all [group] all possible warnings \c{all} is an group alias for \e{all} warning classes. Thus, \c{-w+all} enables all available warnings, and \c{-w-all} disables warnings entirely (since NASM 2.13). db-empty [on] no operand for data declaration Warns about a \c{D}\e{x} declaration with no operands, producing no output. This is permitted, but often indicative of an error. See \k{db}. ea-absolute [on] absolute address cannot be RIP-relative Warns that an address that is inherently absolute cannot be generated with RIP-relative encoding using \c{REL}, see \k{default-rel}. ea-dispsize [on] displacement size ignored on absolute address Warns that NASM does not support generating displacements for inherently absolute addresses that do not match the address size of the instruction. float-denorm [off] floating point denormal Warns about denormal floating point constants. float-overflow [on] floating point overflow Warns about floating point underflow. float-toolong [on] too many digits in floating-point number Warns about too many digits in floating-point numbers. float-underflow [off] floating point underflow Warns about floating point underflow (a nonzero constant rounded to zero.) forward [on] forward reference may have unpredictable results Warns that a forward reference is used which may have unpredictable results, notably in a \c{RESB}-type pseudo-instruction. These would be \i\e{critical expressions} (see \k{crit}) but are permitted in a handful of cases for compatibility with older versions of NASM. This warning should be treated as a severe programming error as the code could break at any time for any number of reasons. implicit-abs-deprecated [on] implicit DEFAULT ABS is deprecated Warns that in a future version of NASM, the 64-bit default addressing form is likely to change from \c{DEFAULT ABS} to \c{DEFAULT REL}. If absolute addressing is indeed intended, it is strongly recommended to specify \c{DEFAULT ABS} explicitly. label-orphan [on] labels alone on lines without trailing \c{:} =orphan-labels Warns about source lines which contain no instruction but define a label without a trailing colon. This is most likely indicative of a typo, but is technically correct NASM syntax (see \k{syntax}.) label-redef [off] label redefined to an identical value Warns if a label is defined more than once, but the value is identical. It is an unconditional error to define the same label more than once to \e{different} values. # # label-redef-late defaults to an error, as this should never # actually happen. Just in case this is a backwards # compatibility problem, still make it a warning so that the # user can suppress or demote it. # label-redef-late [err] label (re)defined during code generation The value of a label changed during the final, code-generation pass. This may be the result of strange use of the preprocessor. This is very likely to produce incorrect code and may end up being an unconditional error in a future version of NASM. number-deprecated-hex [on] $ prefix for hexadecimal is deprecated Warns that the \c{$} prefix for hexadecimal numbers is deprecated, due to the syntactic conflict with \c{$} used as a symbol escape prefix. This syntax may be disabled by default in a future version of NASM. Replace \c{$} with \c{0x} to ensure compatibility with future versions. number-overflow [on] numeric constant does not fit Covers warnings about numeric constants which don't fit in 64 bits. obsolete-nop [on] instruction obsolete and is a noop on the target CPU Warns for an instruction which has been removed from the architecture, but has been architecturally defined to be a noop for future CPUs. obsolete-removed [on] instruction obsolete and removed on the target CPU Warns for an instruction which has been removed from the architecture, and is no longer included in the CPU definition given in the \c{[CPU]} directive, for example \c{POP CS}, the opcode for which, \c{0Fh}, instead is an opcode prefix on CPUs newer than the first generation 8086. obsolete-valid [on] instruction obsolete but valid on the target CPU Warns for an instruction which has been removed from the architecture, but is still valid on the specific CPU given in the \c{CPU} directive. Code using these instructions is most likely not forward compatible. other [on] any warning not assigned to a specific warning class Specifies any warning not included in any specific warning class. phase [off] phase error during stabilization Warns about symbols having changed values during the second-to-last assembly pass. This is not inherently fatal, but may be a source of bugs. pp-else-elif [on] \c{%elif} after \c{%else} Warns that an \c{%elif}-type directive was encountered after \c{%else} has already been encounted. As a result, the content of the \c{%elif} will never be expanded. pp-else-else [on] \c{%else} after \c{%else} Warns that a second \c{%else} clause was found for the same \c{%if} statement. The content of this \c{%else} clause will never be expanded. pp-empty-braces [on] empty \c{%\{\}} construct Warns that an empty \c{%\{\}} was encountered. This expands to a single \c{%} character, which is normally the \c{%} arithmetic operator. pp-environment [on] nonexistent environment variable =environment Warns if a nonexistent environment variable is accessed using the \c{%!