\C{outfmt} \i{Output Formats} NASM is a portable assembler, designed to be able to compile on any ANSI C-supporting platform and produce output to run on a variety of Intel x86 operating systems. For this reason, it has a large number of available output formats, selected using the \i\c{-f} option on the NASM \i{command line}. Each of these formats, along with its extensions to the base NASM syntax, is detailed in this chapter. As stated in \k{opt-o}, NASM chooses a \i{default name} for your output file based on the input file name and the chosen output format. This will be generated by removing the filename \i{extension} (\c{.asm}, \c{.s}, or whatever you like to use) from the input file name, and substituting an extension defined by the output format. The extensions are given with each format below. \H{binfmt} \i\c{bin}: \i{Flat-Form Binary}\I{pure binary} Output The \c{bin} format does not produce object files: it generates nothing in the output file except the code you wrote. Such `pure binary' files are used by \i{MS-DOS}: \i\c{.COM} executables and \i\c{.SYS} device drivers are pure binary files. Pure binary output is also useful for \i{operating system} and \i{boot loader} development. The \c{bin} format supports \i{multiple section names}. For details of how NASM handles sections in the \c{bin} format, see \k{multisec}. Using the \c{bin} format puts NASM by default into 16-bit mode (see \k{bits}). In order to use \c{bin} to write 32-bit or 64-bit code, such as an OS kernel, you need to explicitly issue the \I\c{BITS}\c{BITS 32} or \I\c{BITS}\c{BITS 64} directive. \c{bin} has no default output file name extension: instead, it leaves your file name as it is once the original extension has been removed. Thus, the default is for NASM to assemble \c{binprog.asm} into a binary file called \c{binprog}. It is extremely important to understand that the binary output format is simply nothing other than \e{a linker built into the NASM executable.} As such, NASM behaves just as it does when producing any other output format: notably the list file reflects the code output \e{before} relocation, and the addresses in the list file are addresses relative to the start of the current output section. \S{org} \i\c{ORG}: Binary File \i{Program Origin} The \c{bin} format provides an additional directive to the list given in \k{directive}: \c{ORG}. The function of the \c{ORG} directive is to specify the origin address which NASM will assume the program begins at when it is loaded into memory. For example, the following code will generate the longword \c{0x00000104}: \c org 0x100 \c dd label \c label: Unlike the \c{ORG} directive provided by MASM-compatible assemblers, which allows you to jump around in the object file and overwrite code you have already generated, NASM's \c{ORG} does exactly what the directive says: \e{origin}. Its sole function is to specify one offset which is added to all internal address references within the section; it does not permit any of the trickery that MASM's version does. See \k{proborg} for further comments. \S{binseg} \c{bin} Extensions to the \c{SECTION} Directive\I{section, bin extensions to} The \c{bin} output format extends the \c{SECTION} (or \c{SEGMENT}) directive to allow you to specify the alignment requirements of segments. This is done by appending the \i\c{ALIGN} qualifier to the end of the section-definition line. For example, \c section .data align=16 switches to the section \c{.data} and also specifies that it must be aligned on a 16-byte boundary. The parameter to \c{ALIGN} specifies how many low bits of the section start address must be forced to zero. The alignment value given may be any power of two.\I{section alignment, in bin}\I{segment alignment, in bin}\I{alignment, in bin sections} \S{multisec} \i{Multisection}\I{bin, multisection} Support for the \c{bin} Format The \c{bin} format allows the use of multiple sections, of arbitrary names, besides the "known" \c{.text}, \c{.data}, and \c{.bss} names. \b Sections may be designated \i\c{progbits} or \i\c{nobits}. Default is \c{progbits} (except \c{.bss}, which defaults to \c{nobits}, of course). \b Sections can be aligned at a specified boundary following the previous section with \c{align=}, or at an arbitrary byte-granular position with \i\c{start=}. \b Sections can be given a virtual start address, which will be used for the calculation of all memory references within that section with \i\c{vstart=}. \b Sections can be ordered using \i\c{follows=}\c{
} or \i\c{vfollows=}\c{
} as an alternative to specifying an explicit start address. \b Arguments to \c{org}, \c{start}, \c{vstart}, and \c{align=} are critical expressions. See \k{crit}. For example, in the case of \c{align=(1 << ALIGN_SHIFT)}, \c{ALIGN_SHIFT} must be defined before it is used here. \b Any code which comes before an explicit \c{SECTION} directive is directed by default into the \c{.text} section. \b If an \c{ORG} statement is not given, \c{ORG 0} is used by default. \b The \c{.bss} section will be placed after the last \c{progbits} section, unless \c{start=}, \c{vstart=}, \c{follows=}, or \c{vfollows=} has been specified. \b All sections are aligned on dword boundaries, unless a different alignment has been specified. \b Sections may not overlap. \b NASM creates the \c{section..start} for each section, which may be used in your code. \S{map}\i{Map Files} Map files can be generated in \c{-f bin} format by means of the \c{[map]} option. Map types of \c{all} (default), \c{brief}, \c{sections}, \c{segments}, or \c{symbols} may be specified. Output may be directed to \c{stdout} (default), \c{stderr}, or a specified file. E.g. \c{[map symbols myfile.map]}. No "user form" exists, the square brackets must be used. \H{ithfmt} \i\c{ith}: \i{Intel Hex} Output The \c{ith} file format produces Intel hex-format files. Just as the \c{bin} format, this is a flat memory image format with no support for further relocation or linking. It is usually used with ROM programmers and similar utilities. From a programmer point of view, this behaves identically to the \c{.bin} format; the only difference is the encoding of the output. All extensions supported by the \c{bin} file format is also supported by the \c{ith} file format. \c{ith} provides a default output file-name extension of \c{.ith}. \H{srecfmt} \i\c{srec}: \i{Motorola S-Records} Output The \c{srec} file format produces Motorola S-records files. Just as the \c{bin} format, this is a flat memory image format with no support for relocation or linking. It is usually used with ROM programmers and similar utilities. From a programmer point of view, this behaves identically to the \c{.bin} format; the only difference is the encoding of the output. All extensions supported by the \c{bin} file format is also supported by the \c{srec} file format. \c{srec} provides a default output file-name extension of \c{.srec}. \H{objfmt} \i\c{obj}: \i{Microsoft OMF}\I{OMF} Object Files The \c{obj} file format (NASM calls it \c{obj} rather than \c{omf} for historical reasons) is the one produced by \i{MASM} and \i{TASM}, which is typically fed to 16-bit DOS linkers to produce \i\c{.EXE} files. It is also the format used by \i{OS/2}. \c{obj} provides a default output file-name extension of \c{.obj}. \c{obj} is not exclusively a 16-bit format, though; NASM has full support for the 32-bit extensions to the format. In particular, 32-bit \c{obj} format files are used by \i{Borland's Win32 compilers}, instead of using Microsoft's newer \i\c{win32} object file format. The \c{obj} format does not define any special segment names: you can call your segments anything you like. Typical names for segments in \c{obj} format files are \c{CODE}, \c{DATA} and \c{BSS}. If your source file contains code before specifying an explicit \c{SEGMENT} directive, then NASM will invent its own segment called \i\c{__NASMDEFSEG} for you. When you define a segment in an \c{obj} file, NASM defines the segment name as a symbol as well, so that you can