this post was submitted on 06 Jul 2023
106 points (99.1% liked)
Technology
59143 readers
2297 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Out of curiosity, how is your knowledge?
From a theoretical standpoint, given that it apparently has access to the same physical memory, all that would be needed is to know the correct memory addresses to read a single frame from.
How the kernel allocates that and maps it appropriately is both driver-specific and OS-specific.
Technically speaking, it could be that there are specific pools of physical addresses reserved for such things. The vram for an integrated card has at least a portion which is typically shared, so that's worth taking into account as well, because the boundaries could be device specific or standard specific.
I'd personally have to research more myself, but if we're operating off of Murphy's law, then I'd wager that, yes, it's very possible. Perhaps even if your kernel module for the camera driver is disabled by the host OS.
It's certainly possible that the IME lane access is restricted in terms of what IO will be considered valid, but given that it isn't documented, and given that it runs a multitasking, Unix-OS...that alone says a lot. Maybe not enough to be certain, but a lot.
Previously it was running an RTOS. So, you have constant time tasks. Now that it isn't running an RTOS, there has to be a reason. Maybe it's designed to aid branch prediction - if it is, though, then it probably has access to the instructions that are being executed by the host CPU. If that's the case, then that implies that there are restrictions in the page mapping mechanism with respect to what physical addresses can be virtually mapped - assuming any protection at all.
You see where I'm going with this?