segfault in AES

[zooko]

​https://groups.google.com/forum/?fromgroups=#!topic/cryptopp-users/qGIdqp3MIgg

From that message:

1. You have to make sure that you're taking the SSE2 code path for AES. So, you need to be running on an SSE2 processor, and you either need to disable AESNI when compiling or have a processor that doesn't support it.
2. You have to be running in 32-bit mode.
3. You have to be using signals.

[zooko]

I can't tell if this effects pycryptopp. See my post to the Crypto++ mailing list:

​https://groups.google.com/forum/?fromgroups=#!topic/cryptopp-users/qGIdqp3MIgg

But, if we define CRYPTOPP_DISABLE_ASM always in git/setup.py​, that will almost certainly not hurt and could help.

[daira]

Replying to zooko:

But, if we define CRYPTOPP_DISABLE_ASM always in git/setup.py​, that will almost certainly not hurt and could help.

+1

[zooko]

I posted an update on the Crypto++ mailing list: ​https://groups.google.com/forum/?fromgroups=#!topic/cryptopp-users/qGIdqp3MIgg

#85 appears to be on the way to fixing this bug, but it is somewhat dissatisfying since I don't know _exactly_ why #85 fixes it.

[gdt]

I went to the URL above, and got a blank screen, probably because I don't take google cookies or have a google login. Please consider this a bug report that requiring having a relationship with an advertising company to participate in an open-source project is a bug :-)

[zooko]

gdt: sorry about that. Here's a mailing list archive on an alternate host:

​http://www.mail-archive.com/search?l=cryptopp-users@googlegroups.com&q=subject:%22Re%3A+Segfault+with+core+dump+on+5.6.1+in+Rijndael%3A%3AEnc%3A%3AAdvancedProcessBlocks%22

[zooko]

fixed by #85