Feature #15131

Identify USB Wi-Fi adapters that are easily available and work with Tails

Added by sajolida 2017-12-29 05:49:28 . Updated 2019-11-14 17:16:34 .

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Hardware support
Target version:
Start date:
2018-08-30
Due date:
% Done:

100%

Feature Branch:
Type of work:
Test
Starter:
Affected tool:
Deliverable for:

Description

Model Vendor ID Band Size Note
Adapters that we have
TL-WN725N TP-Link 0bda:8179 2.4 GHz Nano module: r8188eu, MAC spoofing fails
DWA-131 D-Link 2001:3319 2.4 GHz Small module: rtl8xxxu, firmware: rtl8192eu_nic, authentication failure
WUSB6100M Max-Stream AC600 Linksys 13b1:0042 Dual Small no module loading
DWA-171A1 D-Link 2001:3314 Dual Small no module loading
TL-WN823N TP-Link 2357:0109 2.4 GHz Small module: rtl8xxxu, firmware: rtl8192eu_nic, authentication failure
Archer T2U TP-Link 148f:761a Dual Small no module loading
TL-WN821N TP-Link 2357:0107 2.4 GHz Long no module loading
Adapters that we don’t have
AE6000 Linksys Dual Small mt7610
AC600 (A6100) Netgear Dual Small rtl8812au
N300 (WNA3100M) Netgear 2.4 GHz Small rtl8192cu
G54/N150 (WNA1000M) Netgear 2.4 GHz Nano rtl8192cu
N300 F7D2102 Belkin 2.4 GHz Small rtl8192cu r8712u
AC450 (Archer T1U) TP-Link 5 GHz Nano
AC1300 (Archer T4U) TP-Link Dual Long
AC1200 (A6200) Netgear Dual Long
N900 (WNDA4100) Netgear Dual Long
N600 (WNDA3100) Netgear Dual Long
AE2500 N600 Linksys Dual Long Broadcom BCM43236
AE3000 N900 Linksys Dual Long
WUSB6300 AC1200 Linksys Dual Long 0bda:0153
N300 (WNA3100) Netgear 2.4 GHz Long
N150 (WNA1100) Netgear 2.4 GHz Long ath9k_htc
AE1200 N300 Linksys 2.4 GHz Long
N300 F9L1002 Belkin 2.4 GHz Long 050d:845a RTL8192SU
DWA-140 D-Link 2.4 GHz Long rt2870

Files


Subtasks

Feature #15858: Investigate the chips of the Panda wireless adapters Resolved

0


Related issues

Related to Tails - Feature #14544: Spend software developer time on smallish UX improvements In Progress 2018-08-31
Blocks Tails - Bug #11683: Improve installation instructions when Wi-Fi is not working (especially on macOS) Resolved 2017-12-29
Blocks Tails - Feature #15941: Core work 2018Q4 → 2019Q2: Technical writing Resolved 2018-09-11

History

#1 Updated by sajolida 2017-12-29 05:59:00

  • Tracker changed from Bug to Feature

Common vendors:

Other ressources:

I bought 4 adapters (marked with an ‘X’).

