From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933263AbXGRPnV (ORCPT ); Wed, 18 Jul 2007 11:43:21 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1759432AbXGRPnL (ORCPT ); Wed, 18 Jul 2007 11:43:11 -0400 Received: from mail.novatech-llc.com ([64.126.95.172]:65317 "EHLO SERVER1.novatech.novatech-llc.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1759388AbXGRPnK (ORCPT ); Wed, 18 Jul 2007 11:43:10 -0400 X-Greylist: delayed 303 seconds by postgrey-1.27 at vger.kernel.org; Wed, 18 Jul 2007 11:43:10 EDT Thread-Index: AcfJUaJrzFrci7gNQRqqqFp+hxTJyg== X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Importance: normal Content-Class: urn:content-classes:message Message-ID: <469E332C.7050603@novatech-llc.com> Date: Wed, 18 Jul 2007 10:35:08 -0500 From: "George McCollister" User-Agent: Thunderbird 1.5 (X11/20060317) MIME-Version: 1.0 To: Subject: setserial -> serial_core / 8250 interaction Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 18 Jul 2007 15:38:05.0720 (UTC) FILETIME=[A261C980:01C7C951] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I've recently found (using 2.6.21.4) that configuring a serial ports (ST16654) which use the 8250 driver using setserial results in the UART's FIFOs being disabled (unless you specify autoconfig). The problem is that setserial doesn't fill out xmit_fifo_size when calling TIOCSSERIAL to set the port type. There doesn't appear to be any provision to set the port type and have 8250 use the default fifo_size value for the selected FIFO from uart_config. The result is that fifo_size remains 0 and the fifo is disabled. So, I believe that anyone using setserial to configure a serial port (unless autoconfig is used) will have very poor serial port performance (I did, thats how I ran across this). I haven't been able to think of a cure all solution without introducing some pretty nasty changes. You could argue that setserial should set xmit_fifo_size, but it seems unnecessary to require a user level program to have knowledge of the FIFO sizes for all of the different UARTS. I'm going to change my copy of 8250.c to always use fifo_size from uart_config, but some people may not find this acceptable. Maybe I'm missing something here, but I wanted to bring it up in case its actually affecting other people. I'm not subscribed to the list so if you reply please CC to my address. Regards, George McCollister This message is for the designated recipient only and may contain privileged, proprietary, or otherwise private information. If you have received it in error, please notify the sender immediately and delete the original. Any other use of the email by you is prohibited. This message is for the designated recipient only and may contain privileged, proprietary, or otherwise private information. If you have received it in error, please notify the sender immediately and delete the original. Any other use of the email by you is prohibited.