access the segment address of the segment. So, for example: \c segment data \c \c dvar: dw 1234 \c \c segment code \c \c function: \c mov ax,data ; get segment address of data \c mov ds,ax ; and move it into DS \c inc word [dvar] ; now this reference will work \c ret The \c{obj} format also enables the use of the \i\c{SEG} and \i\c{WRT} operators, so that you can write code which does things like \c extern foo \c \c mov ax,seg foo ; get preferred segment of foo \c mov ds,ax \c mov ax,data ; a different segment \c mov es,ax \c mov ax,[ds:foo] ; this accesses `foo' \c mov [es:foo wrt data],bx ; so does this \S{objseg} \c{obj} Extensions to the \c{SEGMENT} Directive\I{SEGMENT, obj extensions to} The \c{obj} output format extends the \c{SEGMENT} (or \c{SECTION}) directive to allow you to specify various properties of the segment you are defining. This is done by appending extra qualifiers to the end of the segment-definition line. For example, \c segment code private align=16 defines the segment \c{code}, but also declares it to be a private segment, and requires that the portion of it described in this code module must be aligned on a 16-byte boundary. The available qualifiers are: \b \i\c{PRIVATE}, \i\c{PUBLIC}, \i\c{COMMON} and \i\c{STACK} specify the combination characteristics of the segment. \c{PRIVATE} segments do not get combined with any others by the linker; \c{PUBLIC} and \c{STACK} segments get concatenated together at link time; and \c{COMMON} segments all get overlaid on top of each other rather than stuck end-to-end. \b \i\c{ALIGN} is used, as shown above, to specify how many low bits of the segment start address must be forced to zero. The alignment value given may be any power of two from 1 to 4096; in reality, the only values supported are 1, 2, 4, 16, 256 and 4096, so if 8 is specified it will be rounded up to 16, and 32, 64 and 128 will all be rounded up to 256, and so on. Note that alignment to 4096-byte boundaries is a \i{PharLap} extension to the format and may not be supported by all linkers.\I{section alignment, in OBJ}\I{segment alignment, in OBJ}\I{alignment, in OBJ sections} \b \i\c{CLASS} can be used to specify the segment class; this feature indicates to the linker that segments of the same class should be placed near each other in the output file. The class name can be any word, e.g. \c{CLASS=CODE}. \b \i\c{OVERLAY}, like \c{CLASS}, is specified with an arbitrary word as an argument, and provides overlay information to an overlay-capable linker. \b Segments can be declared as \i\c{USE16} or \i\c{USE32}, which has the effect of recording the choice in the object file and also ensuring that NASM's default assembly mode when assembling in that segment is 16-bit or 32-bit respectively. \b When writing \i{OS/2} object files, you should declare 32-bit segments as \i\c{FLAT}, which causes the default segment base for anything in the segment to be the special group \c{FLAT}, and also defines the group if it is not already defined. \b The \c{obj} file format also allows segments to be declared as having a pre-defined absolute segment address, although no linkers are currently known to make sensible use of this feature; nevertheless, NASM allows you to declare a segment such as \c{SEGMENT SCREEN ABSOLUTE=0xB800} if you need to. The \i\c{ABSOLUTE} and \c{ALIGN} keywords are mutually exclusive. NASM's default segment attributes are \c{PUBLIC}, \c{ALIGN=1}, no class, no overlay, and \c{USE16}. \S{group} \i\c{GROUP}: Defining Groups of Segments\I{segments, groups of} The \c{obj} format also allows segments to be grouped, so that a single segment register can be used to refer to all the segments in a group. NASM therefore supplies the \c{GROUP} directive, whereby you can code \c segment data \c \c ; some data \c \c segment bss \c \c ; some uninitialized data \c \c group dgroup data bss which will define a group called \c{dgroup} to contain the segments \c{data} and \c{bss}. Like \c{SEGMENT}, \c{GROUP} causes the group name to be defined as a symbol, so that you can refer to a variable \c{var} in the \c{data} segment as \c{var wrt data} or as \c{var wrt dgroup}, depending on which segment value is currently in your segment register. If you just refer to \c{var}, however, and \c{var} is declared in a segment which is part of a group, then NASM will default to giving you the offset of \c{var} from the beginning of the \e{group}, not the \e{segment}. Therefore \c{SEG var}, also, will return the group base rather than the segment base. NASM will allow a segment to be part of more than one group, but will generate a warning if you do this. Variables declared in a segment which is part of more than one group will default to being relative to the first group that was defined to contain the segment. A group does not have to contain any segments; you can still make \c{WRT} references to a group which does not contain the variable you are referring to. OS/2, for example, defines the special group \c{FLAT} with no segments in it. \c{GROUP} is cumulative. The above example can be done like this: \c group dgroup data \c group dgroup bss \S{uppercase} \i\c{UPPERCASE}: Disabling Case Sensitivity in Output Although NASM itself is \i{case sensitive}, some OMF linkers are not; therefore it can be useful for NASM to output single-case object files. The \c{UPPERCASE} format-specific directive causes all segment, group and symbol names that are written to the object file to be forced to upper case just before being written. Within a source file, NASM is still case-sensitive; but the object file can be written entirely in upper case if desired. \c{UPPERCASE} is used alone on a line; it requires no parameters. \S{import} \i\c{IMPORT}: Importing DLL Symbols\I{DLL symbols, importing}\I{symbols, importing from DLLs} The \c{IMPORT} format-specific directive defines a symbol to be imported from a DLL, for use if you are writing a DLL's \i{import library} in NASM. You still need to declare the symbol as \c{EXTERN} as well as using the \c{IMPORT} directive. The \c{IMPORT} directive takes two required parameters, separated by white space, which are (respectively) the name of the symbol you wish to import and the name of the library you wish to import it from. For example: \c import WSAStartup wsock32.dll A third optional parameter gives the name by which the symbol is known in the library you are importing it from, in case this is not the same as the name you wish the symbol to be known by to your code once you have imported it. For example: \c import asyncsel wsock32.dll WSAAsyncSelect \S{export} \i\c{EXPORT}: Exporting DLL Symbols\I{DLL symbols, exporting}\I{symbols, exporting from DLLs} The \c{EXPORT} format-specific directive defines a global symbol to be exported as a DLL symbol, for use if you are writing a DLL in NASM. You still need to declare the symbol as \c{GLOBAL} as well as using the \c{EXPORT} directive. \c{EXPORT} takes one required parameter, which is the name of the symbol you wish to export, as it was defined in your source file. An optional second parameter (separated by white space from the first) gives the \e{external} name of the symbol: the name by which you wish the symbol to be known to programs using the DLL. If this name is the same as the internal name, you may leave the second parameter off. Further parameters can be given to define attributes of the exported symbol. These parameters, like the second, are separated by white space. If further parameters are given, the external name must also be specified, even if it is the same as the internal name. The available attributes are: \b \c{resident} indicates that the exported name is to be kept resident by the system loader. This is an optimization for frequently used symbols imported by name. \b \c{nodata} indicates that the exported symbol is a function which does not make use of any initialized data. \b \c{parm=NNN}, where \c{NNN} is an integer, sets the number of parameter words for the case in which the symbol is a call gate