} preprocessor construct (see \k{getenv}.) Such environment variables are treated as empty (with this warning issued) starting in NASM 2.15; earlier versions of NASM would treat this as an error. pp-macro-def-case-single [on] single-line macro defined both case sensitive and insensitive =macro-def-case-single Warns when a single-line macro is defined both case sensitive and case insensitive. The new macro definition will override (shadow) the original one, although the original macro is not deleted, and will be re-exposed if the new macro is deleted with \c{%undef}, or, if the original macro is the case insensitive one, the macro call is done with a different case. pp-macro-def-greedy-single [on] single-line macro =macro-def-greedy-single Warns that a single-line macro is defined which would match a previously existing greedy definition. The new macro definition will override (shadow) the original one, although the original macro is not deleted, and will be re-exposed if the new macro is deleted with \c{%undef}, and will be invoked if called with a parameter count that does not match the new definition. # The immediately previous versions of NASM considered # this an error, so promote this warning is promoted to # to error by default. pp-macro-def-param-single [err] single-line macro defined with and without parameters =macro-def-param-single Warns if the same single-line macro is defined with and without parameters. The new macro definition will override (shadow) the original one, although the original macro is not deleted, and will be re-exposed if the new macro is deleted with \c{%undef}. pp-macro-defaults [on] macros with more default than optional parameters =macro-defaults Warns when a macro has more default parameters than optional parameters. See \k{mlmacdef} for why one might want to disable this warning. pp-macro-params-legacy [on] improperly calling multi-line macro for legacy support =macro-params-legacy Warns about \i{multi-line macros} being invoked with the wrong number of parameters, but for bug-compatibility with NASM versions older than 2.15, NASM tried to fix up the parameters to match the legacy behavior and call the macro anyway. This can happen in certain cases where there are empty arguments without braces, sometimes as a result of macro expansion. The legacy behavior is quite strange and highly context-dependent, and can be disabled with: \c %pragma preproc sane_empty_expansion true It is highly recommended to use this option in new code. pp-macro-params-multi [on] multi-line macro calls with wrong parameter count =macro-params-multi Warns about \i{multi-line macros} being invoked with the wrong number of parameters. See \k{mlmacover} for an example of why you might want to disable this warning. pp-macro-params-single [on] single-line macro calls with wrong parameter count =macro-params-single Warns about \i{single-line macros} being invoked with the wrong number of parameters. pp-macro-redef-multi [on] redefining multi-line macro Warns that a multi-line macro is being redefined, without first removing the old definition with \c{%unmacro}. pp-open-braces [on] unterminated \c{%\{...\}} Warns that a preprocessor parameter enclosed in braces \c{%\{...\}} lacks the terminating \c{\}} character. pp-open-brackets [on] unterminated \c{%[...]} Warns that a preprocessor \c{%[...]} construct lacks the terminating \c{]} character. pp-open-string [on] unterminated string Warns that a quoted string without a closing quotation mark was encountered during preprocessing. pp-rep-negative [on] regative \c{%rep} count =negative-rep Warns about a negative count given to the \c{%rep} preprocessor directive. pp-sel-range [on] \c{%sel()} argument out of range Warns that the \c{%sel()} preprocessor function was passed a value less than 1 or larger than the number of available arguments. pp-trailing [on] trailing garbage ignored Warns that the preprocessor encountered additional text where no such text was expected. This can sometimes be the result of an incorrectly written expression, or arguments that are inadvertently separated. pragma-bad [off] malformed \c{%pragma} =bad-pragma Warns about a malformed or otherwise unparsable \c{%pragma} directive. pragma-empty [off] empty \c{%pragma} directive Warns about a \c{%pragma} directive containing nothing. This is treated identically to \c{%pragma ignore} except for this optional warning. # Not implemented yet pragma-na [off] \c{%pragma} not applicable to this compilation =not-my-pragma Warns about a \c{%pragma} directive which is not applicable to this particular assembly session. This is not yet implemented. pragma-unknown [off] unknown \c{%pragma} facility or directive =unknown-pragma Warns about an unknown \c{%pragma} directive. This is not yet implemented for most cases. prefix-bnd [on] invalid \c{BND} prefix =bnd Warns about ineffective use of the \c{BND} prefix when the \c{JMP} instruction is converted to the \c{SHORT} form. This should be extremely rare since the short \c{JMP} only is applicable to jumps inside the same module, but if it is legitimate, it may be necessary to use \c{bnd jmp dword}. prefix-hint-dropped [on] invalid branch hint prefix dropped Warns that the \c{{PT}} (predict