Move most arch/i386 files to arch/x86, and adjust the contents of the
Makefiles and the include/bits/*.h headers to reflect the new
locations.

This patch makes no substantive code changes, as can be seen using a
rename-aware diff (e.g. "git show -M5").

This patch does not make the pcbios platform functional for x86_64; it
merely allows it to compile without errors.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Commit c64747d ("[librm] Speed up real-to-protected mode transition
under KVM") rounded down the .text16 segment address calculated in
alloc_basemem() to a multiple of 64 bytes in order to speed up mode
transitions under KVM.

This creates a potential discrepancy between alloc_basemem() and
free_basemem(), meaning that free_basemem() may free less memory than
was allocated by alloc_basemem().

Fix by padding the calculated sizes of both .text16 and .data16 to a
multiple of 64 bytes at build time.

Debugged-by: Yossef Efraim <yossefe@mellanox.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

PCI v3.0 supports a "device list" which allows the ROM to claim
support for multiple PCI device IDs (but only a single vendor ID).
Add support for building such ROMs by scanning the build target
element list and incorporating any device IDs into the ROM's device
list header.  For example:

  make bin/8086153a--8086153b.mrom

would build a ROM claiming support for both 8086:153a and 8086:153b.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Some decompression algorithms (e.g. LZMA) require large amounts of
temporary stack space, which may not be made available by all
prefixes.  Use .bss16 as a temporary stack for the duration of the
calls to install_block (switching back to the external stack before we
start making calls into code which might access variables in .bss16),
and allow the decompressor to define a global symbol to force a
minimum value on the size of .bss16.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Our use of --gc-sections causes the linker to discard the symbols
defined by FILE_LICENCE(), meaning that the resulting licence
determination is incomplete.

We must use the KEEP() directive in the linker script to force the
linker to not discard the licence symbols.  Using KEEP(*(COMMON))
would be undesirable, since there are some symbols in COMMON which we
may wish to discard.

Fix by placing symbols defined by PROVIDE_SYMBOL() (which is used by
FILE_LICENCE()) into a special ".provided" section, which we then mark
with KEEP().  All such symbols are zero-length, so there is no cost in
terms of the final binary size.

Since the symbols are no longer in COMMON, the linker will reject
symbols with the same name coming from multiple objects.  We therefore
append the object name to the licence symbol, to ensure that it is
unique.

Reported-by: Marin Hannache <git@mareo.fr>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

The header of a .mrom image declares its length to be only a few
kilobytes; the remainder is accessed via a sideband mechanism.  This
makes it difficult to append an additional ROM image, such as an EFI
ROM.

Add a second, dummy ROM header covering the payload portion of the
.mrom image, allowing consumers to locate any appended ROM images in
the usual way.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Some versions of binutils have curious concepts of what constitutes
subtraction.  For example:

  0x00000000000000f0 _text16_late = .
  0x0000000000000898 _mtext16 = .
  0x0000000000000898 _etext16 = .
  0x0000000000000898 _text16_late_filesz = ABSOLUTE ((_mtext16 - _text16_late))
  0x00000000000007a8 _text16_late_memsz = ABSOLUTE ((_etext16 - _text16_late))

This has interesting side-effects such as producing sizes for .bss
segments that are negative, causing the majority of addressable memory
to be zeroed out.

Fix by using the form

  ABSOLUTE ( x ) - ABSOLUTE ( y )

rather than

  ABSOLUTE ( x - y )

Reported-by: H. Peter Anvin <hpa@zytor.com>
Suggested-by: H. Peter Anvin <hpa@zytor.com>
Tested-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

Make the allocators used by malloc and linux_umalloc valgrindable.
Include valgrind headers in the codebase to avoid a build dependency
on valgrind.

Signed-off-by: Piotr Jaroszyński <p.jaroszynski@gmail.com>
Modified-by: Michael Brown <mcb30@ipxe.org>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

Some binutils versions will drag in an object to satisfy the entry
symbol; some won't.  Try to cope with this exciting variety of
behaviour by ensuring that all entry symbols are unique.

Remove the explicit inclusion of the prefix object on the linker
command line, since the entry symbol now provides all the information
needed to identify the prefix.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Commit 623469d ("[build] Eliminate unused sections at link-time")
introduced a regression in several build formats, in which the prefix
would end up being garbage-collected out of existence.  Fix by
ensuring that an entry symbol exists in each possible prefix, and is
required by the linker script.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Use -ffunction-sections, -fdata-sections, and --gc-sections to
automatically prune out any unreferenced sections.

This saves around 744 bytes (uncompressed) from the rtl8139.rom build.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Add makefiles, ld scripts and default config for linux platform for
both i386 and x86_64.

