Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The GPU, not the TPM, is the root of hardware DRM

mjg59.dreamwidth.org

Post reply on HN
Print view View original

351–360 of 493 posts

Re: The GPU, not the TPM, is the root of hardware DRM

#351

Post by AnthonyMouse »

devops99 wrote: post #329
AnthonyMouse wrote: post #127

Earlier quoted context omitted.

> The vast majority of users aren't going to have their laptop stolen by the CIA/NSA and have their DIMMs popped and cryofreezed. That's kind of the point. The vast majority of users aren't going to have their laptop stolen at all, if they do it will 99% of the time be by someone who only wants to wipe it and fence it, and attempts to access data are most likely to be by unsophisticated family members who would be de…

> vs. FDE with a boot key stored in some cloud service secured with the user's password instead of a TPM Without secure boot (backed by TPM), I can boot a small USB device that has LEDs on it to indicate to me that the target system has been infected to send me a copy of the target's password, after I already imaged the disk (or when I have another team member steal it or take it by force later). If there's a UEFI pa…

> Without secure boot (backed by TPM), I can boot a small USB device that has LEDs on it to indicate to me that the target system has been infected to send me a copy of the target's password, after I already imaged the disk (or when I have another team member steal it or take it by force later).

Which is the same thing that happens with secure boot, because they just steal the whole device and leave you one that looks the same to enter your password into so it will send it to them.

Meanwhile if you're using tamper-evident materials then you don't need secure boot, because then they can't undetectably remove the cover to get physical access to remove your UEFI password or image the machine.

Re: The GPU, not the TPM, is the root of hardware DRM

#352

Post by advael »

rustcleaner wrote: post #342
braiamp wrote: post #323

Earlier quoted context omitted.

Why not just prohibiting the practice? This isn't weed or alcohol.

Still allow for the multimillion dollar industrial dozen-megamachine makers.

So the idea is to ban the practice for smaller players without the scale to eat the costs?

No thanks, an outright ban is necessary. This will not prevent manufacturers from doing business no matter how they may whine about it, and frankly if this does somehow kill their business it should

Re: The GPU, not the TPM, is the root of hardware DRM

#353

Post by advael »

rustcleaner wrote: post #305
no_time wrote: post #143

The author is correct in that media DRM is tied to GPU vendors on the field right now. But hardware backed DRM can be so much more invasive beyond that. I have no doubts the long term goal of MS is to have a Windows version of Play Integrity.[0] So total control over everything that happens on your device. Just to give an example of what could happen if this becomes reality: https://en.m.wikipedia.org/wiki/Web_Enviro…

I always said a hefty sales tax (50%? 100%? 200%?) on final sale of any product containing just a single Universal Machine which has artificial designs/locks that prevent the owner from replacing any and all firmware/software with versions he has authored, and/or which lacks complete enough documentation of design and interfaces that would enable a knowledgable and capable owner to author his own software/firmware. T…

A "tax" like this is essentially equivalent to a fine, and a fine is a price

Also, companies can just price the additional cost in, blame the government for the price increase, and mislead consumers about the tradeoff being made. A ban is harder to do that about

Re: The GPU, not the TPM, is the root of hardware DRM

#354

Post by slt2021 »

stonogo wrote: post #320
TacticalCoder wrote: post #301

Earlier quoted context omitted.

> the future for personal computing is looking grim I don't know. They could lock up the hardware stack as much as they want, in the end it's pixels being pushed to arrays. It's extremely hard to prevent these pixels from being intercepted. You'll have pirate groups just going deep in the hardware (opening the monitors and soldering and hacking and whatnots) and eventually tap these. As for personal usage: I've got h…

The issue isn't preventing piracy, it is defending GPU market segmentation. In the old days you could flash Quadro firmware to Geforce cards and unlock features or modify clocks. The common thread is artificial scarcity.

it is price discrimination. How to sell the same GPU hardware at different prices based on consumer's wallet:

1. cheaper price for gamers only for games

2. maximum price for crypto/AI bros

Re: The GPU, not the TPM, is the root of hardware DRM

#355

Post by lksaar »

Retr0id wrote: post #65
lksaar wrote: post #54

Earlier quoted context omitted.

WV L1 Keys/ PR SL 3000 keys require breaking into the TEE to steal those decryption keys. Ever wondered why netflix 4k web-dls take a while for less popular shows? Netfliy monitors these more tightly apparently and blacklist keys that are used to download. Then the group needs to buy some new device, the old one is burned.

It's true that known-compromised keys get revoked, but it's possible to avoid them knowing you've compromised a particular device.

Interesting, I hadn't heard about that. But this knowledge is obscure by design I suppose.

Re: The GPU, not the TPM, is the root of hardware DRM

#356

Post by mjg59 »

_yb2s wrote: post #281
Mindwipe wrote: post #146

Earlier quoted context omitted.

> The GPU still has no unique hardware private key, right? GPU's have had unique hardware private keys and secure memory for a decade.

