This is inspired by the FreeBSD code, but nothing but the header and function
names are left of the original version.

Doesn't actually support real asynchronous execution as we're missing the concept
of waitqueues in osv as of now, but that will be added soon.

Created abstract base classes for device and drivers, changed the flow of driver loading / device probing

a tests that uses it and adds plenty of block requests to the host.

this allows the stdio code to figure out if it should go into line
buffered mode.

Bug fixing: parse_pci_config() is called from the device factory already, _have_msix isn't initialized

Longer term it should be nice not to allocate the bars at all

virt_to_phys() no longer works on static variables, so update the code to
allocate kvmclock structures from the heap.

host side reset calls

This will allow for easy addition of the vectored and positional I/O
syscalls, as well as making the structure of the I/O path more similar
to *BSD and Solaris.

Create a central place the managed the console driver (just isa-serial for
now), and exposes them as a device file and to the debug output routines.

This is the low-level console abstraction and should deal in the
char * hardware format.  The upper layer (debug.cc) already handles
all conversion from std::string and boost formats.

Destroy the drivers on OS reset (currently just call it from the loader directly) It fixes the issue of gdb breakage w/ the drivers

It's actually percpu, but we don't have percpu support yet.  So make it
global.

This is intentionally simplistic, supporting only one callback (intended
for waking up threads).

Added pci config capability parsing and moved the virtual call to earlyInitChecks() to the base class

We need only to store the virtual address returned from malloc.
virt_to_phys is the function translating the virtual mem into physical.

In addition, I founf a design error in the virtio PFN code. It will only
affect us on > 44 bit ram..

Now the network driver has 3 queues as it should have