Summary: | ASTERISK-05999: [patch] checking for wrong error value of ast_pthread_create | ||
Reporter: | Tzafrir Cohen (tzafrir) | Labels: | |
Date Opened: | 2006-01-06 17:10:11.000-0600 | Date Closed: | 2006-01-07 00:30:12.000-0600 |
Priority: | Minor | Regression? | No |
Status: | Closed/Complete | Components: | Core/General |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ( 0) pbxspool_pthread_ret.dpatch | |
Description: | ast_pthread_create returns the return of pthread_create . according to its man page This may be -EAGAIN if there are not enough resources to create the thread. EAGAIN defined to be 11 on my system . pbx_spool.c generates a thread and cheks if the result of ast_create_thread is not -1 . Thus any error value other than -1 will go unnoticed. | ||
Comments: | By: Jason Parker (jparker) 2006-01-06 17:23:52.000-0600 Could this not be simplified as follows? if (ast_pthread_create(&t,&attr,attempt_thread, o)) { By: Tzafrir Cohen (tzafrir) 2006-01-06 17:47:54.000-0600 Wouldn't you want to record the error that casued the thread generation to fail? Anyway, is it a good idea to generate exatly the same error message in two separate code locations? Makes tracing the problem from the logs more complicated. By: Jason Parker (jparker) 2006-01-06 17:51:56.000-0600 As for the first question: grep -r ast_pthread_create * | grep "=" | wc -l 3 As for the second...yeah, you're probably right. By: Tilghman Lesher (tilghman) 2006-01-07 00:29:39.000-0600 Committed to 1.2, merged to trunk |