| Summary: | ASTERISK-18479: ast_manager_register_struct attempts to unlock an uninitialized rwlock | ||||||
| Reporter: | Jaco Kroon (jkroon) | Labels: | |||||
| Date Opened: | 2011-09-08 09:32:15 | Date Closed: | 2011-10-11 13:29:25 | ||||
| Priority: | Major | Regression? | |||||
| Status: | Closed/Complete | Components: | Core/ManagerInterface | ||||
| Versions: | 1.8.6.0 | Frequency of Occurrence | Constant | ||||
| Related Issues: |
| ||||||
| Environment: | Attachments: | ||||||
| Description: | Hi, Whilst trying to track another issue, helgrind came across this during asterisk startup. ==9900== Thread #1 unlocked an invalid lock at 0x7E7DB0 ==9900== at 0x4C2B686: pthread_rwlock_unlock (hg_intercepts.c:1592) ==9900== by 0x4DC0E1: ast_manager_register_struct (manager.c:5036) ==9900== by 0x4EF7A4: ast_manager_register2 (manager.c:5088) ==9900== by 0x498A74: ast_data_init (data.c:3290) ==9900== by 0x440F79: main (asterisk.c:3740) ==9900== ==9900== Thread #1 unlocked an invalid lock at 0x7E7DB0 ==9900== at 0x4C2B686: pthread_rwlock_unlock (hg_intercepts.c:1592) ==9900== by 0x4DC0E1: ast_manager_register_struct (manager.c:5036) ==9900== by 0x4DC712: ast_manager_register2.clone.19 (manager.c:5088) ==9900== by 0x4EE491: __init_manager (manager.c:6244) ==9900== by 0x440FB4: main (asterisk.c:3759) ==9900== ==9900== Thread #1 unlocked an invalid lock at 0x7E7DB0 ==9900== at 0x4C2B686: pthread_rwlock_unlock (hg_intercepts.c:1592) ==9900== by 0x4DC0E1: ast_manager_register_struct (manager.c:5036) ==9900== by 0x4DC712: ast_manager_register2.clone.19 (manager.c:5088) ==9900== by 0x4EE4A2: __init_manager (manager.c:6245) ==9900== by 0x440FB4: main (asterisk.c:3759) This repeats for pretty much every call coming into manager.c to register new actions, similar to: ==9900== Thread #1 unlocked an invalid lock at 0x7E7DB0 ==9900== at 0x4C2B686: pthread_rwlock_unlock (hg_intercepts.c:1592) ==9900== by 0x4DC0E1: ast_manager_register_struct (manager.c:5036) ==9900== by 0x4EF7A4: ast_manager_register2 (manager.c:5088) ==9900== by 0x23CC63E4: load_module (chan_sip.c:29567) ==9900== by 0x4D4262: start_resource (loader.c:785) ==9900== by 0x4D48A6: load_resource_list (loader.c:973) ==9900== by 0x4D62D4: load_modules (loader.c:1126) ==9900== by 0x441083: main (asterisk.c:3824) | ||||||
| Comments: | By: Richard Mudgett (rmudgett) 2011-10-11 13:19:50.255-0500 This helgrind report is a false positive. That lock is statically initialized. I have changed the code to not use the AST_RWLIST_TIMEDWRLOCK() anymore. (Added because of ASTERISK-13784) However, in examining the code to see what could be done about it I found several other issues. | ||||||