I've been using a very old version of the MASM32 library for like, forever. Recently, my technical reviewer started complaining that the listings he's reviewing in "The Art of 64-bit Assembly, Vol 2" were getting all kinds of errors.
The problem, it seems, is that the MASM32 library changed quite a bit over the past several years since I first wrote Ao64Av2 (my include files are still present in an "include64" library, for example).
Okay, so I figure I'll download the latest version and see what's going on.
When I include the "Windows.inc" header files, I get assembly errors. Specifically,
DEBUGHOOKINFO STRUCT
hModuleHook DWORD ?
Reserved DWORD ?
lParam DWORD ?
wParam DWORD ?
code DWORD ?
DEBUGHOOKINFO ENDS
it doesn't like the "code DWORD ?" line (isn't "code" a reserved word in MASM?)
Any ideas what is going on here?
Thanks,
Randy Hyde
There have been errors reported with several include files, some for the masm32 SDK and many for the Masm64 SDK, which is missing a lot of include files as well.
My best advice is get a hold of the corresponding C header files for whichever API you are using, for either comparison (to the .inc file) for correctness, or to convert the C header file (.h) to a masm compatible ".inc" file.
I am not sure if "code" is reserved, but ".code" certainly is. Someone else may shed some light here. I am on my iPad so I can't easily check. I am not near a computer at the moment.
Quote from: hyder on December 10, 2024, 06:33:20 AMI've been using a very old version of the MASM32 library for like, forever. Recently, my technical reviewer started complaining that the listings he's reviewing in "The Art of 64-bit Assembly, Vol 2" were getting all kinds of errors.
The problem, it seems, is that the MASM32 library changed quite a bit over the past several years since I first wrote Ao64Av2 (my include files are still present in an "include64" library, for example).
Okay, so I figure I'll download the latest version and see what's going on.
When I include the "Windows.inc" header files, I get assembly errors. Specifically,
DEBUGHOOKINFO STRUCT
hModuleHook DWORD ?
Reserved DWORD ?
lParam DWORD ?
wParam DWORD ?
code DWORD ?
DEBUGHOOKINFO ENDS
it doesn't like the "code DWORD ?" line (isn't "code" a reserved word in MASM?)
Any ideas what is going on here?
Thanks,
Randy Hyde
1st. I like your books :)
2nd. You are doing MASM64, I assume from the context. Hutch's MASM64 SDK does'nt have DEBUGHOOKINFO STRUCT. Searching Windows SDK C++ headers is as follow:
/*
* Structure used by WH_DEBUG
*/
typedef struct tagDEBUGHOOKINFO
{
DWORD idThread;
DWORD idThreadInstaller;
LPARAM lParam;
WPARAM wParam;
int code;
} DEBUGHOOKINFO, *PDEBUGHOOKINFO, NEAR *NPDEBUGHOOKINFO, FAR* LPDEBUGHOOKINFO;
And I assume that LPARAM and WPARAM in 64bit are QWORD's in Windows 64bit. Windows Docs say that
code member is: The value to be passed to the hook in the nCode parameter of the DebugProc callback function.
Tried add the structure to some custom project, It doesnt assemble source with that structure what you did post. I did found that Hutch used as follows in another structure.
NMHDR STRUCT
hwndFrom HWND ?
idFrom UINT_PTR ?
_code UINT ?
UINT ?
NMHDR ENDSMASM32 SDK contains it MASM64 SDK doesnt, it's needed to be translated. That MASM64 for some reason doesnt accept it in source, It's not mentioned in Directives Reference also.
Cheers Petter
DEBUGHOOKINFO 20 14h bytes
idThread +0h 4h
idThreadInstaller +4h 4h
lParam +8h 4h
wParam +Ch 4h
code +10h 4h
DEBUGHOOKINFO 32 20h bytes
idThread +0h 4h
idThreadInstaller +4h 4h
lParam +8h 8h
wParam +10h 8h
code +18h 4h
Don't sure if 64 bit ml still has
Nokeyword, which let you remove keyword status, for example my very old SSE2 macros I used Nokeyword so I could define some macros
It's not that things are missing, it's that existing headers have errors during assembly.