between 32-bit and 16-bit segments. \b An attribute which is just a number indicates that the symbol should be exported with an identifying number (ordinal), and gives the desired number. For example: \c export myfunc \c export myfunc TheRealMoreFormalLookingFunctionName \c export myfunc myfunc 1234 ; export by ordinal \c export myfunc myfunc resident parm=23 nodata \S{dotdotstart} \i\c{..start}: Defining the \i{Program Entry Point} \c{OMF} linkers require exactly one of the object files being linked to define the program entry point, where execution will begin when the program is run. If the object file that defines the entry point is assembled using NASM, you specify the entry point by declaring the special symbol \c{..start} at the point where you wish execution to begin. \S{objextern} \c{obj} Extensions to the \c{EXTERN} Directive\I{EXTERN, obj extensions to} If you declare an external symbol with the directive \c extern foo then references such as \c{mov ax,foo} will give you the offset of \c{foo} from its preferred segment base (as specified in whichever module \c{foo} is actually defined in). So to access the contents of \c{foo} you will usually need to do something like \c mov ax,seg foo ; get preferred segment base \c mov es,ax ; move it into ES \c mov ax,[es:foo] ; and use offset `foo' from it This is a little unwieldy, particularly if you know that an external is going to be accessible from a given segment or group, say \c{dgroup}. So if \c{DS} already contained \c{dgroup}, you could simply code \c mov ax,[foo wrt dgroup] However, having to type this every time you want to access \c{foo} can be a pain; so NASM allows you to declare \c{foo} in the alternative form \c extern foo:wrt dgroup This form causes NASM to pretend that the preferred segment base of \c{foo} is in fact \c{dgroup}; so the expression \c{seg foo} will now return \c{dgroup}, and the expression \c{foo} is equivalent to \c{foo wrt dgroup}. This \I{default-WRT mechanism}default-\c{WRT} mechanism can be used to make externals appear to be relative to any group or segment in your program. It can also be applied to common variables: see \k{objcommon}. \S{objcommon} \c{obj} Extensions to the \c{COMMON} Directive\I{COMMON, obj extensions to} The \c{obj} format allows common variables to be either near\I{near common variables} or far\I{far common variables}; NASM allows you to specify which your variables should be by the use of the syntax \c common nearvar 2:near ; `nearvar' is a near common \c common farvar 10:far ; and `farvar' is far Far common variables may be greater in size than 64Kb, and so the OMF specification says that they are declared as a number of \e{elements} of a given size. So a 10-byte far common variable could be declared as ten one-byte elements, five two-byte elements, two five-byte elements or one ten-byte element. Some \c{OMF} linkers require the \I{element size, in common variables}\I{common variables, element size}element size, as well as the variable size, to match when resolving common variables declared in more than one module. Therefore NASM must allow you to specify the element size on your far common variables. This is done by the following syntax: \c common c_5by2 10:far 5 ; two five-byte elements \c common c_2by5 10:far 2 ; five two-byte elements If no element size is specified, the default is 1. Also, the \c{FAR} keyword is not required when an element size is specified, since only far commons may have element sizes at all. So the above declarations could equivalently be \c common c_5by2 10:5 ; two five-byte elements \c common c_2by5 10:2 ; five two-byte elements In addition to these extensions, the \c{COMMON} directive in \c{obj} also supports default-\c{WRT} specification like \c{EXTERN} does (explained in \k{objextern}). So you can also declare things like \c common foo 10:wrt dgroup \c common bar 16:far 2:wrt data \c common baz 24:wrt data:6 \S{objdepend} Embedded File Dependency Information Since NASM 2.13.02, \c{obj} files contain embedded dependency file information. To suppress the generation of dependencies, use \c %pragma obj nodepend \H{obj2fmt} \i\c{obj2}: \i{OS/2 32-bit OMF}\I{OMF} Object Files The \c{obj2} output format is the same as \c{obj} except: \b Default attributes for a segment are \c{ALIGN=16} and \c{USE32}. \b All 32-bit segment is added to \c{FLAT} group implicitly. \b Support Unix sections such as \c{.text}, \c{.rodata}, \c{.data} and \c{.bss} for compatibility with other Unix platforms. And they are aliased to \c{TEXT32}, \c{CONST32}, \c{DATA32}, \c{BSS32}, respectively. \b Set default classes implicitly for known segments such as TEXT32, CONST32, DATA32, BSS32 and so on. The defaults assumed by NASM if you do not specify the qualifiers are: \c SECTION .text ALIGN=16 USE32 CLASS=CODE FLAT \c SECTION .rodata ALIGN=16 USE32 CLASS=CONST FLAT \c SECTION .data ALIGN=16 USE32 CLASS=DATA FLAT \c SECTION .bss ALIGN=16 USE32 CLASS=BSS FLAT \c SECTION CODE ALIGN=16 USE32 CLASS=CODE FLAT \c SECTION TEXT ALIGN=16 USE32 CLASS=CODE FLAT \c SECTION CONST ALIGN=16 USE32 CLASS=CONST FLAT \c SECTION DATA ALIGN=16 USE32 CLASS=DATA FLAT \c SECTION BSS ALIGN=16 USE32 CLASS=BSS FLAT \c SECTION STACK ALIGN=16 USE32 CLASS=STACK FLAT \c SECTION CODE32 ALIGN=16 USE32 CLASS=CODE FLAT \c SECTION TEXT32 ALIGN=16 USE32 CLASS=CODE FLAT \c SECTION CONST32 ALIGN=16 USE32 CLASS=CONST FLAT \c SECTION DATA32 ALIGN=16 USE32 CLASS=DATA FLAT \c SECTION BSS32 ALIGN=16 USE32 CLASS=BSS FLAT \c SECTION STACK32 ALIGN=16 USE32 CLASS=STACK FLAT \H{win32fmt} \i\c{win32}: Microsoft Win32 Object Files The \c{win32} output format generates Microsoft Win32 object files, suitable for passing to Microsoft linkers such as \i{Visual C++}. Note that Borland Win32 compilers do not use this format, but use \c{obj} instead (see \k{objfmt}). \c{win32} provides a default output file-name extension of \c{.obj}. Note that although Microsoft say that Win32 object files follow the \c{COFF} (Common Object File Format) standard, the object files produced by Microsoft Win32 compilers are not compatible with COFF linkers such as DJGPP's, and vice versa. This is due to a difference of opinion over the precise semantics of PC-relative relocations. To produce COFF files suitable for DJGPP, use NASM's \c{coff} output format; conversely, the \c{coff} format does not produce object files that Win32 linkers can generate correct output from. \S{win32sect} \c{win32} Extensions to the \c{SECTION} Directive\I{SECTION, Windows extensions to} Like the \c{obj} format, \c{win32} allows you to specify additional information on the \c{SECTION} directive line, to control the type and properties of sections you declare. Section types and properties are generated automatically by NASM for the \i{standard section names} \c{.text}, \c{.data} and \c{.bss}, but may still be overridden by these qualifiers. The available qualifiers are: \b \c{code}, or equivalently \c{text}, defines the section to be a code section. This marks the section as readable and executable, but not writable, and also indicates to the linker that the type of the section is code. \b \c{data} and \c{bss} define the section to be a data section, analogously to \c{code}. Data sections are marked as readable and writable, but not executable. \c{data} declares an initialized data section, whereas \c{bss} declares an uninitialized data section. \b \c{rdata} declares an initialized data section that is readable but not writable. Microsoft compilers use this section to place constants in it. \b \c{info} defines the section to be an \i{informational section}, which is not included in the executable file by the linker, but may (for example) pass information \e{to} the linker. For example, declaring an \c{info}-type section called \i\c{.drectve} causes the linker to interpret the contents of the section as command-line options. \b \c{align=}, used with a trailing number as in \c{obj}, gives the \I{section alignment, in win32}\I{alignment, in win32 sections}alignment requirements of the