Rt3070usb - Revision history

Esden at 20:19, 10 March 2011

2011-03-10T20:19:55Z

← Older revision		Revision as of 14:19, 10 March 2011
Line 246:		Line 246:

	</pre>		</pre>

			[[Category:Software]] [[Category:User_Documentation]]

Paulcox: /* Building */

2010-07-30T14:40:54Z

Building

← Older revision		Revision as of 08:40, 30 July 2010
Line 53:		Line 53:

	I haven't looked into it, but I'm guessing the firmware is now included in the module file and perhaps there's firmware for a bunch of different controllers/cards. The file system space isn't too precious, but this doesn't have an impact on load time (and therefore boot time), so it would be good to '''resolve this''' and get it down to a reasonable size.		I haven't looked into it, but I'm guessing the firmware is now included in the module file and perhaps there's firmware for a bunch of different controllers/cards. The file system space isn't too precious, but this doesn't have an impact on load time (and therefore boot time), so it would be good to '''resolve this''' and get it down to a reasonable size.

			Update : after rebuilding and deploying my new SD card image I no longer seem to have a big module (now around 600kB). The first time I rebuilt i had a do_setscene warning about get_kernelmajorversion not working followed by do_strip error (probably due to first). retrying without using the bitbake -b method worked, though. [[User:Paulcox\|Paulcox]] 16:40, 30 July 2010 (CEST)

	To test we can copy the .ko file directly to our overo root filesystem as well as create the /etc/Wireless/blah to put the .dat file or we can simply add the rt3070 recipe to our image recipe, rebuild our omap3-lisa-image. omap3-lisa-image, and create a new sd card.		To test we can copy the .ko file directly to our overo root filesystem as well as create the /etc/Wireless/blah to put the .dat file or we can simply add the rt3070 recipe to our image recipe, rebuild our omap3-lisa-image. omap3-lisa-image, and create a new sd card.

Paulcox: /* Building */

2010-07-30T14:32:19Z

Building

← Older revision		Revision as of 08:32, 30 July 2010
Line 23:		Line 23:
	== Building ==		== Building ==

	The recipe will be fixed ('''patch ~~will be posted~~ here'''), but until then, here are instructions for building :		The recipe will be fixed ('''patch files available [http://enac-lara.googlecode.com/svn/trunk/paul/misc/rt3070_oe_patchfiles.tar.gz here]'''), but until then, here are instructions for building :

	In the overo-oe directory type:		In the overo-oe directory type:

Paulcox: /* Building */

2010-07-30T10:56:08Z

Building

← Older revision		Revision as of 04:56, 30 July 2010
Line 37:		Line 37:
	Patch common/rtmp_init.c with the one line change : replace init_MUTEX as indicated above		Patch common/rtmp_init.c with the one line change : replace init_MUTEX as indicated above

	Then, apply patch (at the bottom of this page or here) as follows :		Then, apply patch (at the bottom of this page or [http://code.google.com/p/enac-lara/source/browse/trunk/paul/misc/rt3070.patch here]) as follows :

	: paul@paul-laptop:~/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/rt3070-2.1.1.0-r80$ patch -p0 < rt3070.patch		: paul@paul-laptop:~/overo-oe/tmp/work/overo-angstrom-linux-gnueabi/rt3070-2.1.1.0-r80$ patch -p0 < rt3070.patch

Paulcox: /* Background */

2010-07-30T10:04:00Z

Background

← Older revision		Revision as of 04:04, 30 July 2010
Line 11:		Line 11:
	My initial search seemed to show the rt2870usb driver was the one to use, and I guess at one time that might have been the case for the chip in this stick. But rt3070usb is now the one to use.		My initial search seemed to show the rt2870usb driver was the one to use, and I guess at one time that might have been the case for the chip in this stick. But rt3070usb is now the one to use.

