On 2016-06-01 23:42:55 +0200, Guus Sliepen wrote: > You're skipping over the check for METHOD, which is set to "none" when > IFACE is "--all": > > case "$METHOD" in > static|dhcp|NetworkManager) ;; > *) exit 0 ;; > esac
Oops, yes, I must be a bit tired (it also checks $ADDRFAM, BTW). However, for the postfix script, the only test is "$IFACE" = "lo". What the script does might be harmless, but it's difficult to say whether that's always the case in all configs. Anyway, this is certainly not what is expected. The ethtool also just tests "$IFACE" != "lo". However, I think that if there are no IF_* variables, this is OK. Now, it would be safer and clearer to exit earlier in the --all case. openssh-server is OK but just by chance: According to the comment, it seems that the goal of the $ADDRFAM test was to reject ipx. So, there's a risk that the code gets incorrectly simplified in the future. -- Vincent Lefèvre <vinc...@vinc17.net> - Web: <https://www.vinc17.net/> 100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/> Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