|   | Vendor  | Model                      | Band    | Price | Driver / Chip              | Size  |
|---+---------+----------------------------+---------+-------+----------------------------+-------|
| X | TP-Link | AC600 (Archer T2U)         | Dual    | $17   | mt7610                     | Small |
|   | Linksys | AE6000                     | Dual    | $30   | mt7610                     | Small |
|   | Netgear | AC600 (A6100)              | Dual    | $34   | rtl8812au                  | Small |
| X | D-Link  | DWA-171                    | Dual    | $15   | rtl8812au                  | Small |
| X | Linksys | WUSB6100M Max-Stream AC600 | Dual    | $15   | rt2800usb                  | Small |
|   |         |                            |         |       |                            |       |
| X | TP-Link | TL-WN823N                  | 2.4 GHz | $13   | rtl8xxxu / rtl8192eu       | Small |
|   | Netgear | N300 (WNA3100M)            | 2.4 GHz | $40   | rtl8192cu                  | Small |
|   | Netgear | G54/N150 (WNA1000M)        | 2.4 GHz | $90   | rtl8192cu                  | Nano  |
|   | Belkin  | N300 F7D2102               | 2.4 GHz | $19   | rtl8192cu r8712u           | Small |
|   | D-Link  | DWA-131                    | 2.4 GHz | $18   | rtl8192cu rtl8188eu r8712u | Small |
|   |         |                            |         |       |                            |       |
|   | TP-Link | AC450 (Archer T1U)         | 5 GHz   |       |                            | Nano  |
|   | TP-Link | TL-WN725N                  | 2.4 GHz |       |                            | Nano  |
|   |         |                            |         |       |                            |       |
|   | TP-Link | AC1300 (Archer T4U)        | Dual    |       |                            | Long  |
|   | Netgear | AC1200 (A6200)             | Dual    |       |                            | Long  |
|   | Netgear | N900 (WNDA4100)            | Dual    |       |                            | Long  |
|   | Netgear | N600 (WNDA3100)            | Dual    |       |                            | Long  |
|   | Linksys | AE2500 N600                | Dual    |       | Broadcom BCM43236          | Long  |
|   | Linksys | AE3000 N900                | Dual    |       |                            | Long  |
|   | Linksys | WUSB6300 AC1200            | Dual    |       | 0bda:0153                  | Long  |
|   | Netgear | N300 (WNA3100)             | 2.4 GHz |       |                            | Long  |
|   | Netgear | N150 (WNA1100)             | 2.4 GHz |       | ath9k_htc                  | Long  |
|   | Linksys | AE1200 N300                | 2.4 GHz |       |                            | Long  |
|   | Belkin  | N300 F9L1002               | 2.4 GHz |       | 050d:845a RTL8192SU        | Long  |
|   | D-Link  | DWA-140                    | 2.4 GHz |       | rt2870                     | Long  |

The 3 dual adapters don’t load any driver in Tails 3.3:

D-Link - AC600 - DWA-171:

Dec 29 00:35:54 amnesia kernel: usb 1-1.1: new high-speed USB device number 6 using ehci-pci
Dec 29 00:35:54 amnesia kernel: usb 1-1.1: New USB device found, idVendor=2001, idProduct=3314
Dec 29 00:35:54 amnesia kernel: usb 1-1.1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Dec 29 00:35:54 amnesia kernel: usb 1-1.1: Product: 802.11n WLAN Adapter
Dec 29 00:35:54 amnesia kernel: usb 1-1.1: Manufacturer: Realtek
Dec 29 00:35:54 amnesia kernel: usb 1-1.1: SerialNumber: 00e04c000001
Dec 29 00:35:54 amnesia systemd-udevd[24613]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1 1 6': No such file or directory
Dec 29 00:35:55 amnesia laptop-mode[24652]: enabled, not active
Dec 29 00:35:55 amnesia laptop-mode[25142]: enabled, not active

Linksys - AC600 - Max-Stream:

Dec 29 00:37:15 amnesia kernel: usb 1-1.2: new high-speed USB device number 7 using ehci-pci
Dec 29 00:37:16 amnesia kernel: Dropped outbound packet: IN= OUT=wlan0 SRC=fe80:0000:0000:0000:f831:837e:b6e9:f912 DST=ff02:0000:0000:0000:0000:0000:0000:0002 LEN=48 TC=0 HOPLIMIT=255 FLOWLBL=495350 PROTO=ICMPv6 TYPE=133 CODE=0 UID=0 GID=0
Dec 29 00:37:16 amnesia kernel: usb 1-1.2: New USB device found, idVendor=13b1, idProduct=0042
Dec 29 00:37:16 amnesia kernel: usb 1-1.2: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Dec 29 00:37:16 amnesia kernel: usb 1-1.2: Product: USBWLAN
Dec 29 00:37:16 amnesia kernel: usb 1-1.2: Manufacturer: Qualcomm Atheros
Dec 29 00:37:16 amnesia kernel: usb 1-1.2: SerialNumber: 12345678
Dec 29 00:37:16 amnesia systemd-udevd[25666]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.2 1 7': No such file or directory
Dec 29 00:37:16 amnesia laptop-mode[25705]: enabled, not active
Dec 29 00:37:16 amnesia laptop-mode[26195]: enabled, not active

TP-Link - AC600 - Archer T2U

Dec 29 00:39:22 amnesia kernel: usb 1-1.2: new high-speed USB device number 8 using ehci-pci
Dec 29 00:39:22 amnesia kernel: usb 1-1.2: New USB device found, idVendor=148f, idProduct=761a
Dec 29 00:39:22 amnesia kernel: usb 1-1.2: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Dec 29 00:39:22 amnesia kernel: usb 1-1.2: Product: WiFi
Dec 29 00:39:22 amnesia kernel: usb 1-1.2: Manufacturer: MediaTek
Dec 29 00:39:22 amnesia kernel: usb 1-1.2: SerialNumber: 1.0
Dec 29 00:39:22 amnesia systemd-udevd[26699]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.2 1 8': No such file or directory
Dec 29 00:39:23 amnesia laptop-mode[26738]: enabled, not active
Dec 29 00:39:23 amnesia laptop-mode[27228]: enabled, not active

The TP-Link TL-WN823N loads the rtl8192eu driver but I can’t authenticate to the access points I have here.

