OpenWRT 24.10

[ZerBea]

OpenWRT 24.10 is available now:

$ git branch -a
  master
  openwrt-23.05
* openwrt-24.10
  remotes/origin/HEAD -> origin/master
  remotes/origin/dgs-1210-28p-luizluca
  remotes/origin/lede-17.01
  remotes/origin/main
  remotes/origin/master
  remotes/origin/openwrt-18.06
  remotes/origin/openwrt-19.07
  remotes/origin/openwrt-21.02
  remotes/origin/openwrt-22.03
  remotes/origin/openwrt-23.05
  remotes/origin/openwrt-24.10

Initial tests have shown that 24.10 is working as fine.

Some notes:
The init process of some devices is slow.
Add "sleep 10" to an init.d script to make sure the device has been initialized before hcxdumptool/hcxlabtool starts.

Running a minimalist OpenWRT, it could be possible that a WiFi USB adapter has not been assigned during boot.
Add "sudo iw phy phyx interface add wlan0 type managed" to an init.d script to make sure the interface has been assigned (hcxdumptool/hcxlabtool does not hijack unassigned interfaces).

[ZerBea]

Update:
Build failed on latest git head (as of today):
openwrt/openwrt#16905

[strasharo]

Damn, I was literally just about to start building it. 😆 What are your general impressions, is it stable enough or it's better to stick with openwrt-23.05 for now?

[ZerBea]

I'm still testing 24.10 (compile last week). It is working fine and stable.

But there are some negative aspects:
OpenWRT Linux kernel 6.6.58 is running fine, but it looks like the rtl8xxxu (kmod-rtl8xxxu) is not the same as the Linux stock kernel driver.
While monitor mode on the Linux stock kernel driver is working fine, it doesn't work on the OpenWRT driver.

[ZerBea]

I switched to the master branch and it (mac80211) compiles fine:

 make[3] -C package/network/config/wifi-scripts compile
 make[3] -C package/kernel/mac80211 compile
 make[3] -C package/libs/ncurses compile

[ZerBea]

Looks like the newly compiled image (from master branch) is working as expected. Now show stopper yet.

[ZerBea]

After some initial hinds, OpenWRT 24.10 performs well (as of today).
Initial tests, running a cheap Realtek device TL-WN722N v2/3
are impressive.

[ZerBea]

To disable the red PWR LED and/or green ACT LED of a Raspberry Pi add this 2 lines to an init.d script:

if [ -f "/sys/class/leds/PWR/brightness" ]; then echo 0 > /sys/class/leds/PWR/brightness; fi
if [ -f "/sys/class/leds/ACT/brightness" ]; then echo 0 > /sys/class/leds/ACT/brightness; fi

Estimated power savings about 5 mA.

[strasharo]

Hey @ZerBea, just noticed the following line in lessons_learned.txt:
"sharing the interface (e.g. iw phy phy1 interface add mon0 type monitor) is counterproductive"

By default in Openwrt I just get phy0 without an interface name assigned to it. What would be the best way to use it in this case?

[ZerBea]

OpenWRT does not assign interfaces and hcxdumptool/hcxlabtool does not hijack unassigned interfaces. A solution could be to assign the interface during boot.
A simple script, placed in /etc/init.d should do the trick:

#!/bin/sh /etc/rc.common
START=99

boot()
	{
	/bin/sleep 10
	/usr/sbin/iw phy phy0 interface add wlan0 type managed
	}

Now hcxdumptool/hcxlabtool sees an assigned managed interface and can set monitor mode.
sleep 10 is useful to make sure OpenWRT got a list of all phy interfaces before we assign it.
#478 (comment)

[strasharo]

Thank you.

[ZerBea]

Full rtw88 USB support has been added to OpenWRT 24.10 (latest git head);

