native: unify and simplify usage of tapsetup script #3182

miri64 · 2015-06-09T21:36:38Z

Started with the intention to make the deactivation of IPv6 optional, ended up overhauling the whole script(s).

From the self-docs:

usage: tapsetup [OPTIONS]

  If no option is given, -c is assumed

  OPTIONS:
  -c [<num>], --create [<num>]: Create <num> tap interfaces (default: 2)
                  -d, --delete: Delete all interface
    -b <name>, --bridge <name>: Give name for the bridge (default: tapbr)
       -t <name>, --tap <name>: name base for the tap interfaces; the
                                generated names will be <name>x
                                (default: tap; ignored on OSX and FreeBSD)
         -6, --deactivate-ipv6: deactivate IPv6 for the interfaces and bridge
                                (ignored on OSX and FreeBSD)
                    -h, --help: prints this text

miri64 · 2015-06-09T21:38:16Z

Should work for Linux, Mac OS X, and FreeBSD. Please test.

LudwigKnuepfer · 2015-06-10T18:41:38Z

cpu/native/tapsetup

+DEACTIVATE_IPV6=""
+
+usage() {
+    echo "usage: $PROGRAM [OPTIONS]" >&2


Please surround all variable names with curly braces, i.e. $PROGRAM -> ${PROGRAM}.

Will do. Any particular reason why?

miri64 · 2015-06-12T07:47:44Z

(this removes the script's dependency on brutils BTW, since it uses iproute2 on Linux for bridge configuration instead of it).

LudwigKnuepfer · 2015-06-13T13:14:04Z

cpu/native/tapsetup

+    echo "                                (default: tap; ignored on OSX and FreeBSD)" >&2
+    echo "         -6, --deactivate-ipv6: deactivate IPv6 for the interfaces and bridge" >&2
+    echo "                                (ignored on OSX and FreeBSD)" >&2
+    echo "                    -h, --help: prints this text" >&2


please make the help options left-aligned

LudwigKnuepfer · 2015-06-13T13:41:14Z

Something fails on FreeBSD:

~/RIOT/examples/default/bin/native/default.elf: open(/dev/tap0): Operation not permitted

LudwigKnuepfer · 2015-06-13T13:44:49Z

set -x on FreeBSD output:

this PR:

+ basename ./cpu/native/tapsetup
+ PROGRAM=tapsetup
+ COUNT=2
+ COMMAND=''
+ BRNAME=tapbr0
+ TAPNAME=tap
+ DEACTIVATE_IPV6=''
+ true
+ BRNAME=bridge0
+ shift 2
+ true
+ break
+ [ -z lo ]
+ [ -z '' ]
+ COMMAND=create
+ uname -s
+ PLATFORM=FreeBSD
+ [ create = create ]
+ create_bridge
+ echo 'creating bridge0'
creating bridge0
+ sudo kldload if_bridge
+ sudo ifconfig bridge0 create
+ begin_tap
+ sudo kldload if_tap
+ seq 0 1
+ create_tap
+ echo 'creating tap0'
creating tap0
+ sudo ifconfig tap0 create
+ sudo chown lo /dev/tap0
+ sudo ifconfig tap0 up
+ sudo ifconfig bridge0 addm tap0
+ create_tap
+ echo 'creating tap1'
creating tap1
+ sudo ifconfig tap1 create
+ sudo chown lo /dev/tap1
+ sudo ifconfig tap1 up
+ sudo ifconfig bridge0 addm tap1
+ up_bridge
+ sudo ifconfig bridge0 up

+ basename ./cpu/native/tapsetup
+ PROGRAM=tapsetup
+ COUNT=2
+ COMMAND=''
+ BRNAME=tapbr0
+ TAPNAME=tap
+ DEACTIVATE_IPV6=''
+ true
+ BRNAME=bridge0
+ shift 2
+ true
+ [ COMMAND = create ]
+ COMMAND=delete
+ shift
+ true
+ break
+ [ -z lo ]
+ [ -z delete ]
+ uname -s
+ PLATFORM=FreeBSD
+ [ delete = create ]
+ [ delete = delete ]
+ delete_bridge
+ echo 'deleting bridge0'
deleting bridge0
+ sudo kldunload if_tap
+ sudo kldunload if_bridge
+ exit 0

master:

