SBSPSS/tools/psyq/bin/egcs/a.txt
2000-08-29 16:26:01 +00:00

342 lines
22 KiB
Plaintext

Usage: CC1PLPSX.EXE input [switches]
Switches:
-ffixed-<register> Mark <register> as being unavailable to the compiler
-fcall-used-<register> Mark <register> as being corrupted by function calls
-fcall-saved-<register> Mark <register> as being preserved across functions
-finline-limit-<number> Limits the size of inlined functions to <number>
-fident Process #ident directives
-fleading-underscore External symbols have a leading underscore
-finstrument-functions Instrument function entry/exit with profiling calls
-fdump-unnumbered Suppress output of instruction numbers and line number notes in debugging dumps
-fprefix-function-name Add a prefix to all function names
-fcheck-memory-usage Generate code to check every memory access
-fstrict-aliasing Assume strict aliasing rules apply
-fargument-noalias-global Assume arguments do not alias each other or globals
-fargument-noalias Assume arguments may alias globals but not each other
-fargument-alias Specify that arguments may alias each other & globals
-fstack-check Insert stack checking code into the program
-fpack-struct Pack structure members together without holes
-foptimize-register-move Do the full regmove optimization pass
-fregmove Enables a register move optimisation
-fgnu-linker Output GNU ld formatted global initialisers
-fverbose-asm Add extra commentry to assembler output
-fdata-sections place data items into their own section
-ffunction-sections place each function into its own section
-finhibit-size-directive Do not generate .size directives
-fcommon Do not put unitialised globals in the common section
-ffast-math Improve FP speed by violating ANSI & IEEE rules
-fbranch-probabilities Use profiling information for branch probabilities
-ftest-coverage Create data files needed by gcov
-fprofile-arcs Insert arc based program profiling code
-fasynchronous-exceptions Support asynchronous exceptions
-fsjlj-exceptions Use setjmp/longjmp to handle exceptions
-fnew-exceptions Use the new model for exception handling
-fexceptions Enable exception handling
-fpic Generate position independent code, if possible
-fbranch-count-reg Replace add,compare,branch with branch on count reg
-fschedule-insns2 Run two passes of the instruction scheduler
-fschedule-insns Reschedule instructions to avoid pipeline stalls
-fpretend-float Pretend that host and target use the same FP format
-frerun-loop-opt Run the loop optimiser twice
-frerun-cse-after-loop Run CSE pass after loop optimisations
-fgcse Perform the global common subexpression elimination
-fdelayed-branch Attempt to fill delay slots of branch instructions
-freg-struct-return Return 'short' aggregates in registers
-fpcc-struct-return Return 'short' aggregates in memory, not registers
-fcaller-saves Enable saving registers around function calls
-fshared-data Mark data as shared rather than private
-fsyntax-only Check for syntax errors, then stop
-fkeep-static-consts Emit static const variables even if they are not used
-finline Pay attention to the 'inline' keyword
-fkeep-inline-functions Generate code for funcs even if they are fully inlined
-finline-functions Integrate simple functions into their callers
-ffunction-cse Allow function addresses to be held in registers
-fforce-addr Copy memory address constants into regs before using
-fforce-mem Copy memory operands into registers before using
-fpeephole Enable machine specific peephole optimisations
-fwritable-strings Store strings in writable data section
-freduce-all-givs Strength reduce all loop general induction variables
-fmove-all-movables Force all loop invariant computations out of loops
-funroll-all-loops Perform loop unrolling for all loops
-funroll-loops Perform loop unrolling when iteration count is known
-fstrength-reduce Perform strength reduction optimisations
-fthread-jumps Perform jump threading optimisations
-fexpensive-optimizations Perform a number of minor, expensive optimisations
-fcse-skip-blocks When running CSE, follow conditional jumps
-fcse-follow-jumps When running CSE, follow jumps to their targets
-fomit-frame-pointer When possible do not generate stack frames
