In certain cases (e.g. toolchain changes) the size of the created EFI
blob can end up triggering a rounding down which breaks the build.
Example:
  debian/util/geniso src/bin-combined/ipxe.iso src/bin-i386-pcbios/ipxe.lkrn \
    src/bin-x86_64-efi/ipxe.efi
  ...
  ++ stat -c %s src/bin-x86_64-efi/ipxe.efi
  + blocks=896
  + mkfs.msdos -C src/bin-combined/iso.dir.uayAjK/efi.img 896
  + mmd -i src/bin-combined/iso.dir.uayAjK/efi.img ::efi
  + mmd -i src/bin-combined/iso.dir.uayAjK/efi.img ::efi/boot
  + mcopy -o -i src/bin-combined/iso.dir.uayAjK/efi.img \
      src/bin-x86_64-efi/ipxe.efi ::efi/boot/bootx64.efi
  Disk full

This is due to:
  echo $(($(stat -c %s "$EFI") / 1024))
  872
  echo $(($(stat -c %s "$EFI") / 1024 + 55))
  927
  echo $(((($(stat -c %s "$EFI") / 1024 + 55)/32*32)))
  896 <- Rounds DOWN

This is due to size+55 => 927 being "just below" the next 32 iteration
which would be at 928. In many cases that works fine (I don't know how
the +55 was invented, I guess that is for headers). But in some cases
it caps too much eventualily ending up in a too small image file.

Rounding up (in bash) usually is "N+1/divisor". Therefore if add +1 before
the rounding for 32 we'd avoid the issue.

I compared the resulting ISO size with former (Ubuntu) builds and it was
still well within the range of past builds (869,947,948 - now 884).

Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>

Patch-Name: util-elf2efi-GNU_SOURCE.patch

Conflicts:
	src/drivers/infiniband/linda.c

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

The Option::ROM module recognizes and checks EFI header of image.  The
disrom.pl utility dumps this header if is present.

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

The Option::ROM module now compares the Code Type in the PCIR header
to 0x00 (PC-AT) in order to check the presence of other header types
(PnP, UNDI, iPXE, etc).  The validity of these headers are checked not
only by offset, but by range and signature checks also.  The image
checksum and initial size also depends on Code Type.

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

GCC 9 warns that abs() may truncate its signed long argument.  Fix by
using labs() instead.

Reported-by: Martin Liška <mliska@suse.cz>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

Fix strcmp() and strncmp() to return proper standard positive/negative
values for unequal strings.  Current implementation is backwards
(i.e. the functions are returning negative when should be positive and
vice-versa).

Currently all consumers of these functions only check the return value
for ==0 or !=0 and so we can safely change the implementation without
breaking things.

Signed-off-by: Aaron Young <Aaron.Young@oracle.com>
Modified-by: Michael Brown <mcb30@ipxe.org>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

Current (simplified):

1. InstallMultipleProtocolInterfaces
       if err goto err_install_protocol_interface;
2. OpenProtocol(efi_nii_protocol_guid)
       if err goto err_open_nii;
3. OpenProtocol(efi_nii31_protocol_guid)
       if err goto err_open_nii31;
4. efi_child_add
       if err goto err_efi_child_add;
...
err_efi_child_add:
   CloseProtocol(efi_nii_protocol_guid) <= should be efi_nii31_protocol_guid
err_open_nii: <= should be err_open_nii31
   CloseProtocol(efi_nii31_protocol_guid) <= should be efi_nii_protocol_guid
err_open_nii31: <= should be err_open_nii
   UninstallMultipleProtocolInterfaces

Signed-off-by: Ignat Korchagin <ignat@cloudflare.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

PCI Configuration Space contains fields prog-if at the offset 0x09,
sub-class at the offset 0x0a and base-class at the offset 0x0b (it
respects little endian).  PCIR structure uses these fields in the same
order.

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

Starting from binutils 2.31.0 (commit bd7ab16b) x86-64 assembler
generates R_X86_64_PLT32 instead of R_X86_64_PC32.

Acked-by: John Jolly <jjolly@suse.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

The first adapters in this family are X2522-10, X2522-25, X2541 and
X2542.

These no longer use PCI BAR 0 for I/O, but use that for memory.  In
other words, BAR 2 on SFN8xxx adapters now becomes BAR 0.

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

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

Devices that support jumbo frames will currently default to the
largest possible MTU.  This assumption is valid for virtual adapters
such as virtio-net, where the MTU must have been configured by a
system administrator, but is unsafe in the general case of a physical
adapter.

Default to the standard Ethernet MTU, unless explicitly overridden
either by the driver or via the ${netX/mtu} setting.

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

Avoid calling rndis_halt() and rndis->op->close() twice if the call to
register_netdev() fails.

Reported-by: Roman Kagan <rkagan@virtuozzo.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

Some versions of gcc seem to silently accept an attempt to disable an
unrecognised warning (e.g. via -Wno-stringop-truncation) but will then
report the unrecognised warning if any other error occurs during the
build, resulting in a potentially misleading error message.

Avoid this potential confusion by using the positive-form tests in
order to determine the workaround CFLAGS.

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

The definition of version_response channel message in Linux doesn't
include version field, so the upcoming VMBus implementation in QEMU
doesn't set it either.  Neither Windows nor Linux had any problem with
this.

The check against this field is redundant because the message is the
response to initiate_contact message containing the specific version
requested, so the response with version_supported=true is unambiguous.

Drop this check and don't rely on the field to be present in the
message.

Signed-off-by: Roman Kagan <rkagan@virtuozzo.com>
Modified-by: Michael Brown <mcb30@ipxe.org>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

register_netdev expects ->hw_addr and ->ll_addr to be already filled,
so move it towards the end of register_rndis, after the respective
fields have been successfully queried from the underlying device.

Signed-off-by: Roman Kagan <rkagan@virtuozzo.com>
Modified-by: Michael Brown <mcb30@ipxe.org>
Signed-off-by: Michael Brown <mcb30@ipxe.org>

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