`Descriptor::is_table_or_page` incorrectly returns `false` if underlying descriptor is in L1 and has NS bit (63) set

[fmckeogh]

Attributes::from_bits returns None if any unknown bits are set. The current Attributes structure is only valid for stage 2 translations, for stage 1 bit 63 is NSTable and may be set, resulting in None being returned.

This alone is likely fine, but the logic of is_table_or_page returns false if self.flags() is None (which should possibly be an unwrap?), which causes a L1 entry that has both the lowest 2 bits set (so is a valid table), but also has any high bits also set to incorrectly return false.

[qwandor]

Adding the missing NSTable bit to Attributes is easy enough to do (along with PXNTable, XNTable and APTable).

What's less clear to me is why Descriptor::flags returns an Option. It looks like this was last touched in #22, before which it returned None for descriptors without the valid bit set. I think it probably makes sense for Descriptor::flags just to return Attributes. I'll send a PR to change that.

[fmckeogh]

@qwandor Awesome, thanks

Given that not all flags are available in different contexts, do you think multiple Attributes types would be appropriate?

Edit: disregard, just saw #59

[qwandor]

Yeah, I was thinking something similar, as you saw in #59. I haven't found a good API for this yet though, and got busy with other things.