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

FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

hannesweissteiner.com

Post reply on HN
Print view View original

1–10 of 17 posts

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#2

Post by Bender »

I see they are testing this on a Mac. I am curious what the test results look like if the users home directory or even the dot directories are tmpfs. On Linux .bash_login can repopulate dot directories from a archive directory think skeleton files and the dot directories can be ephemeral mounted as tmpfs. The person can have a command to commit their ephemeral directories back to the archive if they want to "keep their changes" so to speak. Or automate it on .bash_logout.

    du --max-depth 0 -h -c .cache .config .local
    767M    .cache
    278M    .config
    2.2M    .local
    1.1G    total
It's a bit of space on this CachyOS laptop but it's doable.

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#6

Post by Avamander »

Bender wrote: post #2

I see they are testing this on a Mac. I am curious what the test results look like if the users home directory or even the dot directories are tmpfs. On Linux .bash_login can repopulate dot directories from a archive directory think skeleton files and the dot directories can be ephemeral mounted as tmpfs . The person can have a command to commit their ephemeral directories back to the archive if they want to "keep th…

It's really difficult to reliably separate temporary and persistent browser storage. I tried at some point to reduce HDD noise. But given how neither Firefox or Chrome properly follow the XDG spec, it did not yield the results I wanted without a lot of handcrafted mounts.

In the end I'd guess you can also use some aspects of persistent storage to achieve similar results, even if the rest is actually tmpfs/RAM.

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#7

Post by freedomben »

As much as I love a good backronym, especially one with nested acronyms in it, it could use something self-referentially recursive, preferably with tail-recursion. This is not the solution, but something like FROSTY (Fingerprinting Remotely using OPFS-based SSD Timing with frostY)

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#8

Post by Bender »

Avamander wrote: post #6
Bender wrote: post #2

I see they are testing this on a Mac. I am curious what the test results look like if the users home directory or even the dot directories are tmpfs. On Linux .bash_login can repopulate dot directories from a archive directory think skeleton files and the dot directories can be ephemeral mounted as tmpfs . The person can have a command to commit their ephemeral directories back to the archive if they want to "keep th…

It's really difficult to reliably separate temporary and persistent browser storage. I tried at some point to reduce HDD noise. But given how neither Firefox or Chrome properly follow the XDG spec, it did not yield the results I wanted without a lot of handcrafted mounts. In the end I'd guess you can also use some aspects of persistent storage to achieve similar results, even if the rest is actually tmpfs/RAM.

Indeed. Apps do always seem to keep adding new cruft to the filesystem layout. For a while my entire home directory was tmpfs on a few machines just to stop some of the tracking. I would commit my bookmarks back to persistent storage but that was it. It was a manual process and sometimes I would forget to commit but that's just my laziness. I'm sure others would automate this process.

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#10

Post by ttctciyf »

If browsers have enough low-level access to my storage hardware to carry out timing attacks for fingerprinting, it seems likely they also have enough to maliciously chug the hardware sufficiently to degrade capacity over time and otherwise impact system integrity. I hate the thought of some random website writing and overwriting random bytes in a tight loop in the background while I'm browsing elsewhere to find the cause of my slow disk subsystem.

To that end an option to disable storage access by type would be nice to have. All I see in firefox settings is the ability to block all storage including cookies, and the ability to block persistent storage when the site requests it. It's not clear to me how the OPFS system in TFA relates to either of these, but I'd guess that it's a separate system. There's a bunch of storage quotas in about:config, but nothing obviously related to OPFS (that I can see).

Given the choice I would be happy to allow traditional cookie storage and block everything else with any exceptions I need (none that I can think of) on a per-site basis. If this can be achieved via about:config, I'm all ears!

While looking at my storage data, I see youtube has 174(!) cookies and 57M data stored on my machine. Sigh.

Post reply on HN