~lttng/urcu/trunk

From Revision 761 to 742

expand all

collapse all

Rev		Summary	Authors	Date	Diff	Files
761		Move API.txt to doc/rcu-api.txt, install in system doc/ Move API.txt to doc/rcu-api.txt, install in system doc/ Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
760		rculfhash: document implied memory barriers rculfhash: document implied memory barriers We choose to provide full memory barriers before and after successful hash table update operations. Eventually, new API with weaker semantic can be added, but let's make the basic API as fool-proof as possible. Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Reviewed-by: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>	Mathieu Desnoyers	12 years ago
759		rculfhash: Ensure future-proof memory barrier semantic consi... rculfhash: Ensure future-proof memory barrier semantic consistency Use cmm_smp_mb__before_uatomic_or() prior to the uatomic_or() in _rcu_lfht_del() to ensure correct memory barrier semantic when we relax (in the future) the barrier implementation of some architectures. Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
758		API cleanup: use "uatomic_*" in cmm_smp_mb__ API API cleanup: use "uatomic_*" in cmm_smp_mb__ API Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
757		uatomic: add memory barrier API for and/or/add/sub/inc/sub uatomic: add memory barrier API for and/or/add/sub/inc/sub Implement: cmm_smp_mb__before_and, cmm_smp_mb__after_and cmm_smp_mb__before_or, cmm_smp_mb__after_or cmm_smp_mb__before_add, cmm_smp_mb__after_add cmm_smp_mb__before_sub, cmm_smp_mb__after_sub cmm_smp_mb__before_inc, cmm_smp_mb__after_inc cmm_smp_mb__before_dec, cmm_smp_mb__after_dec For generic and x86. These currently translate into simple compiler barriers on all architectures, but the and/or/add/sub/inc/dec uatomics do not provide memory ordering guarantees (only uatomic_add_return, uatomic_sub_return, uatomic_xchg, and uatomic_cmpxchg provides full memory barrier guarantees before and after the atomic operations). Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
756		rculfhash: add runhash.sh test script rculfhash: add runhash.sh test script Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
755		rculfhash tests: add missing check rculfhash tests: add missing check We need to check if test_ht is NULL. Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
754		rculfhash: fix: race between replace and del operations rculfhash: fix: race between replace and del operations Bug introduced by commit db00ccc36e7fb04ce8044fb1be7964acd1de6ae0 Here is the race: Initially in hash table:  A T0                                         T1 replace A by B                                            del A   read A->next   -> check REMOVED flag, not set yet.                                            read A->next                                            -> check REMOVED flag, not set yet.   cmpxchg A->next to set REMOVED flag   -> cmpxchg succeeds                                            uatomic_or to set REMOVED flag                                            uatomic_xchg to atomically set the REMOVAL_OWNER flag                                            -> first to set the flag.   Replace returns node -> free(A)          Del success -> free(A) With this race, we have a double-free. The problem with the replace code is that it does not set the "REMOVAL_OWNER" flag. Test case to reproduce the bug: test_urcu_hash 0 2 20 -A -s -M 1 -N 1 -O 1 (2 threads, doing replace/del, with a hash table that has only a single key for all values). After just a couple of seconds, either the program hangs, or, more often, it does: *** glibc detected *** /media/truecrypt1/compudj/doc/userspace-rcu/tests/.libs/test_urcu_hash: malloc(): memory corruption (fast): 0x00007ffff3a29e25 *** Program received signal SIGSEGV, Segmentation fault. Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Tested-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
753		rculfhash: replace unneeded rcu_dereference by CMM_LOAD_SHAR... rculfhash: replace unneeded rcu_dereference by CMM_LOAD_SHARED The difference between the two is that CMM_LOAD_SHARED() does not imply a read barrier between the read and following uses of the data pointed to by the pointer read. All sites that only use the pointer load for its bits (never dereference) don't need the read barrier implied by rcu_dereference. Reviewed-by: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
752		rculfhash: use do {} while (0) for dbg_printf() rculfhash: use do {} while (0) for dbg_printf() found by clang(make CC=clang). avoid empty statement. ------------------------- if (condition)         dbg_printf()  /* forget ";", but compiler say nothing if dbg_printf() is empty */ statement; ------------------------- also add printf format check. (we can use gcc extention "__printf(1, 2)" to declare a dummy inline function to do the check, but I use "printf()" directly here) Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com> Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
751		rculfhash: cleanup typo rculfhash: cleanup typo Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
750		rculfhash: update API comments rculfhash: update API comments Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
749		rculfhash: update comments in implementation rculfhash: update comments in implementation Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
748		rculfhash tests: add long hash chains tests rculfhash tests: add long hash chains tests Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
747		rculfhash tests: add uniqueness test rculfhash tests: add uniqueness test Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
746		rculfhash test: print test name rculfhash test: print test name Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
745		rculfhash: stress requirement in documentation rculfhash: stress requirement in documentation Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
744		rculfhash: fix typo rculfhash: fix typo Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
743		rculfhash tests: use array of callbacks to modularize rculfhash tests: use array of callbacks to modularize Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago
742		rculfhash tests: modularize tests rculfhash tests: modularize tests Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>	Mathieu Desnoyers	12 years ago

• « Newer
• Older »

Loggerhead is a web-based interface for Breezy
Version: 2.0.1