Binutils 2.41 now supports efi-app-loongarch64, so to use this we must add a
dummy reloc section and remove the PE header definition from crt0 as this will
conflict with the version added by objcopy
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Sometimes ld emits relocs such as R_RISCV_64 for unwind symbols
these don't need to be handled yet so just can be skipped otherwise
the binary will never load
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
- Make it actually work
- Split text/data/reloc/rodata sections
- Move hash sections to past end of image
(un-needed in PE)
- Correct section & file alignment
- Correct image size
Builds correctly but untested
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
- Make it actually work
- Split text/data/reloc/rodata sections
- Move hash sections to past end of image
(un-needed in PE)
- Correct section & file alignment
- Correct image size
- rename .reloc to .areloc within ELF
as only on ARM32 does this name cause
it to be marked as REL and subsequently
mistakenly added to RELSZ
Builds correctly & tested working in QEMU
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
There is no small data section on ia32,
this ends up being created as ld puts .got/.got.plt
here due to first declaration
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Also add EXEC_PREFIX and INCLUDEDIR variables
This was requested by fwupd-efi, to be able
to retrieve the version number of gnu-efi
so we can detect features
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
This makes setup and teardown functions defined with
__attribute__((__constructor__) and __attribute__((__destructor__)) work
in normal circumstances in EFI binaries.
A couple of notes:
- it implements both the old-style .ctors/.dtors methods and the newer
style .init_array/.fini_array ELF constructor and destructor arrays,
processed in the order:
.init_array[]
.ctors[]
efi_main()
.dtors[]
.fini_array[]
- Destructors will only be called if efi_main() exits using "return";
any call to Exit() will still longjmp() past them.
- InitializeLib() has already been called before constructors run, so
they don't need to call it (and neither does anything else.) For
compatibility, it has been altered so calling it more than once is
safe.
- No attempt is made to handle any constructor or destructor with a
prototype other than "void func(void);", but note that InitializeLib
has been called, so LibImageHandle, ST, BS, and RT are set.
- The init_array/ctor/dtor/fini_array lists aren't the using the GNU
"CONSTRUCTOR" output section command, so they don't start with a size.
- The lists are individually sorted during the link stage via
SORT_BY_NAME() in the linker script.
- The default (empty) init_array/ctor/dtor/fini_array lists are padded
out to 8-byte alignment with ".p2align 3, 0", and each list always has
at least one ".long 0" at the end of it (even if it's completely
empty). As a result, they can have NULLs that need to be skipped.
The sections they're in are mergeable, so the NULLs don't have to be
exclusively at the end.
- The ia64 and mips64el arches have not been tested.
Signed-off-by: Peter Jones <pjones@redhat.com>
The whole gnu-efi project is licensed under BSD license, see [1].
However some of the RISC-V codes have conflict license identifiers:
- Some mention GPL-2.0+ in the SPDX license part, but the long license
header indicates it's actually BSD and GPL-2.0+ dual-licensed
- Some mention GPL-2.0+ in the SPDX license part only
- Some do not have any license indication
To have a matching license with the whole project, this commit updates
all RISC-V codes to have the correct SPDX license identifiers
(GPL-2.0+ or BSD dual-license).
Link: https://sourceforge.net/p/gnu-efi/ [1]
Signed-off-by: Bin Meng <bmeng@tinylab.org>
Acked-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When using function/data sections option from
gcc, some additional un-packed sections get created
move these to there correct placement if not already
included
sections seen on x86/x86_64/aarch64:
.gcc_except_table.efi_main
.bss.debugging
.gcc_except_table.* only appears on objcopy archs
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Objcopy sometimes generates .rela.plt and .rela.data
sections for which their names are longer than the 8
characters limit of PE32
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Date: Tue Feb 7 21:48:27 2023 +0000
ARM32/x86_64 lds: Force data segment align to 4KiB
Some newer versions of binutils can set invalid COMMONPAGESIZE
and MAXPAGESIZE values on these architectures
git fetch git.code.sf.net/u/gmbr3/gnu-efi arm-pagesize
git merge 89d55b364d
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Some newer versions of binutils can set invalid COMMONPAGESIZE
and MAXPAGESIZE values on these architectures
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Binutils 2.38 now supports efi-app-aarch64 so
to use this we must add a dummy reloc section
and remove the PE header definition from crt0
as this will conflict with the version added
by objcopy
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Use ld commands DATA_SEGMENT_ALIGN and DATA_SEGMENT_END
Fixes binutils 2.39 warning:
/usr/bin/ld: warning: efi/fwup.so has a LOAD segment with RWX
permissions
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
The current ld script places .dynamic section right after .text, which
results in that one LOAD segment contains both data and code. Binutils
2.39 and later will throw a warning if RWX segments exist. Code and
data should be placed in different LOAD segments with different
permissions. This patch adds a gap between data and code sections which
is taken from the default ld script in riscv64.
Co-developed-by: Moody Liu <mooodyhunter@outlook.com>
Signed-off-by: Moody Liu <mooodyhunter@outlook.com>
Signed-off-by: Letu Ren <fantasquex@gmail.com>
Acoording to what the spec says about the number of sybols [1]:
"This value should be zero for an image because COFF debugging
information is deprecated."
Changing as if not zero it causes problems to llvm-objcopy. This
affects only the architectures where COFF is not supported by objcopy
and where we build the PE header via assembly code.
[1] https://docs.microsoft.com/en-us/windows/win32/debug/pe-format
As per the PE format specification, relocation blocks must be aligned on 32-bit boundaries. Fix the dummy relocations to obey the alignment constraints.
because of a "make" builtin rule. This breaks make -r, and it is an
inefficient way to build, at least for multiple files.
I have made a git tree which includes a patch for this problem,
and also forces -r by adding it to MAKEFLAGS:
git://git.zytor.com/users/hpa/gnu-efi.git make-r-fixes
Signed-off-by: H. Peter Anvin <hpa@users.sourceforge.net>
Signed-off-by: Nigel Croxon <ncroxon@redhat.com>
Gentoo has slightly modified linker defaults: --hash-style=gnu
This means all ELF files in system have '.gnu.hash' section
but no '.hash' section.
gnuefi's ldscript did not account for it and as a result
one symbol 'ImageBase' did not resolve locally for elilo.so
and caused 'elilo' to fail to load by ia64 EFI:
Loading.: Gentoo (try new elilo)
ImageAddress: pointer is outside of image
ImageAddress: pointer is outside of image
Those two relocations come from crt0-efi-ia64.S PE32 entry point
fdescr:
```
#define IMAGE_REL_BASED_DIR64<->10
.section .reloc, "a"
data4 _start_plabel // Page RVA
data4 12 // Block Size (2*4+2*2)
data2 (IMAGE_REL_BASED_DIR64<<12) + 0 // reloc for plabel's entry point
data2 (IMAGE_REL_BASED_DIR64<<12) + 8 // reloc for plabel's global pointer
```
These refer ImageBase.
The change adds '.gnu.hash' collection (follows existing '.hash'
collection).
Tested on IA-64 by successfully booting elilo-3.16.
Bug: https://bugs.gentoo.org/575300
Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
