Return-path: Received: from mx2.redhat.com ([66.187.237.31]:56915 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752417AbZCWXK4 (ORCPT ); Mon, 23 Mar 2009 19:10:56 -0400 Subject: Re: [PATCH] Marvell CF8381 From: Dan Williams To: Marek Vasut Cc: Holger Schurig , linux-wireless@vger.kernel.org, libertas-dev@lists.infradead.org In-Reply-To: <200903232258.57323.marek.vasut@gmail.com> References: <200903220127.21836.marek.vasut@gmail.com> <200903231327.18890.hs4233@mail.mn-solutions.de> <1237824927.14758.20.camel@localhost.localdomain> <200903232258.57323.marek.vasut@gmail.com> Content-Type: text/plain Date: Mon, 23 Mar 2009 19:08:30 -0400 Message-Id: <1237849710.26204.17.camel@localhost.localdomain> (sfid-20090324_001058_476770_95A85285) Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, 2009-03-23 at 22:58 +0100, Marek Vasut wrote: > On Monday 23 of March 2009 17:15:27 Dan Williams wrote: > > On Mon, 2009-03-23 at 13:27 +0100, Holger Schurig wrote: > > > > - priv->regioncode = le16_to_cpu(cmd.regioncode) & 0xFF; > > > > + priv->regioncode = (le16_to_cpu(cmd.regioncode) & > > > > > > 0xFF00) >> 8; > > > > > > Hmm, the change that this breaks 8385 is quite high, maybe 100 > > > percent. > > > > The cf8385-5.0.16.p0-26306 (gumstix) driver has: > > > > Adapter->RegionCode = wlan_le16_to_cpu(hwspec->RegionCode) >> 8; > > > > So it appears there's precedent for this even on 8385 parts. > > > > Dan > > > > > It could simply be the case that your firmware returns the value > > > in a different order than mine. So we either need to test for the > > > firmware version or for the hardware before getting the value via > > > one or the other method. > > > > > > Maybe it's worthwhile to create some hw_is_8381() function and > > > then do the things that need to be differently based on it's > > > return value? > > > > > > > > > > > > The documentation for CMG_GET_HW_SPEC that I have says > > > > > > RegionCode UINT16 Set to 0 > > > > > > so I see no indicator that we have two 8 bit values here. But > > > we've seen bugs in docs before :-) > > > -- > > > To unsubscribe from this list: send the line "unsubscribe linux-wireless" > > > in the body of a message to majordomo@vger.kernel.org > > > More majordomo info at http://vger.kernel.org/majordomo-info.html > > Ok, so this should be it. I dropped the 8305 support since I found I have some > issues with it still. > > From 3ac8f34db9c4d96197fbdc2776ebbcd571e3fbbe Mon Sep 17 00:00:00 2001 > From: Marek Vasut > Date: Mon, 23 Mar 2009 22:55:51 +0100 > Subject: [PATCH] Add support for CF8381 WiFi card. > A detection function was added for identifying CF8381. > > Signed-off-by: > --- > drivers/net/wireless/libertas/if_cs.c | 24 +++++++++++++++++++++++- > 1 files changed, 23 insertions(+), 1 deletions(-) > > diff --git a/drivers/net/wireless/libertas/if_cs.c > b/drivers/net/wireless/libertas/if_cs.c > index 3f02e6a..56b6475 100644 > --- a/drivers/net/wireless/libertas/if_cs.c > +++ b/drivers/net/wireless/libertas/if_cs.c > @@ -273,7 +273,20 @@ static int if_cs_poll_while_fw_download(struct if_cs_card > *card, uint addr, u8 r > */ > #define IF_CS_PRODUCT_ID 0x0000001C > #define IF_CS_CF8385_B1_REV 0x12 > +#define IF_CS_CF8381_B3_REV 0x04 > > +/* > + * Used to detect other cards than CF8385 since their revisions of silicon > + * doesn't match those from CF8385, eg. CF8381 B3 works with this driver. > + */ > +#define CF8381_MANFID 0x02db > +#define CF8381_CARDID 0x6064 > + > +static inline int if_cs_hw_is_cf8381(struct pcmcia_device *p_dev) > +{ > + return (p_dev->manf_id == CF8381_MANFID && > + p_dev->card_id == CF8381_CARDID); > +} > > /********************************************************************/ > /* I/O and interrupt handling */ > @@ -757,6 +770,7 @@ static void if_cs_release(struct pcmcia_device *p_dev) > static int if_cs_probe(struct pcmcia_device *p_dev) > { > int ret = -ENOMEM; > + unsigned int prod_id; > struct lbs_private *priv; > struct if_cs_card *card; > /* CIS parsing */ > @@ -859,7 +873,14 @@ static int if_cs_probe(struct pcmcia_device *p_dev) > p_dev->io.BasePort1 + p_dev->io.NumPorts1 - 1); > > /* Check if we have a current silicon */ > - if (if_cs_read8(card, IF_CS_PRODUCT_ID) < IF_CS_CF8385_B1_REV) { > + prod_id = if_cs_read8(card, IF_CS_PRODUCT_ID); > + if (if_cs_hw_is_cf8381(p_dev) && prod_id < IF_CS_CF8381_B3_REV) { > + lbs_pr_err("old chips like 8381 rev B3 aren't supported\n"); > + ret = -ENODEV; > + goto out2; > + } > + > + if (prod_id < IF_CS_CF8385_B1_REV) { You'll still want to implement a if_cs_hw_is_cf8385() here though; otherwise, wouldn't 8381 get caught by this 8385 check and get rejected? Dan > lbs_pr_err("old chips like 8385 rev B1 aren't supported\n"); > ret = -ENODEV; > goto out2; > @@ -950,6 +971,7 @@ static void if_cs_detach(struct pcmcia_device *p_dev) > /********************************************************************/ > > static struct pcmcia_device_id if_cs_ids[] = { > + PCMCIA_DEVICE_MANF_CARD(CF8381_MANFID, CF8381_CARDID), > PCMCIA_DEVICE_MANF_CARD(0x02df, 0x8103), > PCMCIA_DEVICE_NULL, > }; > -- > 1.6.2 > > -- > To unsubscribe from this list: send the line "unsubscribe linux-wireless" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html