[Cryptography] Use Linux for its security
phill at hallambaker.com
Wed Sep 28 12:59:06 EDT 2016
On Wed, Sep 28, 2016 at 9:20 AM, Ralf Senderek <crypto at senderek.ie> wrote:
> On Tue, 27 Sep 2016, Jerry Leichter wrote:
> Everyone who complains about this situation should have asked himself:
> "When did I last donate my time and effort to essential code review?"
> (including efforts to reduce
There is no way that I would ever donate a minute of my time to reviewing
Linux code for security issues. And the reason is that I am not prepared to
work with someone who behaves the way that Linus does.
In particular, consider the following case:
The issue that set Linus off in this case was that someone attempted to
use a set of compiler macros that establish guards against certain types of
buffer overrun issue.
“The conflict I get is due to stupid new gcc header file crap,” he writes.
“But what makes me upset is that the crap is for completely bogus reasons.”
No, the reason the programmer made use of the features is that he knows
that the design of IPv6 is a nightmare from the point of view of buffer
overrun errors. Each packet can have multiple options and each option is
described in a length-data delimited structure. Now length-data isn't
necessarily a security issue but nested structures certainly are. Given L1
( Data+ L2 (DataB)), many programmers will allocate a memory buffer using
L1 and entirely forget to check that L2 isn't larger than L1.
This was a good faith attempt by someone trying to solve a real security
issue and they got slapped in the face and ridiculed for their trouble.
That is not the way to get good security and even if it was, that is not an
environment in which I am prepared to work.
Being the slave of an alpha dog does not make you alpha, it makes you a
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the cryptography