Hide Forgot
Description of problem: Broadcast address is wrong for /31 networks, this will cause misconfiguration of ethernet devices. Version-Release number of selected component (if applicable): initscripts-9.03.38-1.el6.centos.1.i686 How reproducible: Always Steps to Reproduce: ipcalc -b 10.10.10.10/31 Actual results: BROADCAST=10.10.10.11 Expected results: BROADCAST=255.255.255.255 Additional info: [root@fmc ~]# cat /etc/sysconfig/network-scripts/ifcfg-eth1 DEVICE="eth1" BOOTPROTO="static" HWADDR="00:0C:29:2F:92:7C" IPV6INIT="no" MTU="1500" NM_CONTROLLED="no" TYPE="Ethernet" ONBOOT=yes IPADDR=10.81.81.81 NETMASK=255.255.255.254 Note that "BROADCAST=255.255.255.255" needs to be manually inserted if you want your ethernet configured correctly, actually it should not require that because the correct result is well defined already. [root@fmc ~]# strace -o /tmp/fff -f /sbin/ifup eth1 [root@fmc ~]# fgrep ipcalc /tmp/fff 9100 execve("/bin/ipcalc", ["/bin/ipcalc", "--prefix", "10.81.81.81", "255.255.255.254"], [/* 18 vars */]) = 0 9101 execve("/bin/ipcalc", ["/bin/ipcalc", "--broadcast", "10.81.81.81", "255.255.255.254"], [/* 18 vars */]) = 0 9106 execve("/bin/ipcalc", ["/bin/ipcalc", "--network", "10.81.81.81", "255.255.255.254"], [/* 18 vars */]) = 0 You can see the script depends on ipcalc to get the answer right... and so it should get the answer right.
http://tools.ietf.org/html/rfc3021
Can you please try this patch: diff --git a/src/ipcalc.c b/src/ipcalc.c index 7316f05..416180b 100644 --- a/src/ipcalc.c +++ b/src/ipcalc.c @@ -141,7 +141,11 @@ struct in_addr calc_broadcast(struct in_addr addr, int prefix) struct in_addr broadcast; memset(&broadcast, 0, sizeof(broadcast)); - broadcast.s_addr = (addr.s_addr & mask.s_addr) | ~mask.s_addr; + + if (mask.s_addr == htonl(0xFFFFFFFE)) + broadcast.s_addr = htonl(0xFFFFFFFF); + else + broadcast.s_addr = (addr.s_addr & mask.s_addr) | ~mask.s_addr; return broadcast; }
I replied to the Bugzilla email, but does not seem to have updated... Yes the above patch worked for me (I tested by rebuilding SRPM). May I suggest a one-line comment in the source code mentioning RFC3021 (so people can find a reference explaining why it works that way, also in case anyone in future greps out the code looking for the RFC number) ?
https://git.fedorahosted.org/cgit/initscripts.git/commit/?h=rhel6-branch&id=7a7ab0c
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2015-1380.html