I’ll try again on other access points.

#2 Updated by sajolida 2018-01-07 19:49:12

  • Assignee changed from sajolida to intrigeri
  • QA Check set to Info Needed

Tried again the TL-WN823N on a different access point and got the same authentication problem.

I’ll also mention here that I tried a TP-Link TL-WN725N in the past (the cheapest and smallest adapter around). It worked but MAC spoofing failed.

The conclusions from my first tests is that out of 5 of most commonly available Wi-Fi USB adapters, having all different Wi-Fi chips, none work out of the box in Tails.

Solving this is out of my skillset, mission, and comfort zone and I need guidance from the Foundations Team on what to do next:

  • Buy more models in hope that we’ll find one that works. But from my initial research we might not be able to find more models that are both small and use a Wi-Fi chip that we haven’t tried yet. Apart from the TP-Link AC450 (Archer T1U) which is 5 GHz only.
  • Debug the ones we already bought. I can hand them over to the Foundations Team in February.

#3 Updated by intrigeri 2018-01-09 14:53:05

  • Assignee changed from intrigeri to sajolida

> The conclusions from my first tests is that out of 5 of most commonly available Wi-Fi USB adapters, having all different Wi-Fi chips, none work out of the box in Tails.

:/

> Solving this is out of my skillset, mission, and comfort zone and I need guidance from the Foundations Team on what to do next:

Deal!

> * Buy more models in hope that we’ll find one that works. But from my initial research we might not be able to find more models that are both small and use a Wi-Fi chip that we haven’t tried yet. Apart from the TP-Link AC450 (Archer T1U) which is 5 GHz only.

> * Debug the ones we already bought. I can hand them over to the Foundations Team in February.

First of all, I think we should focus on Wi-Fi adapters that have a chance to work, i.e. are supported by a driver that’s is in Linux mainline and thus part of Debian:

  • the mt7610 and rtl8812au modules are not in Linux mainline nor in Debian, which explains why no driver is loaded.
  • Wrt. Linksys WUSB6100M Max-Stream AC600, in the current mainline Linux tree I see it should be supported by the ath10k_usb driver (drivers/net/wireless/ath/ath10k/usb.c: {USB_DEVICE(0x13b1, 0x0042)}, /* Linksys WUSB6100M */) rather than by the rt2800usb one as the table above says. The commit that added this support 6 months ago says “WARNING: ath10k USB support is incomplete, don’t expect anything to work!” and the module is not built on Debian yet.

So sadly, there’s nothing to debug for these first 3 adapters. Let’s just keep the them around until Debian supports them (likely for the ath10k one, unlikely for the 2 other ones). Note that Ubuntu has a history of including third-party drivers so their list of supported hardware is probably not a good resource for us (sorry if I let you think otherwise or even pointed you there!).

Wrt. the last adapter (TL-WN823N), the table says rtl8192cu but you wrote “loads the rtl8192eu driver” and I see no rtl8192eu module in Tails 3.3, so I’ll assume you meant “loads the rtl8192cu driver” and/or “load the rtl8192eu firmware”. There are at least a V1 (RTL8192 CU) and a V2 (RTL8192 EU) of this device so it might help to clarify which one we’re talking about here. It would be nice to know if the firmware is correctly loaded. I’d happily take a look at the Journal to investigate a bit (no need to wait until February, I’ve had to learn how to triage hardware issues remotely). According to https://wireless.wiki.kernel.org/en/users/Drivers/rtl819x, if we’re talking about V2 with a RTL8192 EU chipset, the correct driver would be rtl8xxxu, that’s in Tails 3.3; but it might be one of those cases where multiple drivers announce support for the same hardware and one needs to help the kernel pick the right one (even worse when the manufacturer actually kept using the same device ID for both generations); if we’re in this case please unload the driver that’s automatically loaded and try to modprobe rtl8xxxu.

