From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.11) Gecko/20050901 Fedora/1.7.11-4 Description of problem: Could not access network after today's (9/14/05) rawhide update. Question on fedora test list led to cause from Tom London. Backed out iproute-2.6.14-1 and installed iproute-2.6.11-1. Problem fixed. Version-Release number of selected component (if applicable): iproute-2.6.14-1 How reproducible: Always Steps to Reproduce: 1.boot system 2.try to access network, any 3. Actual Results: could not access network Expected Results: should be able to access network Additional info:
NetworkManager runs into this since it uses /sbin/ip to do routing table stuff and a few other things... The result is that NM is unable to add any entries to the routing table. I've yet to figure out what the actual problem here is.
I've tested some ip route options and it seems to work. When exactly it fails? Did you manage to figure this out? I don't use NM here and also I haven't found any problem with iproute so far.
I don't know at which point it fails; however, I reinstalled rawhide iproute-2.6.14-1 and stopped networkmanager and dispatcher, removed them from startup, rebooted, and now the routing table is populated and resolv.conf is correct and the network is reachable. So, seems there is some relationship between networkmanger andf iproute that may be broken.
I now see lots of these messages in /var/log/messages: Sep 15 06:48:58 localhost NetworkManager: nm_ip4_config_get_broadcast: assertion `config != NULL' failed Sep 15 06:48:58 localhost NetworkManager: nm_ip4_config_get_netmask: assertion `config != NULL' failed
Installed today's updates from rawhide (including iproute-2.6.14-1) and I noticed the following behavior: On boot (with networkManager) interface is up (i.e., has address from DHCP, but no routes). If I select the NetowrkManager icon and reselect 'wired network', it now correctly brings up the interface including routes. NetworkManager only sees my wired interface; doesn't see my wireless one. 'iwconfig eth1' says 'no such device', etc. Following messages from /var/log/audit/audit.log: type=AVC msg=audit(1126794443.402:21): avc: denied { read } for pid=2972 comm="killall" name="mls" dev=selinuxfs ino=12 scontext=system_u:system_r:NetworkManager_t:s0 tcontext=system_u:object_r:security_t:s0 tclass=file type=SYSCALL msg=audit(1126794443.402:21): arch=40000003 syscall=5 success=no exit=-13 a0=bfeac3b4 a1=8000 a2=0 a3=8000 items=1 pid=2972 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 comm="killall" exe="/usr/bin/killall" type=CWD msg=audit(1126794443.402:21): cwd="/" type=PATH msg=audit(1126794443.402:21): item=0 name="/selinux/mls" flags=101 inode=12 dev=00:0d mode=0100444 ouid=0 ogid=0 rdev=00:00 type=AVC msg=audit(1126794446.278:22): avc: denied { read } for pid=2991 comm="killall" name="mls" dev=selinuxfs ino=12 scontext=system_u:system_r:NetworkManager_t:s0 tcontext=system_u:object_r:security_t:s0 tclass=file type=SYSCALL msg=audit(1126794446.278:22): arch=40000003 syscall=5 success=no exit=-13 a0=bfafa6b4 a1=8000 a2=0 a3=8000 items=1 pid=2991 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 comm="killall" exe="/usr/bin/killall" type=CWD msg=audit(1126794446.278:22): cwd="/" type=PATH msg=audit(1126794446.278:22): item=0 name="/selinux/mls" flags=101 inode=12 dev=00:0d mode=0100444 ouid=0 ogid=0 rdev=00:00 Not sure if this is related.....
OK, I think I've tracked this down..... NetworkManager is trying to execute 'ip' commands with an argument of 'address'. This appears no longer to be accepted ('addr' is accepted). Looks like either NetworkManager needs to change to use 'addr', or ip needs to change to accept 'address'. Below is a list of the /sbin/ip commands executed by selecting the NetworkManager icon along with any output. route flush dev eth0 address flush dev eth0 Object "address" is unknown, try "ip help". route del default RTNETLINK answers: No such process address flush dev eth0 Object "address" is unknown, try "ip help". route flush dev eth0 Nothing to flush. neigh flush all -6 address add fe80::208:dff:fe89:dbdc/64 dev eth0 Object "address" is unknown, try "ip help". Scripting /sbin/ip to change 'address' to 'addr' makes it work.
Ok, fixed in CVS and we'll try to put out an NM update very very soon. Thanks for the help!
Believe this is closed in RAWHIDE, no?
Yes.