+ COMMAND=create
+ COUNT=''
+ DEFCOUNT=2
+ [ -z lo ]
+ [ -z create ]
+ [ create = create ]
+ [ -z '' ]
+ COUNT=2
+ sudo kldload if_tap
+ sudo kldload if_bridge
+ sudo sysctl net.link.tap.user_open=1
net.link.tap.user_open: 0 -> 1
+ echo 'creating  ...'
creating  ...
+ sudo ifconfig bridge0 create
+ sudo ifconfig bridge0 up
+ seq 0 1
+ sudo ifconfig tap0 create
+ sudo chown lo /dev/tap0
+ sudo ifconfig tap0 up
+ sudo ifconfig bridge0 addm tap0
+ sudo ifconfig tap1 create
+ sudo chown lo /dev/tap1
+ sudo ifconfig tap1 up
+ sudo ifconfig bridge0 addm tap1
+ exit 0

+ COMMAND=delete
+ COUNT=''
+ DEFCOUNT=2
+ [ -z lo ]
+ [ -z delete ]
+ [ delete = create ]
+ [ delete = delete ]
+ sudo sysctl net.link.tap.user_open=0
net.link.tap.user_open: 1 -> 0
+ sudo kldunload if_tap
+ sudo kldunload if_bridge
+ exit 0

miri64 · 2015-06-13T23:26:24Z

Can you explain to me, what this output means?

miri64 · 2015-06-13T23:46:21Z

adapted doc, addressed comments and (hopefully) fixed for OSX and FreeBSD.

miri64 · 2015-06-23T13:20:58Z

@LudwigOrtmann ping?

OlegHahm · 2015-06-23T15:03:06Z

@LudwigOrtmann won't be avaliable until July 4.

haukepetersen · 2015-06-23T15:13:12Z

When @authmillenon and me just talked about this PR, it came to my mind if it makes sense to move this script out of the cpu/native location to somewhere more 'toolish', say dist/tools/tapsetup/ or similar. What do you guys think about this?

OlegHahm · 2015-06-23T15:14:04Z

👍

miri64 · 2015-06-23T15:25:12Z

@LudwigOrtmann won't be avaliable until July 4.

Can someone else test this script then on FreeBSD and OSX?

cgundogan · 2015-06-23T15:26:55Z

cpu/native/tapsetup

+    case "${PLATFORM}" in
+        FreeBSD)
+            sudo kldload if_tap || exit 1;;
+            sudo sysctl net.link.tap.user_open=1


my zsh is complaining here. Please move the ;; to the end of this case

/bin/sh is a symlink to zsh on your system?

no it's not (: But still, this script refuses to work without these fixes

miri64 · 2015-06-23T16:37:57Z

Addressed comments and fixed doc accordingly.

miri64 · 2015-07-01T20:36:38Z

@thomaseichinger and @emmanuelsearch any luck testing this on OSX? @phiros any luck testing this on FreeBSD?

emmanuelsearch · 2015-07-09T14:53:41Z

I tested this on OS X, and it works as well as with #3290 (i.e. it sets up something, but ping6 timeouts 100% of the time in both directions). But I guess this PR was not focused on this issue. Basically: I think I can't test this more right now?

miri64 · 2015-07-09T15:03:46Z

Since I don't have access to a OSX machine a more technical test would be far more helpful. This means: use tests/dev_eth with txtsnd instead of ping and a clarification were the packet exactly gets lost. (Can you sniff the packets using wireshark or similar software? )

miri64 · 2015-07-09T15:05:06Z

Or do you observe the same behavior with the old version of the script too. In that case, you are right: this is not the scope of this PR.

LudwigKnuepfer · 2015-07-11T08:29:02Z

Works on FreeBSD 10

miri64 · 2015-07-16T12:06:49Z

Can somebody ACK then?

LudwigKnuepfer · 2015-07-19T19:00:06Z

AcK

miri64 · 2015-08-14T10:05:53Z

Rebased and squashed

native: unify and simplify usage of tapsetup script

miri64 · 2015-08-14T10:55:30Z

And finally merged.

