[BRLTTY] Focus Keyboard not working

Elias Oltmanns eo at nebensachen.de
Sun Jul 16 09:30:43 EDT 2023


On 2023-07-16 at 06:33:11 (+0200), Nicolas Pitre <nico at fluxnic.net> wrote:
> On Sat, 15 Jul 2023, Dave Mielke wrote:
>> [quoted lines by Nicolas Pitre on 2023/07/13 at 14:41 -0400]
>> 
>> >> Better might be to display a message if TIOCSTI yields EIO.
>> >
>> >Sure. The idea is to let users know why their braille keyboard and 
>> >cut-and-paste might not work.
>> 
>> Do you think we should also make an attempt to "fix" the problem, i.e. to do the sysctl thing?
> 
> Well... I'm not sure. If there is a security hole that comes with that 
> then it would be best not to recreate that hole by default.

Yes, I agree. Users should make an informed decision (well, conscious at
least) when security matters are concerned. After all, kernel developers
considered the situation serious enough to change this behaviour even
many years after concerns have been raised for the first time, as far as
I understand.

So, if BRLTTY could detect the issue and provide a helpful warning
message, that would be great. How about two additional options:
1. enable-tiocsti (off by default),
2. disable-tiocsti-warning (off by default).

BRLTTY would warn by default along these lines:
    Cursor routing and keyboard input disabled due to security concerns,
    see man page for explanation and options.

Setting “enable-tiocsti” to “on” would change that warning to:
    Enabled cursor routing and keyboard input, see man page for
    explanation of security implications.

The option “disable-tiocsti-warning” would shut up both of those
warnings.

Distributors might choose to set “enable-tiocsti“ to “on“ during
installation (if BRLTTY gets started due to a connected braille display
in the first place) yet install the default configuration (setting
turned off) on the target system.

Does that make sense?

> 
> Best would be to fix the kernel.

Well, no dispute there, I suppose. ;)

Best,

Elias


More information about the BRLTTY mailing list