Add more resource types for parsing _CRS packages

[michaelmelanson]

This PR works toward #51 by adding parsers for several descriptor types found in _CRS packages. I basically ran through all the devices that QEMU provided on my machine, and kept adding parsers until it stopped hitting unimplemented!() blocks. This is what I got. In my OS it parses all the Buffer type CRS entries provided by QEMU.

With this, you can parse many Current Resource Settings blocks with code like this:

        if let Ok((name, handle)) = context.namespace.search(&AmlName::from_str("_CRS").unwrap(), &device) {
          let value = context.namespace.get(handle).unwrap();
          let crs = aml::resource::resource_descriptor_list(&value);
          println!("{}: {:X?}", name, crs);
        }

I checked and this appears to merge cleanly against #75 with only minor changes, so it shouldn't be any trouble to merge this either before or after v2.

[michaelmelanson]

@IsaacWoods I'm not quite done yet, but when you get a chance would you mind taking a look at this and letting me know if I'm on the right track here.

[IsaacWoods]

Great stuff, thanks for working on this :)

I think the broad strokes of this are on the right tracks. I've left some review comments from my initial reading, but I'm afraid I won't have time this week to sit down and get my head around resource descriptors again to check this properly. I'll let you finish and try to make some time then.

[IsaacWoods]

The result looks reasonably correct but I don't know how to really validate that it parses these descriptors properly. Some advice for testing would be appreciated there.

The long-term vision for testing is to flesh out our AML test suite with a bunch of stuff. For this, we'd have a bunch of devices with _CRS objects that covered a range of resource descriptors, and their combinations. In aml_tester, we should then enumerate each device and check that its _CRS parses nicely. However, that's a fair way off yet and definitely beyond the scope of this PR.

For now, I'm happy as long as they seem to produce sensible results. You could also either dump the ACPI tables and decompile them using iasl, or look to see if what you're looking for is in this (this won't be exactly the same as your particular tables, but is usually pretty close).

Please ask if anything I've said isn't clear, and I'll try and get back when I can :)

[michaelmelanson]

Thanks for the feedback! I'll work on fixing up the things you've mentioned as soon as I get a chance.

[michaelmelanson]

Thanks for your feedback! I've addressed all your comments from before and I think it's in shape for a proper review now when you get a chance.

[IsaacWoods]

Great, this looks good now. Thanks very much!