section. The maximum you may specify is 64: the Win32 object file format contains no means to request a greater section alignment than this. If alignment is not explicitly specified, the defaults are 16-byte alignment for code sections, 8-byte alignment for rdata sections and 4-byte alignment for data (and BSS) sections. Informational sections get a default alignment of 1 byte (no alignment), though the value does not matter. \b \I{comdat, win32 attribute}\c{comdat=}, followed by a number ("selection"), colon (acting as a separator) and a name, marks the section as a \I{COMDAT section, in win32}"COMDAT section". It allows Microsoft linkers to perform function-level linking, to deal with multiply defined symbols, to eliminate dead code/data. The "selection" number should be one of the \c{IMAGE_COMDAT_SELECT_*} constants from \W{https://github.com/MicrosoftDocs/win32/blob/docs/desktop-src/Debug/pe-format.md#comdat-sections-object-only}\c{COFF format specification}; this value controls if the linker allows multiply defined symbols and how it handles them. The name is the \I{COMDAT symbol, in win32}"COMDAT symbol" - basically a new name for the section. So even though you have one section given by the main name (e.g. \c{.text}), it can actually consist of hundreds of COMDAT sections having their own name (and alignment). When the "selection" is IMAGE_COMDAT_SELECT_ASSOCIATIVE (5), the following name is the "COMDAT symbol" of the associated COMDAT section; this way you can link a piece of code or data only when another piece of code or data gets actually linked. \> So, when linking a NASM-compiled file with some C code, the source may be structured as follows. Note that the default \c{.text} section in handled in a special way and it doesn't work well with \c{comdat}; you may want to append a \c{$} character and an arbitrary suffix to the section name. It will get linked into the \c{.text} section anyway - see the info on \W{https://github.com/MicrosoftDocs/win32/blob/docs/desktop-src/Debug/pe-format.md#grouped-sections-object-only}\c{Grouped Sections}. \c section .text$1 align=16 comdat=1:FirstFnc \c ... ; Code linked only if referenced from C \c \c section .text$1 align=16 comdat=1:SecondFnc \c ... ; Code linked only if referenced from C \c \c section .rdata align=32 comdat=5:FirstFnc \c ... ; Data linked only if the related code (FirstFnc) is linked \c The defaults assumed by NASM if you do not specify the above qualifiers are: \c section .text code align=16 \c section .data data align=4 \c section .rdata rdata align=8 \c section .bss bss align=4 The \c{win64} format also adds: \c section .pdata rdata align=4 \c section .xdata rdata align=8 Any other section name is treated by default like \c{.text}. \S{win32safeseh} \c{win32}: Safe Structured Exception Handling Among other improvements in Windows XP SP2 and Windows Server 2003, Microsoft has introduced the concept of "safe structured exception handling." The general idea is to collect handlers' entry points in a designated read-only table and have SEH entry points verified against this table before exception control is passed to the corresponding handler. In order for an executable module to be equipped with this read-only table, all object modules on linker command line have to comply with certain criteria. If even a single module among them does not, then the table in question is omitted and above mentioned run-time checks will not be performed for the application in question. Table omission is silent by default and therefore can be easily missed. One can instruct the linker to refuse to produce binary without such table by passing the \c{/safeseh} command line option. Without regard to this run-time check, it's natural to expect NASM to be capable of generating modules suitable for \c{/safeseh} linking. From the developer's viewpoint the problem is two-fold: \b how to adapt modules not deploying exception handlers of their own; \b how to adapt/develop modules utilizing custom exception handling; The former can be easily achieved with any NASM version by adding the following line to the source code: \c $@feat.00 equ 1 As of version 2.03 NASM adds this absolute symbol automatically, if it is not already present (in which case the developer can choose to assign another value, if desired, for whatever reason). Registering a custom exception handler on the other hand requires certain "magic." As of version 2.03, an additional \c{safeseh} directive is implemented, which instructs the assembler to produce appropriately formatted input data for the above-mentioned "safe exception handler table." Its typical use would be: \c section .text \c extern _MessageBoxA@16 \c %if __?NASM_VERSION_ID?__ >= 0x02030000 \c safeseh handler ; register handler as "safe handler" \c %endif \c handler: \c push DWORD 1 ; MB_OKCANCEL \c push DWORD caption \c push DWORD text \c push DWORD 0 \c call _MessageBoxA@16 \c sub eax,1 ; incidentally suits as return value \c ; for exception handler \c ret \c global _main \c _main: \c push DWORD handler \c push DWORD [fs:0] \c mov DWORD [fs:0],esp ; engage exception handler \c xor eax,eax \c mov eax,DWORD[eax] ; cause exception \c pop DWORD [fs:0] ; disengage exception handler \c add esp,4 \c ret \c text: db 'OK to rethrow, CANCEL to generate core dump',0 \c caption:db 'SEGV',0 \c \c section .drectve info \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib ' As you might imagine, it's perfectly possible to produce an .exe binary with the "safe exception handler table" and yet invoke an unregistered exception handler. A handler is invoked by manipulating \c{[fs:0]} at run-time, something the linker has no power over. It is therefore important to note that such failure to register a handler's entry point with the \c{safeseh} directive will have undesired side effects at run-time. If an exception is raised and an unregistered handler is to be executed, the application is abruptly terminated without any notification whatsoever. One can argue that the system should at least log some kind of "non-safe exception handler in x.exe at address n" message in the event log, but unfortunately the user is left without any clue as to what might have caused the crash. Finally, all mentions of linker in this paragraph refer to Microsoft linker version 7.x and later. Presence of \c{@feat.00} symbol and input data for "safe exception handler table" causes no backward incompatibilities and "safeseh" modules generated by NASM 2.03 and later can still be linked by earlier versions or non-Microsoft linkers. \S{codeview} Debugging formats for Windows \I{Windows debugging formats} The \c{win32} and \c{win64} formats support the Microsoft \i{CodeView debugging format}. Currently CodeView version 8 format is supported (\i\c{cv8}), but newer versions of the CodeView debugger should be able to handle this format as well. \H{win64fmt} \i\c{win64}: Microsoft Win64 Object Files The \c{win64} output format generates Microsoft Win64 object files, which is nearly 100% identical to the \c{win32} object format (\k{win32fmt}) with the exception that it is meant to target 64-bit code and the x86-64 platform altogether. This object file is used exactly the same as the \c{win32} object format (\k{win32fmt}), in NASM, with regard to this exception. \S{win64pic} \c{win64}: Writing Position-Independent Code While \c{REL} takes good care of RIP-relative addressing, there is one aspect that is easy to overlook for a Win64 programmer: indirect references. Consider a switch dispatch table: \c jmp qword [dsptch+rax*8] \c ... \c dsptch: dq case0 \c dq case1 \c ... Even a novice Win64 assembler programmer will soon realize that the code is not 64-bit savvy. Most notably the linker will refuse to link it, showing: \c 'ADDR32' relocation to '.text' invalid without /LARGEADDRESSAWARE:NO So [s]he will have to split jmp instruction as following: \c lea rbx,[rel dsptch] \c jmp qword [rbx+rax*8] What happens behind the scenes is that the effective address in \c{lea} is encoded