Summary:	ASTERISK-12620: [patch] 'core show sysinfo' on systems that dont have sysinfo but do have sysctl
Reporter:	Michiel van Baak (mvanbaak)	Labels:
Date Opened:	2008-08-20 17:25:16	Date Closed:	2008-09-06 18:40:49
Priority:	Minor	Regression?	No
Status:	Closed/Complete	Components:	Core/Portability
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) 2008082000_core-show-sysinfo-sysctl.diff ( 1) 2008082000_core-show-sysinfo-sysctl-v2.diff
Description:	'core show sysinfo' and the SYSINFO function use the sysinfo call. This call is only present on linux systems. Since I'm using OpenBSD and want at least the memory stats I created a version of 'core show sysinfo' that uses sysctl and swap.h to grab the numbers. ****** ADDITIONAL INFORMATION ****** I tested this on OpenBSD 4.4 (not yet released, but snapshots are available and the CVS tree is tagged) and OpenBSD 4.3 I also tested that the old sysinfo version still works on linux. More test reports needed, as I cannot test it on anything besides linux and openbsd. Right now I do a simple #else in the code that checks for HAVE_SYSINFO. This is not correct and should be some magic #elif If anyone has pointers what it should be please comment on this bug. This is how it looks like on my OpenBSD boxen:      *CLI> core show sysinfo      System Statistics      -----------------        Total Physical memory:     993244 KiB        Total active memory:       218376 KiB        Total used memory:         461752 KiB        Free memory:               506760 KiB        Total Swap Space:          1048005 KiB        Free Swap Space:           1048005 KiB        Number of Processes:       81
Comments:	By: Michiel van Baak (mvanbaak) 2008-08-20 17:59:49 new patch that adds sysctl check to configure.ac thanks to seanbright. TODO: - in main/asterisk.c replace the #if defined(HAVE_SYSINFO).... #else with #if defined(HAVE_SYSINFO)... #elif defined(HAVE_SYSCTL).... - we now have two functions with the same name, depending on the #if #else stuff.  It would be better to seperate the printing and data-gathering and only call the correct data-gathering function based on the defined(HAVE_X) stuff. That way we are not replicating code while there's no use for that. Thanks go to seanbright for the feedback etc.