USB ports - How to tell which type?

[Clarity]

This is NOT a solution. Rather it is a question. Further this is a follow-on to this USB information

USB technology is advancing. PCs have been advancing, sometimes, in step as USB progresses.

ALL of you know of these physical port changes occurring over the past 2 decades. AND, most of you know that the items we plug into those ports are ALL OVER THE MAP in how they perform.

This question, I am raising, is because I have seen 3 "new" PCs from people who call on me with a myriad of USB ports on a single PC. It is NOT just the connector technology, but also the physical behavior of these ports. The PCs come with USB2 ports as well as USB3 ports and well as USB-C as well as Thunderbolt ports.

And most of us know that there are several physical characteristics of these ports for the upper limit protocol that can be inserted into a system's port. Thus USB flash/drive selection is important should one want to get the maximum benefit a port can deliver. Today, the only way one would know is by "eyeballing": either looking at the port for a descriptor or by looking at the owners manual for diagramed descriptors.

This is already leading to various problems is how units are used by the system and has already been seen in unit performance behavior depending upon which port it is plugged.

Most users are naive to this physical connection issues, naive to unit behavior issues, and naive to a need to thoroughly understand what they should know to ask when purchasing a new PC.

This leads to the following 3 Questions:

1. Does anyone know of a Linux manner to tell whether a physical port is USB2 vs USB3?

2. Does anyone know of a Linux manner to tell whether a physical port's protocol is 3.0/3.1/3.2/3.2-2/4.0/thunderbolt?

3. Does anyone know of a Linux manner to tell if a device plugged into a USB physical port is a USB1/USB2/USB3/USB4 unit?

Any knowledge you can bring to the table is helpful info to all of us.

[ozsouth]

@Clarity - terminal commands lsusb -t and lsusb -v will give some information. Pup-SysInfo (under System menu) is also good.

[mikewalsh]

@Clarity :-

I don't know of a specific "Linux" way to tell ports apart. But the accepted way of telling them apart is by "eye-balling" them, certainly.

With the older standard, there seemed to be no consistency. The colour of the internal 'half plastic' tongue could be black or white, or grey. White usually indicated the early USB 1.0 standard (though, not always). USB 2.0, I have seen in grey; black; brown.....even red.

From USB 3.0 onwards, consistency set in.These were always blue, internally....though with 3.0, 3.1 & 3.2, things got a mite confusing.

USB-C are much smaller.....and it doesn't matter which way you plug them in.

USB 4.0, I've heard of.....but so far as I know, no working examples exist in the marketplace as yet. Isn't this what started out life as Apple's "Thunderbolt" ? (It requires at least k5.6 for support, and appears to be mostly for supporting the "tunnelling" of other, earlier protocols...)

Mm-hm. https://en.wikipedia.org/wiki/USB4

STILL confusing..!!!

Mike.

[bigpup]

I think the safest thing is to look at the info provided by the owners manual.
In the first few pages it should provide specific info on what port is what.
Example:

Screenshot(5).jpg (43.02 KiB) Viewed 2549 times

.
Notice the amount of info about the USB port on this side of the laptop.

Screenshot(6).jpg (53.27 KiB) Viewed 2549 times

Usually you can find a pdf of the owners manual to put on the computer.
One laptop I have had it already on the computer.

[Clarity]

I was afraid of this and is the reason I posted. As the technology of display devices (PCs/AIOs/Laptops/TVs/tablets/etc) advance, we users are facing a dilemma which is only going to get worse if you are trying to have a means of identifying the port-device behaviors we are heading into.

So, when I tried to compare device performance using system tools, I ran into to not-so-obvious dilemma which took couple days to wrap my head around. Helping a user, I showed him how to use a couple commands to "see" how a USB performed. He called me later to report an odd behavior; namely when plugged into one type-A port he was getting great performance and in another bad performance. He thought (and I too, at first) that the manufacturer has sold a lemon. As it turns out, NOT SO! And this is where my USB thread that many of you have seen, started.

@bigpup's images produces one example of this phenom.

This phenom is leading to some very interesting behaviors that will only advance. And we have no system-level tools at this juncture to gather an understanding with some intelligence.

Today, when looking at a USB stick, many DO NOT post which USB protocol it supports (1.1/2/3/3.0/3.1/3.2/3.2.1/3.2.2/etc.). These, each, have port performance "targets" but each individual stick may/may-not PERFORM adequate on one of these port=protocols but will be different on another.

