Fix tunslip6

[simonduq]

This PR reverts 6608b62 which seems to break tunslip6 at least for examples/ipv6/rpl-border-router. It also disables by default the IP address inquiry option, not supported by the BR.

[mcr]

Could you please detail how rpl-border-router is broken? It is precisely that example that needs this.

[simonduq]

The symptom is that on linux you get

tunslip6: serial_to_tun: write: Invalid argument
ifconfig tun0 down
netstat -nr | awk '{ if ($2 == "tun0") print "route delete -net "$1; }' | sh
make: *** [connect-router] Error 1

(was reported to me by two different people and I could reproduce).

I didn't dig enough to fully understand why this occurs but the problem I noticed is that with the new conditionals, https://github.com/contiki-os/contiki/blob/master/tools/tunslip6.c#L275 is reached even if uip.inbuf[0] == DEBUG_LINE_MARKER or is_sensible_string(uip.inbuf, inbufptr) (unwanted behavior I believe).

The other problem fixed here is that tunslip6 kept sending ?IPA to which the BR replied Got request message of type I without taking any further action

[mcr]

I will retest things this week: I have been running the code for two years without those problems (on Linux).

[simonduq]

OK, thanks! I hope you can reproduce the same issue.

[fvdnabee]

I also experienced tunslip6 being broken in exactly the same way as Simon. I didn't manage to diagnose the problem (a work around I found was setting verbose > 2).
So, I tested commit 30f69b0 (in cooja with one z1 rpl-border-router) and tunslip6 doesn't crash with "tunslip6: serial_to_tun: write: Invalid argument" and is working fine.

Also, +1 for commit a9a46b2 as per the reference to PR #676.

[bkozak-scanimetrics]

I've also been seeing this exact same problem since 6608b62. Also, since this commit I'm seeing a lot of the following.

*** dropping large 2000 byte packet

[simonduq]

@mcr, any news? More and more people are reporting this issue, would you mind if we merge this now?

[simonduq]

@mcr, merging as tunslip6 currently broken for several users, please open a new PR if you come up with a better fix.

[mcr]

Simon Duquennoy notifications@github.com wrote:
> @mcr, any news? More and more people are reporting this issue, would you mind
> if we merge this now?

Go ahead. I have been ill for a week.
I'll work something out on top of the new code.

] Never tell me the odds! | ipv6 mesh networks [
] Michael Richardson, Sandelman Software Works | network architect [
] mcr@sandelman.ca http://www.sandelman.ca/ | ruby on rails [