Summary: | Ifconfig -alias deletes unspecified IP address | ||
---|---|---|---|
Product: | Base System | Reporter: | Rodney W. Grimes <rgrimes> |
Component: | bin | Assignee: | Devin Teske <dteske> |
Status: | Closed Unable to Reproduce | ||
Severity: | Affects Many People | CC: | db, emaste, melifaro, pat, rgrimes |
Priority: | --- | ||
Version: | CURRENT | ||
Hardware: | Any | ||
OS: | Any | ||
See Also: | https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=228987 |
Description
Rodney W. Grimes
2018-06-12 09:17:14 UTC
I just want to make sure this doesn't get missed. It also caused my RPi3 to panic because of dhclient as dteske noted at BSDCON we have this in dhclient-script delete_old_address() { eval "$IFCONFIG $interface inet -alias $old_ip_address $medium" } (In reply to Diane Bruce from comment #1) Can you provide more details on how I would set up my RPi3 to duplicate your panic. A kernel panic caused by an ifconfig command is probably a seperate bug, as this bug mostly describes bad behavior in the parser of ifconfig that is accepting the -alias command with no address given, and attempting to make some change when it should just return a syntax error with usage(). Argh, now re-reading the initial notes this is why the (maybe two) is there. I have create a seperate bug for the panic, and add cross "see also" pointers on both PR's. (PR228987) Diana can you provide any help information to the new bug I created, which I added you to cc: at PR228987 `ifconfig fooX -alias address` should delete the exact address provided to the ifconfig. IF the `address` argument is empty, `ifconfig` will delete the first IPv4 address set by this interface. Could you please provide more details on the problem happening here? I believe it actually panic'ed my machine at the time but that bug has long been fixed and this PR should be closed IMO. As far as I am concerned this can be closed. I tested this on FreeBSD 13.0-RELEASE yesterday. I am happy to report that everything is fine. Progress! Closing as it is assigned to me. Can always re-file if I hit it again, but going through the comments and historical information (and recent testing on [presumably] supported release [13.0]) tends to support closing. |