The following bugs are fixed in this fork of nicstat.
This fixes http://sourceforge.net/p/nicstat/bugs/1/ by applying the patch provided by Darren in that that bug report.
This bug occurs due to an integer overflow for 10Gb interfaces and is not specific to RHEL.
This fixes http://sourceforge.net/p/nicstat/bugs/3/ by applying the patch I provided in that that bug report.
This patch may not be ideal, and should be revisited for completeness, but does resolve the issue for me.
This fixes http://sourceforge.net/p/nicstat/bugs/4/ by applying the patch I provided in that that bug report.
When utilization is calculated, nicp->duplex checks for the value of "2" for a full duplex link, but on Linux this value is "1" (see /usr/include/linux/ethtool.h). Fix is to use DUPLEX_FULL #define.
This fixes http://sourceforge.net/p/nicstat/bugs/5/ by applying the patch I provided in that that bug report.
An #ifdef for linux vs solaris is not needed as the DUPLEX_ #defines can be used consistently instead.
This fixes http://sourceforge.net/p/nicstat/bugs/6/ by applying the patch I provided in that that bug report.
There were too many arguments to fscanf(), resulting in the following warning when compiling with -Wall:
nicstat.c: In function ‘load_snmp’: nicstat.c:1569:8: warning: too many arguments for format [-Wformat-extra-args]
This fixes http://sourceforge.net/p/nicstat/bugs/7/ by applying the patch I provided in that that bug report.
There are two unused variables in Linux (one is used only for solaris, the other used nowhere), found when compiling with -Wall.