Subversion Repositories Kolibri OS

Compare Revisions

Regard whitespace Rev 1062 → Rev 1115

/programs/develop/fasm/trunk/WHATSNEW.TXT
2,284 → 2,65
Visit http://flatassembler.net/ for more information.
 
 
version 1.67.35 (Mar 09, 2009)
version 1.68 (Jun 13, 2009)
 
[-] Some internal code size reductions.
[+] Added SSSE3 (Supplemental SSE3), SSE4.1, SSE4.2 and SSE4a instructions.
 
[-] Discontinued "-d" switch implementation.
[+] Added the AMD SVM and Intel SMX instructions.
 
[+] Added "rdmsrq", "wrmsrq", "sysexitq" and "sysretq" mnemonics for the
64-bit variants of respective instructions.
 
version 1.67.34 (Feb 22, 2009)
[+] Added "fstenvw", "fstenvd", "fsavew", "fsaved", "frstorw" and "frstord"
mnemonics to allow choosing between 16-bit and 32-bit variants of
structures used by the "fstenv", "fsave" and "frstor" instructions.
 
[-] Fixed small bug with DT directive far pointer syntax.
[+] Added "plt" operator for the ELF output format.
 
[+] Allowed "rva" operator to be used in MS COFF object format, and also
added "static" keyword for the "public" directive.
 
version 1.67.33 (Feb 20, 2009)
[+] Added Intel-style aliases for the additional long mode 8-bit registers.
 
[+] Added ERR directive that allows to signalize error from the source.
[-] The PE formatter now automatically detects whether relocatable labels
should be used, depending on whether the fixups directory is placed
somewhere into executable by programer, or not. This makes possible the
more flexible use of the addressing symbols in case of PE executable fixed
at some position.
 
[-] Added support for outputting the 32-bit address relocations in case of
64-bit object formats and PE executable. This makes some specific
instructions compilable, but it also forces linker to put such
generated code into the low 2 gigabytes of addressing space.
 
version 1.67.32 (Feb 13, 2009)
[+] Added "EFI", "EFIboot" and "EFIruntime" subsystem keywords for PE format.
 
[+] Allowed single quote character to be put inside the number value,
to help improve long numbers readability.
[-] Corrected the precedence of operators of macroinstruction line maker.
The symbol escaping now has always the higher priority than symbol conversion,
and both have higher precedence than concatenation.
 
 
version 1.67.31 (Feb 11, 2009)
 
[-] Fixed floating point converter to no longer generate NaN in some cases,
and corrected denormal numbers generation.
 
 
version 1.67.30 (Feb 08, 2009)
 
[+] Added missing Intel SSE4 instructions.
 
[+] Added SSE4a (EXTRQ/INSERTQ/MOVNTSD/MOVNTSS) instructions.
 
[+] Added FSTENVW/FSTENVD/FSAVEW/FSAVED mnemonics.
 
 
version 1.67.29 (Nov 15, 2008)
 
[+] Added optional symbolic information output, and a set of tools that extract
various kinds of information from it.
 
[+] Allowed RVA operator to be used in COFF object format.
 
[-] Corrected the FIX directive to allow names of preprocessor's directives to be
used as prioritized symbolic constant name.
 
 
version 1.67.28 (Nov 06, 2008)
 
[+] Added EFI/EFIBOOT/EFIRUNTIME subsystem keywords for PE format (experimental
feature, still needs to be tested).
 
[-] Fixed a bug in preprocessor which caused some problems with DISPLAY directive
output in some cases.
 
[-] Fixed a bug that allowed to define a symbol named "eip" or "rip".
 
[-] Fixed a problem with assembling MOVQ RAX,XMM0 instruction.
 
 
version 1.67.27 (Jul 12, 2008)
 
[-] Fixed a bug which caused instructions from JCXZ/LOOP family to be incorrectly
assembled when requiring 67h prefix and provided with "$" operand.
 
[-] Definition of macro with the same name as one of the preprocessor's directives
is no longer allowed.
 
[-] Fixed REPT directive to check out validity of its parameters even when the
count is zero.
 
[-] Fixed VMREAD/VMWRITE syntax in the USE64 mode.
 
