From: Cristian Stoica Subject: Re: [PATCH v2] crypto: caam: fix error reporting Date: Thu, 6 Nov 2014 10:01:08 +0200 Message-ID: <545B2AC4.3090505@freescale.com> References: <20141104105731.ccb0eedea154b9bb6ed195b3@freescale.com> <1415179284-19791-1-git-send-email-cristian.stoica@freescale.com> <20141105104332.bdc904792193d714bd9d4951@freescale.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Cc: , To: Kim Phillips , , , , Return-path: Received: from mail-bn1bon0148.outbound.protection.outlook.com ([157.56.111.148]:23785 "EHLO na01-bn1-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750980AbaKFIBU (ORCPT ); Thu, 6 Nov 2014 03:01:20 -0500 In-Reply-To: <20141105104332.bdc904792193d714bd9d4951@freescale.com> Sender: linux-crypto-owner@vger.kernel.org List-ID: Hi Kim, Herbert, On 11/05/2014 06:43 PM, Kim Phillips wrote: > On Wed, 5 Nov 2014 11:21:24 +0200 > Cristian Stoica wrote: > >> The error code returned by hardware is four bits wide with an expected >> zero MSB. A hardware error condition where the error code can get between >> 0x8 and 0xf will trigger an out of bound array access on the error >> message table. >> This patch fixes the invalid array access following such an error and >> reports the condition. >> >> Signed-off-by: Cristian Stoica >> --- > no v2 change info? All I noticed is the additional string for > "queue manager interface", which, without its implementation fn, Yes. Thanks for review. That and the updated comment. > intoduces an inconsistency wrt NULL checking, so this comment: No. There is no inconsistency in the handling. Please be more specific. > > http://www.mail-archive.com/linux-crypto@vger.kernel.org/msg12105.html > > still applies. Which comment? You refer to a whole email. My rebutal still applies though and I think your version is inferior wrt handling the issue. Feel free to send a competing patch for review though. Herbert, consider that without a fix one way or another, there is a potential kernel attack vector through the error reporting function of the caam driver. Cristian S.