Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Security

Submission + - Why the NSA builds its own hardware

quartertime writes: Remember Reflections on Trusting Trust, the seminal paper describing how to hide a nearly undetectable backdoor in a compiler? Here's a piece about how to hide a nearly undetectable backdoor in hardware. One can install a backdoor in the expansion ROM of a PCI card, which patches the BIOS to patch grub to patch the kernel to grant the attacker remote root access. Even after the victim reinstalls the operating system from CD, the backdoor will still be there. Given that China builds much of the world's computer hardware, I wonder whether this sort of thing is already part of the Chinese espionage playbook?
Security

Submission + - Analysis of a hardware backdoor (ksplice.com)

An anonymous reader writes: Remember Reflections on Trusting Trust? We know we can't trust our compilers, or our operating systems, or our userspace software. Now even our hardware might be out to get us. This post describes how to install a backdoor in the "expansion ROM" of a PCI card, which patches the BIOS to patch GRUB to patch the Linux kernel to give the controller remote root access. The upshot is that even if the compromise is detected and the victim reinstalls the operating from CD, the backdoor will still be there. Now you know why the NSA builds all its own hardware!

Submission + - What Do You Do With A Disruptive Discovery? 3

jcohen writes: Suppose that you've just discovered a way of making a computationally hard bit of math very, very easy. You've written out your proof, you've verified it, you've written code, and now, say, you're factorizing colossal primes at the rate of 1,000 per second. What's next? The consequences could be huge. How do you get another set of eyes on it to make sure that you're not just another crackpot, and that your results are right? Do you disclose your discovery? How? To whom? To your country's intelligence agency? To the public? What are the conceivable answers to these questions that would have the best consequences for you or for the world?
Security

Submission + - Linux kernel exploit aggressively rooting machines (seclists.org)

An anonymous reader writes: Running 64-bit Linux? Haven't updated yet? You're probably being rooted as I type this. CVE-2010-3081, this week's second high-profile local root exploit in the Linux kernel, is compromising machines left and right. Almost all 64-bit machines are affected, and "Ac1db1tch3z" (classy) published code to let any local user get a root shell. Ac1db1tch3z's exploit is more malicious than usual because it leaves a backdoor behind for itself to exploit later even if the hole is patched. Luckily, there's a tool you can run to see if you've already been exploited, courtesy of security company Ksplice, which beat most of the Linux vendors with a "rebootless" version of the patch.
Security

Submission + - Second major hole in Linux being exploited in wild

quartertime writes: CVE-2010-3081, this week's second high-profile local root exploit in the Linux kernel has been quite a doozy! The bug affects all 64-bit kernels going back to 2.6.26 (and was also backported into RHEL 5's 2.6.18 kernel) and wasn't fixed until last week — shortly before "Ac1db1tch3z" published code to let any local user become root. The exploit works on most versions of Red Hat, Debian and Ubuntu. Several vendors, including Ubuntu and Debian but not Red Hat, have rushed out new kernels to address this bug over the last 2 days. Red Hat's recommended workaround, it turns out, didn't actually close the hole — it just makes the published exploit not work. And Ac1db1tch3z's exploit is more malicious than your typical demo exploit: it leaves a backdoor behind for itself to exploit later even if the hole is patched. Hot-updates vendor Ksplice wrote a tool to see if your system has the backdoor installed (meaning you've been exploited) and has rushed out a "rebootless" patch to plug the hole in advance of Red Hat's own fix.

(Today's earlier article on the H-Online on CVE-2010-3301 incorrectly refers to the workaround Red Hat has recommended for CVE-2010-3081 as a workaround for CVE-2010-3301. The workaround is not effective for either vulnerability.)
Wireless Networking

Submission + - The many faces of 3G (ksplice.com)

An anonymous reader writes: Did you ever notice how each new generation of cell-phone tech gets branded "3G", and the previous thing is retroactively downgraded to some lesser number of G's? An MIT engineer explains why in this brilliant essay about "3G" in the last 10 years, showing how the cell carriers have kept offering it and swiping it away to sell more stuff. He cites numerous Cingular/AT&T and Sprint press releases showing how the companies have made "3G" into a brand name ideally suited for amnesiac consumers. Meanwhile, no cell carrier is foolish enough to sell you bottom-line throughput like an ISP in 1996 — you could actually hold them to that.

Submission + - Today is System Administrator Appreciation Day (sysadminday.com)

ArbiterOne writes: The 11th Annual System Administrator Appreciation Day is today. Celebrated worldwide on the last Friday of July, this holiday honors those who fight in the digital trenches to keep the 'Net alive.

OpenDNS offers a way to remind your boss about the holiday, while another blogger shares war stories. The startup Ksplice created an homage to these heroes... in the style of Choose Your Own Adventure.

How are you celebrating Sysadmin Day?

Programming

Submission + - Source Diving for Sysadmins (ksplice.com)

An anonymous reader writes: Developers think sysadmins don't know how to debug. Screw them: here are 7 tips for source diving as a sysadmin.
Unix

Submission + - Writing filesystems now as easy as Web apps (reddit.com)

An anonymous reader writes: Remember the old days of writing Web apps, when you had to parse the CGI arguments separately, do all the safety checks yourself and implement everything manually? Neither do I, but it looks like all the cool stuff from Web apps is making its way to writing filesystems. This guy shows how to writing an entire Linux filesystem in 50 lines of Python using "dispatch" techniques totally stolen from Ruby on Rails. Are we ready to give up the Web and go back to just using the filesystem for everything, the way Unix intended?

Submission + - Blogger shows that cosmic rays are a real problem

Hanji writes: We have discussed the potential effects of and protections against cosmic ray radiation here before, but for the average computer user, it's an obscure threat that doesn't affect them in any real way. Well here's a blog post that describes a strange segfault and, after extensive debugging, traces it down to a single bit flip, probably caused by a stray a cosmic ray. Lots of helpful descriptions of Linux debugging techniques in this one, and a pretty clear demonstration that this can be a real problem. I know I'm never buying a desktop without ECC RAM ever again!

Slashdot Top Deals

Real Programmers don't eat quiche. They eat Twinkies and Szechwan food.

Working...