From owner-freebsd-mobile Tue Dec 12 3:37: 1 2000 From owner-freebsd-mobile@FreeBSD.ORG Tue Dec 12 03:36:59 2000 Return-Path: Delivered-To: freebsd-mobile@freebsd.org Received: from tokyogw.iij.ad.jp (tokyogw.iij.ad.jp [202.232.15.22]) by hub.freebsd.org (Postfix) with ESMTP id E767A37B400 for ; Tue, 12 Dec 2000 03:36:57 -0800 (PST) Received: by tokyogw.iij.ad.jp; id UAA04578; Tue, 12 Dec 2000 20:36:57 +0900 (JST) Received: from mercury.iij.ad.jp(192.168.4.89) by tokyogw.iij.ad.jp via smap (V4.2) id xma004304; Tue, 12 Dec 00 20:36:00 +0900 Received: from localhost (shigeru@localhost [127.0.0.1]) by mercury.iij.ad.jp (8.9.3/8.9.3) with ESMTP id UAA24905 for ; Tue, 12 Dec 2000 20:36:00 +0900 (JST) To: freebsd-mobile@FreeBSD.ORG Subject: Re: patch for wi driver In-Reply-To: <200012120411.VAA26651@harmony.village.org> References: <20001211155532C.shigeru@iij.ad.jp> <200012120411.VAA26651@harmony.village.org> X-Mailer: Mew version 1.94.2 on XEmacs 21.2 (Shinjuku) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <20001212203559B.shigeru@iij.ad.jp> Date: Tue, 12 Dec 2000 20:35:59 +0900 From: YAMAMOTO Shigeru X-Dispatcher: imput version 991025(IM133) Lines: 48 Sender: owner-freebsd-mobile@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >>>>> "Warner" == Warner Losh writes: Warner> : #Current wi driver has initialization and resource allocation Warner> mistakes. Warner> I noticed that you fixed the bus_alloc_resource for the IOPORT to Warner> always be 64 bytes long, aligned on a 64-byte boundary. Yes. Loucent Card and PrismII based cards have 64 bytes I/O space. Warner> Are there other mistakes as well? Currently, I don't know where is other mistakes in wi driver. But I find a same mistake in src/sys/dev/pccbb/pccbb.c. In pccbb_attach(), rid=PCCBBR_SOCKBASE; sc->sc_base_res=bus_alloc_resource(dev, SYS_RES_MEMORY, &rid, 0,~0,1, RF_ACTIVE); ^^ why only one? :-) Warner> If I'm reading this part of the patch collrectly, all wireset does Warner> is put a delay 100000 (100ms) into the compiled in code. Is that Warner> right? Why did you do that? Original code has two initalize and delay parts. But some cards I have sometimes failed at initializing a card. In linux WLAN driver codes, it is wait a WI_CMD_BUSY flag in a routine to send a command to firmware. So I change wi driver like a linux WLAN driver. Warner> Also, is there some reason that tsleep can't be used instead (well, Warner> other than it being soon replaced with msleep)? I think tsleep better than DELAY(). But I don't know wi_init has a process thread when wi_init() is called. Warner> Does this mean that all of your wireless cards now work with Warner> FreeBSD? Or are there still some issues? I will try your patch. I tested my patch for one month or more. There is no problems in no-WEP environment, except 3.3V card support ;-( #some cards work in WEP enviroment when setting promiscas mode, others does #not work in WEP environment.... ------- YAMAMOTO Shigeru Internet Initiative Japan Inc. Network Engineering Div. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-mobile" in the body of the message