From owner-freebsd-bugs@FreeBSD.ORG Mon May 31 21:22:58 2010 Return-Path: Delivered-To: freebsd-bugs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 9B40B10656BD; Mon, 31 May 2010 21:22:58 +0000 (UTC) (envelope-from mayo@oyam.ca) Received: from mail-pw0-f54.google.com (mail-pw0-f54.google.com [209.85.160.54]) by mx1.freebsd.org (Postfix) with ESMTP id 6E0808FC13; Mon, 31 May 2010 21:22:57 +0000 (UTC) Received: by pwj1 with SMTP id 1so279949pwj.13 for ; Mon, 31 May 2010 14:22:57 -0700 (PDT) Received: by 10.115.36.31 with SMTP id o31mr4015909waj.79.1275339171452; Mon, 31 May 2010 13:52:51 -0700 (PDT) Received: from [192.168.167.102] (a221-smpafs01.blockb-142.stargate.ca [208.118.142.221]) by mx.google.com with ESMTPS id 33sm52673832wad.20.2010.05.31.13.52.49 (version=TLSv1/SSLv3 cipher=RC4-MD5); Mon, 31 May 2010 13:52:50 -0700 (PDT) Mime-Version: 1.0 (Apple Message framework v1078) Content-Type: text/plain; charset=us-ascii From: Mayo Jordanov In-Reply-To: Date: Mon, 31 May 2010 13:52:48 -0700 Content-Transfer-Encoding: quoted-printable Message-Id: References: <201005261942.o4QJgacx011479@freefall.freebsd.org> <008FD1A4-F1AB-4683-A3A6-6EAA3C0AECED@oyam.ca> <7353B9CB-BAEF-4A7E-9C89-6B46227DF779@oyam.ca> To: Garrett Cooper X-Mailer: Apple Mail (2.1078) Cc: remko@freebsd.org, freebsd-bugs@freebsd.org Subject: Re: kern/147082: [uart] Serial ports unusable X-BeenThere: freebsd-bugs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Bug reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 31 May 2010 21:22:58 -0000 On 2010-05-26, at 21:46 , Garrett Cooper wrote: > On Wed, May 26, 2010 at 8:01 PM, Mayo Jordanov wrote: >>=20 >> On 2010-05-26, at 19:34 , Garrett Cooper wrote: >>=20 >>> On Wed, May 26, 2010 at 6:42 PM, Mayo Jordanov wrote: >>>>=20 >>>> On 2010-05-26, at 12:42 , remko@FreeBSD.org wrote: >>>>=20 >>>>> Synopsis: [uart] Serial ports unusable >>>>>=20 >>>>> State-Changed-From-To: open->feedback >>>>> State-Changed-By: remko >>>>> State-Changed-When: Wed May 26 19:42:36 UTC 2010 >>>>> State-Changed-Why: >>>>> I send my working serial line to Mayo, set status to feedback >>>>>=20 >>>>> http://www.freebsd.org/cgi/query-pr.cgi?pr=3D147082 >>>>=20 >>>> I'm gona re-reply here, to preserve the history. >>>>=20 >>>>>>> How-To-Repeat: >>>>>> 1) Edit /etc/ttys, do the ttyu0 line reads: >>>>>> ttyu0 "/usr/libexec/getty std.9600" vt100 on secure >>>>>> 2) kill -HUP 1 >>>>>> 3) Try connecting to it with terminal application >>>>>=20 >>>>>=20 >>>>> /etc/ttys >>>>> console "/usr/libexec/getty Pc" cons25 on secure >>>>>=20 >>>>> /boot/loader.conf >>>>> console=3D"comconsole" >>>>>=20 >>>>> That enabled serial access to my Soekris device running on = FreeBSD8(PFsense), it also works for several remote machines I have. >>>>>=20 >>>>> Can you confirm that you have these settings in loader.conf and = rebooted?! >>>>>=20 >>>>> -- >>>>> /"\ Best regards, | remko@FreeBSD.org >>>>> \ / Remko Lodder | remko@EFnet >>>>> X http://www.evilcoder.org/ | >>>>> / \ ASCII Ribbon Campaign | Against HTML Mail and News >>>>=20 >>>> test# cat /boot/loader.conf >>>> console=3D"comconsole" >>>>=20 >>>> looking at the /etc/ttys entry you pasted above, will that enable = serial console? That looks like a regular console on the machine itself. >>>>=20 >>>> I have rebooted numerous times, and I do get the kernel messages = during boot on serial up to the point where getty starts (presumably... = I see portion of the messages as nics are getting setup and then = nothing). >>>>=20 >>>> I have some machines where serial works fine, but I have some = machines where it doesn't and behaves the same way as described above. = If necessary, I can provide access to a troublesome machine. >>>=20 >>> All of this is described in more grusome detail at >>> = http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/serialconsole-se= tup.html >>=20 >> Unfortunately, the handbook does not help in this case. If there is a = solution to the problem I describe in the PR description in the = handbook, please be so kind and point it out to me, because I must be = blind. >>=20 >> Also, why do the gettys get stuck and I have to kill them manually = (or reboot because they can't be killed) to reflect changes to /etc/ttys = instead of simple `kill -HUP 1`? I've never ever had to do it manually = in previous FreeBSD installations/versions. Why are they stuck in ttydtd = and not in ttyin as they are supposed to be? >=20 > The comconsole bit is in the section of /etc/ttys in the > beforementioned page. > Details about debugging your connection is available here: > http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/term.html . > What's the baud rate for your serial connection? > The part about kill -HUP 1 not working sounds like a possible > syscons/tty bug. Can you try killing init, then provide us with the > wait state in ps/top for the relevant getty instance? > Thanks! > -Garrett Garrett, Remko: Sorry to bother you about this again, but do you guys have any other = ideas for me to try? I've been looking through the documentation = (http://oyam.ca/temp/ark3360.pdf - page 14 on paper/page, page 26 in = pdf), and according to the docs all wires in the serial connector are = there, yet putting the port to either std or 3wire mode does not work. = Putting it in std goes to dcd right away, and only makes it in ttyin = when I reboot, followed by killing getty. The new respawned one will be = in ttyin, but only until I try connecting to it. Putting it in 3wire = waits in ttyin, but no messages (other than kernel generated messages - = on vga console, those that are in white) make it though. When booting = though, I get the full boot loader and things work as expected until a = certain port during system startup, where the console goes dead. It gets = revived at some point during shutdown process again and I see the final = shutdown messages. I'm fairly desparate to get FreeBSD going on this device, but this = single thing is preventing me from going forward. I tested couple of = other OSs (Debian, Windows, QNX), and seem to have no issues there, = serial working as expected. I'd be willing to sponsor a dev to look at = it and figure out what's going on with the driver/getty. I can't provide = much, as it would be coming out of my own pocket, but I'll be happy to = help in any way I can and provide any incentives I can :) I don't mind = using 3wire setup, it's all we need, but I need more than just kernel = generated messages making it though. I need to use the serial during and = after boot process. Even if the getty isn't working, I desperately need = it working between two terminal applications (which at this point isn't = working either). I know both of you are really busy and have better things to do than = deal with broken serial, but any suggestions and/or ways to debug (just = to figure out what's going on) would be very welcome. Thank you both for your time, mayo