Date: Tue, 8 Sep 2015 21:52:38 -0700 From: "M. V." <bored_to_death85@yahoo.com> To: Artem Belevich <artemb@gmail.com> Cc: "freebsd-net@freebsd.org" <freebsd-net@freebsd.org> Subject: Re: Problem with receiving packets right after remote-interface is up Message-ID: <1441774358.73226.YahooMailBasic@web165005.mail.bf1.yahoo.com> In-Reply-To: <CAFqOu6ihCTrOiTppYA%2B%2BGkTD_n1e%2BUTiV6x2TznVUp65mno2rA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi guys, On Sun, 9/6/15, Artem Belevich <artemb@gmail.com> wrote: > While you can indeed set speed/duplex manually, you will also need > to make sure both have have proper clock master/slave > =A0selection =A0which is normally done via autonegotiation. On > Freebsd some interfaces support "mediaopt master" > option .. Thanks again for your answer. I continuted tests on this yesterday, and I f= ound root of the problem. As I told you, I checked ifconfig and there was n= o support for changing master/slave option. Yesterday I tested to see if th= is master/slave option affects my tests. I found out every time interface o= f remote system gets down and up, If our interface gets slave, everything i= s ok, but anytime our interface gets master, we don't receive sent packets = in the first second after getting up. So I changed driver code to manualy force interface to get slave (by changi= ng values of 82574's related registers) and my problem was solved completel= y (even with auto-neg. enabled) It seems getting master needs extra work and extra time to setup in hardwar= e level ('cause I didn't find any code in driver to do things differently w= hen we are master).=20 So my question is,=20 - Is there anything I can do to make my scenario work in master-mode too? (= interestingly cisco router doesn't have this problem in both master/slave m= odes) - Is it OK to force interface to always gets slave like this? (I found no p= roblem in my tests. I even checked what happens if 2 connected ports both w= ere forced to be slave. In that situation one of the ports gets master even= though we manually set it to slave, so connection is OK) Thank you. > On Saturday, September 5, 2015, M. V. via freebsd-net > < freebsd-net@freebsd.org> wrote:=20 >> Our product is being tested with Spirent TestCenter, >> and we're facing an unusual problem with the tests. >> We use NICs with intel 82574 and 82576 on FreeBSD 9.2 >> with latest em and igb drivers (we also tested this on=20 >> FreeBSD-10.1) It seems what Spirent TestCenter does to=20 >> start any individual test is, it disables its own interface,=20 >> and at the beginning of the new test, it suddenly "up"s its=20 >> interface and sends test packet right after that. >> This is where we have problem, and we don't receive this >> first packets most of the time (result is vary, in 100 tests,=20 >> we lose about 60~70% of this "first" packets on each test, >> so we FAIL most of tests because apparently we need >> about 0.5~1 seconds after setup and renegotiation >> before we can receive packets) =20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1441774358.73226.YahooMailBasic>