ia32 GNUC: Use __asm__ instead of asm

Causes an error when using strict ISO modes as
then asm isn't defined

Signed-off-by: Callum Farmer <gmbr3@opensuse.org>

*/*.S: add non-executable GNU stack marking on ELF-linux

binutils-2.39 enabed a few warning by default
(https://sourceware.org/pipermail/binutils/2022-August/122246.html):

> The ELF linker will now

*/*.S: add non-executable GNU stack marking on ELF-linux

binutils-2.39 enabed a few warning by default
(https://sourceware.org/pipermail/binutils/2022-August/122246.html):

> The ELF linker will now generate a warning message if the stack is made executable.

Let's suppress the warnings in assembly files by adding non-executables
stack markings. This fixes at least systemd build which uses '-Wl,--fatal-warnings':

https://github.com/systemd/systemd/issues/24226

show more ...

Currently we have DivU64x32 on ia32, but it tries to call
__umoddi3 and __udivdi3 from libgcc, which we don't have.
This fixes it to use our implementation in that case.

Signed-off-by: Peter Jones <

Currently we have DivU64x32 on ia32, but it tries to call
__umoddi3 and __udivdi3 from libgcc, which we don't have.
This fixes it to use our implementation in that case.

Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Nigel Croxon <ncroxon@redhat.com>

show more ...

The default __GNUC__ C code cannot be used with MinGW as it inserts
implicit calls to _umoddi3 and _udivdi3, which are unavailable when
compiling without the standard libraries (-nostdlib).

This pat

The default __GNUC__ C code cannot be used with MinGW as it inserts
implicit calls to _umoddi3 and _udivdi3, which are unavailable when
compiling without the standard libraries (-nostdlib).

This patch addresses this by providing an inline assembly definition
that is an exact conversion of the existing MS one, but for GCC's
AT&T syntax.

Signed-off-by: Pete Batard <pbatard@users.sf.net>
Signed-off-by: Nigel Croxon <nigel.croxon@hpe.com>

show more ...

I also found an issue when compiling setjmp.S using MinGW for both
x86_32 and x86_64, due to the use of ELF specific assembler
extensions (.type setjmp, @function is not comptaible with PE-COFF).

Th

I also found an issue when compiling setjmp.S using MinGW for both
x86_32 and x86_64, due to the use of ELF specific assembler
extensions (.type setjmp, @function is not comptaible with PE-COFF).

This patch addresses that, by making sure that we use the equivalent
.type definition for MinGW. Note that I only addressed x86_32 and
x86_64 as I'm not aware of MinGW developments for any other archs.

Signed-off-by: Pete Batard <pbatard@users.sf.net>
Signed-off-by: Nigel Croxon <nigel.croxon@hpe.com>

show more ...

Commit 3dd05c and subsequent ones introduced the use of the GNU
specific __attribute__((__unused__)), which breaks MSVC compilation...

I'll use this as a reminder that gnu-efi is compiled with more

Commit 3dd05c and subsequent ones introduced the use of the GNU
specific __attribute__((__unused__)), which breaks MSVC compilation...

I'll use this as a reminder that gnu-efi is compiled with more than
GNU toolchains, and that contributors might want to be careful about
adding GNU-specific extensions...

This patch does the following:
Create a new eficompiler.h header that sets all compiler-specific
options and extensions
Create a new EFI_UNUSED macro to be used in lieu of
__attribute__((__unused__)) and applies it to the relevant sources
Move the ALIGN(x) macro, previously defined in efisetjmp.h to the new header

Oh, and it also fixes a bunch of trailing whitespaces (because my code
editor is set to do just that by default).

Signed-off-by: Pete Batard <pbatard@users.sf.net>
Signed-off-by: Nigel Croxon <nigel.croxon@hpe.com>

show more ...

This makes "-Werror=unused-parameter" not fail.

Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hpe.com>

Turns out we actually need setjmp in one of gnu-efi's prominent
users, and it seems to make more sense to put it here than in
the application.

All of these are derived from the Tiano code, but I re-

Turns out we actually need setjmp in one of gnu-efi's prominent
users, and it seems to make more sense to put it here than in
the application.

All of these are derived from the Tiano code, but I re-wrote the
x86_64 one because we use the ELF psABI calling conventions instead
of the MS ABI calling conventions. Which is to say you probably
shouldn't setjmp()/longjmp() between functions with EFIAPI (aka
__attribute__((ms_abi))) and those without.

Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>

show more ...

The gnu-efi-3.0 toplevel subdirectory is really annoying. Kill it.

Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>