Linux 6.6 To Finish Gutting Wireless USB & UWB

[phoronix]

Phoronix: Linux 6.6 To Finish Gutting Wireless USB & UWB

The upcoming Linux 6.6 kernel will finish removing old remnants of Wireless USB support...

https://www.phoronix.com/news/Linux-6.6-Finish-Gutting-WUSB

[M@GOid]

I'm a IT nerd since the 90's, and for the life of me I cannot remember seeing this thing in the wild. Not even on YT channels that deal with unusual PC gadgets.

[muncrief]

I'm a IT nerd since the 90's, and for the life of me I cannot remember seeing this thing in the wild. Not even on YT channels that deal with unusual PC gadgets.

I'd never heard of it either, and I began working in the Silicon Valley in the late 70s. I eventually became an R&D engineer, designing hardware, firmware, and software for myriads of companies and projects, including Bluetooth firmware at Hello Direct, and this is first I've ever even heard the term "Wireless USB."

[szymon_g]

both its users will be devastated

[oiaohm]

I'm a IT nerd since the 90's, and for the life of me I cannot remember seeing this thing in the wild. Not even on YT channels that deal with unusual PC gadgets.

Understandable Wireless USB and matching UWB standard comes into existence in 2010 and is deprecated for "Media Agnostic USB (MA USB)" in 2013. Yes MA USB still has not managed to get into the Linux kernel or Windows.

There is a key reasons why Wireless USB and the matching UWB is died so quickly.
1) Yes FCC did approve the 3ghz to 10ghz frequency usage of Wireless USB and UWB but then china and japan don't approve it at the time. The radio not approval in china equals not allowed to make in china this results in the product not getting into lots of markets. MA USB is like the old USB/IP item where it usb over network so using normal wifi/ethernet so approved radio.
2) Windows(microsoft) and Mac(apple) never added drivers.
3) https://www.techpowerup.com/35699/d-...reless-usb-hub this is 2007 there are many different vendors doing their own wireless usb before the WUSB/UWB stuff all the products end up with horrible reviews. There is a fundamental problem.

Number 2 explains why lots of channels doing odd PC hardware don't cover it. This is odd PC hardware where you need a Linux install to use it or have s stupid large dongle sticking out side of machine.. Retail stocking a Linux only item is not going to work. worse stocking expensive Linux only items is even worse.

https://en.wikipedia.org/wiki/Ultra-wideband#UWB_Products/Chips

Yes none of the current made UWB chips use the protocol the old USB UWB does.
https://cateee.net/lkddb/web-lkddb/UWB.html Yes there is a list of chips/devices that use to exist and were supported.

Number 3 is a clear one of makers not learning for reviews. Every review of wireless USB has the same set of problems. Think about it for one min everyone has to have at least once unplugged a USB that is not safely removed resulting in USB drive being not readable until you do a repair. Wireless connection getting disrupted generates the same problem as being unplugged. Large number of USB devices are not designed to cope with disrupted connections . Lets just say be it MA USB or USB/IP or Wireless USB if you are running over a wireless connection that gets disrupted you are going to end up with unhappy USB devices unless you are careful in your USB device selection.

Of course number 3 that they tried in 2006-2007 and people got badly bitten leading to stores getting returns and having to give out refunds and being stuck with not sell-able stock then the same vendors again attempt new protocol in 2010 with no sign they have fixed anything you can understand stores not stocking the product as well.

Yes writing a formal standard before you have radio usage approval in all regions is major goof up. Yes not getting Microsoft on side another major screw up and they still don't have them on side. Not working out how to address the issues.

USB/IP stuff you see today you can put them over wireless but they also support going over Ethernet cable where radio noise is too much of a problem. Yes vendors selling USB/IP stuff today you still see them glossing over the limitations with USB storage devices and the like with unstable wireless connection so leading to different companies deploying wrong mix of devices then needing someone todo data recovery.

https://www.digi.com/products/networking/infrastructure-management/usb-connectivity/usb-over-ip/anywhereusb

Yes one such vendor.

M@GOid lets just say USB over Wireless has being disaster the complete process all the way up today and it not stopped being a disaster mostly because vendors are not willing to be clear to end users that there are very strict limitations what is safe todo with USB over unstable connection. Those reviewing old devices normally go for products that worked somewhat not odd ball products that are basically broken mess.

I do suspect we are on path for 2 decades of usb over wireless connections being end user harmful due to vendor not being clear on the limitations and this results in very bad word of mouth about the products and limited sales.

[r1348]

I also never saw anything UWB in the wild, and I actually owned a WiMAX modem!

[intelfx]

I also never saw anything UWB in the wild, and I actually owned a WiMAX modem!

I actually used WiMAX on a daily basis for a few years, before LTE was a thing. I even had a laptop with the Intel 5150.

[Ironmask]

>20,000 lines

USB is an invaluable technology but my god do I absolute despise how over-engineered and complex it is. There are like maybe 3 USB driver implementations in existence because it's so unbelievably difficult to implement. Microsoft helped spec it so I wouldn't be surprised if it was almost designed as some sort of bar to entry for making an OS. It's probably one of if not the most esoteric common technologies we use on a daily basis, and thats including modern CPUs and accelerators.

[stormcrow]

>20,000 lines

USB is an invaluable technology but my god do I absolute despise how over-engineered and complex it is. There are like maybe 3 USB driver implementations in existence because it's so unbelievably difficult to implement. Microsoft helped spec it so I wouldn't be surprised if it was almost designed as some sort of bar to entry for making an OS. It's probably one of if not the most esoteric common technologies we use on a daily basis, and thats including modern CPUs and accelerators.

It's less about how difficult it is to implement from the software side than it is about the vast numbers of devices that aren't compliant with the written specs. The problems with compliance and the high degree of brokeness can be partly laid at the feet of the USB spec committees and their negligent enforcement of branding compliance, but also at the flood of no-name mostly non-compliant or utterly broken "USB" devices coming from the PRC flooding markets everywhere. Most of the lines of code are there to deal with "quirks" and other non-compliance in hardware even from supposedly reputable brand names (-cough- Razer -cough-).

[erniv2]

I wonder how many macros and unused code lines actually left from old subsystems