package/firmware/linux-firmware/realtek.mk                   |   14 +
 ...001-v6.12-wifi-rtw88-Set-efuse-ext_lna_5g-fix-typo.patch |   30 +
 ...fi-rtw88-usb-Support-USB-3-with-RTL8822CU-RTL8822B.patch |  284 +++
 ...fi-rtw88-debugfs-support-multiple-adapters-debuggi.patch |  454 ++++
 .../rtl/005-v6.12-wifi-rtw88-select-WANT_DEV_COREDUMP.patch |   25 +
 ...fi-rtw88-8822c-Parse-channel-from-IE-to-correct-in.patch |  151 ++
 ...fi-rtw88-usb-Init-RX-burst-length-according-to-USB.patch |   73 +
 ...fi-rtw88-usb-Update-the-RX-stats-after-every-frame.patch |   31 +
 .../012-v6.12-wifi-rtw88-usb-Support-RX-aggregation.patch   |  118 +
 ...fi-rtw88-Enable-USB-RX-aggregation-for-8822c-8822b.patch |  168 ++
 ...fi-rtw88-assign-mac_id-for-vif-sta-and-update-to-T.patch |  231 ++
 ...-v6.13-wifi-rtw88-Constify-some-arrays-and-structs.patch |  374 ++++
 ...fi-rtw88-Parse-the-RX-descriptor-with-a-single-fun.patch |  560 +++++
 ...12-wifi-rtw88-Fix-the-RX-aggregation-in-USB-3-mode.patch |   70 +
 ...fi-rtw88-Refactor-looping-in-rtw_phy_store_tx_powe.patch |   34 +
 ...fi-rtw88-Report-the-signal-strength-only-if-it-s-k.patch |   39 +
 ...fi-rtw88-Add-some-definitions-for-RTL8821AU-RTL881.patch |  490 +++++
 ...ifi-rtw88-Dump-the-HW-features-only-for-some-chips.patch |   93 +
 ....13-wifi-rtw88-Allow-different-C2H-RA-report-sizes.patch |  175 ++
 ...fi-rtw88-Extend-the-init-table-parsing-for-RTL8812.patch |  165 ++
 ...fi-rtw88-Allow-rtw_chip_info.ltecoex_addr-to-be-NU.patch |   58 +
 ...fi-rtw88-Let-each-driver-control-the-power-on-off-.patch |  272 +++
 ...fi-rtw88-Enable-data-rate-fallback-for-older-chips.patch |  194 ++
 ...-v6.13-wifi-rtw88-Make-txagc_remnant_ofdm-an-array.patch |   85 +
 ...3-wifi-rtw88-Support-TX-page-sizes-bigger-than-128.patch |  115 +
 ...ifi-rtw88-Move-pwr_track_tbl-to-struct-rtw_rfe_def.patch |  297 +++
 ...fi-rtw88-usb-Set-pkt_info.ls-for-the-reserved-page.patch |   27 +
 ...fi-rtw88-Detect-beacon-loss-with-chips-other-than-.patch |   63 +
 ...wifi-rtw88-coex-Support-chips-without-a-scoreboard.patch |   66 +
 ...wifi-rtw88-8821a-Regularly-ask-for-BT-info-updates.patch |   67 +
 .../038-v6.13-wifi-rtw88-8812a-Mitigate-beacon-loss.patch   |   37 +
 .../rtl/039-v6.13-wifi-rtw88-Add-rtw8812a_table.-c-h.patch  | 2862 +++++++++++++++++++++++++
 .../rtl/040-v6.13-wifi-rtw88-Add-rtw8821a_table.-c-h.patch  | 2395 +++++++++++++++++++++
 .../rtl/041-v6.13-wifi-rtw88-Add-rtw88xxa.-c-h.patch        | 2187 +++++++++++++++++++
 .../rtl/042-v6.13-wifi-rtw88-Add-rtw8821a.-c-h.patch        | 1230 +++++++++++
 .../rtl/043-v6.13-wifi-rtw88-Add-rtw8812a.-c-h.patch        | 1135 ++++++++++
 ...4-v6.13-wifi-rtw88-Add-rtw8821au.c-and-rtw8812au.c.patch |   80 +
 ...fi-rtw88-Enable-the-new-RTL8821AU-RTL8812AU-driver.patch |  153 ++
 ...fi-rtw88-8821au-Add-additional-devices-to-the-USB_.patch |   77 +
 .../rtl/047-wifi-rtw88-8812au-Add-more-device-IDs.patch     |   91 +
 .../048-wifi-rtw88-usb-Support-USB-3-with-RTL8812AU.patch   |   69 +
 ...fi-rtw88-usb-Enable-RX-aggregation-for-8821au-8812.patch |   65 +
 ...fi-rtlwifi-rtl8821ae-phy-restore-removed-code-to-f.patch |   32 +
 ...51-wifi-rtw88-Add-additional-USB-IDs-for-RTL8812BU.patch |   31 +

Initial test is here:
#361 (comment)

[strasharo]

Any idea which would be the appropriate package to select for AWUS036AC (RTL8812AU)?
I'm running it on Arch using the rtw88 driver from git, but I'm not sure which one matches it from these in the OpenWRT menuconfig:

[ZerBea]

You need to do a git pull to get the latest drivers (don't forget to update you .config):

Than select kmod-rtw88-8812au.

BTW:
According @morrownr rtl8814au is in development but not ready at this time:
morrownr/USB-WiFi#549

[strasharo]

Guess something was stale on my side, just switched branches to master and then back to openwrt-24.10 and it popped. Thanks a lot.

[ZerBea]

You're welcome.