taken) or \c{{PN}} (predict not taken) branch prediction hint prefixes are specified on an instruction that does not take these prefixes. As these prefixes alias the segment override prefixes, this may be a very serious error, and therefore NASM will not generate these prefixes. To force these prefixes to be emitted, use \c{DS} or \c{CS}, instead, respectively. prefix-hle [on] invalid HLE prefix =hle Warns about invalid use of the HLE \c{XACQUIRE} or \c{XRELEASE} prefixes. prefix-invalid [on] invalid prefix for instruction Warns about an instruction which is only valid with certain combinations of prefixes. The prefix will still be generated as requested, but the result may be a completely different instruction or result in a \c{#UD} trap. prefix-lock-error [on] \c{LOCK} prefix on unlockable instruction =lock Warns about \c{LOCK} prefixes specified on unlockable instructions. prefix-lock-xchg [on] superfluous \c{LOCK} prefix on \c{XCHG} instruction Warns about a \c{LOCK} prefix added to an \c{XCHG} instruction. The \c{XCHG} instruction is \e{always} locking, and so this prefix is not necessary; however, NASM will generate it if explicitly provided by the user, so this warning indicates that suboptimal code is being generated. prefix-opsize [on] invalid operand size prefix Warns that an operand prefix (\c{o16}, \c{o32}, \c{o64}, \c{osp}) invalid for the specified instruction has been specified. The operand prefix will be ignored by the assembler. prefix-seg [on] segment prefix ignored in 64-bit mode Warns that an \c{es}, \c{cs}, \c{ss} or \c{ds} segment override prefix has no effect in 64-bit mode. The prefix will still be generated as requested. ptr [on] non-NASM keyword used in other assemblers Warns about keywords used in other assemblers that might indicate a mistake in the source code. Currently only the MASM \c{PTR} keyword is recognized. If (limited) MASM compatibility is desired, the \c{%use masm} macro package is available, see \k{pkg_masm}; however, carefully note the caveats listed. regsize [on] register size specification ignored Warns about a register with implicit size (such as \c{EAX}, which is always 32 bits) been given an explicit size specification which is inconsistent with the size of the named register, e.g. \c{WORD EAX}. \c{DWORD EAX} or \c{WORD AX} are permitted, and do not trigger this warning. Some registers which \e{do not} imply a specific size, such as \c{K0}, may need this specification unless the instruction itself implies the instruction size: \c KMOVW K0,[foo] ; OK: KMOVW = 16 bits \c KMOV WORD K0,[foo] ; OK: WORD K0 = 16 bits \c KMOV K0,WORD [foo] ; OK: WORD [foo] = 16 bits \c KMOV K0,[foo] ; Error: unknown size reloc-abs-byte [off] 8-bit absolute section-crossing relocation Warns that an 8-bit absolute relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments reloc-abs-dword [off] 32-bit absolute section-crossing relocation Warns that a 32-bit absolute relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments reloc-abs-qword [off] 64-bit absolute section-crossing relocation Warns that a 64-bit absolute relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments reloc-abs-word [off] 16-bit absolute section-crossing relocation Warns that a 16-bit absolute relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments reloc-rel-byte [off] 8-bit relative section-crossing relocation Warns that an 8-bit relative relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments reloc-rel-dword [off] 32-bit relative section-crossing relocation Warns that a 32-bit relative relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments reloc-rel-qword [off] 64-bit relative section-crossing relocation Warns that an 64-bit relative relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments reloc-rel-word [off] 16-bit relative section-crossing relocation Warns that a 16-bit relative relocation that could not be resolved at assembly time was generated in the output format. This is usually normal, but may not be handled by all possible target environments section-alignment-rounded [on] section alignment rounded up Warn if a section alignment is specified which is not supported by the underlying object format, but can be rounded up to a supported value. unknown-warning [off] unknown warning in \c{-W}/\c{-w} or warning directive Warns about a \c{-w} or \c{-W} option or a \c{[WARNING]} directive that contains an unknown warning name or is otherwise not possible to process. user [on] \c{%warning} directives Controls output of \c{%warning} directives (see \k{pperror}). warn-stack-empty [on] warning stack empty A \c{[WARNING POP]} directive was executed when the warning stack is empty. This is treated as a \c{[WARNING *all]} directive. # # This warning is currently issued by backends, but in the future # that code should be centralized. # zeroing [on] \c{RES}\e{x} in initialized section becomes zero A \c{RES}\e{x} directive was used in a section which contains initialized data, and the output format does not support this. Instead, this will be replaced with explicit zero content, which may produce a large output file. zext-reloc [on] relocation zero-extended to match output format Warns that a relocation has been zero-extended due to limitations in the output format.