#4 Updated by sajolida 2018-01-19 18:24:42

#5 Updated by sajolida 2018-01-19 18:24:54

#6 Updated by sajolida 2018-01-19 18:25:24

  • Target version set to 2018

#7 Updated by sajolida 2018-01-19 18:42:59

NB: I’m not sure about what’s the difference between “module”, “firmware”, and “driver”.

Here is a log when pluggin in TL-WN823N

Jan 19 18:27:33 amnesia kernel: usb 1-1.1: new high-speed USB device number 13 using ehci-pci
Jan 19 18:27:33 amnesia kernel: usb 1-1.1: New USB device found, idVendor=2357, idProduct=0109
Jan 19 18:27:33 amnesia kernel: usb 1-1.1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Jan 19 18:27:33 amnesia kernel: usb 1-1.1: Product: 802.11n NIC
Jan 19 18:27:33 amnesia kernel: usb 1-1.1: Manufacturer: Realtek
Jan 19 18:27:33 amnesia kernel: usb 1-1.1: SerialNumber: 00e04c000001
Jan 19 18:27:33 amnesia systemd-udevd[5230]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1 1 13': No such file or directory
Jan 19 18:27:33 amnesia laptop-mode[5269]: enabled, not active
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: Vendor: Realtek
Jan 19 18:27:34 amnesia kernel: [31B blob data]
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: Serial:
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: rtl8192eu_parse_efuse: dumping efuse (0x200 bytes):
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: 00: 29 81 00 7c 01 40 03 00
[...]
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: 1f8: ff ff ff ff ff ff ff ff
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: RTL8192EU rev B (SMIC) 2T2R, TX queues 3, WiFi=1, BT=0, GPS=0, HI PA=0
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: RTL8192EU MAC: 7c:8b:ca:1a:5f:9b
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: rtl8xxxu: Loading firmware rtlwifi/rtl8192eu_nic.bin
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: firmware: direct-loading firmware rtlwifi/rtl8192eu_nic.bin
Jan 19 18:27:34 amnesia kernel: usb 1-1.1: Firmware revision 19.0 (signature 0x92e1)
Jan 19 18:27:35 amnesia kernel: usb 1-1.1: rtl8192eu_rx_iqk_path_a: Path A RX IQK failed!
Jan 19 18:27:35 amnesia kernel: usb 1-1.1: rtl8192eu_rx_iqk_path_a: Path A RX IQK failed!
Jan 19 18:27:35 amnesia NetworkManager[13655]: <info>  [1516386455.8934] (wlan1): using nl80211 for WiFi device control
Jan 19 18:27:35 amnesia systemd[1]: Starting Load/Save RF Kill Switch Status...
Jan 19 18:27:35 amnesia NetworkManager[13655]: <info>  [1516386455.8973] manager: (wlan1): new 802.11 WiFi device (/org/freedesktop/NetworkManager/Devices/4)
Jan 19 18:27:35 amnesia kernel: usbcore: registered new interface driver rtl8xxxu
Jan 19 18:27:35 amnesia laptop-mode[5767]: enabled, not active
Jan 19 18:27:36 amnesia laptop-mode[6217]: Failed to re-set power saving mode for wireless card
Jan 19 18:27:36 amnesia systemd[1]: Started Load/Save RF Kill Switch Status.
Jan 19 18:27:36 amnesia NetworkManager[13655]: <info>  [1516386456.1640] rfkill1: found WiFi radio killswitch (at /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1/1-1.1:1.0/ieee80211/phy1/rfkill1) (driver rtl8xxxu)
Jan 19 18:27:36 amnesia spoof-mac[6227]: Trying to spoof MAC address of NIC wlan1...
Jan 19 18:27:36 amnesia spoof-mac[6243]: Successfully spoofed MAC address of NIC wlan1
Jan 19 18:27:36 amnesia NetworkManager[13655]: <info>  [1516386456.2164] devices added (path: /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1/1-1.1:1.0/net/wlan1, iface: wlan1)
Jan 19 18:27:36 amnesia NetworkManager[13655]: <info>  [1516386456.2165] device added (path: /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1/1-1.1:1.0/net/wlan1, iface: wlan1): no ifupdown configuration found.
Jan 19 18:27:36 amnesia kernel: IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready

So it seems like we’re loading RTL8192EU from rtl8xxxu.

When I do modprobe rtl8xxxu and try to plug the USB again I get the same behavior.