relative to instruction pointer, in a perfectly position-independent manner. But this is only part of the problem! The issue is that in a .dll context, the \c{caseN} relocations will make their way to the final module and might have to be adjusted at .dll load time (specifically, when it can't be loaded at the preferred address). When this occurs, pages with such relocations will be rendered private to current process, which kind of undermines the idea of a shared .dll. But not to worry, it's trivial to fix: \c lea rbx,[rel dsptch] \c add rbx,[rbx+rax*8] \c jmp rbx \c ... \c dsptch: dq case0-dsptch \c dq case1-dsptch \c ... NASM version 2.03 and later provides another alternative, \c{wrt ..imagebase} operator, which returns an offset from base address of the current image, be it .exe or .dll module, hence the name. For those acquainted with PE-COFF format, this base address denotes the start of the \c{IMAGE_DOS_HEADER} structure. Here is how to implement a switch statement with these image-relative references: \c lea rbx,[rel dsptch] \c mov eax,[rbx+rax*4] \c sub rbx,dsptch wrt ..imagebase \c add rbx,rax \c jmp rbx \c ... \c dsptch: dd case0 wrt ..imagebase \c dd case1 wrt ..imagebase That said, the snippet before last works just fine with any NASM version and is not even Windows specific, which makes this operator unnecessary in this case. The real reason for the \c{wrt ..imagebase} operator will become apparent in the next section. It should be noted that \c{wrt ..imagebase} is defined as 32-bit operand only: \c dd label wrt ..imagebase ; ok \c dq label wrt ..imagebase ; bad \c mov eax,label wrt ..imagebase ; ok \c mov rax,label wrt ..imagebase ; bad \S{win64seh} \c{win64}: Structured Exception Handling Structured exception handing in Win64 is completely different compared to Win32. When an exception occurs, the program counter is noted, and a linker-generated table containing start and end addresses of all the functions (in a given executable module) is traversed and compared to the saved program counter. This is used to identify the corresponding \c{UNWIND_INFO} structure. If missing, then the offending subroutine is assumed to be "leaf" and this lookup procedure is instead attempted for its caller. In Win64, a leaf function is a function that does not call any other functions \e{nor} modifies any Win64 non-volatile registers, including the stack pointer. The latter ensures that it's possible to identify a leaf function's caller by simply pulling the value from the top of the stack. While the majority of subroutines written in assembler are not calling any other functions, they may not qualify as "leaf" functions in the Win64 sense. The requirement for non-volatile registers to be unchanged leaves the developer with not more than 7 registers and no stack frame, which is not necessarily what they counted on. Customarily one would meet this requirement by saving non-volatile registers on stack and restoring them upon return. However, if (and only if) an exception is raised at run-time and no \c{UNWIND_INFO} structure is associated with such a "leaf" function, the stack unwind procedure will expect to find the caller's return address on the top of the stack immediately followed by its frame. Given that the developer pushed the caller's non-volatile registers onto the stack, the value on top will no longer point to the right place. The developer can attempt to copy the caller's return address to the top of stack, which would work in some very specific circumstances. But unless the developer can guarantee that these circumstances are always met, it's more appropriate to assume the worst, i.e. the stack unwind procedure goes berserk, abruptly terminating without any notification whatsoever (just like in the the Win32 case). Now that we understand significance of the \c{UNWIND_INFO} structure, let us discuss what is in it and how it is processed. First, it is checked for the presence of a reference to a custom language-specific exception handler. If there is one, then it is invoked. Depending on the return value, execution flow is resumed (exception is said to be "handled"), \e{or} the rest of the \c{UNWIND_INFO} structure is processed as follows. Aside from an optional reference to a custom handler, it carries information about the current callee's stack frame and where non-volatile registers are saved. The information is detailed enough to be able to reconstruct the contents of the caller's non-volatile registers on entry to the current callee. And so the caller's context is reconstructed, at which point the unwind procedure is repeated, using the \c{UNWIND_INFO} structure associated with the caller's instruction pointer. The procedure is repeated recursively until the exception is handled. As a last resort, the system "handles" it by generating a memory dump and terminating the application. As of this writing, NASM unfortunately does not facilitate generation of above mentioned detailed information about stack frame layout. But as of version 2.03, it implements building blocks for generating structures involved in stack unwinding. Here is a simple example showing how to deploy a custom exception handler for a leaf function: \c default rel \c section .text \c extern MessageBoxA \c handler: \c sub rsp,40 \c mov rcx,0 \c lea rdx,[text] \c lea r8,[caption] \c mov r9,1 ; MB_OKCANCEL \c call MessageBoxA \c sub eax,1 ; incidentally suits as return value \c ; for exception handler \c add rsp,40 \c ret \c global main \c main: \c xor rax,rax \c mov rax,QWORD[rax] ; cause exception \c ret \c main_end: \c text: db 'OK to rethrow, CANCEL to generate core dump',0 \c caption:db 'SEGV',0 \c \c section .pdata rdata align=4 \c dd main wrt ..imagebase \c dd main_end wrt ..imagebase \c dd xmain wrt ..imagebase \c section .xdata rdata align=8 \c xmain: db 9,0,0,0 \c dd handler wrt ..imagebase \c section .drectve info \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib ' What you see is that the \c{.pdata} section contains a single-element table, containing function start and end addresses, along with references to associated \c{UNWIND_INFO} structures (only one in this case). The \c{.xdata} section contains the referenced \c{UNWIND_INFO} structure, describing a function with no frame, but with a designated exception handler. These references are \e{required} to be image-relative, which is the real reason for implementing the \c{wrt ..imagebase} operator). It should be noted that \c{rdata align=n}, as well as \c{wrt ..imagebase}, are actually optional in the context of these two segments (they apply even when omitted); \e{all} 32-bit references placed into these two segments will be image-relative. This is important to understand, as the developer is allowed to append handler-specific data to the \c{UNWIND_INFO} structure, and any 32-bit references that are added may require adjustment to obtain the real pointer. As already mentioned, in Win64 terms, a leaf function is one that neither calls any other function \e{nor} modifies any non-volatile registers, including the stack pointer. But it is not uncommon for the programmer to intend to utilize every single register and sometimes even have a variable stack frame, requiring a more complicated \c{UNWIND_INFO} structure than in the example above. Is there anything one can do with these simpler building blocks, and avoid manually composing fully-fledged \c{UNWIND_INFO} structures, which would surely be considered error-prone? Yes, there is. Recall that an exception handler is called first, before the stack layout is analyzed. As it turns out, it is perfectly possible to manipulate current callee's context in a custom handler in a manner that permits further stack unwinding. The general idea is that handler would not actually "handle" the exception, but instead restore the callee's context (restore to state at entry point) and thus mimic a Win64 leaf function. In other words, the handler would effectively undertake part of the unwinding procedure. Consider the