c:\masm32\include\windows.inc(9750) : error A2008:syntax error : dword
c:\masm32\include\windows.inc(10183) : error A2008:syntax error : frame
c:\masm32\include\windows.inc(16333) : error A2008:syntax error : dword
c:\masm32\include\windows.inc(18862) : error A2008:syntax error : Frame
c:\masm32\include\windows.inc(24749) : error A2042:statement too complex
c:\masm32\include\winextra.inc(4647) : error A2008:syntax error : tab
c:\masm32\include\winextra.inc(11052) : error A2026:constant expected
c:\masm32\include\winextra.inc(11053) : error A2026:constant expected
Our current workaround, masm32.inc:
; masm32.inc
if 0 ; 1 to use masm32, 0 for local definitions
include c:\masm32\include\windows.inc
include c:\masm32\include\kernel32.inc
includelib c:\masm32\lib\kernel32.lib
else
STILL_ACTIVE equ 000000103h
INFINITE equ 0FFFFFFFFh
TLS_OUT_OF_INDEXES equ 0FFFFFFFFh
handle TYPEDEF qword
HANDLE TYPEDEF handle
CRITICAL_SECTION struc
DebugInfo qword ?
LockCount dword ?
RecursionCount dword ?
OwningThread handle ?
LockSemaphore handle ?
SpinCount dword ?
dword ?
CRITICAL_SECTION ends
STARTUPINFO STRUCT
cb dword ?
lpReserved qword ?
lpDesktop qword ?
lpTitle qword ?
dwX dword ?
dwY dword ?
dwXSize dword ?
dwYSize dword ?
dwXCountChars dword ?
dwYCountChars dword ?
dwFillAttribute dword ?
dwFlags dword ?
wShowWindow word ?
cbReserved2 word ?
lpReserved2 qword ?
hStdInput handle ?
hStdOutput handle ?
hStdError handle ?
STARTUPINFO ENDS
PROCESS_INFORMATION STRUCT
hProcess HANDLE ?
hThread HANDLE ?
dwProcessId DWORD ?
dwThreadId DWORD ?
PROCESS_INFORMATION ENDS
extrn __imp_CreateThread:ptr proc
extrn __imp_Sleep:ptr proc
extrn __imp_InitializeCriticalSection:ptr proc
extrn __imp_DeleteCriticalSection:ptr proc
extrn __imp_EnterCriticalSection:ptr proc
extrn __imp_LeaveCriticalSection:ptr proc
extrn __imp_GetExitCodeThread:ptr proc
extrn __imp_CreateEventA:ptr proc
extrn __imp_SetEvent:ptr proc
extrn __imp_WaitForSingleObject:ptr proc
extrn __imp_WaitForMultipleObjects:ptr proc
extrn __imp_CloseHandle:ptr proc
extrn __imp_CreateWaitableTimerA:ptr proc
extrn __imp_SetWaitableTimer:ptr proc
extrn __imp_CancelWaitableTimer:ptr proc
extrn __imp_GetLastError:ptr proc
extrn __imp_TlsAlloc:ptr proc
extrn __imp_TlsGetValue:ptr proc
extrn __imp_TlsSetValue:ptr proc
extrn __imp_TlsFree:ptr proc
extrn __imp_CreateProcessA:ptr proc
extrn __imp_GetCurrentThread:ptr proc
extrn __imp_SetThreadAffinityMask:ptr proc
extrn __imp_GetCurrentProcess:ptr proc
extrn __imp_GetProcessAffinityMask:ptr proc
CreateThread equ __imp_CreateThread
Sleep equ __imp_Sleep
InitializeCriticalSection equ __imp_InitializeCriticalSection
DeleteCriticalSection equ __imp_DeleteCriticalSection
EnterCriticalSection equ __imp_EnterCriticalSection
LeaveCriticalSection equ __imp_LeaveCriticalSection
GetExitCodeThread equ __imp_GetExitCodeThread
CreateEvent equ __imp_CreateEventA
SetEvent equ __imp_SetEvent
WaitForSingleObject equ __imp_WaitForSingleObject
WaitForMultipleObjects equ __imp_WaitForMultipleObjects
CloseHandle equ __imp_CloseHandle
CreateWaitableTimer equ __imp_CreateWaitableTimerA
SetWaitableTimer equ __imp_SetWaitableTimer
CancelWaitableTimer equ __imp_CancelWaitableTimer
GetLastError equ __imp_GetLastError
TlsAlloc equ __imp_TlsAlloc
TlsGetValue equ __imp_TlsGetValue
TlsSetValue equ __imp_TlsSetValue
TlsFree equ __imp_TlsFree
CreateProcess equ __imp_CreateProcessA
GetCurrentThread equ __imp_GetCurrentThread
SetThreadAffinityMask equ __imp_SetThreadAffinityMask
GetCurrentProcess equ __imp_GetCurrentProcess
GetProcessAffinityMask equ __imp_GetProcessAffinityMask
endif
Quote from: ttribelli on December 11, 2024, 05:22:15 AMc:\masm32\include\windows.inc
If one day your employer declares C: taboo and forces you to install Masm32 on D:, you will have lots of surprises :cool:
Quote from: C3 on December 10, 2024, 07:32:54 AMQuote from: hyder on December 10, 2024, 06:33:20 AMDEBUGHOOKINFO STRUCT
hModuleHook DWORD ?
