Linux 4.5 Input Updates Bring Changes Even For PS/2 Mice

[devius]

And please, Michael, be more neutral in your articles. It is somewhat expected from a professional journalist like you. Write your articles in a neutral tone and point of view, and share objective facts rather than strongly biased personal opinions.

Welcome to Phoronix. After being a reader for a long time I can assure you that will never happen. That's how he gets all these comments (and clicks/views) on what would otherwise be a pretty innocuous article. Think of this more like a personal blog than a professional news site.

[F1esDgSdUTYpm0iy]

That's how he gets all these comments (and clicks/views)

Pretty bleak though, having to resort to clickbaiting like that. Here's this powerhouse of a benchmarking farm and how is revenue generated to sustain it? Clickbaiting. A better solution is needed, to be sure.

[tegs]

PS/2 is a much better protocol for keyboards. It is interrupt-based rather than polling-based, hence, less strain on the CPU and less latency (can be important for games).

Additionally, PS/2 supports true n-key rollover (you can press as many keys as you want on your keyboard simultaneously and all will register correctly), while USB only supports up to 6 keys being pressed simultaneously. In some scenarios, this can be a deal-breaker.

My keyboard, for instance, has circuitry capable of n-key rollover, but it only works if I connect it to my PC over PS/2. The USB connection limits it to 6. I have seen keyboards that try to do fancy trickery to emulate it on USB (my friend has one), and it sorta works, but breaks other things ...

When I got my mechanical keyboard (Rosewill RK-9000V2 BR) a while back, it came both the PS/2 and USB cables and I went with PS/2. For the reasons you stated in your post, I felt like PS/2 was the best route and I liked that I was freeing up a USB slot. The PS/2 slot was there, so why not use it right? My mouse though is USB since I want to take full advantage of my Sensei RAW and want to use 1000hz for my polling rate.

[caligula]

That is interesting because I have never seen a notebook trackpad which wasn't a PS/2 device. Even my not-too-old Toshiba ultrabook has touchpad with PS/2 interface (AlpsPS/2).

I also have a Haswell Ultrabook with PS/2 Alps touchpad.

[FuturePilot]

Is the "CPU overhead" of USB polling really something we even need to worry about these days?...

[F1esDgSdUTYpm0iy]

Is the "CPU overhead" of USB polling really something we even need to worry about these days?...

No. Obviously there is an overhead but even then, the actual time spent processing the data is obviously marginal. So much so that people have used 1000 Hz polling without noticing any negative effect on CPU performance.

Conversely though, there is actually very little reason to poll at 1000 Hz because, well, the human anatomy cannot process information that fast anyhow, considering the time it takes for signals to travel up/down the spine to "communicate" with the hand alone.

[tigerroast]

Well. Call in the PS/2 cavalry.

While PS/2 keyboards are indeed lovely (I find the buckling spring keyboards nice to use), I really see no advantage of using a PS/2 mouse. The standard polling of USB mice is 125Hz. while the PS/2 microcontroller sends interrupts at 100Hz. Since you can adjust the polling rate on USB, I'd think that the USB mice would be much more precise than PS/2.

[Azrael5]

http://www.gigabyte.com/products/product-page.aspx?pid=5286#ov

see the picture... ps/2 is not legacy is modernity.

[carewolf]

I am pretty sure my laptop still uses the psmouse driver even if it hasn't got any actual PS/2 ports. Isn't it a generic driver these days?