I cannot wait for the accompanying Black Hat talk. Christopher Domas is one of my absolute favorite all-time hackers. He does such a fantastic job of explaining his work. Some of my favorite talks of his:
My introduction to his work was "The future of RE Dynamic Binary Visualization"[0] which completely blew me away. It still feels futuristic today, 13 years later. Novel UI/UX paradigms like this are slow to find widespread adoption, even when they're so clearly demonstrated to be such an ideal fit for their purpose.
I love digraphs. I learned about them from that talk's associated paper... but only about 2 years ago. I've been dumping my system and GPU RAM raw and visualizing via digraphs and it's amazing how such a simple algorithm operating on arbitrary bytes leads to such distinct and consistent image.
If this is the same dude I am thinking of, his wife is also the CISO of Mozilla and do security research together, afair they have a whole book on x86 reverse engineering.
When I started with computers, DRAM was understandable by a teenager: RAS, CAS, read, done.
Ok, the necessary refresh was always a little pain, but still something manageable.
Nowadays, I feel you need three PhD's to even bring up a micro with DRAM and don't get me started on the proprietary binary blobs necessary just for DRAM access. No wonder PSRAM is a thing.
The corollary is that it shouldn't be too surprising that this gigantic attack surface provides many opportunities. (Of course that doesn't mean it is easy to find them, hat tip to Christopher Domas, just that I expect there to be many more).
This is all great to get full unfettered access to your own system, as life should be.
I’m sure Xbox and PlayStation security groups are a little nervous right now though. Getting ring-0 on those machines is near impossible, but once you do then everything else becomes wide open
Not sure if it opens up that much on them as far as their security processors go. Modern consoles already treat DRAM as completely untrusted (an attacker could just sit on the DRAM bus and sniff/issue requests there).
The Xbox One for example encrypts all the DRAM it uses after it gets out of the main CPU die. See this part of Tony Chen's presentation https://youtu.be/U7VwtOrwceo?t=956
In the README it talks about accessing and modifying the code of the Platform Security Processor including accessing the keys, how is what is on an Xbox different?
The Xbox does not use AMD's Platform Security Processor, they have their own custom security chip with Microsoft hardware and their own custom BootROM. This custom security processor actually has its own bit of RAM entirely in the die and doesn't even go out to the DRAM at all.
wouldn't request after DRAM controller be unencrypted ? Would need to have different key per memory type and even then you could do some damage as realistically it won't have number of keys equal to running processes
OK, so this works on AMD Jaguar according to the README. That's a architecture from 2013. There's notes about Zen 3 having a different base address for the memory controller registers, but that's it. What newer CPUs does attack actually work on?
CPU designers have been aware of this attack vector for a while, so many (maybe all) have controls to lock these registers. They are written and locked by vendor firmware before it hands off to the operating system.
You would have first break the firmware or locks before an attack like this on a modern CPU
Zen has completely different memory controller IP (UMC), that's configured at boot by AGESA/PSP. I doubt this exploit applies to modern Zen CPUs, however AMD are the only ones who could really confirm this.
Then publishing it in this incomplete state is just pointless fearmongering and will just make others fill in this information within the next couple of days. And the good guys likely won't be the first ones to do that.
So on an affected system, ring 0 root has access to pretty much everything that was hidden in negative ring territory. The page is pretty quiet about what other processor families might be similar beyond this specific AMD16h (an older AMD low-power family)?
from the GH page:
> Developed and tested on AMD Family 16h CPUs, the last generation whose datasheets document the DRAM controller's translation registers — and show that they can't be locked. 17h and beyond simply leave this information out.
As long as you know the controller's translation registers, it's applicable? Not tested on later one's merely because the information wasn't readily available it seems.
> Developed and tested on AMD Family 16h CPUs, the last generation whose datasheets document the DRAM controller's translation registers — and show that they can't be locked. 17h and beyond simply leave this information out.
Reaching into ring -2 or the TPM allows privilege escalations past traditional "root permissions" and lets attackers defeat the sort of tamper protection that's designed to make escalations to local root manageable. Wipe-resistant malware, falsified cryptographic attestations, all sorts of fun.
AFAIK you can't usually replace CPU microcode, as it is signed. Seizing it here, in RW memory, really begs to toy with it: add instructions, edit them. Could you teach your CPU to understand RISC-V?
This is just one random idea. But altering PSP code is also interesting, to use it for your own purposes, or extract encryption keys / make it lie to clients (breaking DRM, for instance).
The bulk of microcode is still ROM inside the CPU. Also it's mostly used for the more complex instructions, the basic load/store/add/etc. would be decoded and executed more directly.
