A C library and binary for generating machine code of x86_64 assembly language and executing on the fly without invoking another compiler, assembler or linker.
APACHE-2.0 License
Bot releases are hidden (Show)
- fixed bug which caused `asmline -{c,b} <NUM>` to not display the
correct chuks
- asmline {-h,--help} now emits to stdout to be conform with GNU standards
and with a slightly different indentation.
- provide bash and zsh auto completion helper for asmline cli-tool
Published by daviduwu9 almost 2 years ago
- added support for all shift instructions shr/shl/sar/sal as well as all
their respective operand encodings
- fixed operand encoding for xchg instruction when using the return register
- fixed xabort instruction missing 8-bit immediate value
- fixed xbegin instruction rel32 offset value
- fixed a bug that caused a segmentation fault when writing to a file via the -P switch
- added four encodings for {vmovdqu, vmovupd} xmmN {xmmN,m/128} and two for {vmovdqu, vmovupd} m/128, xmmN
- added encoding for movq mv
- added support for rdpmc
Published by dderjoel about 2 years ago
- fixed a bug which caused a segmentation fault when writing to a file via the -P switch
Published by dderjoel about 2 years ago
Published by dderjoel about 2 years ago
Published by daviduwu9 about 2 years ago
- added support for multiple keywords ex: "jmp far word [rax]"
- added support for unconditional far jump
- added support for memory operand without register ex: "[0x7fffffff]"
- added support for push instruction with M and I operand encoding
- implemented thread safe design for global variables
- fixed missing SIB byte for jmp instruction
- fixed a bug which can lead to SIGSEGV's on certain systems using
`strtok_r`. (Thanks to @hugsy for pointing that out)
- fixed bug where the wrong return code is returned if the internal buffer
it too small and assemblyline is compiled for non-Linux systems
- CI: Using GitHub Actions to validate Code Formatting
- CI: Using GitHub Actions to build and run checks
- Repo: Introducing badges to show current stats for the repository
- Tests: run.c now checks for cpu-flags to run the code. (Sometimes it
fails because of missing CPU capabilities on GHActions)
Published by dderjoel over 2 years ago
- added support for neg: Two's Complement Negation
- fixed a bug where shift and rotate instruction produce the wrong opcode
when immediate is zero
- fixed a bug when using sib addressing with no base and a single operand
- asmline can now count the number of instructions that breaks a chunk boundary
- the number of instructions that break a chunk boundary can be counted when
assembling a file.
- adding `cpuid`-instruction
Published by dderjoel over 2 years ago
version 1.2.1-release (2022-02-23)
- Fixed bugs regarding `-imm/8`
Published by dderjoel over 2 years ago
version 1.2.0-release (2022-02-19)
binary file can now be generated from libassemblyline with asm_create_bin_file()
reformatted man page for wide screen
added support for all operand encodings for and
, or
, xor
, sub
, and add
instructions
added support for all setcc
instructions - set byte on condition
added a command-line option in asmline for setting assembly mode to NASM $ asmline --nasm-sib-no-base
where:
in SIB addressing if there is no base register present and scale is equal to 2;
the base register will be set to the index register and the scale will be reduced to 1.
That is: "lea r15, [2rax]" -> "lea r15, [rax+1rax]" -> 4c 8d 3c 00
added a command-line option in asmline for setting assembly mode to STRICT $ asmline --strict-sib-no-base
where:
in SIB addressing when there is no base register present the index and scale would
not change regardless of scale value.
That is: "lea r15, [2rax]" -> "lea r15, [2rax]" -> 4c 8d 3c 45 00 00 00 00
Published by dderjoel over 2 years ago
version 1.1.0-release (2022-01-21)
added support for scaled index addressing mode with syntax [base + scaleindex + offset] or [base + indexscale + offset] ex: "add rax, [rcx+4rbp+0x8]" or "add rax, [rcx+rbp4+0x8]"
added support for MMX, SSE2, AVX, and AVX2 instruction sets
added support for pointers: byte, word, dword, and qword
added support for mov [MEM], IMM
such as mov [rsi], "0xADE1A1DE"
added different modes of register size handling for "mov register, immediate" instruction: NASM, STRICT, SMART
improved 32-bit register supportability: using eax-r15b for memory addressing in vector extension now produces the correct prefix (67H)
improved intel x86_64 instruction supportability: source code is now more inline with the intel manual, therefore more instructions can be supported in the near future
added a command-line option in asmline for setting assembly mode to NASM '$ asmline -n' where: if immediate size for mov is less than or equal to max signed 32-bit asmline will emit code to mov to 32-bit register rather than 64-bit. That is: "mov rax, 0x7fffffff" as "mov eax, 0x7fffffff" -> b8 ff ff ff 7f note: rax got optimized to eax for faster immediate to register transfer and produces a shorter instruction.
added a command-line option in asmline for setting assembly mode to STRICT '$ asmline -t' where: ex: even if immediate size for mov is less than or equal to max signed 32-bit pad the immediate to fit 64bit. That is: "mov rax,0x7fffffff" as "mov rax,0x000000007fffffff" -> 48 b8 ff ff ff 7f 00 00 00 00
added a command-line option in asmline for setting assembly mode to SMART '$ asmline -s' where: asmline will check the immediate value for leading 0's and thus allows manual optimizations. Immediate must be zero padded to 64-bits exactly to ensure it will not optimize. This is currently set as default. ex: "mov rax, 0x000000007fffffff" -> 48 b8 ff ff ff 7f 00 00 00 00 "mov rax, 0x7fffffff" -> b8 ff ff ff 7f
assemblyline functions without asm_ prefix are now deprecated
simplified the process for adding new instructions
improvements to autotools, including use of pkg-config. That enables other autotools-based projects to include assemblyline as such: add to their configure.ac
:
AC_SUBST([LIBASSEMBLYLINE_CFLAGS])
AC_SUBST([LIBASSEMBLYLINE_LIBS])
PKG_CHECK_MODULES([LIBASSEMBLYLINE], [assemblyline >= 1.1.0])
tools/asmline dynamically links to libassemblyline.so
tools/asmline's -r[=LEN] option now runs with six arguments, which point to heap-memory. Each argumenty points to LEN (defaults to ten) 64-bit elements. It also features now --rand, which initializes those pointers to random data.
added asmline 1
and libassemblyline 3
man-page
added man-pages for all the functions from libassemblyline (same content)
Full Changelog: https://github.com/0xADE1A1DE/AssemblyLine/compare/v1.0.5...v1.1.0
Published by dderjoel about 3 years ago
Fixes another bug which caused mov reg, imm to sometimes produce wrong machine code.
Published by dderjoel about 3 years ago
Changelog
This hotfix (-1) adresses the issue of the missing manpage, which results in a failed installation.
Published by dderjoel about 3 years ago
refer to 1.0.4-1
Published by dderjoel about 3 years ago
Published by daviduwu9 about 3 years ago
added support for prefetch instruction
Published by dderjoel about 3 years ago
support for rorx instruction
Published by daviduwu9 about 3 years ago