#8 Updated by intrigeri 2018-01-20 15:10:19

> NB: I’m not sure about what’s the difference between “module”, “firmware”, and “driver”.

OK :) The simplified version is:

  • Module = driver = the code loaded in the Linux kernel that runs on your CPU to support the hardware (modinfo, modprobe, etc.). Tails includes only free software kernel modules that come with the Debian Linux kernel.
  • Firmware = the code that’s uploaded by the Linux kernel from disk into the memory of the Wi-Fi device. Most such firmware is proprietary these days; we include lots of them in Tails (from the non-free section of Debian), hence the controversies about whether Tails is truly® free™ software©.

> Here is a log when pluggin in TL-WN823N […]

> So it seems like we’re loading RTL8192EU from rtl8xxxu.

Right, the kernel detects the device, notices the rtl8xxxu module claims to support it, loads that module which in turn triggers the uploading of the rtl8192eu_nic.bin firmware into the device. So far so good. Now to help more remotely I would need to see the Journal corresponding to the failure mode you’ve reported.

Other things you/we could try:

  • use rfkill to ensure the device is not blocked on that side
  • try without spoofing the MAC address even though someone reports this does not fix the problem you’re facing

#9 Updated by intrigeri 2018-01-20 15:44:17

> Other things you/we could try:

Also, you can try booting with rtl8xxxu.debug=1 on the kernel command line (or a value >> 1) to get more debugging info out of the module.

#10 Updated by sajolida 2018-02-01 23:07:16

  • File journal.pgp added
  • Assignee changed from sajolida to intrigeri
  • QA Check deleted (Info Needed)

I tried without MAC spoofing and with rtl8xxxu.debug=1 and got the same authentication errors. Here is the output.

#11 Updated by sajolida 2018-02-02 09:59:37

But I was very happy to finally learn what a firmware really is (at my age)!

#12 Updated by intrigeri 2018-02-02 14:25:52

  • Assignee changed from intrigeri to sajolida
  • QA Check set to Info Needed

sajolida wrote:
> I tried without MAC spoofing and with rtl8xxxu.debug=1 and got the same authentication errors. Here is the output.

I’d like to see the full Journal, please. Otherwise it’s hard for me to correlate bits and pieces about multiple attempts, scattered around the place on this ticket.

#13 Updated by sajolida 2018-02-19 19:15:18

I thought that /code/attachments/download/1882/journal.pgp contained a full WhisperBack log and that included a “full Journal”. But I didn’t encrypt it for myself so I can’t check.

I’ll upload a new one another day…

#14 Updated by intrigeri 2018-02-19 20:55:37

> I thought that /code/attachments/download/1882/journal.pgp contained a full WhisperBack log and that included a “full Journal”. But I didn’t encrypt it for myself so I can’t check.

I downloaded it again and checked again then. What I see is a 3.6MB except from the Journal that was apparently saved by hand (not by WhisperBack nor by sudo tails-debugging-info).

#15 Updated by sajolida 2018-03-27 17:27:21

I bought 2 new adapters:

  • DWA-131 which loads the same module and firmware as TL-WN823N and suffers from the same authentication failure.
  • TL-WN821N which loads no module.

I’m attaching WhisperBack logs for each of the adapters I have now.

I’m adding to the description of this ticket a table of our findings.

I also found (but didn’t buy) an ASUS USB AC53.

#16 Updated by intrigeri 2018-03-27 19:02:55

  • Assignee changed from intrigeri to sajolida

sajolida wrote:
> I bought 2 new adapters:
>
> * DWA-131 which loads the same module and firmware as TL-WN823N and suffers from the same authentication failure.

That one is idVendor=2001, idProduct=3319. Perhaps we should add such info to the table (“DWA-131” does not mean much as there are several versions of it). The logs don’t tell me more than what you’re reporting.

> * TL-WN821N which loads no module.

This seems to be a TL-WN821N v5 (idVendor=2357, idproduct=0107). I would have thought that commit 5407fd7de69f3352aed659244d4bef18e3cabf5c in linux.git added support for that one in the rtl8xxxu module but apparently it’s not the case (modprobe -c | grep -i "2357.*0107" returns nothing), which is weird because 2357:0108, that’s declared in exactly the same way in drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c, is listed (alias usb:v2357p0108d*dc*dsc*dp*icFFiscFFipFFin* rtl8xxxu).

