From mboxrd@z Thu Jan 1 00:00:00 1970 From: Oliver Neukum Subject: Re: [PATCH] usbnet: ax88179_178a: Add support for writing the EEPROM Date: Thu, 25 Aug 2016 12:08:28 +0200 Message-ID: <1472119708.2877.24.camel@suse.com> References: <20160824135231.21782-1-alban.bedel@avionic-design.de> <1472049039.11480.10.camel@suse.com> <20160824164058.6ec64a8e@avionic-0020> <1472116596.2877.23.camel@suse.com> <20160825120733.3413d395@avionic-0020> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Freddy Xin , linux-kernel@vger.kernel.org, linux-usb@vger.kernel.org, netdev@vger.kernel.org To: Alban Bedel Return-path: In-Reply-To: <20160825120733.3413d395@avionic-0020> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Thu, 2016-08-25 at 12:07 +0200, Alban Bedel wrote: > On Thu, 25 Aug 2016 11:16:36 +0200 > Oliver Neukum wrote: > > > On Wed, 2016-08-24 at 16:40 +0200, Alban Bedel wrote: > > > On Wed, 24 Aug 2016 16:30:39 +0200 > > > Oliver Neukum wrote: > > > > > > > On Wed, 2016-08-24 at 15:52 +0200, Alban Bedel wrote: > > > > > > > + if (block != data) > > > > > + kfree(block); > > > > > > > > And if block == dta, what frees the memory? > > > > > > In this case this function didn't allocate any memory, so there is > > > nothing to free. > > > > Hi, > > > > I see. kfree() has a check for NULL, so you could drop the > > test, but it doesn't matter much either way. > > I think you misunderstand something here. data is the buffer passed > by the caller and block is a local variable. There is two cases: > > 1) The data to write is block aligned, then we use the caller buffer > as is and set block = data. > 2) The requested data is not block aligned, then we kalloc block. > > In both case the writing loop then use the block pointer. Afterwards > we only need to kfree block in case 2, that is when block != data. Thanks for the clarification. Maybe worth a comment in the code? Regards Oliver