[-] Corrected MZ header generation and handling for the case when the last page of
executable contains the round 512 bytes.
 
 
version 1.67.26 (Jan 27, 2008)
 
[+] Added partial SSE4 support (a couple of instructions left to be implemented
in the next releases)
 
[+] Added GETSEC instruction for the SMX functions calling.
 
[-] Some fixes and rearrangements in the DOS version.
 
 
version 1.67.25 (Dec 30, 2007)
 
[-] Fixed a couple of bugs related to undefined symbol error signaling.
 
 
version 1.67.24 (Dec 03, 2007)
 
[+] Added "-d" option to allow predefining symbolic constants from command line.
 
[-] Changed some of the error messages to be more informative.
 
 
version 1.67.23 (Sep 08, 2007)
 
[+] Added "static" keyword for the "public" directive in COFF format.
 
[-] Redirected error information into stderr.
 
 
version 1.67.22 (Aug 05, 2007)
 
[+] Allowed to check "@b" and "@f" symbols with "defined" operator.
 
[+] Allowed "as" operator to specify the output file extension when
placed at the end of the "format" directive line.
 
[-] Fixed a bug with stack size declaration in PE64 format that
caused assembler to crash.
[-] Definition of macro with the same name as one of the preprocessor's directives
is no longer allowed.
 
[-] Corrected the VMREAD and VMWRITE instructions to allow registers
in place of memory operands.
[+] Allowed single quote character to be put inside the number value,
to help improve long numbers readability.
 
[-] Corrected a mistake that made PLT operator not work correctly with
the CALL instruction.
[+] Added optional symbolic information output, and a set of tools that extract
various kinds of information from it.
 
[+] Added "err" directive that allows to signalize error from the source.
 
version 1.67.21 (Feb 17, 2007)
 
[+] Allowed 32-bit relocations in PE64 output format.
 
[-] Fixed a bug that caused "push cr0" error to go unnoticed by assembler.
 
 
version 1.67.20 (Feb 12, 2007)
 
[-] Corrected the range checking of signed address displacements in 64-bit mode.
 
 
version 1.67.19 (Feb 10, 2007)
 
[-] Disallowed labels starting with $ character.
 
[-] Fixed some problems with handling 32-bit relocations in 64-bit formats.
 
 
version 1.67.18 (Dec 27, 2006)
 
[-] Fixed a small mistake in the recent corrections of logical expression parser.
 
 
version 1.67.17 (Dec 20, 2006)
 
[-] Corrected the precedence of operators of macroinstruction line maker.
The symbol escaping now has always the higher priority than symbol conversion,
and both have higher precedence than concatenation.
 
[-] Fixed problems with logical expression parser to handle correctly comparing
values containing round brackets with the EQ or EQTYPE operator.
The only limitation now is that when you use round brackets to enclose some
logical expressions, they must be paired everywhere inside those expressions.
 
 
version 1.67.16 (Dec 17, 2006)
 
[-] Fixed a really old bug in register expression calculator.
 
 
version 1.67.15 (Nov 20, 2006)
 
[-] Some final (?) fixes and cleanup in the relative-offsets calculations.
 
 
version 1.67.14 (Oct 28, 2006)
 
[-] Another small bug fixes.
 
 
version 1.67.13 (Oct 26, 2006)
 
[-] Fixed a bug with handling of invalid expressions by "defined" operator.
 
[-] Fixed a newly-introduced bug with relocations in automatically generated RIP-relative
addressings.
 
 
version 1.67.12 (Oct 1, 2006)
 
[+] Added EIP-relative addressing, and fixed some bugs with RIP-related one aswell.
 
 
version 1.67.11 (Sep 26, 2006)
 
[+] Added PLT operator for the ELF output format.
 
[-] Rearranged and fixed some of the relocation handling routines.
 
 
version 1.67.10 (Sep 23, 2006)
 
[+] Added SSSE3 (Supplemental SSE3) instructions.
 
 
version 1.67.9 (Sep 21, 2006)
 
[+] Added some Win64 headers and examples in the Windows package.
 