Reserved DWORD ?
lParam DWORD ?
wParam DWORD ?
code DWORD ?
DEBUGHOOKINFO ENDS
2nd. You are doing MASM64, I assume from the context. Hutch's MASM64 SDK does'nt have DEBUGHOOKINFO STRUCT. Searching Windows SDK C++ headers is as follow:
Cheers Petter
FWIW, I don't know if "code" is actually reserved. It is used as a class designation in the SEGMENT directive. AFAIKT, I'm not defining "code" anywhere in my header files (that could lead to an error).
Yes, Hutch's old code doesn't have this structure, it's present in the latest downloads.
I wanted to use the latest and greatest include files for "The Art of 64-bit Assembly, Volume 2" (which is currently in the technical review phase), but if the use of said files breaks all my code, then I suppose I'm going to have to package up the definitions I need and stick with that. Of course, this invites all kinds of ugliness as it creates a fork that isn't maintained by the MASM32 site. As such, something I really want to avoid.
Cheers,
Randy Hyde
Regarding the reserved-word-ness of code, I just did a li'l test:
T STRUC
code DD ?
T ENDS
.data
t T<>
code DD ?
It assembled, linked and ran fine. So apparently it's fine to use code as a variable or structure element name. Which I found surprising.
Quote from: hyder on December 12, 2024, 11:16:22 AMFWIW, I don't know if "code" is actually reserved.
Reserved words are case-insensitive. So changing "code" to "CODE" and if the error persists, then it's indeed a reserved word.
With "OPTION NOKEYWORD:<code>" it's possible to remove the keyword and the code should assemble without errors.
With a few experiments one can see that CODE is indeed a directive now: it defines a code section and expects a name as argument:
code abc
mov al,0
Feeding the resulting .OBJ to "DUMPBIN /symbols", one can see that a section named "abc" has been defined, size 2. However, it seems pretty likely that this new CODE directive is due to a misunderstanding at MS - the directive should probably have been .CODE, and that indeed does still work ( and, IIRC, also works with ML.EXE ):
.code abc
mov al,0
Quote from: _japheth on December 14, 2024, 05:13:49 PMReserved words are case-insensitive. So changing "code" to "CODE" and if the error persists, then it's indeed a reserved word.
That doesn't make sense; if, as you say, reserved words are case-insensitive (they are), then it should make no difference whether you write
code or
CODE or
CoDe, right?
Quote from: NoCforMe on December 14, 2024, 05:22:34 PMThat doesn't make sense; if, as you say, reserved words are case-insensitive (they are), then it should make no difference whether you write code or CODE or CoDe, right?
I guess I expressed myself in bad English. The sentence should be read as:
So if your change "code" to "CODE" and the error still persists, then it's indeed a reserved word.
A different test. Using the 64-bit assembler for 64-bit code.
Syntax error rather than undefined symbol.
"code" can replace ".code". I guess someone disliked leading dots.
Microsoft (R) Macro Assembler (x64) Version 14.41.34120.0 12/13/24 14:47:21
test1.asm Page 1 - 1
00000000 .code
add eax,code
test1.asm(2) : error A2008:syntax error : code
00000000 .data
;code dword ?
end
␌Microsoft (R) Macro Assembler (x64) Version 14.41.34120.0 12/13/24 15:58:37
test1.asm Page 1 - 1
00000000 code
00000000 03 05 00000000 R add eax,bcode
00000000 .data
;code dword ?
00000000 00000000 bcode dword ?
end
The code/.code equivalence of ml64 does not extend to the other simplified segment directives.
00000000 code abc
00000000 03 05 00000006 R add eax,bcode
xyz
test1.asm(3) : error A2008:syntax error : xyz
data ;DATA2
test1.asm(4) : error A2008:syntax error : data
00000006 00000032 bcode dword 50
data? ;BSS2
test1.asm(6) : error A2008:syntax error : data?
0000000A 00000000 bdata dword ?
const ;RDATA2
test1.asm(8) : error A2008:syntax error : const
0000000E 61 62 63 00 bconst db "abc",0
end