I need a hand getting a USB Wifi device going.
The device is a "Legaung LG-N120 outdoor ethernet wifi adapter" which is
a cheap USB device with a patch antenna I found on eBay. It's based on
the Ralink 3070 chip and works great under OSX Mavericks 10.9.5, once I
found and installed the Ralink software.
I'm now setting up a Gigabyte Brix GB-BXBT-2807 (Celeron) using the
latest Arch Linux (uname -r reports: 4.2.2-1-ARCH). The Brix comes with
a Realtek wifi card installed. (lspci says: RTL8723BE PCIe Wireless
Network Adapter). This scans and connects just fine using
iw/wpa_supplicant or simply wifi-menu.
I've been using these docs as a starting point:
https://wiki.archlinux.org/index.php/Wireless_network_configuration.
To test, I first shut off the internal wifi with 'ip link set <devname>
down'
Then I plug in the USB Ralink device
it shows up in lsusb,
it can be activated via 'ip link set <devname> up',
it shows up in 'iw dev'
it fails 'iw dev <devname> scan' with the terse message 'scan
aborted'
this also deactivates the device... 'iw dev' says it's no longer
'UP'
In more detail:
I run dmesg -w and open a new pane (using tmux)
I plug in the device
dmesg says:
usb 1-1: new high-speed USB device number 4 using xhci_hcd
usb 1-1: reset high-speed USB device number 4 using xhci_hcd
ieee80211 phy1: rt2x00_set_rt: Info - RT chipset 3070, rev 201
detected
ieee80211 phy1: rt2x00_set_rt: Info - RT chipset 0005 detected
ieee80211 phy1: Selected rate control algorithm 'minstrel_ht'
usbcore: registered new interface driver rt2800usb
rt2800usb: 1-1:1.0 wlp0s20u1: renamed from wlan0
I run 'lsusb'
It says, among other things:
...
Bus 001 Device 004: ID 148f:3070 Ralnk Technology, Corp.
Rt2800/RT3070 Wireless Adapter
...
I run 'ip link'
The new device is now there:
...
4: wlp0s20u1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state
DOWN mode DEFAULT group default qlen 1000
link/ether hh:hh:... ff:ff:ff:ff:ff:ff
I run 'ip link set wlp0s20u1 up'
dmesg says:
ieee80211 phy1: rt2x00lib_request_firmware: Info - Loading
firmware file 'rt2870.bin'
ieee80211 phy1: rt2x00lib_request_firmware: Info - Firmware
detected - version: 0.29
IPv6: ADDRCONF(NETDEV_UP): wlp0s20u1: link is not ready
I note that the firmware filename doesn't seem to be the expected
rt3070.bin, but I press on.
I run 'ip link'
It reports the new device activated:
...
4: wlp0s20u1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc
mq state DOWN mode DEFAULT group default qlen 1000
link/ether hh:hh:... ff:ff:ff:ff:ff:ff
I run 'iw dev wlp0s20u1 scan'
It says:
scan aborted!
demsg says:
usb 1-1: USB disconnect, device number 4
usb 1-1: new high-speed USB device number 5 using xhci_hcd
usb 1-1: reset high-speed USB device number 5 using xhci_hcd
ieee80211 phy2: rt2x00_set_rt: Info - RT chipset 3070, rev 0201
detected
ieee80211 phy2: rt2x00_set_rt: Info - RT chipset 0005 detected
ieee80211 phy2: Selected rate control algorithm 'minstrel_ht'
rt2800usb 1-1:1.0 wlp0s20u1: renamed from wlan0
...so it seems that the 'scan' caused the usb driver to
disconnect/reconnect the device.
The same steps with 'iw event -t -f' running in extra pane give these
messages:
<plug in>
...: phy #1: renamed to phy1
<remaining steps ending with scan>
...: wlp0s20u1 (phy #1): scan sarted
...: (phy #1): scan aborted: 2412 2417 2422 2427 ...
I've also tried skipping the scan, and connecting to a known access
point with wpa_supplicant. I get similar results. I can post details
if it will help.
In researching the surprising firmware filename issue, I find that
/usr/lib/firmware/rt3070.bin
is merely a link to rt2870.bin. A bit of research revealed that this is
normal and correct:
https://wireless.wiki.kernel.org/en/users/drivers/rt2800usb
http://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/commit/?id=ae524638c277301a83daf126aaaf880738e2474b
Here, I would expect dmesg to show me the name of the link instead of
the link target, but I'm not too worried about this. That commit was
four years ago; perhaps the parent driver now knows enough to directly
load rt2070.bin for the 3070.
Just to be sure the firmware file is okay, I cloned the latest
linux-firmware git repo and confirmed that rt2070.bin is a binary match
for what I have. I also checked this against the mediatek source code
(url below), and found that it differed, so I tried the mediatek version
too. It fails as before, except that dmesg reports firmware version
0.32, which is slightly newer than above.
So, next I thought about trying to compile the 'latest' driver code from
MediaTek (who now owns Ralink), but comments in the readme say it's for
2.4-2.6 kernels:
http://mediatek.com/en/downloads1/downloads/rt8070-rt3070-rt3370-rt3572-rt5370-rt5372-rt5572-usb-usb/
...of course it's possible that the code is current and the _comment_
is out-of-date :).
Should I give it a try?
Before resorting to that I reviewed everything I could find in
https://wireless.wiki.kernel.org/, and thought I'd post here first.
Any tips for further troubleshooting? What logs should I be looking at?
-Matt
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
