You are not logged in.
Hello,
recently I started playing with acpid.
read the manual briefly and noticed systemd supposedly handles acpi events.
Upon running acpid. My power button functions as it should, prior to running it the power button did not function on specific machines.
Due to acpi in system firmware being a component of hardware enumeration.
Not always happy with the dmesg results on acpi tables loading up early. I like to imagine watching your first step includes the first couple seconds of boot up. acpi is one of those things.
The bios pushes the tables which get reserved memory by the kernel I believe. That memory as far as I understand cannot be regained by the user without some serious control. I saw some black hat talks where apic registers were movable and exploited but only on one obscure cpu architecture. But also most SMM entry methods I've seen involved very specific timing attacks which ultimately is only practical with in depth knowledge of the hardware.
Most of acpi functions seem trivial. power management and the like. However, because it handles hardware enumeration, as I mention before and after reading a mention of Mark Shuttleworth who compared ACPI with Trojan horses. My thoughts might not be entirely baseless. Even if it's a slightly un-informed opinion. I am just feeling with my gut and reading dmesg outputs every now and again.
With the behavior of certain hardware vendors. Is there even a good reason to trust their code when it comes to acpi?
https://www.markshuttleworth.com/archives/1332
Here is a opinion piece from the Ubuntu dev.
What can we do on the software side of things to mitigate the impact of acpi? Setting up logging might be maybe a half hearted step. Any other reading pieces out there I can divulge in?
Last edited by czeekaj (2023-04-23 10:47:06)
Offline