CPUs have a limited amount of SRAM for holding patches, basically a list of addresses in the microcode ROM, and what their contents should be replaced with. Not enough to totally change the instruction set, but still exciting to potentially get write access to, as indeed the normal update mechanism requires those patches to be cryptographically signed by Intel/AMD.
1) Don't give your guests access to the DCT control registers.
2) On 15h - no obvious way. I don't know the other families.
3) Yes -- this can be verified easily. Pick up the AMD 15h BKDG, look up BankSwizzleMode. It's documented. The one oversight is that this bit is not under the Dram Controller's lock bit.
When Chris Domas left Battelle for Intel years back, shortly after hardware-fuzzing a bank of thin-clients to discover undocumented x86 instructions, I was convinced Intel was basically keeping him on the payroll to shut him up.
One hopes that a hypervisor would not expose hardware control registers directly in the first place, except ones deliberately designed for virtualization support.
Otherwise, the guest is running effectively at the same privilege level as the hypervisor (that's useful sometimes, but probably not intended in most applications).
Ok, on 2. and in general this exploit only works on pre-Zen AMD platforms as the repo states in not-so-clear terms.
Zen changed DTC (DRAM Controller) to UMC (Unified Memory Controller), UMC is programmed at boot, and one would hope they figured that locking access to it makes sense when they were adding confidential compute support; Not clear though because there is no public documentation on it, so best we can hope for is some statement from AMD/3rd party researcher saying "this won't work on Zen because X/Y/Z"
With respect to 2), I don't think this should work architecturally even if the DRAM controller has knobs which can be accessed, because the guest's RAM should be encrypted with keys that can't be recovered in this way.
It's definitely a good research topic because there are a lot of moving pieces and having this kind of primitive might weaken one of them in a useful way, but at least at the top level, you couldn't just swap one guest's DRAM bank with another and get their confidential memory contents back this way.
I don't understand the threat model being attacked here. If you had physical DRAM access you could do all of this anyway right? And I would assume that an unprivileged user would not have write access to the DRAM controller registers?
This doesn't require physical DRAM access, it's all software.
With ring-0 access, this lets you poke "even things walled off and invisible to ring-0 or the CPU itself" including things that the security processor tries hard to wall off.
Which arguably is a good thing. As a owner of the system I really should have complete control over it. But currently there is software I have no control over running at even higher privilege levels.
The only modern silicon that gives me full control over what code is running is some (or most?) microcontrollers.
And this isn't just a question of FOSS principle. Especially SMM is problematic by unpredictably taking CPU cycles away from your workload. This can mess up hard realtime workloads, such as found in CNC controllers. If you are running something like LinuxCNC this something you need to measure to figure out if a given computer is suitable for that job.
Whilst I like getting more control of my hardware, I really prefer if I can contain arbitrary software - sandbox it and compartmentalise it. If arbitrary processes can own your whole system, is anyone truly an owner?
So you might like Qubes OS, which isolates all apps into hardware-assisted VMs. You will still have all the access from the AdminVM, but arbitrary processes won't.
Pragmatically, we all want the MMU to work and for pointers to address the correct logical address. No one is keeping anything valuable from you by preventing this. It's not like a DRM scheme.
You're missing that modern CPUs substantially lock the users out of control of their own computer and include things like hidden additional network connected processors that run their own full on operating systems. ... and may well be used to surveil or remotely access your computers the the behest of powers unknown.
But they still use system dram, so this approach allows looking into those parts of your own computer from which you're normally blocked. At least on some hardware...
It's not fully mitigated by encryption, you can still do a lot of damage without being able to observe plaintexts. For example, you could "rewind" a ciphertext block to an earlier value, and induce a UAF-like condition in the software it belongs to.
Apple’s Secure Enclave has replay protection since Apple A11.
Generally, I don’t see why a modern security platform wouldn’t have its own private SRAM to be used as a root of trust for encrypted blobs stored in shared DRAM.
I really hate to be that guy, but man, as someone who was and is a big Christopher Domas fan (and is way dumber than him, I mean, this stuff is seriously over my head)... it's been really disappointing to see him LLM'ing all the READMEs recently. They used to be a joy to read through, but now the Claudeisms made it such a slog I could barely get through a few paragraphs. I'm glad he's using the new tools to get even more cool stuff done, but I wish he'd have gone for a human writeup at the end.
I came to the comments to complain about how unreadable it was, saw that it was the movfuscator guy, went back to check if I somehow made a mistake but no, if anything that writeup is one of the egregious ones I've seen recently. What a shame, I really liked his talks.
Like.. having no readme or a super technical one doesn't work anymore in the age of LLMs, but having one that hurts to read might?
Because people trying to get an LLM to translate it just get more LLM output. So you actually _have_ to put in manual effort to rip out what the fuck it wants to tell you.