	There's a recipe for the rt3070 driver in openembedded (which I'm pulling from : ). The recipe pulls the source code from the ralink website, patches it, and builds. It's pretty impressive that this driver cross compiles and just works. Well, for me it doesn't compile for me for two reasons :		There's a recipe for the rt3070 driver in openembedded (which I'm pulling from : ). The recipe pulls the source code from the ralink website (or get it [http://www.ralinktech.com/support.php?s=2 from this page]), patches it, and builds. It's pretty impressive that this driver cross compiles and just works. Well, for me it doesn't compile for me for two reasons :

	* 1. My linux kernel has real-time patches so the init_MUTEX() in the driver source can't be found		* 1. My linux kernel has real-time patches so the init_MUTEX() in the driver source can't be found

Paulcox: /* Background */

2010-07-30T09:41:08Z

Background

← Older revision		Revision as of 03:41, 30 July 2010
Line 5:		Line 5:
	== Background ==		== Background ==

	I have an edimax usb wifi stick I needed to get working with Lisa. Its USB id is 7392:7711. My initial search seemed to show the rt2870usb driver was the one to use, and I guess at one time that might have been the case for the chip in this stick. But rt3070usb is now the one to use.		I have an edimax EW-7711UAn usb wifi stick I needed to get working with Lisa. Its USB id is 7392:7711.

			[http://www.edimax.com/en/produce_detail.php?pd_id=277&pl1_id=1&pl2_id=44 Manufacturer] and available at [http://www.amazon.com/EW-7711UAN-Edimax-Wireless-nLite-Adapter/dp/B001KOTDDU/ref=sr_1_17?ie=UTF8&s=electronics&qid=1280482648&sr=8-17 Amazon]

			My initial search seemed to show the rt2870usb driver was the one to use, and I guess at one time that might have been the case for the chip in this stick. But rt3070usb is now the one to use.

	There's a recipe for the rt3070 driver in openembedded (which I'm pulling from : ). The recipe pulls the source code from the ralink website, patches it, and builds. It's pretty impressive that this driver cross compiles and just works. Well, for me it doesn't compile for me for two reasons :		There's a recipe for the rt3070 driver in openembedded (which I'm pulling from : ). The recipe pulls the source code from the ralink website, patches it, and builds. It's pretty impressive that this driver cross compiles and just works. Well, for me it doesn't compile for me for two reasons :

Paulcox: /* Background */

2010-07-30T09:28:48Z

Background

← Older revision		Revision as of 03:28, 30 July 2010
Line 5:		Line 5:
	== Background ==		== Background ==

	I have an edimax usb wifi stick I needed to get working with Lisa. ~~It's~~ USB id is 7392:7711. ~~Initial~~ search seemed to show the rt2870usb driver was the one to use, and I guess at one time that might have been the case for the chip in this stick. But rt3070usb is now the one to use.		I have an edimax usb wifi stick I needed to get working with Lisa. Its USB id is 7392:7711. My initial search seemed to show the rt2870usb driver was the one to use, and I guess at one time that might have been the case for the chip in this stick. But rt3070usb is now the one to use.

	There's a recipe for the rt3070 driver in openembedded (which I'm pulling from : ). It doesn't compile for me for two reasons :		There's a recipe for the rt3070 driver in openembedded (which I'm pulling from : ). The recipe pulls the source code from the ralink website, patches it, and builds. It's pretty impressive that this driver cross compiles and just works. Well, for me it doesn't compile for me for two reasons :

	* 1. My linux kernel has real-time patches so the init_MUTEX() in the driver source can't be found		* 1. My linux kernel has real-time patches so the init_MUTEX() in the driver source can't be found

Paulcox: /* Configuration */

2010-07-30T09:25:07Z

Configuration

@@ Line 16: / Line 16: @@
 * 1. replace init_MUTEX(blah) with sema_init(blah,1)
 * 2. Apply patch from http://lists.rpmfusion.org/pipermail/rpmfusion-commits/2009-August/005943.html (copied below just in case)
 == Configuration ==
 : insmod rt3070sta.ko

Paulcox: Building and Configuring rt3070 driver for Lisa/Overo/OE

2010-07-30T08:35:31Z

Building and Configuring rt3070 driver for Lisa/Overo/OE

New page

== Intro ==

This page explains how to get the rt3070usb (aka rt3070sta) linux driver working in openembedded for the overo on Lisa.

== Background ==

I have an edimax usb wifi stick I needed to get working with Lisa. It's USB id is 7392:7711. Initial search seemed to show the rt2870usb driver was the one to use, and I guess at one time that might have been the case for the chip in this stick. But rt3070usb is now the one to use.