Signed-off-by: Piotr Jaroszyński <p.jaroszynski@gmail.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

Add preprocessor magic to the error definitions to enable every error
usage to be tracked.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Use INT 15,88 to find a suitable temporary decompression area, rather
than a fixed address.  This hopefully gives us a better chance of not
treading on any PMM-allocated areas, in BIOSes where PMM support
exists but tends not to give us the large blocks that we ask for.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

Add a section .text16.early which is always kept inline with the
prefix.  This will allow for some code sharing between the .prefix and
.text16 sections.

Note that the simple solution of just prepending the .prefix section
to the .text16 section will not work, because a bug in Wyse Streaming
Manager server (WLDRM13.BIN) requires us to place a dummy PXENV+ entry
point at the start of .text16.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

The .xrom prefix provides an experimental mechanism for loading ROM
images greater than 64kB in size by mapping the expansion ROM BAR in
at a hopefully-unused address.  This is unreliable, and potentially
dangerous.  In particular, there is no guarantee that any PCI bridges
between the CPU and the device will respond to accesses for the
"unused" memory region that is chosen, and it is possible that the
process of scanning for the "unused" memory region may end up issuing
reads to other PCI devices.  If this ends up trampling on a register
with read side-effects belonging to an unrelated PCI device, this may
cause undefined behaviour.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

The standard option ROM format provides a header indicating the size
of the entire ROM, which the BIOS will reserve space for, load, and
call as necessary. However, this space is strictly limited to 128k for
all ROMs. gPXE ameliorates this somewhat by reserving space for itself
in high memory and relocating the majority of its code there, but on
systems prior to PCI3 enough space must still be present to load the
ROM in the first place. Even on PCI3 systems, the BIOS often limits the
size of ROM it will load to a bit over 64kB.

These space problems can be solved by providing an artificially small
size in the ROM header: just enough to let the prefix code (at the
beginning of the ROM image) be loaded by the BIOS. To the BIOS, the
gPXE ROM will appear to be only a few kilobytes; it can then load
the rest of itself by accessing the ROM directly using the PCI
interface reserved for that task.

There are a few problems with this approach. First, gPXE needs to find
an unmapped region in memory to map the ROM so it can read from it;
this is done using the crude but effective approach of scanning high
memory (over 0xF0000000) for a sufficiently large region of all-ones
(0xFF) reads. (In x86 architecture, all-ones is returned for accesses
to memory regions that no mapped device can satisfy.) This is not
provably valid in all situations, but has worked well in practice.
More importantly, this type of ROM access can only work if the PCI ROM
BAR exists at all. NICs on physical add-in PCI cards generally must
have the BAR in order for the BIOS to be able to load their ROM, but
ISA cards and LAN-on-Motherboard cards will both fail to load gPXE
using this scheme.

Due to these uncertainties, it is recommended that .xrom only be used
when a regular .rom image is infeasible due to crowded option ROM
space. However, when it works it could allow loading gPXE images
as large as a flash chip one could find - 128kB or even higher.

Signed-off-by: Marty Connor <mdc@etherboot.org>

REQUIRE_SYMBOL() formerly used a formulation of symbol requirement
that would allow a link to succeed despite lacking a required symbol,
because it did not introduce any relocations. Fix by renaming it to
REQUEST_SYMBOL() (since the soft-requirement behavior can be useful)
and add a REQUIRE_SYMBOL() that truly requires.

Add EXPORT_SYMBOL() and IMPORT_SYMBOL() for REQUEST_SYMBOL()-like
behavior that allows one to make use of the symbol, by combining a
weak external on the symbol itself with a REQUEST_SYMBOL() of a second
symbol.

Signed-off-by: Marty Connor <mdc@etherboot.org>

The option ROM header contains a one-byte field indicating the number
of 512-byte sectors in the ROM image.  Currently it is linked to
contain the number of uncompressed sectors, with an instruction to the
compressor to correct it.  This causes link failure when the
uncompressed size of the ROM image is over 128k.

Fix by replacing the SUBx compressor fixup with an ADDx fixup that
adds the total compressed output length, scaled as requested, to an
addend stored in the field where the final length value will be
placed.  This is similar to the behavior of ELF relocations, and
ensures that an overflow error will not be generated unless the
compressed size is still too large for the field.

This also allows us to do away with the _filesz_pgh and _filesz_sect
calculations exported by the linker script.

Output tested bitwise identical to the old SUBx mechanism on hd, dsk,
lkrn, and rom prefixes, on both 32-bit and 64-bit processors.

Modified-by: Michael Brown <mcb30@etherboot.org>
Signed-off-by: Michael Brown <mcb30@etherboot.org>

elf2efi converts a suitable ELF executable (containing relocation
information, and with appropriate virtual addresses) into an EFI
executable.  It is less tightly coupled with the gPXE build process
and, in particular, does not require the use of a hand-crafted PE
image header in efiprefix.S.

elf2efi correctly handles .bss sections, which significantly reduces
the size of the gPXE EFI executable.

_filesz was incorrectly forced to be aligned up to MAX_ALIGN.  In a
non-compressed build, this would cause a build failure unless _filesz
happened to already be aligned to MAX_ALIGN.