So, we are at a point where it is becoming apparent that some tools are needed to determine the device protocol and the port protocols when "married" (connected) to each other.

Only then, will the user, using system level tools, be able to have a means of knowing if the marriage is what one would expect.

Thus, when running performance tests, we would have a clearer understanding of what is actual in occurence. Further, when helping forum users, this adds another level of consideration when complaints arise because of boot issues from slow/fast to cannot find SFS.

P.S. This also true as Thunderbolt usage increases simultaneous with USB. (remember Betamax vs VHS?)

[mikewalsh]

In my view, the real problem that exists is quite simple. Each new "generation" of USB.....each "improvement" of the standard.....introduces new, and additional protocols & behaviours, while at the same time attempting to maintain compatibility with every single previous generation. Over time, the list of 'protocols' supported is eventually going to become, quite frankly, ridiculous.

Eventually, that huge list is going to HAVE to be rationalised, and 'simplified'.....or else 'understanding' is going to end up taking a back seat to confusion.

(*shrug*)

I think what you're after, Clarity, is some kind of easy-to-use 'system' tool that will inform users, of all skill-levels (be they 'noob' OR 'expert'), in the simplest terms possible, of exactly what any given port is capable of doing/supporting at any given juncture....yes? Am I correct?

AFAIK, such a thing does not exist outside, perhaps, of manufacturer's specialist tools for their own R & D people. Could be an interesting project for someone, though; after all, if somebody can initially write/code it, somebody else can at some point duplicate it.....right?

However, I'm not at all sure that the necessary system libraries yet exist to extract quite the level of detail you're indicating. I don't think even the kernel itself reports that kind of detail.....though I could, of course, be wrong.

It has been known..!

Mike.

[mikewalsh]

@Clarity - terminal commands lsusb -t and lsusb -v will give some information. Pup-SysInfo (under System menu) is also good.

@ozsouth :-

Interesting, Oz.....especially the 'verbose' option.

Mike.

[Clarity]

Most forum members know my disdain for USB. It is severly problematiic at best even though useable.

Thunderbolt is better by comparison for some reasons I wont go into here, but its problem is that it is NOT ubiquitous in almost all 64bit platforms ... except Apples.

But USB graduates to USB4 in some/many/most/all 2023 platforms while predictably dragging new issues along.

Here is a video, though comedic, has a large measure of accuracy pertaining to modern USBs.

Enjoy the knowledge

[Jafadmin]

You will find all you USB controller information at:

Code: Select all

#ls /sys/bus/usb/devices/usb?
#ls /sys/bus/usb/devices/usb?/speed

Regards

[bigpup]

Today, when looking at a USB stick, many DO NOT post which USB protocol it supports (1.1/2/3/3.0/3.1/3.2/3.2.1/3.2.2/etc.). These, each, have port performance "targets" but each individual stick may/may-not PERFORM adequate on one of these port=protocols but will be different on another.

If they do not provide the info on which USB protocol it supports (1.1/2/3/3.0/3.1/3.2/3.2.1/3.2.2/etc.)
I do not buy it.

[mikewalsh]

@Clarity:-

Further to earlier points made, I'm now cognizant of what a Thunderbolt connector actually looks like. Strangely, it's not so unique as everybody seems to make out....

If you know what the port connector for an internal PCI-e x 1 slot looks like, an early Thunderbolt connector (gen-1 or -2) is rather like an "inverted" or negative image of one of these. A USB connector usually has the 'pins' or connectors arranged around the middle, pointing outwards.....whereas Thunderbolt has them arranged around the outside of the port, pointing inwards.

Later Thunderbolt connectors (gen-3 or -4) are far more similar to either USB 3.0, or USB-C. Just to muddy the waters still further, like..!

The step Apple took to make it as fast as it is was all in the actual port-control chipset.....whatever it is that Apple uses as an equivalent to the traditional 'southbridge', being as how this is where USB controllers always used to be situated before everything migrated onto the CPU die itself. And what they've done with the ARM-based 'Apple Silicon' CPUs is anybody's guess....

---------------------------------

I treated myself to another SanDisk USB 3.1 'Ultra' Fit yesterday; already have a pair of these @32 GB. The new purchase is @ 128 GB.

