[code] Testers wanted: Bluetooth dead man's switch

76 views
Skip to first unread message

Andrew

unread,
Jul 23, 2016, 10:29:16 PM7/23/16
to qubes-users
Hello,

I have made proper RPMs out of the dead man's switch code posted to the
list a few years back. The code can be found here:
https://github.com/kbrn/qubes-app-dms. Pre-built RPMs are attached.
These packages have already been tested on my machine and Work for Me™.

The code comes in three packages:
-qubes-app-dms: for Dom0, contains the core files for dead man's switch
action
-qubes-app-dms-timeout: for Dom0, contains files and config necessary to
trigger the dead man's switch after (by default) 2h of a locked screen
-qubes-app-dms-bluetooth: for a TemplateVM, configured inside AppVM,
contains files and config necessary to trigger the dead man's switch by
absence of a specific previously-seen Bluetooth device.

These both need configuration post-install. See the README.md for more
information. Patches for .deb building or other dead man's switch
triggers are very appreciated!

Marek, what are the prospects of getting this in the Qubes repos? What
else needs to be done?

Andrew
qubes-app-dms-1.0.0-1.fc23.x86_64.rpm
qubes-app-dms-bluetooth-1.0.0-1.fc23.x86_64.rpm
qubes-app-dms-timeout-1.0.0-1.fc23.x86_64.rpm
signature.asc

Manuel Amador (Rudd-O)

unread,
Jul 31, 2016, 8:15:15 PM7/31/16
to qubes...@googlegroups.com
On 07/24/2016 02:29 AM, Andrew wrote:
> Hello,
>
> I have made proper RPMs out of the dead man's switch code posted to the
> list a few years back. The code can be found here:
> https://github.com/kbrn/qubes-app-dms. Pre-built RPMs are attached.
> These packages have already been tested on my machine and Work for Me™.

You need to document what the dead man's switch does on the readme. I
had to go read the source to understand what it does.

Also, frankly, trying to shut down the VMs gracefully is a life-dooming
mistake. Kill them using low-level xl commands or qvm-kill, then
proceed to memory wipe.

--
Rudd-O
http://rudd-o.com/

Andrew

unread,
Jul 31, 2016, 8:25:54 PM7/31/16
to qubes...@googlegroups.com
Manuel Amador (Rudd-O):
Thanks for looking. I'll improve the documentation.

I disagree about the prudence of qvm-kill'ing everything. One must
strike a balance between the consequences of killing everything on a
false alarm and the consequences of delaying a bit when really needed.

I'll make this configurable in the next version so each user can make
the decision for his or herself.

Thanks,
Andrew
Reply all
Reply to author
Forward
0 new messages