[-] Fixed another operand size checking bug with MOVQ instruction.
 
 
version 1.67.8 (Sep 17, 2006)
 
[-] Fixed a bug in address processing that caused disallowed combinations like
[qword eax] to be accepted.
 
 
version 1.67.7 (Jul 31, 2006)
 
[-] Fixed a bug that caused over-repeated processing of macro operators when
many embedded macros were placed in one line
 
[+] Added SYSRETQ mnemonic.
 
 
version 1.67.6 (Jul 21, 2006)
 
[+] Added (not yet documented) RDMSRQ/WRMSRQ/SYSEXITQ mnemonics for the 64-bit
variants of respective instructions.
 
[+] Added information about memory allocation to the command line interfaces.
This should make it easier to decide when to use the -m switch.
 
 
version 1.67.5 (Jul 7, 2006)
 
[-] Fixed encoding of MOV (E)AX,[WORD ADDR] instruction in 32-bit mode.
 
 
version 1.67.4 (Jul 5, 2006)
 
[+] Added R8L-R15L (Intel-style) aliases for R8B-R15B registers.
 
[-] Further optimizations in the parser.
 
 
version 1.67.3 (Jul 2, 2006)
 
[-] Improved an instruction searching algorithm a bit.
 
 
version 1.67.2 (Jul 1, 2006)
 
[-] Fixed a recently introduced bug with ELF relocations in case of 32-bit
object.
 
 
version 1.67.1 (Jun 30, 2006)
 
[+] Added the support for the new AMD SVM technology instructions.
 
 
version 1.67.0 (Jun 28, 2006)
 
[-] The PE formatter now automatically detects whether relocatable labels
should be used, depending on whether the fixups directory is placed
somewhere into executable by programer, or not. This makes possible the
more flexible use of the addressing symbols in case of PE executable fixed
at some position.
 
[-] Added support for outputting the 32-bit address relocations in case of
64-bit object formats. This makes some specific instructions compilable,
but it also forces linker to put such generated code into the low 2
gigabytes of addressing space.
 
 
version 1.66 (May 7, 2006)
 
[+] Added DEFINE directive to preprocessor, which defines symbolic constants,
the same kind as EQU directive, however there's an important difference
that DEFINE doesn't process symbolic constants in the value before
[+] Added "define" directive to preprocessor, which defines symbolic constants,
the same kind as "equ" directive, however there's an important difference
that "define" doesn't process symbolic constants in the value before
assigning it. For example:
 
a equ 1
302,7 → 83,7
nop
end if
 
is parsed into just NOP instruction, since parser is able to determine
is parsed into just "nop" instruction, since parser is able to determine
that the condition is true, even though one of the logical values makes no
sense - but since this is none of the "eq", "eqtype" and "in" expressions,
the parser doesn't investigate.
325,7 → 106,7
[-] The size operator applied to jump no longer applies to the size of relative
displacement - now it applies to the size of target address.
 
[-] The RET instruction with 0 parameter is now assembled into short form,
[-] The "ret" instruction with 0 parameter is now assembled into short form,
unless you force using the 16-bit immediate with "word" operator.
 
[+] Added missing extended registers for the 32-bit addressing in long mode.
350,7 → 131,7
[-] Some of the internal structures have been extended to provide the
possibility of making extensive symbol dumps.
 
[-] Corrected FIX directive to keep the value intact before assigning it to the
[-] Corrected "fix" directive to keep the value intact before assigning it to the
prioritized constant.
 
