAX, TRG_ACTION_MAX instead of magic values.
Added get_trigger_info() method for obtaining trigger's meta-data.
Added drop_all_triggers() method which drops all triggers for table.
Added declarations of trg_action_time_type_names/trg_event_type_names
arrays which hold names of triggers action time types and event types.
sql/sql_yacc.yy:
Changed grammar for CREATE/DROP TRIGGER to support database wide trigger
namespace. Added new SHOW TRIGGERS statement.
sql/table.h:
enum enum_schema_tables:
Added constant for new INFORMATION_SCHEMA.TRIGGERS view.