There's a recipe for the rt3070 driver in openembedded (which I'm pulling from : ). It doesn't compile for me for two reasons :

* 1. My linux kernel has real-time patches so the init_MUTEX() in the driver source can't be found
* 2. Some sort of device structure change has occurred in the kernel since who knows when

The fixes are:

* 1. replace init_MUTEX(blah) with sema_init(blah,1)
* 2. Apply patch from http://lists.rpmfusion.org/pipermail/rpmfusion-commits/2009-August/005943.html (copied below just in case)

== Configuration ==

: insmod rt3070sta.ko
:: at this point iwconfig shows an ra0 interface
: ifconfig ra0 up
: iwconfig ra0 essid ap_h16

=== wpa_supplicant ===

As usual, the wpa key goes in /etc/wpa_supplicant.conf and started wpa_supplicant with
: wpa_supplicant -ira0 -c/etc/wpa_supplicant.conf &

=== DHCP ===

: dhclient ra0

=== Console output ===

<pre>
root@overo:/home/paul# iwconfig ra0
ra0 RT2870 Wireless ESSID:"ap_h16" Nickname:"RT2870STA"
Mode:Managed Frequency=2.412 GHz Access Point: 00:25:4B:06:E3:9F
Bit Rate=65 Mb/s
RTS thr:off Fragment thr:off
Encryption key:447F-01CB-C235-6BF1-0F8C-0E61-5C88-EEEA
Link Quality=100/100 Signal level:-39 dBm Noise level:-83 dBm
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0 Missed beacon:0

root@overo:/home/paul# lsusb
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 002: ID 7392:7711
root@overo:/home/paul# ping bill

PING bill.recherche.enac.fr (10.31.4.5) 56(84) bytes of data.
64 bytes from bill.recherche.enac.fr (10.31.4.5): icmp_seq=1 ttl=64 time=367 ms
^C
--- bill.recherche.enac.fr ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 367.279/367.279/367.279/0.000 ms
root@overo:/home/paul# WPA: Group rekeying completed with 00:25:4b:06:e3:9f [GTK=TKIP]
</pre>

== Patch ==

<pre>
diff -rupN 2009_0525_RT3070_Linux_STA_v2.1.1.0.old/include/rtmp_os.h 2009_0525_RT3070_Linux_STA_v2.1.1.0/include/rtmp_os.h
--- 2009_0525_RT3070_Linux_STA_v2.1.1.0.old/include/rtmp_os.h 2009-05-20 23:33:56.000000000 -0400
+++ 2009_0525_RT3070_Linux_STA_v2.1.1.0/include/rtmp_os.h 2009-08-04 00:00:46.000000000 -0400
@@ -53,6 +53,7 @@
*/
typedef struct _RTMP_OS_NETDEV_OP_HOOK_
{
+ /*
void *open;
void *stop;
void *xmit;
@@ -63,6 +64,8 @@ typedef struct _RTMP_OS_NETDEV_OP_HOOK_
unsigned char devAddr[6];
unsigned char devName[16];
unsigned char needProtcted;
+ */
+ struct net_device_ops *netdev_ops;
}RTMP_OS_NETDEV_OP_HOOK, *PRTMP_OS_NETDEV_OP_HOOK;