I've just run the DriveSpeed! speed tester on it a few minutes ago. These always run through the HP's USB 3.0/3.1/3.2-enabled USB 3.0 ports. Write speed around 50 MB/s.....read speed around 130 MB/s. For a 'nano'-sized USB thumb drive, those are not speeds to be "sniffed at".....I think many would be quite happy with that level of performance. I know I am.

Mike.

[Phoenix]

This is NOT a solution. Rather it is a question. Further this is a follow-on to this USB information

USB technology is advancing. PCs have been advancing, sometimes, in step as USB progresses.

ALL of you know of these physical port changes occurring over the past 2 decades. AND, most of you know that the items we plug into those ports are ALL OVER THE MAP in how they perform.

This question, I am raising, is because I have seen 3 "new" PCs from people who call on me with a myriad of USB ports on a single PC. It is NOT just the connector technology, but also the physical behavior of these ports. The PCs come with USB2 ports as well as USB3 ports and well as USB-C as well as Thunderbolt ports.

And most of us know that there are several physical characteristics of these ports for the upper limit protocol that can be inserted into a system's port. Thus USB flash/drive selection is important should one want to get the maximum benefit a port can deliver. Today, the only way one would know is by "eyeballing": either looking at the port for a descriptor or by looking at the owners manual for diagramed descriptors.

This is already leading to various problems is how units are used by the system and has already been seen in unit performance behavior depending upon which port it is plugged.

Most users are naive to this physical connection issues, naive to unit behavior issues, and naive to a need to thoroughly understand what they should know to ask when purchasing a new PC.

This leads to the following 3 Questions:

1. Does anyone know of a Linux manner to tell whether a physical port is USB2 vs USB3?

2. Does anyone know of a Linux manner to tell whether a physical port's protocol is 3.0/3.1/3.2/3.2-2/4.0/thunderbolt?

3. Does anyone know of a Linux manner to tell if a device plugged into a USB physical port is a USB1/USB2/USB3/USB4 unit?

Any knowledge you can bring to the table is helpful info to all of us.

You can use lsusb but it will spit out an address. This is not very useful for physically locating or identifying it.

[greengeek]

I have often encountered the situation where a "fast" usb drive is plugged into a "fast" usb port - but copying of data is very slow.

It appears that this occurs when the usb port driver that Puppy has allocated - is NOT the fastest one available.

I have never established why this is the case - or how to get around the problem. But I think it shows that any kind of theory suurounding which port is the fastest - is secondary to actually plugging a given usb drive into each port of your computer and running speed tests.

The results can be surprising (and vary from Pup to Pup) - but the data will be helpful in selecting the best port.

[robertstein]

To determine the type of USB ports and devices in Linux, you can use several commands. Running `lsusb -t` shows the USB hierarchy and speeds, helping you distinguish between USB2 and USB3 ports. To check the protocol version (e.g., 3.0, 3.1, or 3.2), use `lsusb -v | grep -i bcdUSB`, which displays version details for connected devices. Additionally, the command `dmesg | grep -i usb` provides information about the port's speed and type when a device is connected. Using `usb-devices` also gives a concise overview of all USB devices and their speeds.

[rcrsn51]

Is it just me or do all of these posts from robertstein sound like they are AI-generated?

[mikewalsh]

Is it just me or do all of these posts from robertstein sound like they are AI-generated?

My thoughts exactly.....though nothing, as yet, has actually broken any of the forum guidelines.

We shall keep an eye on him (her?/it?)

Mike.

[wizard]

@greengeek

I have often encountered the situation where a "fast" usb drive is plugged into a "fast" usb port - but copying of data is very slow.

Keep in mind that the claimed vs actual speed of flash drives can be all over the map. Have tested some USB 3 drives that were much slower than USB 1.0 , especially write speeds.

wizard

[greengeek]

@greengeek

I have often encountered the situation where a "fast" usb drive is plugged into a "fast" usb port - but copying of data is very slow.

Keep in mind that the claimed vs actual speed of flash drives can be all over the map. Have tested some USB 3 drives that were much slower than USB 1.0 , especially write speeds.

wizard

I do wonder if that type of problem could come down to the software choosing an incorrect driver for the device?
ie, if puppy doesn't recognize how to properly drive (or recognize) a particular usb3 device - then the port won't transfer at the speed the drive should theoretically handle.
I've had some pretty bad results from some "usb3" stuff - even good name brands sometimes.

