Commit graph

53 commits

Author SHA1 Message Date
Vladislav Vaintroub
031b3dfc22 MDEV-25123 support MSVC ASAN 2021-03-12 08:44:55 +01:00
Marko Mäkelä
b6ec1e8bbf MDEV-20377 post-fix: Introduce MEM_MAKE_ADDRESSABLE
In AddressSanitizer, we only want memory poisoning to happen
in connection with custom memory allocation or freeing.

The primary use of MEM_UNDEFINED is for declaring memory uninitialized
in Valgrind or MemorySanitizer. We do not want MEM_UNDEFINED to
have the unwanted side effect that AddressSanitizer would no longer
be able to complain about accessing unallocated memory.

MEM_UNDEFINED(): Define as no-op for AddressSanitizer.

MEM_MAKE_ADDRESSABLE(): Define as MEM_UNDEFINED() or
ASAN_UNPOISON_MEMORY_REGION().

MEM_CHECK_ADDRESSABLE(): Wrap also __asan_region_is_poisoned().
2020-07-02 17:59:28 +03:00
Monty
65f831d17c Fixed bugs found by valgrind
- Some of the bug fixes are backports from 10.5!
- The fix in innobase/fil/fil0fil.cc is just a backport to get less
  error messages in mysqld.1.err when running with valgrind.
- Renamed HAVE_valgrind_or_MSAN to HAVE_valgrind
2020-07-02 17:57:34 +03:00
Marko Mäkelä
1df1a63924 Merge 10.2 into 10.3 2020-07-02 06:17:51 +03:00
Marko Mäkelä
c36834c832 MDEV-20377: Make WITH_MSAN more usable
MemorySanitizer (clang -fsanitize=memory) requires that all code
be compiled with instrumentation enabled. The only exception is the
C runtime library. Failure to use instrumented libraries will cause
bogus messages about memory being uninitialized.

In WITH_MSAN builds, we must avoid calling getservbyname(),
because even though it is a standard library function, it is
not instrumented, not even in clang 10.

Note: Before MariaDB Server 10.5, ./mtr will typically fail
due to the old PCRE library, which was updated in MDEV-14024.

The following cmake options were tested on 10.5
in commit 94d0bb4dbe:

cmake \
-DCMAKE_C_FLAGS='-march=native -O2' \
-DCMAKE_CXX_FLAGS='-stdlib=libc++ -march=native -O2' \
-DWITH_EMBEDDED_SERVER=OFF -DWITH_UNIT_TESTS=OFF -DCMAKE_BUILD_TYPE=Debug \
-DWITH_INNODB_{BZIP2,LZ4,LZMA,LZO,SNAPPY}=OFF \
-DPLUGIN_{ARCHIVE,TOKUDB,MROONGA,OQGRAPH,ROCKSDB,CONNECT,SPIDER}=NO \
-DWITH_SAFEMALLOC=OFF \
-DWITH_{ZLIB,SSL,PCRE}=bundled \
-DHAVE_LIBAIO_H=0 \
-DWITH_MSAN=ON

MEM_MAKE_DEFINED(): An alias for VALGRIND_MAKE_MEM_DEFINED()
and __msan_unpoison().

MEM_GET_VBITS(), MEM_SET_VBITS(): Aliases for
VALGRIND_GET_VBITS(), VALGRIND_SET_VBITS(), __msan_copy_shadow().

InnoDB: Replace the UNIV_MEM_ macros with corresponding MEM_ macros.

ut_crc32_8_hw(), ut_crc32_64_low_hw(): Use the compiler built-in
functions instead of inline assembler when building WITH_MSAN.
This will require at least -msse4.2 when building for IA-32 or AMD64.
The inline assembler would not be instrumented, and would thus cause
bogus failures.
2020-07-01 17:23:00 +03:00
Marko Mäkelä
6a6bcc53b8 Merge 10.2 into 10.3 2020-05-15 17:55:01 +03:00
Marko Mäkelä
ff66d65a09 Amend af784385b4: Avoid vtable overhead
When neither MSAN nor Valgrind are enabled, declare
Field::mark_unused_memory_as_defined() as an empty inline function,
instead of declaring it as a virtual function.
2020-05-15 17:23:08 +03:00
Monty
af784385b4 Fix for using uninitialized memory
MDEV-22073 MSAN use-of-uninitialized-value in
collect_statistics_for_table()