Disappointing indeed. Us security folk have already earned a poor reputation for ineffective technical communication, and LLMs are even worse, not better. This is especially disappointing in this case because we know xoreaxeaxeax is one of the few who are actually capable of effectively communicating beautifully cursed low-level hackery, but is now choosing to outsource the most impactful part of his work (since most people will only read the abstract, I mean, README) to an LLM. I'll still take new projects with sloppy READMEs over the previous years of radio silence, but I really wish xoreaxeaxeax would recognize the value in spending the time to write a single page of text in his own voice.
I'm not the person you are replying to, but the readme is very clearly written by an AI, and it sounds nothing like his older work. Sometimes it's just super clear to people something is written with AI without you getting some sort of singular "gotcha" word or indicator. It's just writing patterns that would be hard to clearly establish rules for here in an HN comment, but it's incredibly obvious when you learn to spot it.
How do you know when it's not written by LLM except to take someone's word? If I got caught slopping I'd be too embarassed to admit it. "I wrote it all myself! I actually majored in slick writing and minored in tenuous metaphor."
I'd just like to address this real quick because some people seem to think this is just a "hunch" that has some probability of being false; there is absolutely nothing more certain on planet Earth than the LLM involvement in this writing. It is difficult to come up with things that are certain enough to compare this to to convey the lack of doubt that exists.
I am not going to make fun of you for not being able to tell, although I do find it surprising that people seem to struggle in both directions with telling AI and human writing apart (are our brains really that different?) - I just want it to be clear that some of us can pick up Claudisms within just a couple of sentences with no effort. A Claude-generated sentence, in isolation, may not ring any alarm bells. A few of them in a row, however, that's a load-bearing smoking gun right there.
We can certainly argue to what extent undisclosed LLM involvement is an issue or not, though frankly I don't like reading LLM writeups so I would greatly prefer if people would stop using LLMs for public facing documents. But, it is at least worth making this much clear: we can tell.
The only other reasonable explanation would be that he has consumed so much LLM content that the machine has erased his voice and replaced it with its own.
This blows my mind just as badly and I see it reasonably often: I talk nothing like any AI model that has ever been, it isn't even close. I'm going to be bold and say that even without checking, more analytical approaches like stylometry would prove me right with little doubt. To mistake my writing for being "AI-esque", you would pretty much have to be entirely blind to the stylistic aspects of the text and fall back on even more superficial details like how verbose it is.
Go ahead and throw the comment into your favorite unreliable AI detector. Even though I suspect they're mostly garbage, my writing is just so far away from what AI models do that it doesn't even matter.
edit: I caved into temptation and checked. Big fat zero on GPTZero.
The em dashes are the most obvious stereotypical tell, but that doesn't really matter that much (I actually like them and occasionally used them pre-AI). It's hard to put a finger on, but the most annoying LLMism to me is the overdramatic, staccato, almost "epic" way they talk. It feels like a 2009 lens flare effect over everything, it sounds like a stereotypical hacker in a CSI show.
> the last generation whose datasheets document the DRAM controller's translation registers — and show that they can't be locked
> When your code dereferences *p, it appears to access the DRAM at p. It does not — p is a virtual address
> Physical addresses are really more of a suggestion.
> That's the exploit. All of it.
The worst part is that this stuff is genuinely cool and deserves to be dramatic. And I like stereotypical, campy hacker speak! But LLMs are, IDK... bad at it? Or maybe it just becomes a bore to read the same. Exact. Dramatic. Voice. From literally everyone. After you've heard it enough times.
None of this is against Mr. Domas. He seems like a cool person, with a cool voice, and I want to read his voice, not Claude's.
> What if you're wrong?
I definitely could be! Apologies if I am. But with all the em dashes and such, and having read his previous work, I felt confident enough to mention it. And as the sibling comment says, it really is something you just learn to spot over time.
It's the ASCII diagrams for me. Although I hesitate to point it out because they are usually helpful unlike the em dashes. Before LLMs, ASCII diagrams in readmes were a half baked mess, if one were present at all. Now every project has at least one perfectly made ASCII diagram - instant LLM tell.
> Although I hesitate to point it out because they are usually helpful unlike the em dashes.
Likewise, I don't mind the diagrams. Though they do often have the same flaw as other text, being that the LLM throws in EVERYTHING, vs. a handmade one that'd generally have more taste and discretion to it. That can kind of work in its favor here, since the point is just to show the complexity of the stack, but on the other hand the reader lacks confidence that every item in there is "really" a part of the stack (which I would be fully confident in for this author, had he written it by hand) and not just some process related to memory/DRAM that the LLM decided to toss in.
