From mboxrd@z Thu Jan 1 00:00:00 1970 From: Abhishek Sahu Subject: Re: [PATCH 1/2] i2c: qup: Cleared the error bits in ISR Date: Mon, 20 Jun 2016 18:18:41 +0530 Message-ID: <7c979c4e5b5c1652211405ce1e93fe72@codeaurora.org> References: <1462797871-8595-1-git-send-email-absahu@codeaurora.org> <1462797871-8595-2-git-send-email-absahu@codeaurora.org> <000701d1ab9d$df9b3100$9ed19300$@codeaurora.org> <934365f2bd30f242d8548e2a27483679@codeaurora.org> <20160512051347.GE8453@hector.attlocal.net> <20160512061843.GA6187@chewinlnx07.qualcomm.com> <20160512175830.GA32019@hector> <20160512193229.GA10499@chewinlnx07.qualcomm.com> <20160512202755.GA3680@chewinlnx07.qualcomm.com> <20160618163228.GG3017@tetsubishi> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20160618163228.GG3017@tetsubishi> Sender: linux-arm-msm-owner@vger.kernel.org To: Wolfram Sang Cc: Andy Gross , Sricharan , Andy Gross , Archit Taneja , linux-arm-msm , ntelkar@codeaurora.org, Linux Kernel list , linux-i2c@vger.kernel.org, dmaengine@vger.kernel.org, linux-arm-kernel@lists.infradead.org List-Id: linux-i2c@vger.kernel.org On 2016-06-18 22:02, Wolfram Sang wrote: >> We run the command i2cdetect for address 0x3 to 0x77. The QUP >> generates >> write error for address 0x3 to 0x7 apart from other bus errors since >> these are reserved addresses. I was getting the crash in non DMA mode >> and BAM hang in DMA mode before putting the fix. >> >> Also we have connected the I2C TPM and run the following script >> overnight for both DMA and Non DMA mode. The script checks for >> all transfer length and we are generating multiple NACK and >> Non NACK error before each valid transfer. >> >> a=1 >> >> while [ $a -lt 4096 ] >> do >> echo $a >> i2cdetect -y -a -r 0 0x03 0x77 >> tpm-manager get_random $a >> i2cdetect -y -a -r 1 0x03 0x77 >> a=`expr $a + 1` >> done > > So, what is the outcome of this discussion? Discussion was regarding resetting the QUP state in ISR and it is the part of existing code itself. The current patch only added the error checking for bus errors so we can go ahead with this patch. I have shared the script which we are using for testing error path.