From mboxrd@z Thu Jan 1 00:00:00 1970 From: Junliang Li Subject: SCSI error handle trace Date: Wed, 23 Oct 2013 15:01:47 +0800 Message-ID: <1382511707.2348.36.camel@ljl> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-pa0-f43.google.com ([209.85.220.43]:46158 "EHLO mail-pa0-f43.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751222Ab3JWHB5 (ORCPT ); Wed, 23 Oct 2013 03:01:57 -0400 Received: by mail-pa0-f43.google.com with SMTP id hz1so564579pad.2 for ; Wed, 23 Oct 2013 00:01:57 -0700 (PDT) Received: from [192.168.20.26] ([119.39.124.239]) by mx.google.com with ESMTPSA id oj6sm38724545pab.9.2013.10.23.00.01.54 for (version=SSLv3 cipher=RC4-SHA bits=128/128); Wed, 23 Oct 2013 00:01:56 -0700 (PDT) Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: linux-scsi@vger.kernel.org Hi, all Recently, fnic driver patches a feature that gathers various important stats for debugging issues. Sysadm can find problems via this trace utility, but just for fnic driver. In Linux SCSI subsystem, only a few trace points are set and we can not trace deeply while middle layer handles error. AFAIK, scsi_dispatch_cmd_error is available for running command error in LLD, but how about Fibre Channel error or something like that? A RAS tool named rasdaemon can use trace info to record errors. Now it supports cpu, memory and PCIe error recording. If SCSI subsystem exports more trace points, I think we can support SCSI error recording in rasdaemon. Best regards, Junliang Li