How does the remote streaming server know a key is an authentic hardware GPU that hasn't been compromised, and not something you just generated in software, to enable software level decryption of the media? It seems like you'd need some central SSL like certificate authority to verify and revoke credentials that were universally implemented in the same way by all GPU manufacturers.... surely there is no such thing?

There doesn't need to be a central CA, you just need to establish trust with the DRM vendor. The GPU vendors coordinate with Microsoft to make Playready work, Android devices have certs that can be validated by Google for Widevine, Apple just does their own thing.

Re: The GPU, not the TPM, is the root of hardware DRM

#357

Post by NotPractical »

MBCook wrote: post #126

I fully get the DRM hate. Now I don’t really follow the Windows world but I thought the goal of the newer TPM stuff was to be able to provide a trusted boot chain the way Apple does. I’m under the impression that some of the earlier versions allowed the TPM module to be a separate piece of hardware from the CPU and thus exposed an hardware attack path where someone could snoop or man in the middle. If you have a full…

> to provide a trusted boot chain the way Apple does

Your flaw is assuming that Apple's only doing that for your security and has no ulterior motives. But iOS apps are disabled and Netflix reduces to a lower resolution when you disable System Integrity Protection on a Mac (among other things?). The trusted boot chain is clearly a DRM enforcement tool in addition to being a security feature.

https://github.com/cormiertyshawn895/RecordingIndicatorUtili...

Re: The GPU, not the TPM, is the root of hardware DRM

#358

Post by mjg59 »

cryptonector wrote: post #231
mjg59 wrote: post #139

Earlier quoted context omitted.

How would the GPU verify it's speaking to a real TPM? You'd need to bake the full set of legitimate EK cert CAs into it somehow (charitably let's say that's a signed blob that the driver pushes in at startup), but that's still going to be a terrible user experience because you won't get media playback if your machine has a TPM that's too new or from too niche a vendor.

> How would the GPU verify it's speaking to a real TPM? Option 1: as I said, the GPU could have its own, and yes in that case the EK cert would be known to the GPU (or it could have a platform-like cert issued by the GPU OEM). Option 2: the platform vendor can teach the GPU the EK cert (or the public key for some primary key anyways). Option 3: the GPU could learn it on first use. > charitably let's say that's a sign…

> What do you mean "too new"? Like, you replaced your TPM? That's a thing on servers, but not laptops.

I buy a GPU in 2025. I buy a new motherboard in 2026 and plug the GPU into it. How does the GPU learn about the new EK CA? These are devices that can be moved between systems, you can't delegate this to the platform vendor or TOFU, the GPU would need to generate independent trust in the TPM.

Re: The GPU, not the TPM, is the root of hardware DRM

#359

Post by hex4def6 »

TacticalCoder wrote: post #301
no_time wrote: post #143

The author is correct in that media DRM is tied to GPU vendors on the field right now. But hardware backed DRM can be so much more invasive beyond that. I have no doubts the long term goal of MS is to have a Windows version of Play Integrity.[0] So total control over everything that happens on your device. Just to give an example of what could happen if this becomes reality: https://en.m.wikipedia.org/wiki/Web_Enviro…

> the future for personal computing is looking grim I don't know. They could lock up the hardware stack as much as they want, in the end it's pixels being pushed to arrays. It's extremely hard to prevent these pixels from being intercepted. You'll have pirate groups just going deep in the hardware (opening the monitors and soldering and hacking and whatnots) and eventually tap these. As for personal usage: I've got h…

I'm not so optimistic.

Yes, you can never "plug the analog hole" completely, but you can definitely lock stuff down to the point it's impractical for 95% of people.

For instance, imagine some sort of audio / video fingerprint system that resides in Intel and/or nVidia's GPU drivers. Content gets played through the on-GPU HEVC / h.264 decoders already. Doesn't seem like a huge stretch to add a fingerprint authentication system to that stage.

Have a list of content IDs that are protected, and require a valid license to play.

Yes, your source file is unprotected (video camera in front of monitor), but all of your devices are unable to play it. Yes, your ancient, circa 2024 desktop PC will still play it, but your new 2030 model TV implements this fingerprint system as well so you can't just cast this file to your 100" display in your living room.

This is to say nothing of other forms of content (applications / games / web pages) that actually could require attestation / DRM HW / always-on internet to run.

Re: The GPU, not the TPM, is the root of hardware DRM

#360

Post by mjg59 »

sidewndr46 wrote: post #204
p_ing wrote: post #200

Earlier quoted context omitted.

I'm not groking what you're saying. Replace what "entire device"?

You're at an industry conference. I want the data on your laptop's hard drive. You leave your laptop in the hotel room. Which one is easier: 1. Go into your room and screw around with the boot loader to somehow give me unencrypted access to your laptop after you login next time. 2. Go into your room. Take your laptop. Put an identical looking laptop in place that runs software that boots and looks identical. Have it…

Joanna Rutkowska described a way to avoid this back in 2011 (https://theinvisiblethings.blogspot.com/2011/09/anti-evil-ma...), I extended it to be more usable in 2015 (https://mjg59.dreamwidth.org/35742.html). Both solutions make use of a TPM.
Post reply on HN