> When your code dereferences *p, it appears to access the DRAM at p. It does not — p is a virtual address
I hate these especially much: It's at the same both both overly dramatic, it's presented as some great reveal that will change everything, while at the same time being completely trivial and only detracts from the explanation. If you have no idea that memory addresses are translated you will understand absolutely nothing from the text or even what this is all about. If you want to explain what an MMU is, just do that instead and don't present it as some great revelation.
But some equally dramatic phrasings could just as well be something that leaves you astonished. You never know. You have to skim the text to find what is useful information and what is just filler. The signal-to-noise is low.
This is so cool. Outside of a cool demo, and maybe some black hat type stuff, this is surely dangerous, a bad idea, and shouldn't be done in prod. But pure hacker ethos at its heart.
Without any swizzling, certain common access patterns can end up with subpar performance, for example walking the columns of a 2d array with a certain stride - if it ends up directing every access to the same bank on the same channel, the throughput is much lower than if the load was evenly distributed across multiple banks/channels.
Swizzling "randomizes" bank/rank/channel distribution, which makes unlucky access patterns less likely. (Something I'd like to research is microbenchmarking different access patterns to infer the swizzle pattern and defeat physical ASLR)
Holy shit, Christopher Domas is back. I remember watching his Defcon talks on x86 shenanigans[^1][^2] and being amazed at what he's been able to discover. Then he got whisked away by Intel and now drops this. I'm excited.
Nah, it's not just that, literally all the stuff they've posted this year is obvious LLM writing, none of the stuff from previous years is. To get this close to LLM writing style without actually using an LLM, you would pretty much have to be purposefully trying.
But I have a new favorite way of demonstrating this:
Like, who cares? In today's present, I wouldn't bother writing the article myself neither besides giving the instructions and auditing the output. Substance is what matters
I struggled to understand the article, it doesn't properly explain what the exploit is in a way I could immediately understand, after staring at it for a bit and reading comments here I think it's a hardware register that's exposed to user software that shouldn't be but I really didn't understand that at first glance ("Poke the DRAM controller" made me think a hardware exploit) and I'm still not sure if I'm right about that (if it is, just start by explaining that instead of all the unnecessary filler and diagrams and animations of the Mona Lisa being scrambled and talk of bath salts which I still don't understand the relevance of)
Ultimately, yes, this seems to be a memory controller register that should not be exposed to ring 0 or above.
Ring -1 needs DRAM, so it tells the memory controller to give it some blocks. The memory controller hands back a “physical” address, and promises not to let anything but ring -1 access that address.
The exploit takes advantage of that control register to remap the same DRAM blocks to a different physical address. Since the memory controller only promised to protect the physical address it handed back, that protection is bypassed when using the new address.
There are several theoretical ways to mitigate this exploit, but it remains to be seen if the system is sufficiently field-upgradeable to defend.
Assuming this is a genuine question, here is why people care, and why this sort of writing is a waste of everyone's time.
LLMs can of course generate a lot of text about a subject, but they are still quite bad at generating a piece of writing with a coherent point. Remember how in grade school they teach you that your writing should have stuff like "introductions" a "thesis" and "topic sentences" and "conclusions"? How these things give structure to your writing, communicating to your reader both what they are reading about and why you are telling them about it? LLMs still don't seem have a model for the why part of writing. They can generate large homogeneous blobs of text on the topic at hand, but fail to differentiate the important parts from the details.
For example: all those LinkedIn cliches that LLMs are so fond of - "it's not X, it's Y", rule of 3, etc. - these are tools for bringing focus to the most important points. Even terrible LinkedIn posters implicitly know to use these cliches to drive home their (usually anodyne) messages. LLMs don't understand this, so they just use linkedin cliches everywhere, turning the whole thing into a breathless monotone.
Besides all of that - unless you've somehow missed the constant parade of people begging others to stop sending them LLM-generated prose, and all the reasons they've given for why it's actively bad for everyone involved - it seems like "who cares?" is a bit of a disingenuous question, and one you could have easily answered yourself.
> I wouldn't bother writing the article myself neither besides giving the instructions and auditing the output.
The article would be better with just the instructions and audited output. All the LLM added bloat is tiring and distracting; it's like an article from New-Yorker or Wired.
Well for one thing I also trust most people I respect to not suddenly have someone else speaking on their behalf in their voice with no disclosure in most contexts where it would not be normal, yet that seems to have happened here. So while I personally do still trust and even respect the author, I can't help but empathize with someone who is suddenly a bit more skeptical.
Speaking as someone who posts a lot of stuff like this, the question isn't "Do I use AI or not?" The question is, "Do I have time to deal with writing this up for public consumption or not?"
