typical build will now include 880 bytes of PCI support code, compared to
2327 bytes in Etherboot 5.4.

(There is a slight cost of around 5 extra bytes per access to a
non-constant config space address; this should be an overall win.
Driver-specific accesses will usually be to constant addresses, for
which there is no additional cost.)

Generic PCI code now handles 64-bit BARs correctly when setting
"membase"; drivers should need to call pci_bar_start() only if they want
to use BARs other than the first memory or I/O BAR.

Split rarely-used PCI functions out into pciextra.c.

Core PCI code is now 662 bytes (down from 1308 bytes in Etherboot 5.4).
284 bytes of this saving comes from the pci/pciextra split.

Cosmetic changes to lots of drivers (e.g. vendor_id->vendor in order to
match the names used in Linux).

for use by malloc().

This breaks the image-loading code (which previously used the heap to
allocate the buffer for downloading the image), but that's not a major
concern since I'm going to tear out all the image formats within the next
couple of days anyway.  Byebye, NBI!  :)

linear calls to load the MS-DOS boot sector in my test setup.

don't conform to BBS (e.g. the one created by mkdosfs).

Shaved around 100 bytes off vsprintf.o.  It's now 50 bytes smaller than
the old implementation and provides much more conformant semantics,
including the ability to return the number of characters that would have
been printed to the string had the buffer been big enough.  (iSCSI needs
this functionality).

stack, even if we don't yet print it out.

At some point, vsprintf() needs to be fixed up so that it can correctly
cope with limited-sized buffers (i.e. vsnprintf), long longs, and
standard format specifiers (e.g. "%04x").  We should also remove the
special types (MAC addresses and IP addresses).  This would then enable
us to use gcc's ability to type-check printf format strings.

interrupts hooked.

devices).

implementation.

Moved to net/

Add method for hooking real-mode interrupt vectors.

Use .text16.data section with "aw" attributes, to avoid section type
conflicts when placing both code and data into .text16.

Add __from_{text16,data16}.

Avoid optimisation errors under -fstrict-aliasing.

between the low half stored in the static variable rm_sp, and the high
half stored on the prot_call() stack, because:

  Just using the stack would screw up when a prot_call()ed routine
  executes a real_call(); it would have no way to find the current top of
  the RM stack.

  Extending rm_sp to rm_esp would not be safe, because the guarantee that
  rm_sp must return to the correct value by the time an external
  real-mode call returns applies only to %sp, not to %esp.

interrupt vectors, for example, will be easiest to handle if placed in
.text16.

from protected-mode code.

Set up %ds to point to .data16 in prot_to_real, so that code specified
via REAL_EXEC() and friends can access variables in .data16.

Move most real-mode librm variables from .text16 to .data16.

.data16.  librm will need to supply "char *data16", i.e. the virtual
address of the start of .data16.

portions such as the stack) during the installation process.

init_fns() are called from main().  We need to think about initial
control flow some more.

install_prealloc.  I *think* these will suffice for all the 16-bit
prefixes.

However, it's worth adding a .size directive so that objdump reports the
correct sizes for the code fragments.