on Chromium they should state. its a combo of GPU and the app failing to isolate cross-domain data… leaking it.
Firefox is not vulnerable… just chrome/edge, etc.
Too bad this is not included in the title (or subtitle) !
Not my pixels!
MINE
Our
deleted by creator
the pixel is the just the base unit… expand the exploit and you get ‘images’. any image on the remote site… and from there you could target sites that use imaging for password/username stuff (as a method of preventing text-based exploits).
the one pixel leads to lots of nonsense
its a teeny tiny hole, but thats all you need
That and apparently a lot of time. Am I right in reading it could take hours to leak enough pixels to form an image? So to get a password the password would need to be plain text, visible on the target website, and not be moved, removed or otherwise changed for hours.
deleted by creator
For GPU.zip to work, a malicious page must be loaded into the Chrome or Edge browsers. Under-the-hood differences in the way Firefox and Safari work prevent the attack from succeeding when those browsers process an attack page.
Lol, amazing.
Firefox ftw!!!
This is the best summary I could come up with:
The researchers found that data compression that both internal and discrete GPUs use to improve performance acts as a side channel that they can abuse to bypass the restriction and steal pixels one by one.
“We found that modern GPUs automatically try to compress this visual data, without any application involvement,” Yingchen Wang, the lead author and a researcher at the University of Texas at Austin, wrote in an email.
Most websites restrict the cross-origin embedding of pages displaying user names, passwords, or other sensitive content through X-Frame-Options or Content-Security-Policy headers.
All of the GPUs analyzed use proprietary forms of compression to optimize the bandwidth available in the memory data bus of the PC, phone, or other device displaying the targeted content.
The insights yielded a method that uses the SVG, or the scalable vector graphics image format, to maximize differences in DRAM traffic between black and white target pixels in the presence of compression.
Our proof-of-concept attack succeeds on a range of devices (including computers, phones) from a variety of hardware vendors with distinct GPU architectures (Intel, AMD, Apple, Nvidia).
The original article contains 832 words, the summary contains 181 words. Saved 78%. I’m a bot and I’m open source!
GPUs from all six of the major suppliers
Wait, what? Six? There’s AMD, Nvidia, and Intel. Who are the other three? Are they counting mobile chips made by Apple, Qualcomm, and Samsung as GPUs?
On top of my head there is AMD, Nvidia, Intel, ARM, Qualcomm, Broadcom, Apple. Samsung licenses their GPU’s from ARM and AMD as far as I know. Also why wouldn’t you count the other manufacturers? There are certainly more ARM IP GPU’s in use than AMD and NVIDIA and Apple is probably up there too, especially with the M1 and M2 launch.
Does VIA still make onboard GPUs?
Yeah for Zhaoxin, but that’s for the chinese market.
The attack works on GPUs provided by Apple, Intel, AMD, Qualcomm, Arm, and Nvidia.
Even new(ish) GPUs from Apple. Sounds like a flaw in the product category, not just certain implementations.