I’m not sure what you’re expecting from me. Do you want to tell me what you should do next? If so, maybe buy one that’s flagged as supported out of the box on https://help.ubuntu.com/community/HardwareSupportComponentsWirelessNetworkCardsTP-Link? Beware of those that have multiple versions with differing support levels, unless you can be sure of what version you’re buying.

#17 Updated by sajolida 2018-04-08 09:58:01

  • Description updated

#18 Updated by sajolida 2018-04-08 10:27:15

  • Assignee changed from sajolida to intrigeri

>> * DWA-131 which loads the same module and firmware as TL-WN823N and suffers from the same authentication failure.
>
> That one is idVendor=2001, idProduct=3319. Perhaps we should add such info to the table (“DWA-131” does not mean much as there are several versions of it). The logs don’t tell me more than what you’re reporting.

I added the vendor and product IDs to the table and fixed a bit the
formatting.

>> * TL-WN821N which loads no module.
>
> This seems to be a TL-WN821N v5 (idVendor=2357, idproduct=0107). I would have thought that commit 5407fd7de69f3352aed659244d4bef18e3cabf5c in linux.git added support for that one in the rtl8xxxu module but apparently it’s not the case (modprobe -c | grep -i "2357.*0107" returns nothing), which is weird because 2357:0108, that’s declared in exactly the same way in drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c, is listed (alias usb:v2357p0108d*dc*dsc*dp*icFFiscFFipFFin* rtl8xxxu).

Should we investigate this more and report it upstream?

> I’m not sure what you’re expecting from me.

Me neither! :(

> Do you want to tell me what you should do next? If so, maybe buy one that’s flagged as supported out of the box on https://help.ubuntu.com/community/HardwareSupportComponentsWirelessNetworkCardsTP-Link?

Great idea!

I checked the list against:

  • The TP-Link website (some models are end-of-life and some others are
    not produced anymore).
  • Amazon.com to see if they were sold.

I’m tempted to buy:

  • TL-WN321G which is not produced anymore but still sold on Amazon and
    only has positive reviews (and many).
  • TL-WN723N which is still produced and has one positive review.

Shall I do that?

> Beware of those that have multiple versions with differing support levels, unless you can be sure of what version you’re buying.

That’s a shame because I’m afraid it might be very hard for our users to
shop for a given model of an adapter. That’s why I filtered the ones who
only have positive reviews, as a beginning…

#19 Updated by sajolida 2018-04-08 10:28:00

From https://help.ubuntu.com/community/HardwareSupportComponentsWirelessNetworkCardsTP-Link

Model Works tp-link.com amazon.com Size
TL-WN321G Yes Not listed $12 Long
TL-WN721N Yes End of life $15 Long
TL-WN722N Yes Listed $22 Long with antenna
TL-WN723N Yes Listed $7 Small
TL-WN727N Yes & No
TL-WN821N Yes & No
TL-WN822N Yes & No

#20 Updated by intrigeri 2018-04-10 08:11:30

  • related to Feature #14544: Spend software developer time on smallish UX improvements added

#23 Updated by sajolida 2018-05-14 19:29:21

  • Assignee changed from intrigeri to goupille

> When setting such priorities […]

My wild guess (and personal experience) is that the support for internal Wi-Fi interfaces is much better than what I found out for easily available USB Wi-Fi adapters. I guess that that the few people who have a broken internal Wi-Fi, either give up on Tails or Wi-Fi in Tails but that didn’t found a USB Wi-Fi adapter (or never tried to get one) to replace it.

> So perhaps we should ask them?

Help desk: roughly how frequent are complains about Wi-Fi not working compaired to graphics card not working?

>> It’s also on our roadmap.
>
> Absolutely. Note however that when we added it to our roadmap, it was supposed to be done as volunteer work and it was not clear that engineering work would be required.

I added it (git grep 11683) in 2016 with a ‘$’ when I was alone on the issue. In 2017 BitingBird put her name with a ‘♥’. It might be one of these occurrences where I thought that it would be weird to work in a mixed team of paid/unpaid people and joined the unpaid side by default. The fact now is that:

  • I’ve never heard back from BitingBird.
  • This happened to be much more work than expected because of issues I don’t really have the skills to debug.

So yeah… I moved it to my UX plate without waiting for the summit to clarify that I’m not ready to do this as volunteer alone. Maybe I should have been more explicit or I should stop working on this for now seeing that it’s much more work than expected.