-fdefer-pop Defer popping functions args from stack until later
-fvolatile-static Consider all mem refs to static data to be volatile
-fvolatile-global Consider all mem refs to global data to be volatile
-fvolatile Consider all mem refs through pointers as volatile
-ffloat-store Do not store floats in registers
-O[number] Set optimisation level to [number]
-Os Optimise for space rather than speed
-pedantic Issue warnings needed by strict compliance to ANSI C
-pedantic-errors Like -pedantic except that errors are produced
-w Suppress warnings
-W Enable extra warnings
-Winline Warn when an inlined function cannot be inlined
-Wuninitialized Warn about unitialized automatic variables
-Wcast-align Warn about pointer casts which increase alignment
-Waggregate-return Warn about returning structures, unions or arrays
-Wswitch Warn about enumerated switches missing a specific case
-Wshadow Warn when one local variable shadows another
-Wunused Warn when a variable is unused
-Wid-clash-<num> Warn if 2 identifiers have the same first <num> chars
-Wlarger-than-<number> Warn if an object is larger than <number> bytes
-p Enable function profiling
-o <file> Place output into <file>
-G <number> Put global and static data smaller than <number>
bytes into a special section (on some targets)
-gcoff Generate COFF format debug output
-gstabs+ Generate extended STABS format debug output
-gstabs Generate STABS format debug output
-ggdb Generate default extended debug format output
-g Generate default debug format output
-aux-info <file> Emit declaration info into <file>.X
-quiet Do not display functions compiled or elapsed time
-version Display the compiler's version
-d[letters] Enable dumps from specific passes of the compiler
-dumpbase <file> Base name to be used for dumps from specific passes
-sched-verbose-<number> Set the verbosity level of the scheduler
--help Display this information
Language specific options:
-ansi Compile just for ANSI C
-fallow-single-precisio Do not promote floats to double if using -traditional
-std= Determine language standard
-funsigned-bitfields Make bitfields by unsigned by default
-fsigned-char Make 'char' be signed by default
-funsigned-char Make 'char' be unsigned by default
-traditional Attempt to support traditional K&R style C
-fno-asm Do not recognise the 'asm' keyword
-fno-builtin Do not recognise any built in functions
-fhosted Assume normal C execution environment
-ffreestanding Assume that standard libraries & main might not exist
-fcond-mismatch Allow different types as args of ? operator
-fdollars-in-identifier Allow the use of $ inside identifiers
-fshort-double Use the same size for double as for float
-fshort-enums Use the smallest fitting integer to hold enums
-Wall Enable most warning messages
-Wbad-function-cast Warn about casting functions to incompatible types
-Wmissing-noreturn Warn about functions which might be candidates for attribute noreturn
-Wcast-qual Warn about casts which discard qualifiers
-Wchar-subscripts Warn about subscripts whose type is 'char'
-Wcomment Warn if nested comments are detected
-Wcomments Warn if nested comments are detected
-Wconversion Warn about possibly confusing type conversions
-Wformat Warn about printf format anomalies
-Wimplicit-function-dec Warn about implicit function declarations
-Wimplicit-int Warn when a declaration does not specify a type
-Wimport Warn about the use of the #import directive
-Wno-long-long Do not warn about using 'long long' when -pedantic
-Wmain Warn about suspicious declarations of main
-Wmissing-braces Warn about possibly missing braces around initialisers
-Wmissing-declarations Warn about global funcs without previous declarations
-Wmissing-prototypes Warn about global funcs without prototypes
-Wmultichar Warn about use of multicharacter literals
-Wnested-externs Warn about externs not at file scope level
-Wparentheses Warn about possible missing parentheses
-Wpointer-arith Warn about function pointer arithmetic
-Wredundant-decls Warn about multiple declarations of the same object
-Wsign-compare Warn about signed/unsigned comparisons
-Wunknown-pragmas Warn about unrecognised pragmas
-Wstrict-prototypes Warn about non-prototyped function decls
