[Server-devel] [PATCH] Detect school network, fix sleep and race, fix AC

Martin Langhoff martin.langhoff at gmail.com
Mon Jun 16 18:34:15 EDT 2008


On Mon, Jun 16, 2008 at 6:09 PM, Michael Stone <michael at laptop.org> wrote:
> Are you saying that you think the overhead of starting a single extra
> python interpreter is too high?

Yes. This is a cronjob that we want to run cheaply and often on a
power-constrained machine. If we keep the costs low, then we can be
more eager on limited wattage. What I want to do is to evaluate cheap
variables up front and from bash/dash.

This is what I mean:

$ time bash -c 'exit'
real    0m0.015s
user    0m0.010s
sys     0m0.000s
$ time python -c 'exit'

real    0m0.110s
user    0m0.090s
sys     0m0.010s

> Incidentally, I'm nominally the owner of olpc-netstatus (and, more
> generally, of olpc-netutils / olpc-utils) so please send me some
> patches. I suggest we break it up into a smaller shell programs or a
> D-Bus object, then rewrite both olpc-netstatus and your test in terms of
> those primitives.

Will do if I get back on to relying on network topology - atm it looks
like a dead end :-/

>> In any case, this isn't compatible with Uy or with NYC, so it has to
>> go. We are back to trying to resolve "schoolserver" via DNS and
>> claiming victory if the RSA keys of the server match our expectations.
>
> Ick. Emiliano was asking me how we could include UY-XS detection in our
> initramfs so that we can make our theft-deterrence protocol work for
> them. Thoughts?

Not many at the moment  - but maybe we can talk about it and flesh
something out :-/



m
-- 
 martin.langhoff at gmail.com
 martin at laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff


More information about the Server-devel mailing list