> I did, mostly because at the summit, this was deemed to be a blocker for the “Spend less time on handling individual support requests” change in help desk’s mission. Otherwise I would not have dared counting this as FT work unless there was budget left for this.

Thanks for the clarification.

>> For me this issue is much more critical than a “smallish UX improvement”.
>
> Ouch, I think there’s been a misunderstanding! I really did not meant this ticket is not important / not critical. The way I meant it when proposing Feature #14544 for our roadmap, “smallish” meant “small amount of dev work required”, i.e. low hangin fruits, and not “small impact / not very important”. It looks like you instead mean (and perhaps meant) “smallish” applies to the impact/importance, not to the amount of work required. In this case, I don’t know about the impact (I’d like help desk to help us define it) but wrt. amount of dev work needed, what we’ve seen so far on this ticket leads me to think that it definitely won’t be “smallish”.

We’re already discussing on Feature #14544 about the meaning that I had in mind when I wrote “smallish” on this post-it. Both times your version of “smallish” seemed much smaller than what I had in mind originally. For me “smallish” could mean pretty much anything that wouldn’t be worth justifying on its own in a grant application :)

#24 Updated by sajolida 2018-05-15 09:37:14

Highlighting the info we need from help desk:

Please don’t read everything but tell us roughly how frequent are complains about Wi-Fi not working compaired to graphics card not working.

#25 Updated by goupille 2018-05-15 12:05:29

  • Assignee changed from goupille to sajolida

I can only give you a rough estimation, I’d say there are between one and three reports a month about wi-fi not working, and at least ten reports a month for graphic issues…

#26 Updated by sajolida 2018-05-18 20:36:52

  • QA Check deleted (Info Needed)

Thanks for the quick info goupille.

I think that I’ll go on and buy one TL-WN723N online. If it doesn’t work either, then I’ll stop working on this ticket until we have more engineering resources to dedicate to it.

This ticket would have proven much more difficult than expected and we should reconsider it’s cost/benefit when updating our roadmap this summer.

#27 Updated by intrigeri 2018-05-25 14:58:31

> or I should stop working on this for now seeing that it’s much more work than expected.

I trust you to manage your UX budget in an excellent way, taking your skills into account, so that’s your call :)

In any case, even if you stop working on this now, we’ll have learnt something important: this requires much more work than we thought initially, so perhaps it’s not the best strategy to address the underlying UX problem and we should go back to the drawing board. Or at least we’ll know better how to integrate this into our roadmap next time! So all the time we’ve both spent on it is not wasted time IMO.

> We’re already discussing on Feature #14544 about the meaning that I had in mind when I wrote “smallish” on this post-it. Both times your version of “smallish” seemed much smaller than what I had in mind originally. For me “smallish” could mean pretty much anything that wouldn’t be worth justifying on its own in a grant application :)

OK. Then please clarify this on Feature #14544 because there I only see my own attempts at defining “smallish” :)

#28 Updated by intrigeri 2018-05-25 14:58:48

> I think that I’ll go on and buy one TL-WN723N online. If it doesn’t work either, then I’ll stop working on this ticket until we have more engineering resources to dedicate to it.

> This ticket would have proven much more difficult than expected and we should reconsider it’s cost/benefit when updating our roadmap this summer.

Makes sense to me!

#29 Updated by sajolida 2018-06-25 16:27:02

  • blocks Feature #15411: Core work 2018Q2 → 2018Q3: Technical writing added

#30 Updated by cbrownstein 2018-08-29 12:41:04

I own a TP-Link TL-WN722N[1] v1.x which I purchased specifically for
use with Tails. It works flawlessly. I emphasize that the adapter I own
is a version 1.x because there are multiple versions of this adapter.
The different versions use different chipsets: version 1.x uses the
Atheros AR9271 chipset; and versions 2.x and up use the Realtek
RTL8188EUS chipset[2]. I do not have experience with other than version
1.x. But, quick research tells me that the other versions are not as
well-supported by Linux as version 1.x is[3].

If multiple versions of an adapter exist, there is a likelihood that the
user will buy the “incorrect” version of the adapter. For example,

https://www.amazon.com/gp/customer-reviews/RDCADVGT5FXLM/

> After purchasing this, I found it does not properly work with some
> Linux setups that were usually guaranteed to work (Kali). After some
> reaserch, I found that the adapter has two versions - one with an
> Atheros Chipset, and one with a RealTek Chipset.

