From mboxrd@z Thu Jan 1 00:00:00 1970 From: Petr Vorel Date: Mon, 25 Jun 2018 16:01:58 +0200 Subject: [LTP] [PATCH 2/2] network/route: Rewrite route-change-gw into new API In-Reply-To: <20180518110110.zeabusjbiq2lq263@dell5510> References: <20180517103527.23542-1-pvorel@suse.cz> <20180517103527.23542-2-pvorel@suse.cz> <20180518110110.zeabusjbiq2lq263@dell5510> Message-ID: <20180625140158.rvjasl7jq2m4blbc@dell5510> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ltp@lists.linux.it Hi Alexey, > > don't remember why the rewritten test is > > not using background traffic and not changing gateway anymore? > The current implementation was based on your suggestion for route-change-dst [1] (I'll > mention that in commit message), which uses ns-udpsender. The other patch [2] is > changing gateway. > Tests used ns-udpsender in previous implementation and we decided to keep it as use > background traffic isn't sufficient enough [3]. > But ns-udpsender isn't good enough, as it doesn't check whether destination receive it [4]. > And indeed you can pass it to any destination and it will not fail, e.g.: > ns-udpsender -f 4 -D 1.2.3.4 -p 1 -o -s 1472 > I guess we need to write another helper, which will listen for packets on remote side. > Any recommendations/ideas on it? Ping, please. I suppose we want these gateway tests do the real client server communication even it's done on localhost, via tst_ipaddr_un(). I guess we want to keep UDP, as it's faster than TCP. So I'll create some server, which listens the packets, counts them and reports it. Kind regards, Petr > Kind regards, > Petr > [1] http://lists.linux.it/pipermail/ltp/2017-August/005367.html > [2] https://lists.linux.it/pipermail/ltp/2018-May/008152.html > [3] http://lists.linux.it/pipermail/ltp/2017-July/005125.html > [4] http://lists.linux.it/pipermail/ltp/2017-August/005336.html