following example: \c function: \c mov rax,rsp ; copy rsp to volatile register \c push r15 ; save non-volatile registers \c push rbx \c push rbp \c mov r11,rsp ; prepare variable stack frame \c sub r11,rcx \c and r11,-64 \c mov QWORD[r11],rax ; check for exceptions \c mov rsp,r11 ; allocate stack frame \c mov QWORD[rsp],rax ; save original rsp value \c magic_point: \c ... \c mov r11,QWORD[rsp] ; pull original rsp value \c mov rbp,QWORD[r11-24] \c mov rbx,QWORD[r11-16] \c mov r15,QWORD[r11-8] \c mov rsp,r11 ; destroy frame \c ret The key is that until \c{magic_point}, the original \c{rsp} value remains in the chosen volatile register, and no non-volatile register except for \c{rsp} is modified. After \c{magic_point}, \c{rsp} remains constant till the very end of the \c{function}. In this case a custom language-specific exception handler would look like this: \c EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame, \c CONTEXT *context,DISPATCHER_CONTEXT *disp) \c { ULONG64 *rsp; \c if (context->Rip<(ULONG64)magic_point) \c rsp = (ULONG64 *)context->Rax; \c else \c { rsp = ((ULONG64 **)context->Rsp)[0]; \c context->Rbp = rsp[-3]; \c context->Rbx = rsp[-2]; \c context->R15 = rsp[-1]; \c } \c context->Rsp = (ULONG64)rsp; \c \c memcpy (disp->ContextRecord,context,sizeof(CONTEXT)); \c RtlVirtualUnwind(UNW_FLAG_NHANDLER,disp->ImageBase, \c dips->ControlPc,disp->FunctionEntry,disp->ContextRecord, \c &disp->HandlerData,&disp->EstablisherFrame,NULL); \c return ExceptionContinueSearch; \c } As this custom handler allows the example function to mimic a Win64 leaf function, the corresponding \c{UNWIND_INFO} structure does not need to contain any information about the stack frame and its layout. \H{cofffmt} \i\c{coff}: \i{Common Object File Format} The \c{coff} output type produces \c{COFF} object files suitable for linking with the \i{DJGPP} linker. \c{coff} provides a default output file-name extension of \c{.o}. The \c{coff} format supports the same extensions to the \c{SECTION} directive as \c{win32} does, except that the \c{align} qualifier and the \c{info} section type are not supported. \H{machofmt} \I{Mach-O}\i\c{macho32} and \i\c{macho64}: \i{Mach Object File Format} The \c{macho32} and \c{macho64} output formts produces Mach-O object files suitable for linking with the \i{MacOS X} linker. \i\c{macho} is a synonym for \c{macho32}. \c{macho} provides a default output file-name extension of \c{.o}. \S{machosect} \c{macho} extensions to the \c{SECTION} Directive \I{SECTION, macho extensions to} The \c{macho} output format specifies section names in the format "\e{segment}\c{,}\e{section}". No spaces are allowed around the comma. The following flags can also be specified: \b \c{data} - this section contains initialized data items \b \c{code} - this section contains code exclusively \b \c{mixed} - this section contains both code and data \b \c{bss} - this section is uninitialized and filled with zero \b \c{zerofill} - same as \c{bss} \b \c{no_dead_strip} - inhibit dead code stripping for this section \b \c{live_support} - set the live support flag for this section \b \c{strip_static_syms} - strip static symbols for this section \b \c{debug} - this section contains debugging information \b \c{align=}\e{alignment} - specify section alignment The default is \c{data}, unless the section name is \c{__text} or \c{__bss} in which case the default is \c{text} or \c{bss}, respectively. For compatibility with other Unix platforms, the following standard names are also supported: \c .text = __TEXT,__text text \c .rodata = __DATA,__const data \c .data = __DATA,__data data \c .bss = __DATA,__bss bss If the \c{.rodata} section contains no relocations, it is instead put into the \c{__TEXT,__const} section unless this section has already been specified explicitly. However, it is probably better to specify \c{__TEXT,__const} and \c{__DATA,__const} explicitly as appropriate. \S{machotls} \i{Thread Local Storage in Mach-O}\I{TLS}: \c{macho} special symbols and \i\c{WRT} Mach-O defines the following special symbols that can be used on the right-hand side of the \c{WRT} operator: \b \c{..tlvp} is used to specify access to thread-local storage. \b \c{..gotpcrel} is used to specify references to the Global Offset Table. The GOT is supported in the \c{macho64} format only. \S{macho-ssvs} \c{macho} specific directive \i\c{subsections_via_symbols} The directive \c{subsections_via_symbols} sets the \c{MH_SUBSECTIONS_VIA_SYMBOLS} flag in the Mach-O header, that effectively separates a block (or a subsection) based on a symbol. It is often used for eliminating dead codes by a linker. This directive takes no arguments. This is a macro implemented as a \c{%pragma}. It can also be specified in its \c{%pragma} form, in which case it will not affect non-Mach-O builds of the same source code: \c %pragma macho subsections_via_symbols \S{macho-ssvs} \c{macho} specific directive \i\c{no_dead_strip} The directive \c{no_dead_strip} sets the Mach-O \c{SH_NO_DEAD_STRIP} section flag on the section containing a a specific symbol. This directive takes a list of symbols as its arguments. This is a macro implemented as a \c{%pragma}. It can also be specified in its \c{%pragma} form, in which case it will not affect non-Mach-O builds of the same source code: \c %pragma macho no_dead_strip symbol... \S{macho-pext} \c{macho} specific extensions to the \c{GLOBAL} Directive: \i\c{private_extern} The directive extension to \c{GLOBAL} marks the symbol with limited global scope. For example, you can specify the global symbol with this extension: \c global foo:private_extern \c foo: \c ; codes Using with static linker will clear the private extern attribute. But linker option like \c{-keep_private_externs} can avoid it. \S{macho-bver} \c{macho} specific directive \i\c{build_version} The directive \c{build_version} generates a \c{LC_BUILD_VERSION} load command in the Mach-O header, which allows specifying a target platform, minimum OS version and optionally SDK version. Newer Xcode linker versions warn if this is not present in object files. This directive takes the target platform name and minimum OS version as arguments, in this form: \c build_version macos,10,7 Platform names that make sense for x86 code are \c{macos}, \c{iossimulator}, \c{tvossimulator} and \c{watchossimulator}. Optionally, a trailing version number and minimum SDK version can also be specified with this syntax: \c build_version macos, 10, 14, 0 sdk_version 10, 14, 0 This is a macro implemented as a \c{%pragma}. It can also be specified in its \c{%pragma} form, in which case it will not affect non-Mach-O builds of the same source code: \c %pragma macho build_version ... This latter form is also useful on the command line when using the \c{--pragma} command-line switch: \c nasm -f macho64 --pragma "macho build_version macos,10,9" ... \H{elffmt} \i\c{elf32}, \i\c{elf64}, \i\c{elfx32}: \I{ELF}\I{linux, elf}Executable and Linkable Format Object Files The \c{elf32}, \c{elf64} and \c{elfx32} output formats generate \c{ELF32} and \c{ELF64} (Executable and Linkable Format) object files, as used by \i{Linux} as well as \i{Unix System V}, including \i{Solaris x86}, \i{UnixWare} and \i{SCO Unix}. ELF provides a default output file-name extension of \c{.o}. \c{elf} is a synonym for \c{elf32}. The \c{elfx32} file format is an ELF32 file containing 64-bit x86 code, and is used for the \i{x32} ABI, which runs the CPU in 64-bit mode while using 32-bit values for pointers to reduce memory footprint. Thus, code intended to be used with the x32 ABI should be assembled with \c{BITS 64}. \S{abisect} ELF specific directive \i\c{osabi} The ELF header specifies the application binary interface for the target operating system (OSABI). This field can be set by using the \c{osabi} directive with the numeric value (0-255) of the target system. If this directive is not used, the default value will be "UNIX System V ABI" (0) which will work on most systems which support ELF. \S{elfsect} ELF extensions to