From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mx1.redhat.com ([66.187.233.31]:45523 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759147AbYDCD20 (ORCPT ); Wed, 2 Apr 2008 23:28:26 -0400 Subject: Re: [RFC] libertas: convert libertas driver to use an event/cmdresp queue From: Dan Williams To: Holger Schurig Cc: linux-wireless@vger.kernel.org In-Reply-To: <1207191923.9250.6.camel@localhost.localdomain> References: <200803271700.58477.hs4233@mail.mn-solutions.de> <1206722403.3392.9.camel@localhost.localdomain> <200804011439.54618.hs4233@mail.mn-solutions.de> <1207191923.9250.6.camel@localhost.localdomain> Content-Type: text/plain Date: Wed, 02 Apr 2008 23:26:46 -0400 Message-Id: <1207193206.9435.0.camel@localhost.localdomain> (sfid-20080403_042829_804189_1FB71FC6) Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Wed, 2008-04-02 at 23:05 -0400, Dan Williams wrote: > On Tue, 2008-04-01 at 14:39 +0200, Holger Schurig wrote: > > > It's been tested lightly on USB, SDIO, and CF. I'd like wider > > > testing on CF as this patch causes SSH transfers to stall > > > (only on CF!) > > > > I've modified if_cs.c somewhat and did > > > > $ scp inkscape inkview emacs-snapshot-x mtn \ > > emacs-snapshot-nox pgadmin3 quanta gs-gpl \ > > root@10.2.1.4:/tmp > > inkscape 100% 5856KB 650.7KB/s 00:09 > > inkview 100% 5853KB 585.3KB/s 00:10 > > emacs-snapshot-x 100% 5333KB 666.6KB/s 00:08 > > mtn 100% 5263KB 657.9KB/s 00:08 > > emacs-snapshot-nox 100% 4913KB 614.1KB/s 00:08 > > pgadmin3 100% 3925KB 654.2KB/s 00:06 > > quanta 100% 3597KB 719.5KB/s 00:05 > > gs-gpl 100% 3597KB 599.4KB/s 00:06 > > > > I didn't notice any stall so far. What test did you make? > > Just tried a 300MB scp; got up to 37MB @ 715KB/s and then stalled. I > think my card is somehow "special" or the controller sucks (Ricoh > RL5c476 II) or something. I'll try to find another machine with a CF > slot. No amount of ifconfig or iwconfig magic will help it, the card > accepts a TX packet with if_cs_send_data() and then apparently never > sends an interrupt to say the TX is done. The driver then does the RSSI > thing to poke the firmware, which of course times out, and then all is > lost. > > Maybe I have dodgy firmware? Can you point me to some known working > stuff? I'm running 5.0.16.p0 BTW. Dan