On 2026-05-16, Henrik Ahlgren <pablo@seestieto.com> wrote:

Anyway, today, modern powerful GaN chargers are tiny, so bring your own,
not just a cable.

What's a GaN charger?

--- PyGate Linux v1.5.14
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

On Sat, May 16, 2026 at 04:07:01PM -0000, CGS wrote:

On 2026-05-16, Henrik Ahlgren <pablo@seestieto.com> wrote:

Anyway, today, modern powerful GaN chargers are tiny, so bring your own, not just a cable.

What's a GaN charger?

A switching power supply. The switching transistors are gallium nitride,
which allows for higher switching frequencies (at pretty low losses) and
thus for smaller inductive components and smaller packages. This makes
for surprisingly small power supplies
But they'll only help you if you find a mains outlet -- many public transport vehicles and places (rail stations, e.g.) in our area do have USB-A charging sockets but no mains. Which kinda makes sense.
Cheers
--
t


--- PyGate Linux v1.5.14
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

On 2026-05-16, <tomas@tuxteam.de> <tomas@tuxteam.de> wrote:

Anyway, today, modern powerful GaN chargers are tiny, so bring your own, >> > not just a cable.

What's a GaN charger?

A switching power supply. The switching transistors are gallium nitride, which allows for higher switching frequencies (at pretty low losses) and
thus for smaller inductive components and smaller packages. This makes
for surprisingly small power supplies

But they'll only help you if you find a mains outlet -- many public transpo= rt
vehicles and places (rail stations, e.g.) in our area do have USB-A charging sockets but no mains. Which kinda makes sense.

I was interested in Andy Smith's comment about power/data usb cables.

I'm dumbfounded how GaN chargers got into this. My question was actually non-rhetorical.

Someone said all modern charging cables support data transfer. Where can I acquire one that doesn't?

I guess I'll have to figure it out by myself.

Cheers

--- PyGate Linux v1.5.14
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

CGS <etphonehomefrance@gmail.com> writes:

I'm dumbfounded how GaN chargers got into this. My question was actually non-rhetorical.

The easiest way to protect yourself against "juice jacking" attacks is
to use your own cables and chargers. This is less of a hassle with
modern chargers that are small enough to fit in your pocket (25W
models). On the other hand, juice jacking is mostly theoretical; I don't believe there have been any documented real-world cases. Of course,
it all depends on your personal threat model.

https://en.wikipedia.org/wiki/Juice_jacking

Someone said all modern charging cables support data transfer. Where can I acquire one that doesn't?

There are several products available, like

https://plugable.com/products/usbc-cc1m

I believe those must have some active electronics to allow for Power
Delivery negotation while blocking all other data transfer.

--- PyGate Linux v1.5.14
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

On Sat, May 16, 2026 at 10:32:20PM -0500, David Wright wrote:

On Fri 15 May 2026 at 21:52:36 (+0200), tomas@tuxteam.de wrote:

[...]

https://en.wikipedia.org/wiki/BadUSB

Who needs automount?

OK, I see now that you're extending the discussion from charging ports
to inserting random USB sticks into your computer when you don't know
their provenance. I guess the techies that are likely to encounter
these devices are employed way above my paygrade. I'd be flattered
to be targeted by the people who make these devices.
(Likewise if I was sent a white powder in the mail?I don't have
the means to distinguish flour from anthrax.)

Not necessarily, see below.

I don't work for a company where they block your USB ports or harden
their machines to that extent. Whether hardened versions of Debian
can determine if an attached keyboard is genuine before accepting its keystrokes, IDK.

USB devices identify themselves with a couple of numbers: the device
class, the vendor ID and the product ID [1],as defined by the vendor.
The device can do whatever it wants, it's just firmware pushing bits,
so no -- it can tell your computer whatever it wants.
The operating system then uses these IDs to decide what to do (e.g.
load a kernel driver, whatnot). Udev is the one responsible for
that in our countries.
But Stefan's approach went another way: ask the user (they are, after
all, those sticking the thing into the port). If you stick your device
to a charger and it asks you "is connecting to this keyboard OK?",
it's on you to say "HELL, NO!" :-)
Having that as an option makes sense.
Cheers
[1] http://www.linux-usb.org/usb-ids.html
--
t


--- PyGate Linux v1.5.14
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

On Sun 17 May 2026 at 09:49:13 (+0200), tomas@tuxteam.de wrote:

USB devices identify themselves with a couple of numbers: the device
class, the vendor ID and the product ID [1],as defined by the vendor.

The device can do whatever it wants, it's just firmware pushing bits,
so no -- it can tell your computer whatever it wants.

The operating system then uses these IDs to decide what to do (e.g.
load a kernel driver, whatnot). Udev is the one responsible for
that in our countries.

I agree. The more that udev can inform us about, the better.
My own udev scripts are a tiny part of that. Obviously you
have to be careful with keyboards, as you need a device for
submitting your response!

But Stefan's approach went another way: ask the user (they are, after
all, those sticking the thing into the port). If you stick your device
to a charger and it asks you "is connecting to this keyboard OK?",
it's on you to say "HELL, NO!" :-)

Having that as an option makes sense.

Sure. But I thought charging ports were done and dusted about
five posts upthread (power-only cable/power bank/mains adapter).
The approach is akin to assuming all fuel pumps have had a
credit-card skimmer installed (but more practical).

Cheers,
David.

--- PyGate Linux v1.5.14
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

But Stefan's approach went another way: ask the user (they are, after
all, those sticking the thing into the port). If you stick your device
to a charger and it asks you "is connecting to this keyboard OK?",
it's on you to say "HELL, NO!" :-)
Having that as an option makes sense.

Sure. But I thought charging ports were done and dusted about
five posts upthread (power-only cable/power bank/mains adapter).
The approach is akin to assuming all fuel pumps have had a
credit-card skimmer installed (but more practical).

We've learned to make it tolerable for random end users in the context
of Bluetooth (rather than only for those paranoid ones which walk around
with power-banks and mains chargers), so I wish we could do the same
for USB. ?

It's not like people usually spend a large amount of time connecting
USB devices.


=== Stefan

--- PyGate Linux v1.5.14
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)