This was a relatively complicated post of the sort that we are lucky to get in any form, AI-assisted or otherwise. Does it meet my personal stylistic standards? No, it's too LLM-ish. Assuming I cared about the presentation at all -- which I don't always, but would here -- I wouldn't be able to stop myself from fixing that in the process of reviewing it. Is it the usual bucket of slop? Emphatically no.
I wonder if the author asked an LLM to turn his Black Hat talk track into a blog post. I can’t imagine prompting an LLM to produce a post this organized and fact-dense (if stilted in tone) would be any quicker than just writing it up myself.
You are literally asking this to people who clearly care...
> In today's present, I wouldn't bother writing the article myself neither besides giving the instructions and auditing the output.
Then why exactly are you even bothering to reply to me instead of having Claude do it and auditing the output? If HN didn't have a rule against it, would you even bother replying yourself?
But really, there's a fair bit more to unpack here than just that. Why wouldn't you bother? Is writing a README.md about some project you worked on really that hard? Even with heavy LLM assistance, I'd wager to guess this project, which clearly involved working on real hardware, was more than just prompting. So clearly there was human effort other than prompting. And I do respect that, but I want people who write things to respect my time. I'm not asking them to disclose every tool they use, I'm asking them to not waste our time with crappy irritating Claude writeups. Whether it's explictly specified or not, we know.
Frankly I struggle to believe that people don't really mind if someone else speaks for them in their own voice, just because they're too fucking lazy to speak for themselves anymore. We've had competent GenAI for like a year or two, at this rate people are going to forget their potty training in another few months.
> Substance is what matters
Substance matters, which is not great for LLMs, because they put out text that has far more fluff than substance. What, however, is far worse for LLMs, is the fact that kick and scream and cry all you want, but: style and presentation matters, too.
It is absolutely true that if you just dropped a very brief blurb that all AMD CPUs from a certain generation can be pwned it would have a decent chance to hit the HN frontpage just out of sheer interestingness. That is not because the style and presentation doesn't matter, it's just that the substance is significant in spite of the bad presentation and style.
And absolutely, we can easily forgive someone for simply not being very good at the presentation and style part, certainly I'm not really an expert at it. But this author has released plenty of great hits before, so I damn well know they can. It's a serious disappointment to see them downgrade to irritating, grating Claude garbage output.
Guardrails are a product-quality smell. Boycott guardrailed models. Punish guardrailed model providers with reduced revenue and bankruptcy. "I'm sorry Dave" must become a subscription-cancelling response or the nannying will never stop!
Likewise -- also had zero issues going over this with Sol. Seemed to give solid advice for how to test it -- use an expendable bare-metal AMD family 16h test system w/ usual standard checks that apply.
> Run `platform_check` first and do not use `SKITTER_FORCE=1` casually. Start with the read-only `dram_state` and `dram_carveouts`, then `dram_dump --dry-run`. Avoid `dram_poke` until maps have been freshly collected and calibrated. Do not bypass fingerprint checks, calibration, fencing, or verification.
Claude's (apparently externally-mandated?) lobotomization continues to be concerning. :-/
This is the level of access the rightful owner of a computer should have to his own system.
He should also be able to fuse away this access forever, to be fair. But out of the box, when I get a new laptop, I should be able to read and write every byte of DRAM.
Humpty Dumpty said in rather a scornful tone, “it means just what I choose it to mean—neither more nor less.” “The question is,” said Alice, “whether you can make words mean so many different things.” “The question is,” said Humpty Dumpty, “which is to be master—that's all.”
The hardware DRAM controller maps "physical addresses" approximately to: {DRAM slot number, chip number in slot, bank number in chip, row number in bank, byte number in row} via a complex map for various irrelevant reasons. All permission checks are before this mapping. So if you change the mapping, you can access shit you should not be able to, like TPM and SMM memory. OP found a way to change the mapping.
Skitter creek bath salts... Or SCBS
Guess there'll be a talk called Secure Computing BullShit in the next Blackhat conf! I'll be eagerly waiting for it! :)
- Psychological Warfare in Reverse Engineering https://www.youtube.com/watch?v=HlUe0TUHOIc
- The MoVfuscator https://www.youtube.com/watch?v=R7EEoWg6Ekk
- Hardware Backdoors in redacted x86 https://www.youtube.com/watch?v=jmTwlEh8L7g
0: https://www.youtube.com/watch?v=4bM3Gut1hIk&pp=ygURY2hyaXN0b...
Very cool!
https://onlinelibrary.wiley.com/doi/book/10.1002/97813942771...
He did a fantastic job of explaining his work.
I had to look at the comments here to understand what was going on
So maybe he USED to explain things well, but that's not on display
https://www.youtube.com/watch?v=iOq8O_phwbA
He looks so different.
