[BRLTTY] [PATCH] don't steal abused USB ids

[Dave Mielke]

[quoted lines by Stanislav Brabec on 2018/05/14 at 19:23 +0200]

>> This turns out to be a matter of opinion. The users of braille devices with 
>> those identifiers wouldn't like it very much if their devices weren't detected.
>
>I am aware of this problem. But there is also a reverse side of the
>problem: brltty is a part of default installation of many Linux
>distributions (including all SUSE and openSUSE products). These rules
>break many computers equipped with devices like USB-serial converters,
>GPSes, AVR ISP programmers, Arduinos etc.

I understand, and am not trying to ignore it.

>Auto detection should be supported by the manufacturers first in the order. 
>Some UART bridge vendors even have a special program to support unique USB id 
>for commercial products based on their devices.

I completely agree, but, sadly, there are those, including braille device 
manufacturers, who don't.

>But we have to live with such a broken devices.

Yes. It's particularly annoying when it's a braille device as being able to 
autodetect it is, in my opinion, critical.

>Can you detect these devices by the serial communication somehow?

In general, yes, but I don't like the idea because that can confuse, if not 
entirely brick, a device. And, as you might guess, braille devices are not only 
very important but also very expensive.

>If not, we are lost and there is no automatic solution at all.

Maybe not. It's just that we haven't thought of it yet. That's how I face these 
issues, anyway. Let's keep discussing it, and, as I always do, encourage anyone 
else who has a possible idea to speak up.

>brltty is not only part of the default installation. It is even part of
>the installer.

It took years (and that's an understatement) to get to this point, and it's an 
excellent point to have reached so we must find a solution.

>None of typical UART devices has to be supported by the installer.
>
>Installer can relatively safely suppose that any USB UART bridge is a
>Braille device. (With the above mentioned exception.)

Mightn't it be possible, if not fairly probable, that a system with a lot of 
those devices already connected to it is being reinstalled? I suppose, though, 
that what you mean is that it'd be okay to assume a braille device since those 
other devices, even if connected, wouldn't be useful when installing.

>1. Distribute udev rules without generic USB UART ids.

Well, as I think we should start doing, we should include both files and each 
distribution can decide which one it'd like to distribute. We could also 
include a comment at the top of each that clearly describes why one may or may 
not wish to use it.

>2. Installer will suppose Braille device.

I assume you mean that the installer would use the rules that include all of 
the devices (not just the non-generic ones).

>3. In the beginning of the installation something like
>brltty-uart-detect will be called. If any UART device is detected as
>Braille device, it will generate e. g. /etc/udev/rules.d/68-brltty.rules
>with proper rules.

Would it be sufficient if the number is high enough to be after the rules for 
the other devices? We currently recommend 90. I don't know if distributions are 
using it or picking their own number. I also don't know if 90 is high enough.

>If we reach the worst case (no way to detect Braille device at all),
>then brltty-uart-detect has to be interactive: ask user to reply,
>whether it is a Braille device. In case of no response, it will be
>considered as not Braille device.

This would be a problem because a braille user wouldn't be able to read the 
prompt until his braille device is working.

>Serial communication with an unknown serial device is a bit unsafe.
>Example: Serial UPS connected to USB UART Bridge. Some old UPS devices
>have a single letter commands, including Power Off. If you send the
>letter, computer will immediately turn off.

Yes. As mentioned above, I don't like serial probing.

>The above mentioned solution will still make impossible to use Arduino,
>some GPS devices and generic USB to serial converters by blind people.
>To solve this problem, it would be needed to limit the driver to a
>particular USB port.

How would a user specify the port, and how would we find out which port the 
device is on?

>The (rejected) patch is not sufficient. Kernel still hooks to this
>device. Here is an example of my GPS Qstarz BT-1000P with openSUSE
>Leap 15.0:

So the kernel isn't using enough information for the other devices to be sure
either?

>  looking at device '/devices/virtual/input/input10':
>    KERNEL=="input10"
>    SUBSYSTEM=="input"
>    DRIVER==""
>    ATTR{name}=="BRLTTY 5.6 Linux Screen Driver Keyboard"
>    ATTR{phys}=="pid-11913/brltty/10"
>    ATTR{properties}=="0"
>    ATTR{uniq}==""

This is a device that brltty creates with uinput. How can we create it 
differently so that it won't cause a problem?

-- 
I believe the Bible to be the very Word of God: http://Mielke.cc/bible/
Dave Mielke            | 2213 Fox Crescent | WebHome: http://Mielke.cc/
EMail: Dave at Mielke.cc  | Ottawa, Ontario   | Twitter: @Dave_Mielke
Phone: +1 613 726 0014 | Canada  K2A 1H7   |