News for hackers. Stuff that matters.
Re:Ending the anomaly: achieving low latency and airtime fai... by jabl on Tuesday August 01, 2017 @06:40AM (#14899325)
Awesome work! It's just too bad ath9k equipped routers are starting to be a bit hard to find..

And since ath9k is the only blob-free driver, prospects for fixing other drivers are not that bright.

Disclaimer: I'm not involved in any of this work myself, the above is my imperfect understanding of the situation. I'm just a happy owner and user of a ath9k router.

[ Parent | on HN ]
Re:Ending the anomaly: achieving low latency and airtime fai... by Maakuth on Tuesday August 01, 2017 @08:12AM (#14899828)
I think the possibility of fixing existing hardware in bufferbloat project was always limited at best. What can happen, though, is that good research results would find their way to vendors. Then, slowly, the issues would be fixed in new firmware blobs.

edit: typo

[ Parent | on HN ]
Re:Ending the anomaly: achieving low latency and airtime fai... by pedrocr on Tuesday August 01, 2017 @07:21AM (#14899536)
I had assumed the situation was better as I've had great results with the TP-Link AC1750 and Openwrt. Everything is well supported and works fast. But that has moved on to the ath10k which is now blob-based like the other drivers so these kinds of fixes are now normal hardware bugs and not something someone can hack independently.
[ Parent | on HN ]
Re:Ending the anomaly: achieving low latency and airtime fai... by mangix on Tuesday August 01, 2017 @01:40PM (#14903289)
Most of this is limited to the ath9k driver. ath10k is blob based and seems to have poor quality control as well as poor maintenance judging from all the bugs that go unanswered.
[ Parent | on HN ]
Re:Ending the anomaly: achieving low latency and airtime fai... by kayali on Tuesday August 01, 2017 @07:13AM (#14899481)
For the uninitiated: how long before we can expect the results of this research to be used in mainstream products?
[ Parent | on HN ]
Re:Ending the anomaly: achieving low latency and airtime fai... by legulere on Tuesday August 01, 2017 @09:55AM (#14900587)
From the abstract: "The implementation has been accepted into the mainline kernel distribution, making it available for deployment on billions of devices running Linux today."
[ Parent | on HN ]
Re:Ending the anomaly: achieving low latency and airtime fai... by kayali on Tuesday August 01, 2017 @10:25AM (#14900914)
Perhaps you misunderstood my question. How long would it take for routers that run that version of the kernel to be available?
[ Parent | on HN ]
Re:Ending the anomaly: achieving low latency and airtime fai... by JPLeRouzic on Tuesday August 01, 2017 @03:43PM (#14904585)
As mangix said, it is not a question of "having this version in router's kernel" because it depends strictly on the kind of chip the hardware uses. Most Wi-Fi chips use their own program which is called "a binary blob", they do not use the Wi-Fi stack of Linux except as a wrapper around their own code, which is not accessible in source, only in binary.

This is why for example in Ubuntu (but also in most other Linux distributions) there are "third party codes" that are "not free".

[ Parent | on HN ]