While researching Wi-Fi adapters that work well with Tails, I came
across adapters made by Panda Wireless[4]. Their Wi-Fi adapters are
expressly advertised as being compatible with Linux. And, their Wi-Fi
adapters are readily available on Amazon[5]. If desired, I can purchase
one or two of these adapters to confirm that they work with Tails.

If the user wishes to entirely avoid possible Wi-Fi adapter issues, it
makes sense for the user to purchase a wireless bridge. See, e.g.:

http://us.dlink.com/resource-centre/how-to-guides/how-to-extend-your-network-with-a-wireless-bridge/

With a wireless bridge, the user does not have to be concerned with
buying a supported Wi-Fi adapter. As long as the bridge has a working
Wi-Fi connection, all the user has to do is plug into the bridge with an
Ethernet cable. This being said, however, many newer laptops do not have
an Ethernet port, for example, the ThinkPad X1 Carbon[6] (6th gen) and
MacBook Pro[7].

[1] https://www.tp-link.com/us/products/details/TL-WN722N.html

[2] https://wikidevi.com/wiki/TP-LINK_TL-WN722N_v1.x

[3]
https://www.wirelesshack.org/best-kali-linux-compatible-usb-adapter-dongles-2016.html

[4] http://pandawireless.com/

[5] https://www.amazon.com/s/field-keywords=panda+wireless

[6]
https://www.lenovo.com/us/en/laptops/thinkpad/thinkpad-x/ThinkPad-X1-Carbon-6th-Gen/p/22TP2TXX16G

[7] https://www.apple.com/macbook-pro/specs/

#31 Updated by sajolida 2018-08-30 05:58:42

Thanks Cody for the detailed report!

I’m also worried about advertising adapters that might have different vendor versions and might or might not work for the user in the end.

I didn’t know about Panda wireless and it looks very promising!

My next step is to ask them more info about their chips. I’d like to know the chips inside before buying them. Note that they are not all advertised as being compatible with Linux, for example: http://www.pandawireless.com/pandaAC600.htm.

#32 Updated by sajolida 2018-09-02 19:46:00

I tested again my 9 Wi-Fi adapters with Tails 3.9 and they work the same.

#33 Updated by sajolida 2018-09-09 19:20:05

  • blocks Bug #11683: Improve installation instructions when Wi-Fi is not working (especially on macOS) added

#34 Updated by sajolida 2018-09-11 17:57:15

  • blocked by deleted (Feature #15411: Core work 2018Q2 → 2018Q3: Technical writing)

#35 Updated by sajolida 2018-09-11 17:57:18

  • related to Feature #15941: Core work 2018Q4 → 2019Q2: Technical writing added

#36 Updated by sajolida 2018-09-11 17:57:26

  • related to deleted (Feature #15941: Core work 2018Q4 → 2019Q2: Technical writing)

#37 Updated by sajolida 2018-09-11 17:57:28

  • blocks Feature #15941: Core work 2018Q4 → 2019Q2: Technical writing added

#38 Updated by sajolida 2018-11-18 17:42:31

  • Target version changed from 2018 to Tails_3.11

#39 Updated by sajolida 2018-12-10 15:47:07

  • Target version changed from Tails_3.11 to Tails_3.12

#40 Updated by sajolida 2018-12-17 13:19:43

  • Assignee deleted (sajolida)

I think we should document only the Panda Wireless adapters while still keeping an eye on adapters commonly available on brick and mortar shops.

#41 Updated by sajolida 2018-12-17 13:19:56

  • Status changed from Confirmed to Resolved

#42 Updated by sajolida 2019-01-16 18:00:12

  • Blueprint set to https://tails.boum.org/blueprint/wi-fi_adapters/

I summarizes all this in a blueprint: https://tails.boum.org/blueprint/wi-fi_adapters/.

#43 Updated by emmapeel 2019-02-14 08:05:03

reported by a user as not working:

> 1. ASUS USB-AC56
> 2. TP-Link TL-WN722N v3

Not recognized at all by Tails.

#44 Updated by sajolida 2019-02-19 18:11:38

Thanks! I think I’m only interested in knowing about adapters that work, so I can provide more useful information.

#45 Updated by numbat 2019-11-07 11:02:58

A user just reported that the Netgear WG111v2 USB wifi card works well with Tails.

#46 Updated by sajolida 2019-11-14 17:16:34

> A user just reported that the Netgear WG111v2 USB wifi card works well with Tails.

Thanks! I’ll try to get one.