[+] The ` operator now works with any kind of symbol; when used with quoted
/programs/develop/fasm/trunk/formats.inc
481,11 → 481,11
and eax,1111b
sub ecx,eax
mov edx,edi
xor al,al
xor eax,eax
rep stos byte [edi]
mov dword [org_origin],edi
mov dword [org_origin+4],0
mov [org_registers],0
mov dword [org_origin+4],eax
mov [org_registers],eax
mov [org_start],edi
mov eax,edx
call undefined_data
668,14 → 668,14
mov eax,40h+default_stub_end-default_stub
mov cx,100h+default_stub_end-default_stub
mov word [edx],'MZ'
mov word [edx+4],1
mov byte [edx+4],1
mov word [edx+2],ax
mov word [edx+8],4
mov word [edx+0Ah],10h
mov byte [edx+8],4
mov byte [edx+0Ah],10h
mov word [edx+0Ch],0FFFFh
mov word [edx+10h],cx
mov word [edx+3Ch],ax
mov word [edx+18h],40h
mov byte [edx+18h],40h
lea edi,[edx+40h]
mov esi,default_stub
mov ecx,default_stub_end-default_stub
804,7 → 804,7
mov [esi+0Ah],ax
binary_heap_ok:
mov word [esi],'MZ'
mov word [esi+8],4
mov byte [esi+8],4
mov ax,0FFFFh
mov [esi+0Ch],ax
dec ax
813,7 → 813,7
mov [esi+0Eh],ax
mov [esi+16h],ax
mov word [esi+14h],100h
mov word [esi+18h],40h
mov byte [esi+18h],40h
mov eax,[display_buffer]
sub eax,ecx
cmp edi,eax
977,9 → 977,9
mov word [edx],'PE' ; signature
mov ax,[machine]
mov word [edx+4],ax
mov dword [edx+38h],1000h ; section alignment
mov dword [edx+3Ch],200h ; file alignment
mov word [edx+40h],1 ; OS version
mov byte [edx+38h+1],10h ; section alignment
mov byte [edx+3Ch+1],2 ; file alignment
mov byte [edx+40h],1 ; OS version
mov eax,[subsystem_version]
mov [edx+48h],eax
mov ax,[subsystem]
993,28 → 993,26
mov word [edx+1Ah],VERSION_MAJOR + VERSION_MINOR shl 8
test [format_flags],8
jnz init_peplus_specific
mov dword [edx+14h],0E0h ; size of optional header
mov byte [edx+14h],0E0h ; size of optional header
mov dword [edx+16h],10B010Eh; flags and magic value
mov eax,[image_base]
mov dword [edx+34h],eax
mov dword [edx+60h],1000h ; stack reserve
mov dword [edx+64h],1000h ; stack commit
mov dword [edx+68h],10000h ; heap reserve
mov dword [edx+6Ch],0 ; heap commit
mov dword [edx+74h],16 ; number of directories
mov [edx+34h],eax
mov byte [edx+60h+1],10h ; stack reserve
mov byte [edx+64h+1],10h ; stack commit
mov byte [edx+68h+2],1 ; heap reserve
mov byte [edx+74h],16 ; number of directories
jmp pe_header_ok
init_peplus_specific:
mov dword [edx+14h],0F0h ; size of optional header
mov byte [edx+14h],0F0h ; size of optional header
mov dword [edx+16h],20B002Eh; flags and magic value
mov eax,[image_base]
mov dword [edx+30h],eax
mov [edx+30h],eax
mov eax,[image_base_high]
mov dword [edx+34h],eax
mov dword [edx+60h],1000h ; stack reserve
mov dword [edx+68h],1000h ; stack commit
mov dword [edx+70h],10000h ; heap reserve
mov dword [edx+78h],0 ; heap commit
mov dword [edx+84h],16 ; number of directories
mov [edx+34h],eax
mov byte [edx+60h+1],10h ; stack reserve
mov byte [edx+68h+1],10h ; stack commit
mov byte [edx+70h+2],1 ; heap reserve
mov byte [edx+84h],16 ; number of directories
pe_header_ok:
bsf ecx,[edx+3Ch]
imul ebx,[number_of_sections],28h
1089,11 → 1087,11
mov [org_start],edi
bt [format_flags],8
jnc dll_flag_ok
or dword [edx+16h],2000h
or byte [edx+16h+1],20h
dll_flag_ok:
bt [format_flags],9
jnc wdm_flag_ok
or word [edx+5Eh],2000h
or byte [edx+5Eh+1],20h
wdm_flag_ok:
jmp format_defined
pe_section:
1234,7 → 1232,6
mov eax,[code_start]
sub eax,[stub_size]
sub [ebx+14h],eax
 
mov ecx,[ebx+10h]
test byte [ebx+24h],20h
jz pe_code_sum_ok
1254,7 → 1251,6
mov eax,[ebx+0Ch]
mov [edx+30h],eax
pe_data_sum_ok:
 
mov eax,[ebx+8]
or eax,eax
jz udata_ok
1263,7 → 1259,6
or byte [ebx+24h],80h
add [edx+24h],ecx
udata_ok:
 
