Realtek 802.11ax WiFi Driver "rtw89" Queued Ahead Of Linux 5.16

[phoronix]

Phoronix: Realtek 802.11ax WiFi Driver "rtw89" Queued Ahead Of Linux 5.16

Arriving in the wireless-drivers-next branch this week is the "rtw89" driver as a Realtek-contributed open-source 802.11ax WiFi driver...

https://www.phoronix.com/scan.php?pa...802.11ax-rtw89

[uid313]

Do you prefer Intel Wi-Fi or Realtec Wi-Fi or Atheros Wi-Fi or Wi-Fi from some other manufacturer?
And why?

[kiffmet]

Realtek drivers seem a bit strange to me. Some of them are upstreamed, whereas others are an out-of-tree mess that is provided as a snapshot from some internal repo and given to hardware vendors (i.e. producers of a USB wireless sticks that use a Realtek chip, RTL8812AU comes to mind).

Realtek LAN: Upstream, Realtek sound: Upstream, Realtek Wireless: A freaking minefield
Please just get it over with and upstream everything. This erratic behaviour is lazy, aswell as hostile towards customers as compatibility with newer kernels may randomly break for the users of some of your devices.

[kiffmet]

uid313 Intel and Atheros do have in-tree drivers for nearly all of their products that are mature and work very well, aswell as day-one Linux support for new devices. Mediatek has a reputation for being buggy to the extent of creating severe security holes and as for Realtek, read my other post above.

[-MacNuke-]

Had this chip in my new Lenovo notebook with the out-of-tree driver. It was a pretty big mess. Driver was constantly crashing, no IPV6 support and bad power management. I hope this driver is better for other people. I switched to an Intel module and Lenovo did too with newer revisions of this notebook.

[Gusar]

Realtek drivers seem a bit strange to me. Some of them are upstreamed, whereas others are an out-of-tree mess that is provided as a snapshot from some internal repo and given to hardware vendors (i.e. producers of a USB wireless sticks that use a Realtek chip, RTL8812AU comes to mind).

Realtek's PCIe and USB driver teams are literally in different countries. Very different parts of the company. That's why support for PCIe chips is generally ok, while drivers for USB devices are the mess you've described.

[kiffmet]

Gusar Imagine if there was some kind of software that allows developers to work together on a codebase in different "branches" and merge those into one fully operational driver stack, once everyone involved is happy with the results. It would be a game changer :P

[Gusar]

Gusar Imagine if there was some kind of software that allows developers to work together on a codebase in different "branches" and merge those into one fully operational driver stack, once everyone involved is happy with the results. It would be a game changer :P

It's not about code management, but about corporate culture. Some companies haven't yet figured out the open source development model. It's not just about engineering, it has to be part of the culture all the way from the top execs. This simply doesn't exist in the Realtek branch that does USB. And it's not something simple that can change overnight.

[cbxbiker61]

Great to see another AX wifi module added to the kernel!

I'm currently using this out of tree driver with great success in conjunction with kernel 5.14.x. It seems to behave/perform as well as the Intel drivers for me. That's high praise since I consider the Intel wifi drivers the gold standard on Linux.

https://github.com/lwfinger/rtw89

[moriel5]

I'm curious, since I have been hearing (inconsistent) reports about Realtek's Wireless-AX cards having fewer issues than their predecessors (OS-agnostic, since generally Windows users have it worse than Linux users with Realtek, stability-wise), is there any truth to this?