The section name seems to have significance for some versions of
binutils.

There is no way to instruct gcc that sections such as .bss16 contain
uninitialised data; it will emit them with contents explicitly set to
zero.  We therefore have to rely on the linker script to force these
sections to become uninitialised-data sections.  We do this by marking
them as NOLOAD; this seems to be the closest semantic equivalent in the
linker script language.

However, this gets ignored by some versions of ld (including 2.17 as
shipped with Debian Etch), which mark the resulting sections with
(CONTENTS,ALLOC,LOAD,DATA).  Combined with the fact that this version of
ld seems to ignore the specified LMA for these sections, this means that
they end up overlapping other sections, and so parts of .prefix (for
example) get obliterated by .data16's bss section.

Rename the .bss sections from .section_bss to .bss.section; this seems to
cause these versions of ld to treat them as uninitialised data.

Not fully understood, but it seems that the LMA of bss sections matters
for some newer binutils builds.  Force all bss sections to have an LMA
at the end of the file, so that they don't interfere with other
sections.

The symptom was that objcopy -O binary -j .zinfo would extract the
.zinfo section from bin/xxx.tmp as a blob of the correct length, but
with zero contents.  This would then cause the [ZBIN] stage of the
build to fail.

Also explicitly state that .zinfo(.*) sections have @progbits, in case
some future assembler or linker variant decides to omit them.

Some versions of ld choke on the "AT ( _xxx_lma )" in efi.lds with an
error saying "nonconstant expression for load base".  Since these were
only explicitly setting the LMA to the address that it would have had
anyway, they can be safely omitted.

We have EFI APIs for CPU I/O, PCI I/O, timers, console I/O, user
access and user memory allocation.

EFI executables are created using the vanilla GNU toolchain, with the
EXE header handcrafted in assembly and relocations generated by a
custom efilink utility.

Reduce the number of sections within the linker script to match the
number of practical sections within the output file.

Define _section, _msection, _esection, _section_filesz, _section_memsz,
and _section_lma for each section, replacing the mixture of symbols that
previously existed.

In particular, replace _text and _end with _textdata and _etextdata, to
make it explicit within code that uses these symbols that the .text and
.data sections are always treated as a single contiguous block.

Wyse Streaming Manager server (WLDRM13.BIN) assumes that the PXENV+
entry point is at UNDI_CS:0000; apparently, somebody at Wyse has
difficulty distinguishing between the words "may" and "must"...

Add a dummy entry point at UNDI_CS:0000, which just jumps to the
correct entry point.

IBM's iSCSI Firmware Initiator checks the UNDIROMID pointer in the
!PXE structure that gets created by the UNDI loader.  We didn't
previously fill this value in.

Allow for an arbitrary number of splits of the system memory map via
INT 15,e820.

Features of the new map-mangling algorithm include:

  Supports random access to e820 map entries.

  Requires only sequential access support from the underlying e820
  map, even if our caller uses random access.

  Empty regions will always be stripped.

  Always terminates with %ebx=0, even if the underlying map terminates
  with CF=1.

  Allows for an arbitrary number of hidden regions, with underlying
  regions split into as many subregions as necessary.

Total size increase to achieve this is 193 bytes.

Define a list of N allowed memory regions, and split each underlying
e820 region into up to N subregions.  Strip resulting empty regions
out of the map, avoiding using the "return with CF set to strip last
empty region" trick, because it seems that bootmgr.exe in Win2k8 gets
upset if the memory map is terminated with CF set.

This is an intermediate checkin that defines a single allowed memory
region covering the entire 64-bit address space, and uses the existing
map-mangling code on top of the new region-splitting code.  This
sanitises the memory map to the point that Win2k8 is able to boot even
on a system that defines a final zero-length region at the 4GB mark.

I'm checking this in because it may be useful for future debugging
efforts to be able to run with the existing and known-working map
mangling code together with the map sanitisation capabilities of the
new map mangling code.

ROM initialisation vector now attempts to allocate a 2MB block using
PMM.  If successful, it copies the ROM image to this block, then
shrinks the ROM image to allow for more option ROMs.  If unsuccessful,
it leaves the ROM as-is.

ROM BEV now attempts to return to the BIOS, resorting to INT 18 only
if the BIOS stack has been corrupted.

_textdata_link_addr, _load_addr and _max_align in the linker scripts.
A bug in some versions of ld causes segfaults if the DEFINED() macro
is used in a linker script *and* the -Map option to ld is present.

We don't currently need to override any of these values; if we need to
do so in future then the solution will probably be to always specify
the values on the ld command line, and have the linker script not
define them at all.

compressor fixup section.

start of the non-prefix blocks in the loaded image, and rely on the
image ordering.  This should make introducing compression much easier.

image file.

causing the serial console to ignore input, because it happened to end up
linked with serial_ischar() at address 0, which core/console.c decided was
invalid).

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

--nulltrap to the make target at the moment.)

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.

thinks they are.
Other bus drivers modified for consistency.