mov ecx,[edx+38h]
dec ecx
add eax,ecx
2261,10 → 2256,11
flat_section_flags_ok:
mov dword [ebx+14h],eax
mov [current_section],ebx
mov [number_of_sections],0
xor eax,eax
mov [number_of_sections],eax
mov dword [org_origin],edi
mov dword [org_origin+4],0
mov [org_registers],0
mov dword [org_origin+4],eax
mov [org_registers],eax
mov [org_start],edi
mov [org_symbol],ebx
mov [labels_type],2
2287,8 → 2283,8
mov [ebx],al
mov [ebx+8],edi
mov dword [org_origin],edi
mov dword [org_origin+4],0
mov [org_registers],0
mov dword [org_origin+4],eax
mov [org_registers],eax
mov [org_start],edi
mov [org_symbol],ebx
mov [labels_type],2
2935,8 → 2931,8
mov [ebx],al
mov [ebx+8],edi
mov dword [org_origin],edi
mov dword [org_origin+4],0
mov [org_registers],0
mov dword [org_origin+4],eax
mov [org_registers],eax
mov [org_start],edi
mov [org_symbol],ebx
test [format_flags],8
3977,7 → 3973,7
 
dump_symbols:
mov ebx,[code_start]
mov dword [ebx],'fas'+27 shl 24
mov dword [ebx],'fas'+1Ah shl 24
mov dword [ebx+4],VERSION_MAJOR + VERSION_MINOR shl 8 + 38h shl 16
add ebx,38h
mov edi,ebx
/programs/develop/fasm/trunk/tables.inc
963,6 → 963,8
dw loop_instruction_16bit-assembler
db 'loopz',0E1h
dw loop_instruction-assembler
db 'lzcnt',0BDh
dw popcnt_instruction-assembler
db 'maxpd',5Fh
dw sse_pd_instruction-assembler
db 'maxps',5Fh
1324,6 → 1326,8
dw sse_pd_instruction-assembler
db 'hsubps',07Dh
dw cvtpd2dq_instruction-assembler
db 'invept',80h
dw vmx_inv_instruction-assembler
db 'invlpg',0
dw invlpg_instruction-assembler
db 'lfence',0E8h
1615,6 → 1619,8
dw insertq_instruction-assembler
db 'invlpga',0DFh
dw invlpga_instruction-assembler
db 'invvpid',81h
dw vmx_inv_instruction-assembler
db 'ldmxcsr',10b
dw fxsave_instruction-assembler
db 'loopned',0E0h
/programs/develop/fasm/trunk/version.inc
1,5 → 1,5
 
; flat assembler version 1.67
; flat assembler version 1.68
; Copyright (c) 1999-2009, Tomasz Grysztar.
; All rights reserved.
;
33,7 → 33,7
; cannot simply be copied and put under another distribution licence
; (including the GNU Public Licence).
 
VERSION_STRING equ "1.67.38"
VERSION_STRING equ "1.68"
 
VERSION_MAJOR = 1
VERSION_MINOR = 67
VERSION_MINOR = 68
/programs/develop/fasm/trunk/x86_64.inc
6146,8 → 6146,6
lods byte [esi]
cmp al,','
jne invalid_operand
call vmread_check_size
mov [operand_size],0
lods byte [esi]
call get_size_operator
cmp al,10h
6165,7 → 6163,6
lods byte [esi]
cmp al,','
jne invalid_operand
mov [operand_size],0
lods byte [esi]
call get_size_operator
cmp al,10h
6196,8 → 6193,6
lods byte [esi]
call convert_register
mov [postbyte_register],al
call vmread_check_size
mov [operand_size],0
lods byte [esi]
cmp al,','
jne invalid_operand
6216,6 → 6211,33
mov bl,al
mov [base_code],0Fh
jmp nomem_instruction_ready
vmx_inv_instruction:
mov [opcode_prefix],66h
mov [extended_code],38h
mov [supplemental_code],al
lods byte [esi]
call get_size_operator
cmp al,10h
jne invalid_operand
lods byte [esi]
call convert_register
mov [postbyte_register],al
call vmread_check_size
mov [operand_size],0
lods byte [esi]
cmp al,','
jne invalid_operand
lods byte [esi]
call get_size_operator
cmp al,'['
jne invalid_operand
call get_address
mov al,[operand_size]
or al,al
jz vmx_size_ok
cmp al,16
jne invalid_operand_size
jmp vmx_size_ok
simple_svm_instruction:
push eax
mov [base_code],0Fh