Other things:
innodb.analyze_table was changed to mainly test statistic
collection. Was discussed with Marko.
2020-05-15 15:10:58 +03:00
Marko Mäkelä
65d48b4a7b Merge 10.2 to 10.3 2019-08-13 19:28:51 +03:00
Marko Mäkelä
be33124c9d Merge 10.1 into 10.2 2019-08-12 18:25:35 +03:00
Marko Mäkelä
b2a387a3f1 Document TRASH_FILL, TRASH_ALLOC, TRASH_FREE 2019-08-12 14:14:17 +03:00
Marko Mäkelä
90a9193685 Merge 10.2 into 10.3 2019-05-29 11:32:46 +03:00
Marko Mäkelä
d59e15bdb9 Merge 10.1 into 10.2 2019-05-28 15:56:24 +03:00
Sergei Golubchik
5034b31b01 MDEV-17799 Add ASAN-poisoned redzones for MEM_ROOT
post-merge changes

Closes #954
2019-05-20 18:24:49 +02:00
Eugene Kosov
7b59ec6f34 MDEV-17799 Add ASAN-poisoned redzones for MEM_ROOT and mem_heap_t
This patch is for MEM_ROOT only.
In debug mode add 8 byte of poisoned memory before every allocated chunk.
On the right of every chunk there will be either 1-7 trailing poisoned bytes, or
next chunk's redzone, or poisoned non allocated memory or redzone of a
malloc()ed buffer.
2019-05-20 18:18:23 +02:00
Marko Mäkelä
be85d3e61b Merge 10.2 into 10.3 2019-05-14 17:18:46 +03:00
Marko Mäkelä
26a14ee130 Merge 10.1 into 10.2 2019-05-13 17:54:04 +03:00
Vladislav Vaintroub
0c188d5efc Make TRASH_FREED_MEMORY a cmake option, similar to SAFEMALLOC 2019-05-13 10:08:42 +00:00
Michal Schorm
17b4f99928 Update FSF address
This commit is based on the work of Michal Schorm, rebased on the
earliest MariaDB version.

Th command line used to generate this diff was:

find ./ -type f \
  -exec sed -i -e 's/Foundation, Inc., 59 Temple Place, Suite 330, Boston, /Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, /g' {} \; \
  -exec sed -i -e 's/Foundation, Inc. 59 Temple Place.* Suite 330, Boston, /Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, /g' {} \; \
  -exec sed -i -e 's/MA.*.....-1307.*USA/MA 02110-1335 USA/g' {} \; \
  -exec sed -i -e 's/Foundation, Inc., 59 Temple/Foundation, Inc., 51 Franklin/g' {} \; \
  -exec sed -i -e 's/Place, Suite 330, Boston, MA.*02111-1307.*USA/Street, Fifth Floor, Boston, MA 02110-1335 USA/g' {} \; \
  -exec sed -i -e 's/MA.*.....-1307/MA 02110-1335/g' {} \;
2019-05-10 20:52:00 +03:00
Oleksandr Byelkin
65c5ef9b49 dirty merge 2019-02-07 13:59:31 +01:00
Marko Mäkelä
a249e57b68 Merge 10.1 into 10.2
Temporarily disable a test for
commit 2175bfce3e
because fixing it in 10.2 requires updating libmariadb.
2019-02-03 17:22:05 +02:00
Marko Mäkelä
213ece2f2e Merge 10.1 into 10.1
This is joint work with Oleksandr Byelkin.
2019-02-02 13:00:15 +02:00
Vladislav Vaintroub
e292d1a800 Avoid noisy Clang 7 warning about unused variable.
Patch by Eugene Kosov.
2019-01-19 14:01:09 +01:00
Eugene Kosov
5b25dc6fa4 MDEV-17248 Improve ASAN memory pool instrumentation
alloc_root(): unpoison only requested amount of bytes instead of a
possible bigger aligned-sized buffer.
2018-09-21 10:17:37 +03:00
Sergei Golubchik
b1818dccf7 Merge branch '10.2' into 10.3 2018-03-28 17:31:57 +02:00
Marko Mäkelä
3d7915f000 Merge 10.1 into 10.2 2018-03-21 22:58:52 +02:00
Marko Mäkelä
613be24b7a Merge 10.0 into 10.1 2018-03-20 19:25:08 +02:00
Marko Mäkelä
0492100059 Merge 5.5 into 10.0 2018-03-20 18:36:03 +02:00
Eugene Kosov
5a8f8f89d6 honor alignment rules and xtradb too 2018-03-20 10:46:57 +03:00
Alexander Barkov
3cad31f2a7 Merge remote-tracking branch 'origin/10.2' into bb-10.2-ext 2018-02-08 19:06:25 +04:00
Sergei Golubchik
4771ae4b22 Merge branch 'github/10.1' into 10.2 2018-02-06 14:50:50 +01:00
Sergei Golubchik
d4df7bc9b1 Merge branch 'github/10.0' into 10.1 2018-02-02 10:09:44 +01:00
Monty
ffcedfab46 Added TRASH_FREED_MEMORY compilation option
One can use -DTRASH_FREED_MEMORY to enable TRASH
macros. Useful to do when one suspects that there
is accesses to freed memory.

