From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Boyd Subject: Re: [RFC PATCH] tty: serial: msm_serial: Don't reset uart on set_termios Date: Wed, 15 Jun 2016 18:11:33 -0700 Message-ID: <5761FCC5.3010709@codeaurora.org> References: <1465844571-1229-1-git-send-email-bjorn.andersson@linaro.org> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1465844571-1229-1-git-send-email-bjorn.andersson@linaro.org> Sender: linux-kernel-owner@vger.kernel.org To: Bjorn Andersson , Andy Gross , David Brown Cc: Greg Kroah-Hartman , Jiri Slaby , "open list:ARM/QUALCOMM SUPPORT" , "open list:ARM/QUALCOMM SUPPORT" , "open list:SERIAL DRIVERS" , open list List-Id: linux-arm-msm@vger.kernel.org On 06/13/2016 12:02 PM, Bjorn Andersson wrote: > Upon opening the tty, uart_open() ends up calling msm_set_baud_rate() > which resets the uart block. If this happens as we're coming out of > msm_console_write() a full fifo worth of console output will be > discarded. > > Cc: Stephen Boyd > Signed-off-by: Bjorn Andersson > --- > > As reported here: > https://bugs.96boards.org/show_bug.cgi?id=378 Urgh. As mentioned in commit a12f1b406f2d (tty: serial: msm: Reset uartdm after baud rate change, 2014-10-29) we actually need to reset the hardware sometimes. Perhaps as discussed over IRC we need to take a different approach here and only reset the hardware if the baud actually changes? One way to test this would be to try running a getty on the console and see if input still works. -- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project