Ok, the necessary refresh was always a little pain, but still something manageable.
Nowadays, I feel you need three PhD's to even bring up a micro with DRAM and don't get me started on the proprietary binary blobs necessary just for DRAM access. No wonder PSRAM is a thing.
The corollary is that it shouldn't be too surprising that this gigantic attack surface provides many opportunities. (Of course that doesn't mean it is easy to find them, hat tip to Christopher Domas, just that I expect there to be many more).
I’m sure Xbox and PlayStation security groups are a little nervous right now though. Getting ring-0 on those machines is near impossible, but once you do then everything else becomes wide open
The Xbox One for example encrypts all the DRAM it uses after it gets out of the main CPU die. See this part of Tony Chen's presentation https://youtu.be/U7VwtOrwceo?t=956
Also see this bit on the Apple Secure Enclave in the "Memory Protection Engine" section which also explains how they encrypt stuff stored in DRAM: https://support.apple.com/guide/security/the-secure-enclave-...
See the slide on Tony Chen's presentation about Xbox One security https://youtu.be/U7VwtOrwceo?t=843 and this video from the person who hacked the Xbox One https://youtu.be/FTFn4UZsA5U?t=299
You would have first break the firmware or locks before an attack like this on a modern CPU
> Developed and tested on AMD Family 16h CPUs, the last generation whose datasheets document the DRAM controller's translation registers
> Developed and tested on AMD Family 16h CPUs, the last generation whose datasheets document the DRAM controller's translation registers — and show that they can't be locked. 17h and beyond simply leave this information out.
This is just one random idea. But altering PSP code is also interesting, to use it for your own purposes, or extract encryption keys / make it lie to clients (breaking DRM, for instance).
The bulk of microcode is still ROM inside the CPU. Also it's mostly used for the more complex instructions, the basic load/store/add/etc. would be decoded and executed more directly.
CPUs have a limited amount of SRAM for holding patches, basically a list of addresses in the microcode ROM, and what their contents should be replaced with. Not enough to totally change the instruction set, but still exciting to potentially get write access to, as indeed the normal update mechanism requires those patches to be cryptographically signed by Intel/AMD.
Its not to say that its not impressive, but its fairly isolated to a specific family of processors from 2013.
And whether it's really real in the first place.
2) On 15h - no obvious way. I don't know the other families.
3) Yes -- this can be verified easily. Pick up the AMD 15h BKDG, look up BankSwizzleMode. It's documented. The one oversight is that this bit is not under the Dram Controller's lock bit.
Otherwise, the guest is running effectively at the same privilege level as the hypervisor (that's useful sometimes, but probably not intended in most applications).
So far seems this is about right:
1. You need platform register access, so seems can't KVM-escape with just this
2. Big question is what about breaking Confidential SEV-SNP guests from the host?
Zen changed DTC (DRAM Controller) to UMC (Unified Memory Controller), UMC is programmed at boot, and one would hope they figured that locking access to it makes sense when they were adding confidential compute support; Not clear though because there is no public documentation on it, so best we can hope for is some statement from AMD/3rd party researcher saying "this won't work on Zen because X/Y/Z"
It's definitely a good research topic because there are a lot of moving pieces and having this kind of primitive might weaken one of them in a useful way, but at least at the top level, you couldn't just swap one guest's DRAM bank with another and get their confidential memory contents back this way.
With ring-0 access, this lets you poke "even things walled off and invisible to ring-0 or the CPU itself" including things that the security processor tries hard to wall off.
The only modern silicon that gives me full control over what code is running is some (or most?) microcontrollers.
And this isn't just a question of FOSS principle. Especially SMM is problematic by unpredictably taking CPU cycles away from your workload. This can mess up hard realtime workloads, such as found in CNC controllers. If you are running something like LinuxCNC this something you need to measure to figure out if a given computer is suitable for that job.
But they still use system dram, so this approach allows looking into those parts of your own computer from which you're normally blocked. At least on some hardware...
Generally, I don’t see why a modern security platform wouldn’t have its own private SRAM to be used as a root of trust for encrypted blobs stored in shared DRAM.
Like.. having no readme or a super technical one doesn't work anymore in the age of LLMs, but having one that hurts to read might?
Because people trying to get an LLM to translate it just get more LLM output. So you actually _have_ to put in manual effort to rip out what the fuck it wants to tell you.
That would be clever.
> the Claudeisms
This is hand-waving. Please be more specific.
> made it such a slog
On the flip-side, I didn't find it a slog at all. What if you're wrong?