the \c{SECTION} Directive \I{SECTION, ELF extensions to} Like the \c{obj} format, \c{elf} allows you to specify additional information on the \c{SECTION} directive line, to control the type and properties of sections you declare. Section types and properties are generated automatically by NASM for the \i{standard section names}, but may still be overridden by these qualifiers. The available qualifiers are: \b \i\c{alloc} defines the section to be one which is loaded into memory when the program is run. \i\c{noalloc} defines it to be one which is not, such as an informational or comment section. \b \i\c{exec} defines the section to be one which should have execute permission when the program is run. \i\c{noexec} defines it as one which should not. \b \i\c{write} defines the section to be one which should be writable when the program is run. \i\c{nowrite} defines it as one which should not. \b \i\c{progbits} defines the section to be one with explicit contents stored in the object file: an ordinary code or data section, for example. \b \i\c{nobits} defines the section to be one with no explicit contents given, such as a BSS section. \b \i\c{note} indicates that this section contains ELF notes. The content of ELF notes are specified using normal assembly instructions; it is up to the programmer to ensure these are valid ELF notes. \b \i\c{preinit_array} indicates that this section contains function addresses to be called before any other initialization has happened. \b \i\c{init_array} indicates that this section contains function addresses to be called during initialization. \b \i\c{fini_array} indicates that this section contains function pointers to be called during termination. \b \I{align, ELF attribute}\c{align=}, used with a trailing number as in \c{obj}, gives the \I{section alignment, in elf}\I{alignment, in elf sections}alignment requirements of the section. \b \c{byte}, \c{word}, \c{dword}, \c{qword}, \c{tword}, \c{oword}, \c{yword}, or \c{zword} with an optional \c{*}\i{multiplier} specify the fundamental data item size for a section which contains either fixed-sized data structures or strings; it also sets a default alignment. This is generally used with the \c{strings} and \c{merge} attributes (see below.) For example \c{byte*4} defines a unit size of 4 bytes, with a default alignment of 1; \c{dword} also defines a unit size of 4 bytes, but with a default alignment of 4. The \c{align=} attribute, if specified, overrides this default alignment. \b \I{pointer, ELF attribute}\c{pointer} is equivalent to \c{dword} for \c{elf32} or \c{elfx32}, and \c{qword} for \c{elf64}. \b \I{strings, ELF attribute}\c{strings} indicate that this section contains exclusively null-terminated strings. By default these are assumed to be byte strings, but a size specifier can be used to override that. \b \i\c{merge} indicates that duplicate data elements in this section should be merged with data elements from other object files. Data elements can be either fixed-sized objects or null-terminated strings (with the \c{strings} attribute). A size specifier is required unless \c{strings} is specified, in which case the size defaults to \c{byte}. \b \i\c{tls} defines the section to be one which contains thread local variables. The defaults assumed by NASM if you do not specify the above qualifiers are: \I\c{.text} \I\c{.rodata} \I\c{.lrodata} \I\c{.data} \I\c{.ldata} \I\c{.bss} \I\c{.lbss} \I\c{.tdata} \I\c{.tbss} \I\c\{.comment} \c section .text progbits alloc exec nowrite align=16 \c section .rodata progbits alloc noexec nowrite align=4 \c section .lrodata progbits alloc noexec nowrite align=4 \c section .data progbits alloc noexec write align=4 \c section .ldata progbits alloc noexec write align=4 \c section .bss nobits alloc noexec write align=4 \c section .lbss nobits alloc noexec write align=4 \c section .tdata progbits alloc noexec write align=4 tls \c section .tbss nobits alloc noexec write align=4 tls \c section .comment progbits noalloc noexec nowrite align=1 \c section .preinit_array preinit_array alloc noexec nowrite pointer \c section .init_array init_array alloc noexec nowrite pointer \c section .fini_array fini_array alloc noexec nowrite pointer \c section .note note noalloc noexec nowrite align=4 \c section other progbits alloc noexec nowrite align=1 (Any section name other than those in the above table is treated by default like \c{other} in the above table. Please note that section names are case sensitive.) \S{elfwrt} \i{Position-Independent Code}\I{PIC}: ELF Special Symbols and \i\c{WRT} Since \c{ELF} does not support segment-base references, the \c{WRT} operator is not used for its normal purpose; therefore NASM's \c{elf} output format makes use of \c{WRT} for a different purpose, namely the PIC-specific \I{relocations, PIC-specific}relocation types. \c{elf} defines five special symbols which you can use as the right-hand side of the \c{WRT} operator to obtain PIC relocation types. They are \i\c{..gotpc}, \i\c{..gotoff}, \i\c{..got}, \i\c{..plt} and \i\c{..sym}. Their functions are summarized here: \b Referring to the symbol marking the global offset table base using \c{wrt ..gotpc} will end up giving the distance from the beginning of the current section to the global offset table. (\i\c{_GLOBAL_OFFSET_TABLE_} is the standard symbol name used to refer to the \i{GOT}.) So you would then need to add \i\c{$$} to the result to get the real address of the GOT. \b Referring to a location in one of your own sections using \c{wrt ..gotoff} will give the distance from the beginning of the GOT to the specified location, so that adding on the address of the GOT would give the real address of the location you wanted. \b Referring to an external or global symbol using \c{wrt ..got} causes the linker to build an entry \e{in} the GOT containing the address of the symbol, and the reference gives the distance from the beginning of the GOT to the entry; so you can add on the address of the GOT, load from the resulting address, and end up with the address of the symbol. \b Referring to a procedure name using \c{wrt ..plt} causes the linker to build a \i{procedure linkage table} entry for the symbol, and the reference gives the address of the \i{PLT} entry. You can only use this in contexts which would generate a PC-relative relocation normally (i.e. as the destination for \c{CALL} or \c{JMP}), since ELF contains no relocation type to refer to PLT entries absolutely. \b Referring to a symbol name using \c{wrt ..sym} causes NASM to write an ordinary relocation, but instead of making the relocation relative to the start of the section and then adding on the offset to the symbol, it will write a relocation record aimed directly at the symbol in question. The distinction is a necessary one due to a peculiarity of the dynamic linker. A fuller explanation of how to use these relocation types to write shared libraries entirely in NASM is given in \k{picdll}. \S{elftls} \i{Thread Local Storage in ELF}\I{TLS}: \c{elf} Special Symbols and \i\c{WRT} \b In ELF32 mode, referring to an external or global symbol using \c{wrt ..tlsie} \I\c{..tlsie} causes the linker to build an entry \e{in} the GOT containing the offset of the symbol within the TLS block, so you can access the value of the symbol with code such as: \c mov eax,[tid wrt ..tlsie] \c mov [gs:eax],ebx \b In ELF64 or ELFx32 mode, referring to an external or global symbol using \c{wrt ..gottpoff} \I\c{..gottpoff} causes the linker to build an entry \e{in} the GOT containing the offset of the symbol within the TLS block, so you can access the value of the symbol with code such as: \c mov rax,[rel tid wrt ..gottpoff] \c mov rcx,[fs:rax] \S{elfglob} \c{elf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL, elf extensions to}\I{GLOBAL, aoutb extensions to} \c{ELF} object files can contain more information about a global symbol than just its address: they can contain the \I{symbols, specifying sizes}\I{size, of symbols}size of the symbol and its \I{symbols, specifying