Questions about my idea (for gsoc)

Bharath Vedartham linux.bhar at gmail.com
Mon Apr 1 14:53:29 EDT 2019


On Mon, Apr 01, 2019 at 08:37:51PM +0200, Andrea Laisa wrote:
> I'm not sure what project i'm looking to apply, maybe a netdev driver? Or
> two network protocol driver with a userspace daemon/utilities to control the
> antenna? Or just a userspace daemon/utilities.
> I don't have a clear idea where I should write the code that controls the
> antenna.
>
I think you want to apply for Gsoc for the linux foundation. Go check
out this website for more details. 
https://wiki.linuxfoundation.org/gsoc/google-summer-code-2019
The deadline for submitting your application is April 9th!
> Il 01/04/19 20:28, Andrea Laisa ha scritto:
> >Il 01/04/19 19:16, Greg KH ha scritto:
> >>Who is "you" here?  What project specifically are you looking to apply
> >>to that asked you to post on the kernelnewbies list?
> >
> >As "you" I mean all linux kernel hacker.
"You" should mean an organisation like the Linux foundation, strace etc.
Please be more specific on that.
> >
> >
> >Il 01/04/19 19:16, Greg KH ha scritto:
> >>   - do you have documentation for how this protocol works?
> >>   - do you have access to hardware where you could test these kernel
> >>     changes?
> >
> >Yes, I have access to my home antenna (and access point), but I don't have
> >any tyoe of official documentation because it doesn't exist or I doesnt't
> >found it,
> >
> >I have sniffed my network traffic between the access point and the antenna
> >and I discovered(with wireshark) that they use a custom protocol over LCC
> >(dsap 0x40 and 0x10) to setup the connection to my ISP via LTE. After the
> >setup the AP routes the LAN network traffic to the antenna as expected
> >over a 802.1q-tagged IP packet(with VLAN ID=1121).
> >
> >Although I don't have access to the protocol specifications, I'm going to
> >reverse engineering the protocol!
> >
You have stated that the protocol is proprietary? Could you give more
info on your wireshark output. If the protocol is proprietary, then it
would be really difficult for you to write a driver for it. Reversing
engineering a protocol would take you months to figure it out. I don't
Gsoc would be possible for you right now on this idea. If your interested check out
the other ideas in the link above. 
Also the link above has explained on how to propose your own idea for a
Gsoc project. Please do check it out!
Also try reaching out to the company to see if they already support this
protocol on linux.(Probably not because it is proprietary).

Thanks! 
Bharath
> 
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies at kernelnewbies.org
> https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies



More information about the Kernelnewbies mailing list