A question about staged implementations?

[ethindp]

So, I really want to use this for my ACPI library, it looks great! However, I curious how stage-based implementation might work. That is, if I want to pull in the library and only implement parts at a time (for instance, maybe I don't have a scheduler, or no concept of threads). What's the best strategy for handling cases like this one? Is it required that an OS have all of this functionality before implementing uACPI?

[d-tatianin]

Hey there and thanks for the feedback, I'm glad you like it!

As for your question: you can stub out a lot of the helpers and it will work just fine! There's a UACPI_STAUTS_UNIMPLEMENTED error code for such cases.

Mutexes might be required though, but you can totally "fake" them if you don't have a scheduler nor any sort of concurrency.
Example:

uacpi_handle uacpi_kernel_create_mutex(void)
{
    // any non-null value should work
    return (uacpi_handle)1;
}

uacpi_bool uacpi_kernel_acquire_mutex(uacpi_handle, uacpi_u16)
{
     return true;
}

void uacpi_kernel_release_mutex(uacpi_handle)
{
}

Hope this answers your question.

[ethindp]

Mutexes and such I believe I have implemented via my languages stdlib, it's just threads I'm concerned about. Was thinking I could make threads as processor-based for now, but I need to flesh out that implementation strategy (was thinking of a virtual ISA of sorts... Might get complex though lol). Thanks so much, I love this library -- so, so much simpler than ACPICA! :)