Same title, new chapter. Because nothing has changed since yesterday as regards the merits. And this is not, as we warned repeatedly, if Intel could have played dirty or not, not being possible to prove it. The bottom line is blurred what is all the Internet world by the excesses of the NSA and other U.S. government agencies.
But back to pick up the story yesterday, as there is additional information that may interest you. Without going any further, the angry response Linus Torvalds to this issue, specifically to a petition initiated and concluded yesterday at Change.org to remove RdRand of / dev / random:
Where do I start a petition to raise the IQ and the kernel of people? Guys go to read drivers / char / random.c. Then, learn about cryptography. Finally, come back here and to the world: be instructed that you were wrong. Short answer: we know what we are doing. You do not. Long answer: RdRand used as one of many “entries in the entropy pool” [Ndt: Refers to a method for random numbers], and used it as a way to improve that group at random. So even if the NSA put a backdoor in RdRand, our use of RdRand actually improves the quality of the random numbers you get from / dev / random. Short answer: you are ignorant. “
As always, Torvalds responds blunt and sharp. However, it is his word the last one that we will meet today. Theodore Ts’o , a veteran and award-winning developer of Linux also expressed his confidence in this matter, via Google+:
I am very happy to have resisted pressure from Intel engineers that / dev / random is based only on the instruction RdRand.
judgment solely on the random number generator hardware you are using an application within a chip seal that is impossible to analyze, is a bad idea.
So begins a conversation which is now in the 180-odd comments where you will find a little of everything, from asking other veteran users of Linux, Alan Cox, giving his opinion or explaining technical details. And fabric . A very juicy discussion thread, if you are interested in the subject. Also just published an article highly recommended Linux Space reading, where they explain what the component of discord, RdRand.
In conclusion, however vehement response was Linus Torvalds and no reason to take what you say because you (not inconsistent with any of the above), there are other peer experts are not so comfortable with this. In the references we refer.
In fact, ipso facto mend the mistake of not publishing an appointment really descriptive Matt Mackall , the protagonist of the story from yesterday, the commentary of the controversy. But only the end, which, in farewell and, in brackets, said: “ And meanwhile, my distrust in cryptography Intel has moved from” standard professional paranoia “to” legitimate real concern “.
Finally, we repeat once again that it is not confirmed that RdRand-or any other proprietary components that are used or distributed with the kernel, as we are-contains a “trap” of the NSA (there is also code SELinux NSA). And even if it did, as indicated by either Torvalds and linked sources, it is a complementary element only work with Intel chips (from Ivy Bridge) on which also can be disabled.
Beyond this particular case, the great unease that permeates these days, as suggested by the quoted sentence Mackall, is the loss of confidence in all that goes by the United States . And the United States spends. That took her in a general sense, because it points to a much more specific goal.
In any case, the holder of this article and its first part, A ‘back door’ in Linux? It could be , was not exclusively by RdRand, as I suppose many of our readers will have understood by now and before. No need to dig too to meet with “back doors” in Linux, here Skype. Nor is demoralizing issue worldwide. Simply, things are as they are. Should know and, if necessary, go debugging information until it is understood.