USB4 Support Being Introduced With Linux 5.6 Kernel

[Lizintacer]

The recent kernels should have solved this and now there exists a Linux Thunderbolt 3 Device Manager (bolt), but I have not tested this, because I boot Linux from a Thunderbolt 3 SSD and after that I do not connect other Thunderbolt 3 / USB 4 devices.

I'm using Thunderbolt with all security options and with bolt on latest kernel. It works very well (boltctl cli is easy to use). There are some minor issues when one is dual-booting with Windows and there were key provisioning problems with earlier versions but currently I don't have any troubles.

[FireBurn]

Is Sage Sharp working on this? They helped do the initial bring up of USB3

[torsionbar28]

If the shipping [Chinese] solution boots Windows, and they can sell the solution to the masses, they likely will continue not to care.

Sounds like if IOMMU is a requirement to have feature complete USB4, the days of Chinese vendors omitting it may be numbered.

[zyxxel]

You've found the connections to be reliable? On my HP laptop sometimes if I even type too hard the connection goes away, and sometimes I have to try a couple of plug ins to even make an initial connection.

One issue with USB Type C connectors is that dust can result in the cable not plugging in fully. You don't get the expected "click" effect and touching the cable can have one or more of the signals disconnecting. This is mostly an issue with mobile phones carried in pant pockets, but having a laptop close to an open window - or carried in a dusty bag - can result in enough dust so the cable doesn't insert fully.

[kcrudup]

I tried cleaning both sides- ports and cables- with my HP, made no difference in the lack of reliability in the cables and I eventually had to send it back to HP to get the motherboard replaced.

[zyxxel]

I tried cleaning both sides- ports and cables- with my HP, made no difference in the lack of reliability in the cables and I eventually had to send it back to HP to get the motherboard replaced.

It's quite hard to clean the port on the phone/computer since the relevant parts are hidden inside the connector and because of the small dimensions. The normal issue is that the bottom of the connector is packed very tightly with dust - packed harder and harder every time you connect a cable.

I have created myself a tiny "hook" with scissors and plastic from a blister pack. The plastic is very stiff so it works well to insert the thin needle-like point inside the connector and scrape the bottom of the connector to fish out any dust that has jammed tight at the bottom. Two minutes of "fishing" and the cables will start to "click" as expected again when inserting and if done on a phone it stops complaining about "check cable connection" and stays with fastest possible charging.

I wish there were nice rubber plugs available that could be placed in the host connectors to keep dust out when a cable isn't connected. It's easy to replace a cable, but no fun when the dust blocks the port on a really expensive phone or laptop. There are 24 signals + shield on the connector so the connector can't be replaced without good solder equipment - it's basically a task for Louis Rossman et al to fix. Not sure how it looks on laptops but at least on phones, the host side is normally waterproof so they can't have an opening where you can push the dust straight through the connector and into the device itself.

[kcrudup]

Yes I did everything you did. But why did neither my phone nor my tablet (both USB-C) require these kinds of measures? And there are rubber plugs, I got some from Amazon.

Bottom line is HP picked an unreliable connector (other such machines at work started showing these same issues within months of purchase), which I fixed by getting a Dell

[zyxxel]

Yes I did everything you did. But why did neither my phone nor my tablet (both USB-C) require these kinds of measures? And there are rubber plugs, I got some from Amazon.

Bottom line is HP picked an unreliable connector (other such machines at work started showing these same issues within months of purchase), which I fixed by getting a Dell

I don't think HP intentionally goofed but it isn't impossible that they did select bad connectors or had some issues with the soldering process. Nokia N900 was a rather nice phone - but a phone that made it obvious to Nokia how silly it is to use a USB connector that is surface mounted with no mechanical pins to take up the forces. End result was that a huge number of users ended up with the connector following the charger cable out. Shit happens. The difference between good and mad manufacturers is that the good manufacturers learns their lesson and makes sure future models don't repeat the same mistakes.

[kcrudup]

I don't think HP intentionally goofed but it isn't impossible that they did select bad connectors or had some issues with the soldering process.

Oh, I'm sure. What's unfortunate is apparently the port in my replacement motherboard hasn't been ECNed (so is the same as the defective one).

[zyxxel]

Oh, I'm sure. What's unfortunate is apparently the port in my replacement motherboard hasn't been ECNed (so is the same as the defective one).

With some luck it isn't. Some problems can be batch-specific. So maybe a batch with problematic soldering or a batch with defective connectors. Hopefully the new motherboard jumped quite a bit in serial number. In the end, the connector manufacturer may have had to issue an important ECN.