| Bug #43152 | Assertion `bitmap_is_set_all(&table->s->all_set)' failed in handler::ha_reset | ||
|---|---|---|---|
| Submitted: | 24 Feb 2009 15:20 | Modified: | 6 Apr 2011 1:32 |
| Reporter: | Philip Stoev | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Server | Severity: | S1 (Critical) |
| Version: | 6.0-bzr,5.4, next-mr, 5.6.1, 5.1, 5.0 | OS: | Any |
| Assigned to: | Jon Olav Hauglid | CPU Architecture: | Any |
[24 Feb 2009 15:20]
Philip Stoev
[24 Feb 2009 18:04]
Philip Stoev
This bug happens across a wide variety of workloads.
[27 Feb 2009 14:47]
Peter Gulutzan
Compare Bug#43129 "failed assertion `bitmap_is_set_all(&table->s->all_set)' with concurrent inserts"
[5 Mar 2009 15:18]
MySQL Verification Team
I got different crash output than the one reported on Ubuntu 8.10 64-bit with latest source mysql-6.0-falcon-team. Could you please see if still the original stack trace happens on your side?. I attached a file with the stack trace. Thanks in advance.
[28 Jul 2009 11:19]
Philip Stoev
This bug is no longer repeatable, likely a duplicate of the other bugs that have been closed.
[18 Jan 2010 11:47]
Philip Stoev
This is again repeatable, with the following RQG command (please pull fresh RQG from Launchpad): perl pb2gentest.pl /build/bzr/mysql-next-mr /tmp/vardir - rqg_innodb_stress This is going to cause the crash in about 50% of the cases, within a few minutes after takeoff.
[1 Feb 2010 14:46]
Mattias Jonsson
I have tested with latest mysql-next-mr-bugfixing (alik@sun.com-20100111131051-ksxfc3lnu2wgcj7r) on ubuntu 9.10 x86_64 using both disk and ramdisk without repeating the bug. Could you please be more specific on how to repeat? (hardware, os, mysql-version/rev-id...)
[2 Mar 2010 0:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
[6 Oct 2010 12:18]
Michael Widenius
The bug is still existing in MySQL 5.0 and forwards. This is caused by some bitmap test functions are changing the bitmap other threads are using. A bug fix for this can be found in MariaDB 5.1.51.
[19 Nov 2010 21:07]
Sveta Smirnova
Michael, thank you for the feedback. Could you please send us instructions how to repeat the problem and link to the fix: I found nothing related to bitmap in MariaDB 5.1.51 changelog.
[13 Jan 2011 15:23]
Jon Olav Hauglid
Non-deterministic MTR test case:
CREATE TABLE t1 (a TEXT, b TEXT, PRIMARY KEY(a(1))) ENGINE=InnoDB;
INSERT INTO t1 VALUES('a',REPEAT('b',10000));
connect (con1, localhost, root);
let $try = 1000;
while ($try)
{
connection default;
--send UPDATE t1 SET a='a'
connection con1;
--send SELECT COUNT(*) FROM t1
connection default;
--reap
connection con1;
--reap
dec $try;
}
connection default;
disconnect con1;
DROP TABLE t1;
[14 Jan 2011 13:04]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/128743 3244 Jon Olav Hauglid 2011-01-14 Bug #43152 Assertion `bitmap_is_set_all(&table->s->all_set)' failed in handler::ha_reset This assertion could be triggered if two connections simultaneously executed two bitmap test functions on the same bitmap. Even if these functions have read-only schematics and have const bitmaps as parameters, several of them modified the internal state of the bitmap. With interleaved execution of two such functions it was possible for one function to modify the state of the same bitmap that the other function had just modified. This lead to an inconsistent state and could trigger the assert. Internally the bitmap uses 32 bit words for storage. Since bitmaps can contain any number of bits, the last word in the bitmap may not be fully used. A 32 bit mask is maintained where a bit is set if the corresponding bit in the last bitmap word is unused. The problem was that several test functions applies this mask to the last word. Some used bitwise AND and some used bitwise OR. This meant that if a function first used bitwise AND and another function then used bitwise OR before the first function tested the result, the result would be invalid. This patch fixes the problem by changing the implementation of 8 bitmap functions that modified the bitmap state even if the bitmap was declared const. These functions now preserves the internal state of the bitmap. This makes it possible for two connections to concurrently execute two of these functions on the same bitmap without issues. The patch also reactivates the built-in test of my_bitmap. No test case added as this would require adding several sync points to the bitmap functions. The patch has been tested with a non-deterministic test case posted on the bug report.
[14 Jan 2011 14:08]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/128772 3244 Jon Olav Hauglid 2011-01-14 Bug #43152 Assertion `bitmap_is_set_all(&table->s->all_set)' failed in handler::ha_reset This assertion could be triggered if two connections simultaneously executed two bitmap test functions on the same bitmap. Even if these functions have read-only schematics and have const bitmaps as parameters, several of them modified the internal state of the bitmap. With interleaved execution of two such functions it was possible for one function to modify the state of the same bitmap that the other function had just modified. This lead to an inconsistent state and could trigger the assert. Internally the bitmap uses 32 bit words for storage. Since bitmaps can contain any number of bits, the last word in the bitmap may not be fully used. A 32 bit mask is maintained where a bit is set if the corresponding bit in the last bitmap word is unused. The problem was that several test functions applies this mask to the last word. Some used bitwise AND and some used bitwise OR. This meant that if a function first used bitwise AND and another function then used bitwise OR before the first function tested the result, the result would be invalid. This patch fixes the problem by changing the implementation of 8 bitmap functions that modified the bitmap state even if the bitmap was declared const. These functions now preserves the internal state of the bitmap. This makes it possible for two connections to concurrently execute two of these functions on the same bitmap without issues. The patch also reactivates the built-in test of my_bitmap. No test case added as this would require adding several sync points to the bitmap functions. The patch has been tested with a non-deterministic test case posted on the bug report.