My gut feeling agrees. The rule of three is one of the stronger signals, can't stamp that out of the AI even if you wanted :-)
I'll push it to GH later, it's nothing fancy but it has been quite good in my experience. Here is highlights which it used
``` tricolon coordinated VERB run: “…break / on them collapse / unlock everything .” (3 members) tricolon coordinated VERB run: “…guard physical addresses / not DRAM coordinates / you rearrange the” (3 members) tricolon coordinated NOUN run: “…handful of data / it to z3 / the translation matrix” (3 members) tricolon coordinated NOUN run: “…view / the elaborate fences / locks / security checks the” (4 members) tricolon coordinated VERB run: “…Read it / the alias map / pipe” (3 members) ```
I'd just like to address this real quick because some people seem to think this is just a "hunch" that has some probability of being false; there is absolutely nothing more certain on planet Earth than the LLM involvement in this writing. It is difficult to come up with things that are certain enough to compare this to to convey the lack of doubt that exists.
I am not going to make fun of you for not being able to tell, although I do find it surprising that people seem to struggle in both directions with telling AI and human writing apart (are our brains really that different?) - I just want it to be clear that some of us can pick up Claudisms within just a couple of sentences with no effort. A Claude-generated sentence, in isolation, may not ring any alarm bells. A few of them in a row, however, that's a load-bearing smoking gun right there.
We can certainly argue to what extent undisclosed LLM involvement is an issue or not, though frankly I don't like reading LLM writeups so I would greatly prefer if people would stop using LLMs for public facing documents. But, it is at least worth making this much clear: we can tell.
Go ahead and throw the comment into your favorite unreliable AI detector. Even though I suspect they're mostly garbage, my writing is just so far away from what AI models do that it doesn't even matter.
edit: I caved into temptation and checked. Big fat zero on GPTZero.
The em dashes are the most obvious stereotypical tell, but that doesn't really matter that much (I actually like them and occasionally used them pre-AI). It's hard to put a finger on, but the most annoying LLMism to me is the overdramatic, staccato, almost "epic" way they talk. It feels like a 2009 lens flare effect over everything, it sounds like a stereotypical hacker in a CSI show.
> the last generation whose datasheets document the DRAM controller's translation registers — and show that they can't be locked
> When your code dereferences *p, it appears to access the DRAM at p. It does not — p is a virtual address
> Physical addresses are really more of a suggestion.
> That's the exploit. All of it.
The worst part is that this stuff is genuinely cool and deserves to be dramatic. And I like stereotypical, campy hacker speak! But LLMs are, IDK... bad at it? Or maybe it just becomes a bore to read the same. Exact. Dramatic. Voice. From literally everyone. After you've heard it enough times.
None of this is against Mr. Domas. He seems like a cool person, with a cool voice, and I want to read his voice, not Claude's.
> What if you're wrong?
I definitely could be! Apologies if I am. But with all the em dashes and such, and having read his previous work, I felt confident enough to mention it. And as the sibling comment says, it really is something you just learn to spot over time.
Likewise, I don't mind the diagrams. Though they do often have the same flaw as other text, being that the LLM throws in EVERYTHING, vs. a handmade one that'd generally have more taste and discretion to it. That can kind of work in its favor here, since the point is just to show the complexity of the stack, but on the other hand the reader lacks confidence that every item in there is "really" a part of the stack (which I would be fully confident in for this author, had he written it by hand) and not just some process related to memory/DRAM that the LLM decided to toss in.
Damn. Now my artist-mode in Emacs skills are useless.
I hate these especially much: It's at the same both both overly dramatic, it's presented as some great reveal that will change everything, while at the same time being completely trivial and only detracts from the explanation. If you have no idea that memory addresses are translated you will understand absolutely nothing from the text or even what this is all about. If you want to explain what an MMU is, just do that instead and don't present it as some great revelation.
But some equally dramatic phrasings could just as well be something that leaves you astonished. You never know. You have to skim the text to find what is useful information and what is just filler. The signal-to-noise is low.
It's called slop for a reason.
https://news.ycombinator.com/from?site=github.com/xoreaxeaxe...
Swizzling "randomizes" bank/rank/channel distribution, which makes unlucky access patterns less likely. (Something I'd like to research is microbenchmarking different access patterns to infer the swizzle pattern and defeat physical ASLR)
[^1]: https://www.youtube.com/watch?v=XH0F9r0siTI
[^2]: https://www.youtube.com/watch?v=jmTwlEh8L7g
But why on earth do they have to use AI to write their writeups?!
But I have a new favorite way of demonstrating this:
https://github.com/search?q=owner%3Axoreaxeaxeax+load-bearin...
Guess how many of these are from before 2025.
Ring -1 needs DRAM, so it tells the memory controller to give it some blocks. The memory controller hands back a “physical” address, and promises not to let anything but ring -1 access that address.