-Wtraditional Warn about constructs whose meaning change in ANSI C
-Wtrigraphs Warn when trigraphs are encountered
-Wwrite-strings Mark strings as 'const char *'
Options for Objective C:
-gen-decls Dump decls to a .decl file
-fgnu-runtime Generate code for GNU runtime environment
-fnext-runtime Generate code for NeXT runtime environment
-Wselector Warn if a selector has multiple methods
-Wno-protocol Do not warn if inherited methods are unimplemented
-print-objc-runtime-inf Generate C header of platform specific features
Options for C++:
-fno-access-control Do not obey access control semantics
-fall-virtual Make all member functions virtual
-falt-external-template Change when template instances are emitted
-fcheck-new Check the return value of new
-fconserve-space Reduce size of object files
-fno-const-strings Make string literals `char[]' instead of `const char[]'
-fno-default-inline Do not inline member functions by default
-fno-rtti Do not generate run time type descriptor information
-fno-for-scope Scope of for-init-statement vars extends outside
-fguiding-decls Implement guiding declarations
-fno-gnu-keywords Do not recognise GNU defined keywords
-fhandle-signatures Handle signature language constructs
-fhonor-std Treat the namespace `std' as a normal namespace
-fhuge-objects Enable support for huge objects
-fno-implement-inlines Export functions even if they can be inlined
-fno-implicit-templates Only emit explicit template instatiations
-fno-implicit-inline-te Only emit explicit instatiations of inline templates
-finit-priority Handle the init_priority attribute
-flabels-ok Labels can be used as first class objects
-fnew-abi Enable experimental ABI changes
-fno-nonnull-objects Do not assume that a reference is always valid
-foperator-names Recognise and/bitand/bitor/compl/not/or/xor
-fno-optional-diags Disable optional diagnostics
-fpermissive Downgrade conformance errors to warnings
-frepo Enable automatic template instantiation
-fsquangle Enable squashed name mangling
-fstats Display statistics accumulated during compilation
-fno-strict-prototype Do not assume that empty prototype means no args
-ftemplate-depth- Specify maximum template instantiation depth
-fthis-is-variable Make 'this' not be type '* const'
-fvtable-gc Discard unused virtual functions
-fvtable-thunks Implement vtables using thunks
-fweak Emit common-like symbols as weak symbols
-fxref Emit cross referencing information
-Wreturn-type Warn about inconsistent return types
-Woverloaded-virtual Warn about overloaded virtual function names
-Wno-ctor-dtor-privacy Don't warn when all ctors/dtors are private
-Wnon-virtual-dtor Warn about non virtual destructors
-Wextern-inline Warn when a function is declared extern, then inline
-Wreorder Warn when the compiler reorders code
-Wsynth Warn when synthesis behaviour differs from Cfront
-Wno-pmf-conversions Don't warn when type converting pointers to member functions
-Weffc++ Warn about violations of Effective C++ style rules
-Wsign-promo Warn when overload promotes from unsigned to signed
-Wold-style-cast Warn if a C style cast is used in a program
-Wno-non-template-frien Don't warn when non-templatized friend functions are declared within a template
-Wno-deprecated Don't announce deprecation of compiler features
Options for Fortran:
-fversion Print g77-specific compiler version info, run internal tests
-ff66 Program is written in typical FORTRAN 66 dialect
-ff77 Program is written in typical Unix f77 dialect
-fno-f77 Program does not use Unix-f77 dialectal features
-ff90 Program is written in Fortran-90-ish dialect
-fno-automatic Treat local vars and COMMON blocks as if they were named in SAVE statements
-fdollar-ok Allow $ in symbol names
-fno-f2c f2c-compatible code need not be generated
-fno-f2c-library Unsupported; do not generate libf2c-calling code
-fflatten-arrays Unsupported; affects code-generation of arrays
-ffree-form Program is written in Fortran-90-ish free form
-fpedantic Warn about use of (only a few for now) Fortran extensions
-fvxt Program is written in VXT (Digital-like) FORTRAN
-fno-ugly Disallow all ugly features
-fno-ugly-args Hollerith and typeless constants not passed as arguments
-fugly-assign Allow ordinary copying of ASSIGN'ed vars