[17 Jan 2011 11:37]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/128936 3258 Jon Olav Hauglid 2011-01-17 Bug #43152 Assertion `bitmap_is_set_all(&table->s->all_set)' failed in handler::ha_reset This assertion could be triggered if two connections simultaneously executed two bitmap test functions on the same bitmap. Even if these functions have read-only schematics and have const bitmaps as parameters, several of them modified the internal state of the bitmap. With interleaved execution of two such functions it was possible for one function to modify the state of the same bitmap that the other function had just modified. This lead to an inconsistent state and could trigger the assert. Internally the bitmap uses 32 bit words for storage. Since bitmaps can contain any number of bits, the last word in the bitmap may not be fully used. A 32 bit mask is maintained where a bit is set if the corresponding bit in the last bitmap word is unused. The problem was that several test functions applies this mask to the last word. Some used bitwise AND and some used bitwise OR. This meant that if a function first used bitwise AND and another function then used bitwise OR before the first function tested the result, the result would be invalid. This patch fixes the problem by changing the implementation of 8 bitmap functions that modified the bitmap state even if the bitmap was declared const. These functions now preserves the internal state of the bitmap. This makes it possible for two connections to concurrently execute two of these functions on the same bitmap without issues. The patch also removes dead testing code from my_bitmap.c. These tests have already been moved to unittest/mysys/bitmap-t.c. No test case added as this would require adding several sync points to the bitmap functions. The patch has been tested with a non-deterministic test case posted on the bug report.
[4 Feb 2011 11:18]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/130391 3301 Jon Olav Hauglid 2011-02-04 Bug #43152 Assertion `bitmap_is_set_all(&table->s->all_set)' failed in handler::ha_reset This assertion could be triggered if two connections simultaneously executed two bitmap test functions on the same bitmap. Even if these functions have read-only schematics and have const bitmaps as parameters, several of them modified the internal state of the bitmap. With interleaved execution of two such functions it was possible for one function to modify the state of the same bitmap that the other function had just modified. This lead to an inconsistent state and could trigger the assert. Internally the bitmap uses 32 bit words for storage. Since bitmaps can contain any number of bits, the last word in the bitmap may not be fully used. A 32 bit mask is maintained where a bit is set if the corresponding bit in the last bitmap word is unused. The problem was that several test functions applies this mask to the last word. Sometimes the mask is negated and used to zero out the remainder of the last word and sometimes the mask is used as-is to fill the remainder of the last word with 1's. This meant that if a function first used the negated mask and another function then used the mask as-is (or vice-versa), the result would be invalid. This patch fixes the problem by changing the implementation of 9 bitmap functions that modified the bitmap state even if the bitmap was declared const. These functions now preserves the internal state of the bitmap. This makes it possible for two connections to concurrently execute two of these functions on the same bitmap without issues. The patch also removes dead testing code from my_bitmap.c. These tests have already been moved to unittest/mysys/bitmap-t.c. Existing test coverage of my_bitmap has been extended. No MTR test case added as this would require adding several sync points to the bitmap functions. The patch has been tested with a non-deterministic test case posted on the bug report. @ include/my_bit.h Removed now unused function. @ unittest/mysys/bitmap-t.c Extended test coverage of my_bitmap.
[10 Feb 2011 10:36]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/131004 3321 Jon Olav Hauglid 2011-02-10 Bug #43152 Assertion `bitmap_is_set_all(&table->s->all_set)' failed in handler::ha_reset This assertion could be triggered if two connections simultaneously executed two bitmap test functions on the same bitmap. For example, the assertion could be triggered if one connection executed UPDATE while a second connection executed SELECT on the same table. Even if bitmap test functions have read-only schematics and have const bitmaps as parameter, several of them modified the internal state of the bitmap. With interleaved execution of two such functions it was possible for one function to modify the state of the same bitmap that the other function had just modified. This lead to an inconsistent state and could trigger the assert. Internally the bitmap uses 32 bit words for storage. Since bitmaps can contain any number of bits, the last word in the bitmap may not be fully used. A 32 bit mask is maintained where a bit is set if the corresponding bit in the last bitmap word is unused. The problem was that several test functions applies this mask to the last word. Sometimes the mask wass negated and used to zero out the remainder of the last word and sometimes the mask wass used as-is to fill the remainder of the last word with 1's. This meant that if a function first used the negated mask and another function then used the mask as-is (or vice-versa), the first function would then get the wrong result. This patch fixes the problem by changing the implementation of 9 bitmap functions that modified the bitmap state even if the bitmap was declared const. These functions now preserve the internal state of the bitmap. This makes it possible for two connections to concurrently execute two of these functions on the same bitmap without issues. The patch also removes dead testing code from my_bitmap.c. These tests have already been moved to unittest/mysys/bitmap-t.c. Existing test coverage of my_bitmap has been extended. No MTR test case added as this would require adding several sync points to the bitmap functions. The patch has been tested with a non-deterministic test case posted on the bug report.
[6 Apr 2011 1:32]
Paul DuBois
Noted in 5.1.57, 5.5.11, 5.6.2 changelogs. Bitmap functions used in one thread could change bitmaps used by other threads, causing an assertion to be raised. CHANGESET - http://lists.mysql.com/commits/131946
