Lennart Poettering
Lennart Poettering
When you pass it to syscalls you specify the size in a separate parameter, and you want to specify only the size you actually make use of. See our SOCKADDR_UN_LEN()...
> I'm still trying to figure out exactly how I want to make it, and exactly what to name it, but I think an extension library that augments musl with...
i'd use the new-style initrd loading by the linux kernel as template (aka LoadFile2), see src/boot/efi/initrd.c
@bluca you might want to follow here.
> `some-kernel` would be loaded via `LoadImage` and checked over the current security policy, i.e. SecureBoot. Then `some-initrd` would be or won't be (depend on what is the security policy...
i guess we are all technically on the same page. just be careful with naming stuff right. i.e. to comply with @bluca's request if you do sd-stub based kernels but...
maybe call them "type 1.5"? ;-)
did some reviewing. could you squash your commits so that individual commits are logical steps and individually "perfect" rather than reflecting some historical progress? it makes it a ton easier...
i don't grok this. This boolean field should be something we serialize, no? i.e. with your patch if the timer is supposed to trigger eventually, and you do a daemon...
> I'm not sure about the point of this test... but it seems to me that we always want to rebase such timer on `inactive_exit_timestamp`, no ? `inactive_exit_timestamp` is filled...