This reverts commit 03ca71c38d.

(0x7ef4) tends to really screw up anything that loads at 0x7c00...

ISA 3c509 is currently non-functional, although the EISA (3c509-eisa) and
MCA (3c529) variants should build OK.

None of this code is yet tested.

clean separation between loading and starting, but does mean that more
PXE stacks survive the process.

do it.

allocation.

addresses if the signature checks fail.

we can never return to the PXE stack and must always use INT 18.

Record all information required for populating a struct undi_device.

Make debugging output more human-readable.

and call UNDI_SHUTDOWN first to make sure the NIC is in a safe state.

memory (unless we get an error while stopping the base code).  Leave UNDI
resident (though stopped) for .kpxe.

Still need to add code to record the device identification parameters
prior to stopping UNDI.

having to edit PXELOADER_KEEP_ALL as a config option.

Fix bug that caused over-allocation of .text16 and .data16 memory areas
by a factor of 16.

flow of control through this code.

We now always add PCI and PnP headers even for ISA ROMs; there's no harm
in doing so.

UNDI loader is still missing.

on the stack.

install() now calls relocate(), moves the protected-mode code to the new
location, and calls hide_etherboot().

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).

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.

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.

advantage of the fact that we have to have a permanently-resident block
in base memory.

highmem data, so decompress16 will be able to unpack blocks bigger
than 64kB.

running the decompresser in 16:16 protected mode using the real-mode
stack.  However, there's an even simpler way to do it...

I want to get to the point where any header in include/ reflects a
standard user-level header (e.g. a POSIX header), while everything that's
specific to gPXE lives in include/gpxe/.  Headers that reflect a Linux
header (e.g. if_ether.h) should also be in include/gpxe/, with the same
name as the Linux header and, preferably, the same names used for the
definitions.

mechanism has changed anyway, and so must be rewritten, but at least doing
this makes the output of "make symcheck" more legible.

initial device, and will also allow for e.g. a device menu to be presented
to the user.

 callback mechanism.