Running tcpdump on a non-existing or unconfigured device crashes phone (only in terminal app, not ssh)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Ubuntu Terminal App |
New
|
Undecided
|
Unassigned |
Bug Description
Running tcpdump twice on a network device that doesn't exist or is simply configured down crashes the phone:
tcpdump -ni abc
On first attempt, it exits right away
On second attempt, the phone freezes, then reboots
This only happens when commands issued in terminal, not in an ssh session.
After the first attempt, the following is logged to the kernel log:
[ 161.951328] type=1400 audit(145410049
[ 161.951428] type=1400 audit(145410049
[ 161.951537] type=1400 audit(145410049
[ 162.087330] type=1400 audit(145410049
[ 162.087433] type=1400 audit(145410049
[ 162.087519] type=1400 audit(145410049
[ 162.368152] [lcm_esd_
root@alains-
Actually, this happens also on existing and configured interfaces, with the following difference in the first call:
1. If device doesn't exist, first call exits right away, without any error message (... but with an exit status of 1).
2. If the device *does* exist, commands keeps running (as it should), but doesn't show any traffic, even if there is actually traffic.
Second call always leads to crash (in case #2, interrupt first call with Control-C)
Again, this only happens in a terminal session. In an ssh session you do get an error message when trying to tcpdump a non-existing device, and you do see traffic if you tcpdump an existing device which has traffic.