From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from server19320154104.serverpool.info ([193.201.54.104]:33692 "EHLO hauke-m.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751927Ab2ECLrE (ORCPT ); Thu, 3 May 2012 07:47:04 -0400 Message-ID: <4FA2702C.8060803@hauke-m.de> (sfid-20120503_134709_127378_6CD8D83E) Date: Thu, 03 May 2012 13:46:52 +0200 From: Hauke Mehrtens MIME-Version: 1.0 To: Arend van Spriel CC: linville@tuxdriver.com, linux-wireless@vger.kernel.org Subject: Re: [PATCH 00/19] brcmsmac: use sprom and PCIe code from bcma References: <1335660643-26368-1-git-send-email-hauke@hauke-m.de> <4F9CF839.1040401@broadcom.com> <4FA0CF03.9060307@broadcom.com> <4FA1B66A.4060901@hauke-m.de> <4FA22325.2030104@broadcom.com> In-Reply-To: <4FA22325.2030104@broadcom.com> Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: On 05/03/2012 08:18 AM, Arend van Spriel wrote: > On 05/03/2012 12:34 AM, Hauke Mehrtens wrote: >> >> Hi Arend, >> >> thanks for testing. >> >> Have you tried suspend/resume and hibernate/resume? I think some of the >> workaround in drivers/bcma/driver_pci.c should be called in the resume >> path and not only in the initialization path. >> >> Hauke >> > > Regarding using brcmsmac on bcm4718 soc, I think we cleaned up a bit too > much during mainline process. There was some chip specific stuff in > coreinit function and quite a few things in phy code, but those had to > do with the write flush, I believe. However, the tx fifo size mismatch > is detected before the chip specific setting in coreinit so I am not > sure if it is going to help us. I will create a patch for you to try. > > Gr. AvS > Hi Arend, I looked into the code of the initial brcm80211 commit to staging and it contains some code which should only be executed on bcm4718 SoCs and not on bcm43224 chips. I haven't seen any oblivious part which could trigger the error I am getting, but I will try out the other parts. Did you remove any significant code before the initial commit to staging? I will also look what b43 does different from brcmsmac. /* * FIFOSZ fixup * 1) core5-9 use ucode 5 to save space since the PSM is the same * 2) newer chips, driver wants to controls the fifo allocation */ /* For old ucode, txfifo sizes needs to be modified(increased) for Corerev >= 9 */ These two comments from brcmsmac are looking interesting, but that should not affect the bcm4718 as it is uses a rev 17 core. Hauke