types}\I{type, of symbols}type as well. These are not merely debugger conveniences, but are actually necessary when the program being written is a \I{elf shared library}shared library. NASM therefore supports some extensions to the \c{GLOBAL} directive, allowing you to specify these features. You can specify whether a global variable is a function or a data object by suffixing the name with a colon and the word \i\c{function} or \i\c{data}. (\i\c{object} is a synonym for \c{data}.) For example: \c global hashlookup:function, hashtable:data exports the global symbol \c{hashlookup} as a function and \c{hashtable} as a data object. Optionally, you can control the ELF visibility of the symbol. Just add one of the \I{elf visibility}visibility keywords: \I{default, elf}\c{default}, \I{internal, elf}\c{internal}, \I{hidden, elf}\c{hidden}, or \I{protected, elf}\c{protected}. The default is \c{default} of course. For example, to make \c{hashlookup} hidden: \c global hashlookup:function hidden Since version 2.15, it is possible to specify symbols binding. The keywords are: \i\c{weak} to generate weak symbol or \i\c{strong}. The default is \i\c{strong}. You can also specify the size of the data associated with the symbol, as a numeric expression (which may involve labels, and even forward references) after the type specifier. Like this: \c global hashtable:data (hashtable.end - hashtable) \c \c hashtable: \c db this,that,theother ; some data here \c .end: This makes NASM automatically calculate the length of the table and place that information into the \c{ELF} symbol table. Declaring the type and size of global symbols is necessary when writing shared library code. For more information, see \k{picglobal}. \S{elfextrn} \c{elf} Extensions to the \c{EXTERN} Directive\I{EXTERN, elf extensions to}\I{EXTERN, elf extensions to} Since version 2.15 it is possible to specify keyword \i\c{weak} to generate weak external reference. Example: \c extern weak_ref:weak \S{elfcomm} \c{elf} Extensions to the \c{COMMON} Directive \I{COMMON, elf extensions to} \c{ELF} also allows you to specify alignment requirements \I{common variables, alignment in elf}\I{alignment, of elf common variables}on common variables. This is done by putting a number (which must be a power of two) after the name and size of the common variable, separated (as usual) by a colon. For example, an array of doublewords would benefit from 4-byte alignment: \c common dwordarray 128:4 This declares the total size of the array to be 128 bytes, and requires that it be aligned on a 4-byte boundary. \S{elf16} 16-bit code and ELF \I{ELF, 16-bit code} Older versions of the \c{ELF32} specification did not provide relocations for 8- and 16-bit values. It is now part of the formal specification, and any new enough linker should support them. ELF has currently no support for segmented programming. \S{elfdbg} Debug formats and ELF \I{ELF, debug formats} ELF provides debug information in \c{STABS} and \c{DWARF} formats. Line number information is generated for all executable sections, but please note that only the ".text" section is executable by default. \H{aoutfmt} \i\c{aout}: Linux \I{a.out, Linux version}\I{linux, a.out}\c{a.out} Object Files The \c{aout} format generates \c{a.out} object files, in the form used by early \i{Linux} systems (current Linux systems use ELF, see \k{elffmt}.) These differ from other \c{a.out} object files in that the magic number in the first four bytes of the file is different; also, some implementations of \c{a.out}, for example NetBSD's, support position-independent code, which Linux's implementation does not. \c{a.out} provides a default output file-name extension of \c{.o}. \c{a.out} is a very simple object format. It supports no special directives, no special symbols, no use of \c{SEG} or \c{WRT}, and no extensions to any standard directives. It supports only the three \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}. \H{aoutfmt} \i\c{aoutb}: \i{NetBSD}/\i{FreeBSD}/\i{OpenBSD} \I{a.out, BSD version}\c{a.out} Object Files The \c{aoutb} format generates \c{a.out} object files, in the form used by the various free \c{BSD Unix} clones, \c{NetBSD}, \c{FreeBSD} and \c{OpenBSD}. For simple object files, this object format is exactly the same as \c{aout} except for the magic number in the first four bytes of the file. However, the \c{aoutb} format supports \I{PIC}\i{position-independent code} in the same way as the \c{elf} format, so you can use it to write \c{BSD} \i{shared libraries}. \c{aoutb} provides a default output file-name extension of \c{.o}. \c{aoutb} supports no special directives, no special symbols, and only the three \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}. However, it also supports the same use of \i\c{WRT} as \c{elf} does, to provide position-independent code relocation types. See \k{elfwrt} for full documentation of this feature. \c{aoutb} also supports the same extensions to the \c{GLOBAL} directive as \c{elf} does: see \k{elfglob} for documentation of this. \H{as86fmt} \c{as86}: \i{Minix}/Linux\I{linux, as86} \i\c{as86} Object Files The Minix/Linux 16-bit assembler \c{as86} has its own non-standard object file format. Although its companion linker \i\c{ld86} produces something close to ordinary \c{a.out} binaries as output, the object file format used to communicate between \c{as86} and \c{ld86} is not itself \c{a.out}. NASM supports this format, just in case it is useful, as \c{as86}. \c{as86} provides a default output file-name extension of \c{.o}. \c{as86} is a very simple object format (from the NASM user's point of view). It supports no special directives, no use of \c{SEG} or \c{WRT}, and no extensions to any standard directives. It supports only the three \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}. The only special symbol supported is \c{..start}. \H{dbgfmt} \i\c{dbg}: Debugging Format The \c{dbg} format does not output an object file as such; instead, it outputs a text file which contains a complete list of all the transactions between the main body of NASM and the output-format back end module. It is primarily intended to aid people who want to write their own output drivers, so that they can get a clearer idea of the various requests the main program makes of the output driver, and in what order they happen. For simple files, one can easily use the \c{dbg} format like this: \c nasm -f dbg filename.asm which will generate a diagnostic file called \c{filename.dbg}. However, this will not work well on files which were designed for a different object format, because each object format defines its own macros (usually user-level forms of directives), and those macros will not be defined in the \c{dbg} format. Therefore it can be useful to run NASM twice, in order to do the preprocessing with the native object format selected: \c nasm -e -f elf32 -o elfprog.i elfprog.asm \c nasm -a -f dbg elfprog.i This preprocesses \c{elfprog.asm} into \c{elfprog.i}, keeping the \c{elf32} object format selected in order to make sure ELF special directives are converted into primitive form correctly. Then the preprocessed source is fed through the \c{dbg} format to generate the final diagnostic output. This workaround will still typically not work for programs intended for \c{obj} format, because the \c{obj} \c{SEGMENT} and \c{GROUP} directives have side effects of defining the segment and group names as symbols; \c{dbg} will not do this, so the program will not assemble. You will have to work around that by defining the symbols yourself (using \c{EXTERN}, for example) if you really need to get a \c{dbg} trace of an \c{obj}-specific source file. \c{dbg} accepts any section name and any directives at all, and logs them all to its output file. \c{dbg} accepts and logs any \c{%pragma}, but the specific \c{%pragma}: \c %pragma dbg maxdump where \c{} is either a number or \c{unlimited}, can be used to control the maximum size for dumping the full contents of a \c{rawdata} output object.