miri64 added Type: enhancement The issue suggests enhanceable parts / The PR enhances parts of the codebase / documentation Platform: native Platform: This PR/issue effects the native platform labels Jun 9, 2015

miri64 added this to the Release NEXT MAJOR milestone Jun 9, 2015

miri64 assigned LudwigKnuepfer Jun 9, 2015

LudwigKnuepfer reviewed Jun 10, 2015
View reviewed changes

LudwigKnuepfer reviewed Jun 13, 2015
View reviewed changes

miri64 modified the milestones: POSIX compliance, Release NEXT MAJOR, Release 2015.06 Jun 15, 2015

cgundogan reviewed Jun 23, 2015
View reviewed changes

miri64 force-pushed the native/enh/tapsetup branch from 99ad86a to f05efed Compare June 23, 2015 16:39

miri64 mentioned this pull request Jul 6, 2015

native: tapsetup script fails on OS X #3320

Closed

miri64 mentioned this pull request Jul 10, 2015

netdev2_tap: crashes when hammered #3341

Closed

OlegHahm added the CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR label Jul 20, 2015

miri64 added the Community: Hack'n'ACK candidate This PR is a candidate for review and discussion during one of RIOT's monthly Hack'n'ACK parties label Jul 27, 2015

miri64 and others added 2 commits August 14, 2015 12:05

native: unify and simplify usage of tapsetup script

5760f90

doc: adapt documentation on tapsetup script

664f51e

miri64 force-pushed the native/enh/tapsetup branch from f05efed to 664f51e Compare August 14, 2015 10:05

miri64 added a commit that referenced this pull request Aug 14, 2015

Merge pull request #3182 from authmillenon/native/enh/tapsetup

99adcbe

native: unify and simplify usage of tapsetup script

miri64 merged commit 99adcbe into RIOT-OS:master Aug 14, 2015

miri64 deleted the native/enh/tapsetup branch August 14, 2015 10:55

miri64 mentioned this pull request Aug 21, 2015

native: create tap interfaces ad-hoc #1473

Closed

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

native: unify and simplify usage of tapsetup script #3182

native: unify and simplify usage of tapsetup script #3182

miri64 commented Jun 9, 2015

miri64 commented Jun 9, 2015

LudwigKnuepfer Jun 10, 2015

miri64 Jun 11, 2015

miri64 Jun 11, 2015

miri64 commented Jun 12, 2015

LudwigKnuepfer Jun 13, 2015

LudwigKnuepfer commented Jun 13, 2015

LudwigKnuepfer commented Jun 13, 2015

miri64 commented Jun 13, 2015

miri64 commented Jun 13, 2015

miri64 commented Jun 23, 2015

OlegHahm commented Jun 23, 2015

haukepetersen commented Jun 23, 2015

OlegHahm commented Jun 23, 2015

miri64 commented Jun 23, 2015

cgundogan Jun 23, 2015

OlegHahm Jun 23, 2015

cgundogan Jun 23, 2015

miri64 commented Jun 23, 2015

miri64 commented Jul 1, 2015

emmanuelsearch commented Jul 9, 2015

miri64 commented Jul 9, 2015

miri64 commented Jul 9, 2015

LudwigKnuepfer commented Jul 11, 2015

miri64 commented Jul 16, 2015

LudwigKnuepfer commented Jul 19, 2015

miri64 commented Aug 14, 2015

miri64 commented Aug 14, 2015

native: unify and simplify usage of tapsetup script #3182

native: unify and simplify usage of tapsetup script #3182

Conversation

miri64 commented Jun 9, 2015

miri64 commented Jun 9, 2015

LudwigKnuepfer Jun 10, 2015

Choose a reason for hiding this comment

miri64 Jun 11, 2015

Choose a reason for hiding this comment

miri64 Jun 11, 2015

Choose a reason for hiding this comment

miri64 commented Jun 12, 2015

LudwigKnuepfer Jun 13, 2015

Choose a reason for hiding this comment

LudwigKnuepfer commented Jun 13, 2015

LudwigKnuepfer commented Jun 13, 2015

miri64 commented Jun 13, 2015

miri64 commented Jun 13, 2015

miri64 commented Jun 23, 2015

OlegHahm commented Jun 23, 2015

haukepetersen commented Jun 23, 2015

OlegHahm commented Jun 23, 2015

miri64 commented Jun 23, 2015

cgundogan Jun 23, 2015

Choose a reason for hiding this comment

OlegHahm Jun 23, 2015

Choose a reason for hiding this comment

cgundogan Jun 23, 2015

Choose a reason for hiding this comment

miri64 commented Jun 23, 2015

miri64 commented Jul 1, 2015

emmanuelsearch commented Jul 9, 2015

miri64 commented Jul 9, 2015

miri64 commented Jul 9, 2015

LudwigKnuepfer commented Jul 11, 2015

miri64 commented Jul 16, 2015

LudwigKnuepfer commented Jul 19, 2015

miri64 commented Aug 14, 2015

miri64 commented Aug 14, 2015