2 ms·
While I can understand both arguments for and against a gradual rollout, this is the main issue: why do these things need to be processed in kernel? And if ther
by vladvasiliu 2y ago
While I can understand both arguments for and against a gradual rollout, this is the main issue: why do these things need to be processed in kernel? And if there’s a good reason to do it, why isn’t there some kind of circuit breaker?
- lotr5 2y agoprobably they didn't find solution where they fully trust information coming from usermode process
- notepad0x90 2y agobecause the thing that uses them is in kernel mode, and the sensor needs to be performant. at some point, the content must be consumed by the kernel mode sensor. user mode edr's exist but bypassing them is trivial, intercepting syscalls rootkit style and monitoring kernel+usermode memory is the best and most performant way to monitor the whole system.
- calf 2y agoApple documentation argues the opposite: "Developers can use frameworks such as DriverKit and NetworkExtension to write USB and human interface drivers, endpoint security tools (like data loss prevention or other endpoint agents), and VPN and network tools, all without needing to write kexts. Third-party security agents should be used only if they take advantage of these APIs or have a robust road map to transition to them and away from kernel extensions." Specifically the 2nd sentence above says security software should use the APIs, not Apple's kernel extensions.
- notepad0x90 2y agowell, this is windows not macos. I don't know what you can do with driverkit for example. maybe microsoft should learn from apple?
- calf 2y agoYour prior argument was about sensors being performant having to reside within the kernel -- a very general argument -- of which the macos provides one counterexample in its official documentation. So the problem is in your original argument.