From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sebastian Andrzej Siewior Subject: Re: [PATCH 2/2] tty: serial: 8250_core: Check that port->line is >=0 Date: Fri, 16 Jan 2015 12:33:09 +0100 Message-ID: <54B8F6F5.1030204@linutronix.de> References: <3293c93689f5d8c684133a9d9952411a201609dc.1421333585.git.michal.simek@xilinx.com> <54B8D32E.5000805@linutronix.de> <54B8ED20.60308@linutronix.de> <3ecb2012312c4af5bbd395391aa82921@BY2FFO11FD005.protection.gbl> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 8bit Return-path: In-Reply-To: <3ecb2012312c4af5bbd395391aa82921@BY2FFO11FD005.protection.gbl> Sender: linux-kernel-owner@vger.kernel.org To: Michal Simek , linux-kernel@vger.kernel.org, monstr@monstr.eu, Olof Johansson Cc: Greg Kroah-Hartman , linux-serial@vger.kernel.org, Jiri Slaby , Peter Hurley , Alan Cox , Tony Lindgren , Ricardo Ribalda Delgado , Andy Shevchenko , Ingo Molnar List-Id: linux-serial@vger.kernel.org On 01/16/2015 12:02 PM, Michal Simek wrote: >>> Origin patch looks good to me but this checking will be good to add. >>> Are you using of_serial.c because I didn't find any of_alias_get_id call >>> for 8250? >> >> I'm using of_alias_get_id() in 8250_omap.c which made it into >> v3.19-rc1. I think the change you mention made it one release earlier. > > I see > serial: of-serial: fetch line number from DT > commit 1bd8324535ec1ff44aef55c0e40b9e7d56b310fb > > but it was reverted > > Revert "serial: of-serial: fetch line number from DT" > commit e4c787dac21e50ac9ef53fac91b9182d4fede465 > > Olof: what was the reason for that breakage? > Is it missing alias list on Tegra? > which caused that port->line number is bogus. I doubt it, because a missing alias would return -1 and would not be considered for port->line. I think it is more what happens if you use the second UART and the primary is unused (as in not probed). With this change the console will be on ttyS1 - without it it remains on ttyS0. The former is what we wanted. And it seems they use for "nvidia,tegra[23]0-hsuart" the serial-tegra.c driver which makes use of of_alias_get_id() but the of_serial() one which is used by nvidia,tegra20-uart() does not. And since this change does not notify anyone during upgrade but removes the console everyone gets a little grumpy*. * and I've been looking for a smooth solution for the ttyOx -> ttySx change in OMAP just to avoid people complaining and risking a revert :) > Thanks, > Michal Sebastian