On 11 Jul 2016, at 16:27, 김성민 <dallas1004@gmail.com> wrote:

Hi, I'm Ph.D. Student in KAIST, South Korea. I'm implementing a modified Tor on top of Intel SGX CPU to improve the security issues of Tor as a research. (Here is a related paper work [1, 2].) For this, we port the Tor source code on the VS2012 to build an SGX project. To make Tor running on top of SGX, we modified several lines of code of Tor and I think there is some mistake while printing a platform string.

Tor shouldn't produce output like the output we're seeing, unless the platform name retrieved by tor has been localised. (Which seems unlikely, as the platform name is reported correctly some times, and incorrectly other times.) What do you get if you compile a simple program that only contains the function that retrieves the Windows platform name? It looks like you might be corrupting memory or mangling string processing. But it's very hard to help you unless we know what lines of code you changed, and how you changed them.

Currently, We successfully compiled a SGX-Tor binary and we ran it as a Tor relay for testing compatibility with existing Tor network. We tested a Tor client with our SGX-Tor and it successfully works. Is there any issues or rules to be considered for testing our SGX-Tor?

It should pass all the unit tests run using "make check". It should pass the integration tests run using chutney and "make test-network-all". https://gitweb.torproject.org/chutney.git/ If you do find that it's a programming error in tor, feel free to submit a patch, or a unit test that would catch similar errors in the platform string. https://trac.torproject.org/projects/tor Tim

Thanks, Seongmin

[1] http://ina.kaist.ac.kr/~dongsuh/paper/kim-hotnets2015.pdf [2] http://ina.kaist.ac.kr/~dongsuh/paper/opensgx.pdf

2016-07-08 9:07 GMT+09:00 Tim Wilson-Brown - teor <teor2345@gmail.com>:

...

On 8 Jul 2016, at 09:48, Tim Wilson-Brown - teor <teor2345@gmail.com> wrote:

...

On 8 Jul 2016, at 09:41, nusenu <nusenu@openmailbox.org> wrote:

Hi Seongmin,

out of curiosity I was wondering whether your so called tor "platform" string ("??B`?\u0001") or your tor relay [1] was generated by a modified tor installation on Windows 8 or if we are looking at some bug in vanilla tor?

We plan on authorities rejecting descriptors with non-ASCII characters in 0.2.9, if we get it implemented before the code freeze. https://trac.torproject.org/projects/tor/ticket/18938

Hmm, I can't find an actual descriptor with these characters in it. The latest descriptor I can find for this relay has a normal platform line: https://collector.torproject.org/recent/relay-descriptors/server-descriptors...

Could this be a bug in Atlas?

...

...

thanks!

[1] https://atlas.torproject.org/#details/21E84B294794821E2898E8ED18402E45E4FC35...

descriptor containing that platform string: https://collector.torproject.org/recent/relay-descriptors/server-descriptors...

_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays

Tim Wilson-Brown (teor)

teor2345 at gmail dot com PGP C855 6CED 5D90 A0C5 29F6 4D43 450C BA7F 968F 094B ricochet:ekmygaiu4rzgsk6n

Tim Wilson-Brown (teor)

teor2345 at gmail dot com PGP C855 6CED 5D90 A0C5 29F6 4D43 450C BA7F 968F 094B ricochet:ekmygaiu4rzgsk6n

Tim Wilson-Brown (teor) teor2345 at gmail dot com PGP C855 6CED 5D90 A0C5 29F6 4D43 450C BA7F 968F 094B OTR 8F39BCAC 9C9DDF9A DF5FAE48 1D7D99D4 3B406880 ricochet:ekmygaiu4rzgsk6n