diff -rupN 2009_0525_RT3070_Linux_STA_v2.1.1.0.old/os/linux/rt_linux.c 2009_0525_RT3070_Linux_STA_v2.1.1.0/os/linux/rt_linux.c
--- 2009_0525_RT3070_Linux_STA_v2.1.1.0.old/os/linux/rt_linux.c 2009-05-20 23:33:56.000000000 -0400
+++ 2009_0525_RT3070_Linux_STA_v2.1.1.0/os/linux/rt_linux.c 2009-08-04 00:04:15.000000000 -0400
@@ -1506,20 +1506,21 @@ int RtmpOSNetDevAttach(
if (pDevOpHook)
{
PRTMP_ADAPTER pAd = RTMP_OS_NETDEV_GET_PRIV(pNetDev);
-
+ pNetDev->netdev_ops = pDevOpHook->netdev_ops;
+ /*
pNetDev->open = pDevOpHook->open;
pNetDev->stop = pDevOpHook->stop;
pNetDev->hard_start_xmit = (HARD_START_XMIT_FUNC)(pDevOpHook->xmit);
pNetDev->do_ioctl = pDevOpHook->ioctl;
-
+ */
/* if you don't implement get_stats, just leave the callback function as NULL, a dummy
function will make kernel panic.
*/
- if (pDevOpHook->get_stats)
+ /*if (pDevOpHook->get_stats)
pNetDev->get_stats = pDevOpHook->get_stats;
-
+ */
/* OS specific flags, here we used to indicate if we are virtual interface */
- pNetDev->priv_flags = pDevOpHook->priv_flags;
+ // pNetDev->priv_flags = pDevOpHook->priv_flags;

#if (WIRELESS_EXT < 21) && (WIRELESS_EXT >= 12)
pNetDev->get_wireless_stats = rt28xx_get_wireless_stats;
@@ -1544,15 +1545,15 @@ int RtmpOSNetDevAttach(
#endif // CONFIG_APSTA_MIXED_SUPPORT //

// copy the net device mac address to the net_device structure.
- NdisMoveMemory(pNetDev->dev_addr, &pDevOpHook->devAddr[0], MAC_ADDR_LEN);
+ // NdisMoveMemory(pNetDev->dev_addr, &pDevOpHook->devAddr[0], MAC_ADDR_LEN);

- rtnl_locked = pDevOpHook->needProtcted;
+ // rtnl_locked = pDevOpHook->needProtcted;
}
-
+ /*
#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,24)
pNetDev->validate_addr = NULL;
#endif
-
+ */
if (rtnl_locked)
ret = register_netdevice(pNetDev);
else
diff -rupN 2009_0525_RT3070_Linux_STA_v2.1.1.0.old/os/linux/rt_main_dev.c 2009_0525_RT3070_Linux_STA_v2.1.1.0/os/linux/rt_main_dev.c
--- 2009_0525_RT3070_Linux_STA_v2.1.1.0.old/os/linux/rt_main_dev.c 2009-05-20 23:33:56.000000000 -0400
+++ 2009_0525_RT3070_Linux_STA_v2.1.1.0/os/linux/rt_main_dev.c 2009-08-04 00:06:08.000000000 -0400
@@ -504,6 +504,21 @@ err:
return (-1);
} /* End of rt28xx_open */

+static const struct net_device_ops rt2870_netdev_ops = {
+ .ndo_open = MainVirtualIF_open,//
+ .ndo_stop = MainVirtualIF_close,//
+ .ndo_do_ioctl = rt28xx_ioctl,//
+ .ndo_get_stats = RT28xx_get_ether_stats,//
+ .ndo_validate_addr = NULL,
+ .ndo_set_mac_address = eth_mac_addr,
+ .ndo_change_mtu = eth_change_mtu,
+#ifdef IKANOS_VX_1X0
+ .ndo_start_xmit = IKANOS_DataFramesTx,//
+#else
+ .ndo_start_xmit = rt28xx_send_packets,//
+#endif
+};
+

PNET_DEV RtmpPhyNetDevInit(
IN RTMP_ADAPTER *pAd,
@@ -520,6 +535,8 @@ PNET_DEV RtmpPhyNetDevInit(
}

NdisZeroMemory((unsigned char *)pNetDevHook, sizeof(RTMP_OS_NETDEV_OP_HOOK));
+ pNetDevHook->netdev_ops = &rt2870_netdev_ops;
+ /*
pNetDevHook->open = MainVirtualIF_open;
pNetDevHook->stop = MainVirtualIF_close;
pNetDevHook->xmit = rt28xx_send_packets;
@@ -531,7 +548,7 @@ PNET_DEV RtmpPhyNetDevInit(
pNetDevHook->get_stats = RT28xx_get_ether_stats;

pNetDevHook->needProtcted = FALSE;
-
+ */
RTMP_OS_NETDEV_SET_PRIV(net_dev, pAd);
//net_dev->priv = (PVOID)pAd;
pAd->net_dev = net_dev;

</pre>