The exploit takes advantage of that control register to remap the same DRAM blocks to a different physical address. Since the memory controller only promised to protect the physical address it handed back, that protection is bypassed when using the new address.
There are several theoretical ways to mitigate this exploit, but it remains to be seen if the system is sufficiently field-upgradeable to defend.
Assuming this is a genuine question, here is why people care, and why this sort of writing is a waste of everyone's time.
LLMs can of course generate a lot of text about a subject, but they are still quite bad at generating a piece of writing with a coherent point. Remember how in grade school they teach you that your writing should have stuff like "introductions" a "thesis" and "topic sentences" and "conclusions"? How these things give structure to your writing, communicating to your reader both what they are reading about and why you are telling them about it? LLMs still don't seem have a model for the why part of writing. They can generate large homogeneous blobs of text on the topic at hand, but fail to differentiate the important parts from the details.
For example: all those LinkedIn cliches that LLMs are so fond of - "it's not X, it's Y", rule of 3, etc. - these are tools for bringing focus to the most important points. Even terrible LinkedIn posters implicitly know to use these cliches to drive home their (usually anodyne) messages. LLMs don't understand this, so they just use linkedin cliches everywhere, turning the whole thing into a breathless monotone.
Besides all of that - unless you've somehow missed the constant parade of people begging others to stop sending them LLM-generated prose, and all the reasons they've given for why it's actively bad for everyone involved - it seems like "who cares?" is a bit of a disingenuous question, and one you could have easily answered yourself.
The article would be better with just the instructions and audited output. All the LLM added bloat is tiring and distracting; it's like an article from New-Yorker or Wired.
This was a relatively complicated post of the sort that we are lucky to get in any form, AI-assisted or otherwise. Does it meet my personal stylistic standards? No, it's too LLM-ish. Assuming I cared about the presentation at all -- which I don't always, but would here -- I wouldn't be able to stop myself from fixing that in the process of reviewing it. Is it the usual bucket of slop? Emphatically no.
You are literally asking this to people who clearly care...
> In today's present, I wouldn't bother writing the article myself neither besides giving the instructions and auditing the output.
Then why exactly are you even bothering to reply to me instead of having Claude do it and auditing the output? If HN didn't have a rule against it, would you even bother replying yourself?
But really, there's a fair bit more to unpack here than just that. Why wouldn't you bother? Is writing a README.md about some project you worked on really that hard? Even with heavy LLM assistance, I'd wager to guess this project, which clearly involved working on real hardware, was more than just prompting. So clearly there was human effort other than prompting. And I do respect that, but I want people who write things to respect my time. I'm not asking them to disclose every tool they use, I'm asking them to not waste our time with crappy irritating Claude writeups. Whether it's explictly specified or not, we know.
Frankly I struggle to believe that people don't really mind if someone else speaks for them in their own voice, just because they're too fucking lazy to speak for themselves anymore. We've had competent GenAI for like a year or two, at this rate people are going to forget their potty training in another few months.
> Substance is what matters
Substance matters, which is not great for LLMs, because they put out text that has far more fluff than substance. What, however, is far worse for LLMs, is the fact that kick and scream and cry all you want, but: style and presentation matters, too.
It is absolutely true that if you just dropped a very brief blurb that all AMD CPUs from a certain generation can be pwned it would have a decent chance to hit the HN frontpage just out of sheer interestingness. That is not because the style and presentation doesn't matter, it's just that the substance is significant in spite of the bad presentation and style.
And absolutely, we can easily forgive someone for simply not being very good at the presentation and style part, certainly I'm not really an expert at it. But this author has released plenty of great hits before, so I damn well know they can. It's a serious disappointment to see them downgrade to irritating, grating Claude garbage output.
https://jxself.org/titanic.shtml
He did it well. On "security", the author loves more to own his code/adata than anything. as did the PDP10/ITS hackers.
> Run `platform_check` first and do not use `SKITTER_FORCE=1` casually. Start with the read-only `dram_state` and `dram_carveouts`, then `dram_dump --dry-run`. Avoid `dram_poke` until maps have been freshly collected and calibrated. Do not bypass fingerprint checks, calibration, fencing, or verification.
Claude's (apparently externally-mandated?) lobotomization continues to be concerning. :-/
He should also be able to fuse away this access forever, to be fair. But out of the box, when I get a new laptop, I should be able to read and write every byte of DRAM.
Odd, he had an American accent.
-- Alice in Wonderland
The hardware DRAM controller maps "physical addresses" approximately to: {DRAM slot number, chip number in slot, bank number in chip, row number in bank, byte number in row} via a complex map for various irrelevant reasons. All permission checks are before this mapping. So if you change the mapping, you can access shit you should not be able to, like TPM and SMM memory. OP found a way to change the mapping.