https://bugzilla.novell.com/show_bug.cgi?id=235244 dieter.jurzitza@t-online.de changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dieter.jurzitza@t-online.de ------- Comment #1 from dieter.jurzitza@t-online.de 2007-01-17 11:59 MST ------- I (think) I found the root cause for the issues I have discovered with the ralink Wlan card (other people suffer from similar problems as can be found in the internet). Debugging yielded the following results: - "/lib/udev/rename_netiface ra0" is always called on start. This differs from the call for a "normal" ethernet card "/lib/udev/rename_netiface eth0 eth0". Please note the missing second argument in the first call example. - the further steps in the script detect a valid MACADDRESS that is written into $DEV_ID. Finally check_if_name_is_free() is being called with the name ra0 as parameter. There you read: grep -qs "^[^#].*rename_netiface %k $NEWNAME\" $RULE_FILE && return 1 A persistent rule had had been put into the rule - file /etc/udev/rules.d/30-net_persistent_names.rules during configuration of the interface via YaST (by the way, this happens for each interface that is being configured, isn't it?): SUBSYSTEM=="net", ACTION=="add", SYSFS{address}=="00:0e:2e:b0:c1:9b", IMPORT="/lib/udev/rename_netiface %k ra0" This appears meaningful and right. Now exactly this entry is being used, to decide, that the interface of this card should not be used, because the "grep" command above returns 1 telling the script that the ID of this card should be incremented and the call should be repeated. This must be an error. By any means this rule leads to the fact that at every boot new entries are being written into the rule file, each of them referring to the very same interface. IMHO this cannot be wanted this way. However, the scripts expects network cards that do not come with two interfacenames when calling "/lib/udev/rename_netiface", thus the behaviour of my network card appears to be a regular case. To fix this, a patch has been created. This patch does basically two things: 1.) I check whether there is an entry in the rule file pointing to exactly this mac-address with exactly this interface name. If so, I assume that everything is ok and no further action is required, the call to check_if_name_is_used() returns 0 then and the further processing terminates silently. 2.) Given this (retval 0 of check_if_name_is_used() ) the new old rule should not be appended to the rulefile. Therefore I set the variable $WRITE_RULE to "no". Please take a look into the patch attached to the bug report for further details. Some more comments: In line 143 there is an inconsistency: the line 93 sets NEWNAME = $2. In line 132 $NEWNAME is checked for "". In line 143 I see the call get_device_id "$OLDNAME" "$2" hence with $2 == $NEWNAME == "" as parameter - that seems quite meaningless and irritating, isn't it? In the comment in line 138 the bug devil had bitten someone, it should say "30-net..." instead of "60-net ...". -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.