Extended my_free() to TRASH freed memory
2018-01-28 15:54:17 +02:00
Vicențiu Ciorbaru
d833bb65d5 Merge remote-tracking branch '5.5' into 10.0 2018-01-24 12:29:31 +02:00
Sergei Golubchik
8539e4b1b6 improve ASAN instrumentation: clang
translate clang __has_feature to gcc macros
2018-01-22 16:36:39 +01:00
Sergei Golubchik
a966d422ca improve ASAN instrumentation: TRASH
mark freed memory as not accessible, not merely undefined
2018-01-22 11:39:54 +01:00
Sergei Golubchik
22ae3843db Correct TRASH() macro usage
TRASH was mapped to TRASH_FREE and was supposed to be used for memory
that should not be accessed anymore, while TRASH_ALLOC() is to be
used for uninitialized but to-be-used memory.

But sometimes TRASH() was used in the latter sense.

Remove TRASH() macro, always use explicit TRASH_ALLOC() or TRASH_FREE().
2018-01-22 11:39:54 +01:00
Sergei Golubchik
5e7593add4 add support for ASAN instrumentation 2018-01-16 22:57:28 +01:00
Sergei Golubchik
da4d71d10d Merge branch '10.1' into 10.2 2017-03-30 12:48:42 +02:00
iangilfillan
f0ec34002a Correct FSF address 2017-03-10 18:21:29 +01:00
Marko Mäkelä
89d80c1b0b Fix many -Wconversion warnings.
Define my_thread_id as an unsigned type, to avoid mismatch with
ulonglong.  Change some parameters to this type.

Use size_t in a few more places.

Declare many flag constants as unsigned to avoid sign mismatch
when shifting bits or applying the unary ~ operator.

When applying the unary ~ operator to enum constants, explictly
cast the result to an unsigned type, because enum constants can
be treated as signed.

In InnoDB, change the source code line number parameters from
ulint to unsigned type. Also, make some InnoDB functions return
a narrower type (unsigned or uint32_t instead of ulint;
bool instead of ibool).
2017-03-07 19:07:27 +02:00
Sergei Golubchik
4368efe870 valgrind failures
and a couple of collaterals:
* debugging assert in my_valgrind.h
* trivial cleanup in sql_union.cc
2016-09-21 10:03:16 +02:00
Sergei Golubchik
65d69c84cf cmake: remove unused checks, options, and symbols 2015-11-23 16:02:56 +01:00
Sergei Golubchik
0dc23679c8 10.0-base merge 2014-02-26 15:28:07 +01:00
Sergei Golubchik
95f4bf1857 MDEV-5613 m_string.h exports generic function names without a namespace prefix, like str2int
don't include m_string. in my_valgrind.h
2014-02-17 11:09:57 +01:00
Vladislav Vaintroub
ae325ec6ce Compile 10.0 on Windows 2012-08-30 11:36:24 +02:00
Michael Widenius
f1159b18d9 More fixes 2012-08-17 16:46:34 +03:00
Michael Widenius
1d0f70c2f8 Temporary commit of merge of MariaDB 10.0-base and MySQL 5.6 2012-08-01 17:27:34 +03:00
unknown
c895e235f3 Do not use Valgrind client requests in a normal release build, they have a small but noticable performance impact. 2012-03-22 12:31:09 +01:00
Sergei Golubchik
db39910f08 disable safemalloc for valgrind builds.
always try to use valgrind headers in debug builds.
define HAVE_valgrind for --with-valgrind
fix valgrind check in my_valgrind.h
2012-02-23 15:31:24 +01:00