Arch Linux User Repository

To all having an issue with this driver: please try https://aur.archlinux.org/packages/rtl8814au-aircrack-dkms-git alternatively.

Interesting, looks like an evolution of 5.2.20 and that they have integrated aircrack injection code. Possibly the future for all rtl8812/14/21 series?

https://github.com/aircrack-ng/rtl8812au/issues/180 :)

Unfortunately I cannot help due to the fact I do not own a 8814au device. As you write, you would need some metrics. But you may notice some bugs being removed (disconnection, etc.). Finally, there may be some security fixes/enhancements that are not visible.

For sure I can test it, but why should I? I mean, even with rtl8812au-aircrack-dkms-git, the only thing I noticed is lower ping (3ms instead of 4ms with your driver. How can I validate it works "better"? :)

In addition, on the Edimax website, there is a EW-7833UAC_Linux_Driver_1.0.1.0.zip file which decompresses as rtl8814AU_linux_v4.3.21.1_20171108.tar.gz. This is the same version number than for the TP-link, but the timestamp changes by 9 days. I did a quick comparison with T9UH_linux_v4.3.21.1_24835.20171031.zip and indeed they are almost identical. The Edimax archive has an Edimax specific entry while the TP-link has, as expected, its own specific T9UH UIDs. Thus vendors only adapt the code to their adapter names, that's all. Otherwise the rest of code is mostly the same. This shows how Realtek just provides boiler plates to vendors that make their own archives, hence the mess. That said, it is probably worth trying v4.3.21.1 over v4.3.21 as it may have upstream corrections from Realtek.

I did further searches, and found out this version: T9UH_linux_v4.3.21.1_24835.20171031.zip from https://www.tp-link.com/us/download/Archer-T9UH.html#Driver. Just found out that it is available in Github with kernel patches there: https://github.com/coolshou/rtl8814au or https://github.com/francoism90/rtl8814au. Worth a try maybe? Do you want to try this code base and report? If they do work and fix bugs then it could be worth using them.

@erkexzcx: this is the reason why I am very reluctant to use these sources. The 3 PKGBUILD I maintain so far use the latest code version from Realtek (allegedly) for each chipset, repectively. The only modifications are code to make them compatible with the latest Linux kernel versions. Other projects often use hybrid code bases to make the drivers compatible with many chipsets, and end up with drivers that are not up to date and carry a lot of hacks. But in the end it is Realtek to blame; they should be more open and transparent with their code and their multiple rewrites.

Actually using AUR package rtl8812au-aircrack-dkms-git I got lower latency and overall better performance when using with my AWUS1900 but then suddenly it started to randomly "reload" itself and connection is getting lost about every 1-2 minutes. I then got back to this (your) AUR package/driver and all the issues went away. So I assume - rtl8812au-aircrack-dkms-git is a more complete driver with more pull requests than in original driver, but again - it was unstable to me.

@erkexzcx: fantastic, many thanks. Besides, do you know if this is still the latest release from Realtek? It is infuriating the driver is much behind the rtl8812au version (if the version numbering is indeed significant, that is.)

@zebulon I've updated this package. Nothing has changed for x86_64 and i686 architectures, just a very small patch for ARM devices to get it work. Just double checked it by installing this package on my ARM device - works perfectly. Thanks a lot!