Re: [PATCH/RFC 0/7] wl12xx SDIO interface
From: Florian Fainelli <florian@openwrt.org>
Date: 2009-06-11 15:08:32
Le Thursday 11 June 2009 17:04:01 Gábor Stefanik, vous avez écrit :
On Thu, Jun 11, 2009 at 4:59 PM, Florian Fainelli[off-list ref]
wrote:
quoted
Le Thursday 11 June 2009 16:55:20 Samuel Ortiz, vous avez écrit :quoted
On Thu, 2009-06-11 at 16:38 +0200, Gábor Stefanik wrote:quoted
On Thu, Jun 11, 2009 at 4:02 AM, Bob Copeland[off-list ref]
wrote:
quoted
quoted
quoted
quoted
Here's v2 of the SDIO interface for wl12xx, still with a FIXME or two. For my part, this is mainly a port of the Google/TI Android SDIO interface logic to the existing SPI driver. I have confirmed that the driver associates and sends some traffic. Pierre, hoping to get your perspective on the SDIO bits in patch 7, since I very likely don't know what I'm doing :) Especially regarding the platform code setup, in which the correct GPIOs have to be written to turn on the device before probe will work. For the SPI interface, we use spi_register_board_info() for a similar purpose, but I'm not sure if there is a standard way to do that for SDIO. This copies Google's platform driver for now. Conversion to readl/writel etc where appropriate is TODO. Bob Copeland (7): wl12xx: separate bus i/o code into io.c wl12xx: use wiphy_dev instead of wl->spi->dev wl12xx: introduce wl12xx_if_operations struct wl12xx: make wl12xx_set_partition bus agnostic wl12xx: move module probe methods into spi.c wl12xx: split spi interface into separate module wl12xx: add sdio support drivers/net/wireless/wl12xx/Kconfig | 24 +++- drivers/net/wireless/wl12xx/Makefile | 10 +- drivers/net/wireless/wl12xx/acx.c | 3 +- drivers/net/wireless/wl12xx/boot.c | 3 +- drivers/net/wireless/wl12xx/cmd.c | 15 +- drivers/net/wireless/wl12xx/event.c | 6 +- drivers/net/wireless/wl12xx/io.c | 181 +++++++++++++++++++++ drivers/net/wireless/wl12xx/io.h | 62 +++++++ drivers/net/wireless/wl12xx/main.c | 152 ++++-------------- drivers/net/wireless/wl12xx/ps.c | 3 +- drivers/net/wireless/wl12xx/rx.c | 10 +- drivers/net/wireless/wl12xx/sdio.c | 250 +++++++++++++++++++++++++++++ drivers/net/wireless/wl12xx/spi.c | 289 ++++++++++++++------------------- drivers/net/wireless/wl12xx/spi.h | 40 ----- drivers/net/wireless/wl12xx/tx.c | 64 ++++---- drivers/net/wireless/wl12xx/wl1251.c | 9 +- drivers/net/wireless/wl12xx/wl12xx.h | 14 ++- 17 files changed, 752 insertions(+), 383 deletions(-) create mode 100644 drivers/net/wireless/wl12xx/io.c create mode 100644 drivers/net/wireless/wl12xx/io.h create mode 100644 drivers/net/wireless/wl12xx/sdio.cCan we expect vlynq any time soon?I guess we need: 1) A linux vlynq driver for that, and last thing I saw [1], it wasnt really pretty/readyThat driver is perfectly functionnal on a TI AR7 and we have used it in OpenWrt for months with the acx100/111 driver. What makes you think it is not yet ready ? There was a couple of suggestions from a TI employee, which are perfectly valid, but I have no hardware to implement what he is suggesting (no daisy chaining ...).Better forget the acx1xx driver. There is legal controversy around it - the devs never released any documentation during the reverse engineering process, and development was quite slow until TI's copyrighted ti_ap.o driver's source code leaked...
This is why I ask if wl12xx can actually drive an acx111 chipset or the HW is completely different. -- Best regards, Florian Fainelli Email : florian@openwrt.org http://openwrt.org -------------------------------