[greengeek]

I wonder if there are any speed test utilities that spit out some data regarding how the system sees each device under test.
Not sure what is controlled by the port hardware / firmware of the port and what exactly is controlled by the software (is choice of EHCI or other driver)

[wizard]

@greengeek

I do wonder if that type of problem could come down to the software choosing an incorrect driver for the device?

Don't have a good answer for that. Always validate my bad test by running in both Pup and W10.

wizard

[Clarity]

@greengeek you are showing, accurately, the issues of this thread started few years ago.

Over time, I have found many issues in USB use, ports, devices, and technology differences. USB "IS" a convenient technology and is useful. But, what is seen on one port may/will differ on another port on same PC. For example a USB-A port that manufacturer spec says is version 3 could behave as a version 2 or version 3.0 or version 3.1 or ...

Some manufacturers will notch a speed (such as "10) to indicate a difference of one on the USB-A port vs another.
Further when we get to USB-C ports, the varied behaviors, again, are as inconsistent as they are on USB-A ports. And when troubleshooting, should one choose to do so, how do you distinguish if the physical port has a problem, the USB device controller has a problem, the USB memory has a problem, or even an external USB hub which may/may-not have varied port behaviors.

So, WE, the user-members are left with manual efforts to know which is which. This is true if you have several PCs you use for home efforts.

Lack of a consistent tools and lack of expectation in use, is the problem I have noticed and up to now have found little methods to apply across the board for to establishing consistency in measurements.

Anyone ever tested a USB-C device on your phone (or a single-port tablet) and later run the same test with the device plugged into a hub plugged into your phone? Fascinating!

[greengeek]

The title is - "USB ports - How to tell which type? - so I wonder if there is some way to write a utility that pops up a yad/xmessage etc - that alerts the user to the following when they plug in a USB device:

- which physical port you just plugged in
- which other devices share that hub
- which driver is selected for the device you just plugged in
- what speed is max potential for this bus.

Yes that is asking a lot.

But it would be helpful...

Because a usb bus only runs at the speed of the slowest device "deemed" to be present on that bus.
And can only run at the speed determined by the driver (ohci, ehci etc) allocated to that bus (or device).

I am guessing any such utility would have to be triggered by UDEV. (so probably an impossibility??)

Being a "shared bus" which has extremely variable operating speed - it would be nice to have access to more info...

[Clarity]

Hello @greengeek I think you are asking for some sort of a USB port configuration utility that can be run on each of your USB ports on your PC. Like you and others, this remains a desire.

You also state

...Because a usb bus only runs at the speed of the slowest device "deemed" to be present on that bus ...

But, as the state of USB has had multiple changes over the years, I am not so sure this statement is relevant now. Notice that device speeds are different depending on port and device and hubs, if present. Even when you connect hub1 to hub2 to device on the same USB PC port (a configuration I've seen students do).

FYI

[greengeek]

Keep in mind that the claimed vs actual speed of flash drives can be all over the map. Have tested some USB 3 drives that were much slower than USB 1.0 , especially write speeds.

wizard

Yeah, the performance of modern drives really concerns me.
I have been really upset by the performance of a Sandisk 64GB usb stick that I bought several months ago - as it is is always red hot - even when it is not in use.
What on earth could its' controller be doing to make it so hot - all of the time ??
Even when it is not mounted it is still too hot to touch. (Can't be good for my laptop's usb port...)

So today I bought 5 different usb drives to see how they performed.

I have tried to copy all of the data off the Sandisk 64GB drive and i find that all of the new usb drives run way cooler than the Sandisk.

Unfortunately the write speeds to those new drives is totally different to their advertised read speeds. Actually - most of them do not even stipulate their write speeds. It is left to guesswork.

The testing continues...

EDIT : In the past I had very bad results from HP usb sticks - but that was with a different Pup and a different laptop. Today - I seem to be getting the best results from HP usb sticks. Sandisk no good. Go figure..

Of course - this is not really about identifying which USB port is which - but more about performance of usb drives that are plugged in...

[wizard]

@greengeek

Think your test outcomes will be valuable for USB flash users. Suggest starting a new topic and posting the results.

Thanks
wizard

[bigpup]

There are fake USB sticks out there.

They take a slower, maybe even smaller stick, and repackage it, saying it is something much bigger and faster.

I am not even sure you are safe with name brand packaging.

No reason, they could not make the repackage look exactly like, the name brand package.

It is what it is.

So far, buying from a well known retailer, seems to be keeping me from getting one of these fake USB sticks!

I assume they get them directly from the manufactures.