mirror of
https://github.com/MariaDB/server.git
synced 2026-10-23 18:49:29 +03:00
378d5bd10e
When a select containing 32 ranges is made on a table containing a
compound key with 32 parts, the range optimizer can run off the end
of a stack variable, invalidly overwriting subsequent stack variables.
In the struct st_sel_arg_range_seq, we have an array
RANGE_SEQ_ENTRY stack[MAX_REF_PARTS];
MAX_REF_PARTS is 32.
check_quick_select
/ sel_arg_range_seq_init
initialises stack[0] as NOT a key part
/ sel_arg_range_seq_next
iterates through the key parts, adding key part n to stack[n+1]
key part #32 gets referenced by step_down_to(), setting seq->i off the
end of the array.
Fix: RANGE_SEQ_ENTRY stack[MAX_REF_PARTS+1];
The above change exposed an issue with key length calculation on
MS Windows. Calling make_prev_keypart_map(32) caused the resultant
bitmap to be calculated as (1UL << 32) - 1. Using the MSVC compiler
this resulted in an empty key length calculation during
handler::index_read_map, causing an assertion in ha_innobase::index_read().
As we only need 32 bits to represent our key map, we change the type thus
-typedef ulong key_part_map;
+typedef uint32 key_part_map;
We correct make_keypart_map() and make_prev_keypart_map() to call our
overflow safe my_set_bits(). We also correct bka_range_seq_next()
and bkah_range_seq_next() to use make_prev_keypart_map().
We also add some DBUG_ASSERTS in key_part_map processing elsewhere,
exposing some issues in our BNLH implementation. We cap the number of
keyuse parts here, altering the explain output of 2 of our tests.
Numerous places needed bit shift operations altered to use
make*keymap_part and various format strings needed to be corrected.
== Summary ==
This directory contains RocksDB-based Storage Engine (RDBSE) for MySQL,
also known as "MyRocks".
== Resources ==
https://github.com/facebook/mysql-5.6/wiki/Getting-Started-with-MyRocks
https://www.facebook.com/groups/MyRocks/
== Coding Conventions ==
The baseline for MyRocks coding conventions for the code in storage/rocksdb/
is based on the default clang format with a few minor changes. The file
storage/rocksdb/.clang-format describes conventions and can be integrated
with Vim or Emacs as described here:
http://releases.llvm.org/3.6.0/tools/clang/docs/ClangFormat.html#vim-integration
All code outside of storage/rocksdb/ should conform to the MySQL coding
conventions:
http://dev.mysql.com/doc/internals/en/coding-guidelines.html.
Several refinements:
0. There is an umbrella C++ namespace named "myrocks" for all MyRocks code.
1. We introduced "RDB" as the super-short abbreviation for "RocksDB". We will
use it as a name prefix, with different capitalization (see below), to ease
up code navigation with ctags and grep.
N.B. For ease of matching, we'll keep the variables and functions dealing
with sysvars as close as possible to the outside visible names of
sysvars, which start with "rocksdb_" prefix, the outward storage
engine name.
2. The names for classes, interfaces, and C++ structures (which act as
classes), start with prefix "Rdb_".
NB: For historical reasons, we'll keep the "ha_<storage_engine_name>" class
name for ha_rocksdb class, which is an exception to the rule.
3. The names for global objects and functions start with prefix "rdb_".
4. The names for macros and constants start with prefix "RDB_".
5. Regular class member names start with "m_".
6. Static class member names start with "s_".
7. Given the 80 character per line limit, we'll not always use full English
words in names, when a well known or easily recognizable abbreviation
exists (like "tx" for "transaction" or "param" for "parameter" etc).
8. When needing to disambiguate, we use different suffixes for that, like
"_arg" for a function argument/parameter, "_arr" for a C style array, and
"_vect" for a std::vector etc.
== Running Tests ==
To run tests from rocksdb, rocksdb_rpl or other rocksdb_* packages, use the
following parameters:
--default-storage-engine=rocksdb
--skip-innodb
--default-tmp-storage-engine=MyISAM
--rocksdb