From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mtiwmhc13.worldnet.att.net ([204.127.131.117]:44001 "EHLO mtiwmhc13.worldnet.att.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751476AbYFYSYc (ORCPT ); Wed, 25 Jun 2008 14:24:32 -0400 Message-ID: <48628D9A.5090502@lwfinger.net> (sfid-20080625_202434_996532_36ED50E8) Date: Wed, 25 Jun 2008 13:25:30 -0500 From: Larry Finger MIME-Version: 1.0 To: Pavel Roskin CC: linville@tuxdriver.com, hintak_leung@yahoo.co.uk, linux-wireless@vger.kernel.org Subject: Re: [RFC] rtl8187: Realtek 8187B wireless support with product id0x8197/0x8189 References: <4862796a.pmrQX9GXIZ1wDm2k%Larry.Finger@lwfinger.net> <1214415411.7901.25.camel@dv> <4862840F.4030409@lwfinger.net> <1214416662.9237.6.camel@dv> In-Reply-To: <1214416662.9237.6.camel@dv> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Sender: linux-wireless-owner@vger.kernel.org List-ID: Pavel Roskin wrote: > On Wed, 2008-06-25 at 12:44 -0500, Larry Finger wrote: > I'm not sure I can stand behind that patch quality-wise. It's 40k of > intrusive changes that I could not even test on an old rtl8187. If I > had more time and a original rtl8187 device, I would try to split more > parts or restructure the code first. > > Also, the startup time (ifconfig up) is over 20 seconds, which is > absolutely unacceptable. Either some delays are over-conservative, or > we are doing something wrong. > For the review process, I have split the patch into 5 parts. Each step compiles, thus bisection will be preserved. With this change, the largest of the pieces is a little under 15KiB. Is this small enough? Further splitting would be a lot more difficult. As to the quality of the patch, it barely works on my device. RX seems to be OK except for the wireless statistics, but I can associate only about 1 try in 10. Once I get that problem squared away, then I can work on the timing. I really would like to get this into wireless-testing ASAP. I don't have an 8187 device either, and will have to rely on someone else for that part of the testing. Larry