BTW:
No need to switch to master and back.
set feeds to 24.10 (last one is my local hcxlabtool feed);

src-git packages https://git.openwrt.org/feed/packages.git;openwrt-24.10
src-git luci https://git.openwrt.org/project/luci.git;openwrt-24.10
src-git routing https://git.openwrt.org/feed/routing.git;openwrt-24.10
src-git telephony https://git.openwrt.org/feed/telephony.git;openwrt-24.10
src-link mypackages /home/zerobeat/Source/raspberry/openwrt_packages/

do make clean
update and install feeds

./scripts/feeds clean
./scripts/feeds update -a
./scripts/feeds install -a

do make menuconfig
do make -j $(($(nproc)+1))

[ZerBea]

The new (updated) drivers rtl8xxxu and rtw88 usb are real game changers since they joined the bandwagon.
Running hcxlabtool/hcxdumptool, almost every tested (cheap) rtl8xxxu or rtw88 usb device is performing better than (expensive) mt76 devices.

[morrownr]

@ZerBea @strasharo

Full rtw88 USB support has been added to OpenWRT 24.10 (latest git head)

I'm glad this is happening quickly. I did not expect it to happen this quickly. One thing to be aware of is that not all patches are in mainline yet so are likely not in OpenWRT. Examples: USB3 detect and autoswitch support for 8812bu just landed in kernel 6.13 rc2. USB3 detect and autoswitch for 8812au is scheduled to go into kernel 6.14. There are more. The amount of patches necessary to get USB support up to speed in rtw88 was large and took over a year.

It appears that I need to update some of my documentation at github.com/morrownr/USB-WiFi . It would help me if I see reports on the quality of the rtw88 support. Thanks.

[ZerBea]

This is all you need:

Left hand side:
Raspberry Pi Zero (1) & TP-Link TL-WN823N (rtl8xxxu)
Right hand side:
Raspberry Pi Zero (1) & TP-Link Archer T2UV3 (rtw88 usb)
Bottom:
Powerbank

[ZerBea]

Or

Raspberry Pi Zero (1) & TP-Link TL-WN822N (rtl8xxxu)

[ZerBea]

Or, if you prefer an external antenna:

Raspberry Pi Zero (1) & EW-7811USC (rtw88 usb)

[ZerBea]

All penetration systems (RPI + WiFi adapter + USB cables + power bank) mentioned above are cheaper than a single ALFA AWUS036AXML - and in view of the current driver/firmware situation, they are performing much better!

[ZerBea]

@morrownr
basic driver tests are here (as soon as I can put my hands on a device)
penetration test results only pgp encrypted

[morrownr]

@ZerBea

USB3 is still a mess due to massive interferes. The same applies to WiFi/BT devices.

I agree completely. Knowledge of the problems that USB3 creates is not widespread. Even some adapter makers are not aware of it. There are certain guidelines that I emphasize on my USB-WiFi site:

• Buy only single-function adapters (no onboard bluetooth support)
• Buy only single-state adapters (no onboard storage for a Windows driver)

My Plug and Play List only shows adapters that meet the above criteria. I probably need to go further and explain in detail the interference problems. Even using a modern dual or tri band adapter in band 1 (2.4 GHz) can be an issue and I recommend that users in this situation only use USB2.

[ZerBea]

BTW:
hcxdumptool and hcxtools are part of OpenWRT:

and of course they are part of Arch Linux, too:
https://archlinux.org/packages/?q=hcx

[strasharo]

Packages in OpenWRT are slightly behind currently. Opened PR's to bump them to 6.3.5, however they've asked to include my name in the commits, so I closed them. :/

[ZerBea]

Unfortunately, this is the case in many distributions.
hcxlabtool/hcxdumptool and hcxtools are kind of a rolling release. The release cycle depends on daily analysis of wpa-sec uploads, the Linux kernel cycle and of course the release cycle of hashcat and JtR.
Unfortunately most of the package maintainers are very busy and they can't follow this cycle.
Luckily all hcx tools compile well on all distributions if the dependencies are met (libopenssl, zlib, libcurl and pkgconfig for hcxtools).
And they should work as expected if the system is updated (WiFi network drivers for hcxlabtool/hcxdumptool).

[strasharo]

I would gladly volunteer to do the maintenance of the Openwrt packages, however that real name requirement they have is an overkill for me. 😄 Guess I'll go with local feeds instead.

[ZerBea]

Well, let's see how long it takes until OpenWRT receives full rtl8192fu support.
This device is really worth it.

[ZerBea]

Missing firmware rtl8192fufw.bin (e.g. Edimax EW-7722UTn V3)
openwrt/openwrt#17286

Driver/device test is here:
#361 (comment)