-fugly-assumed Dummy array dimensioned to (1) is assumed-size
-fugly-comma Trailing comma in procedure call denotes null argument
-fugly-complex Allow REAL(Z) and AIMAG(Z) given DOUBLE COMPLEX Z
-fno-ugly-init Initialization via DATA and PARAMETER is type-compatible
-fugly-logint Allow INTEGER and LOGICAL interchangeability
-fxyzzy Print internal debugging-related info
-finit-local-zero Initialize local vars and arrays to zero
-fno-backslash Backslashes in character/hollerith constants not special (C-style)
-femulate-complex Have front end emulate COMPLEX arithmetic to avoid bugs
-fno-underscoring Disable the appending of underscores to externals
-fno-second-underscore Never append a second underscore to externals
-fintrin-case-initcap Intrinsics spelled as e.g. SqRt
-fintrin-case-upper Intrinsics in uppercase
-fintrin-case-any Intrinsics letters in arbitrary cases
-fmatch-case-initcap Language keywords spelled as e.g. IOStat
-fmatch-case-upper Language keywords in uppercase
-fmatch-case-any Language keyword letters in arbitrary cases
-fsource-case-upper Internally convert most source to uppercase
-fsource-case-preserve Internally preserve source case
-fsymbol-case-initcap Symbol names spelled in mixed case
-fsymbol-case-upper Symbol names in uppercase
-fsymbol-case-lower Symbol names in lowercase
-fcase-strict-upper Program written in uppercase
-fcase-strict-lower Program written in lowercase
-fcase-initcap Program written in strict mixed-case
-fcase-upper Compile as if program written in uppercase
-fcase-lower Compile as if program written in lowercase
-fcase-preserve Preserve all spelling (case) used in program
-fbadu77-intrinsics-del Delete libU77 intrinsics with bad interfaces
-fbadu77-intrinsics-dis Disable libU77 intrinsics with bad interfaces
-fbadu77-intrinsics-hid Hide libU77 intrinsics with bad interfaces
-ff2c-intrinsics-delete Delete non-FORTRAN-77 intrinsics f2c supports
-ff2c-intrinsics-disabl Disable non-FORTRAN-77 intrinsics f2c supports
-ff2c-intrinsics-hide Hide non-FORTRAN-77 intrinsics f2c supports
-ff90-intrinsics-delete Delete non-FORTRAN-77 intrinsics F90 supports
-ff90-intrinsics-disabl Disable non-FORTRAN-77 intrinsics F90 supports
-ff90-intrinsics-hide Hide non-FORTRAN-77 intrinsics F90 supports
-fgnu-intrinsics-delete Delete non-FORTRAN-77 intrinsics g77 supports
-fgnu-intrinsics-disabl Disable non-FORTRAN 77 intrinsics F90 supports
-fgnu-intrinsics-hide Hide non-FORTRAN 77 intrinsics F90 supports
-fmil-intrinsics-delete Delete MIL-STD 1753 intrinsics
-fmil-intrinsics-disabl Disable MIL-STD 1753 intrinsics
-fmil-intrinsics-hide Hide MIL-STD 1753 intrinsics
-funix-intrinsics-delet Delete libU77 intrinsics
-funix-intrinsics-disab Disable libU77 intrinsics
-funix-intrinsics-hide Hide libU77 intrinsics
-fvxt-intrinsics-delete Delete non-FORTRAN-77 intrinsics VXT FORTRAN supports
-fvxt-intrinsics-disabl Disable non-FORTRAN-77 intrinsics VXT FORTRAN supports
-fvxt-intrinsics-hide Hide non-FORTRAN-77 intrinsics VXT FORTRAN supports
-fzeros Treat initial values of 0 like non-zero values
-fdebug-kludge Emit special debugging information for COMMON and EQUIVALENCE
-fonetrip Take at least one trip through each iterative DO loop
-fno-silent Print names of program units as they are compiled
-fno-globals Disable fatal diagnostics about inter-procedural problems
-ftypeless-boz Make prefix-radix non-decimal constants be typeless
-fbounds-check Generate code to check subscript and substring bounds
-ffortran-bounds-check Fortran-specific form of -fbounds-check
-Wno-globals Disable warnings about inter-procedural problems
-Wsurprising Warn about constructs with surprising meanings
-I Add a directory for INCLUDE searching
-ffixed-line-length- Set the maximum line length
Target specific options:
-mno-check-range-divisi Don't trap on integer divide overflow
-mcheck-range-division Trap on integer divide overflow
-mno-check-zero-divisio Don't trap on integer divide by zero
-mcheck-zero-division Trap on integer divide by zero
-m3900 Optimize for 3900
-m4650 Optimize for 4650
-mno-fix4300 Don't work around early 4300 hardware bug
-mfix4300 Work around early 4300 hardware bug
-mno-mad Don't use multiply accumulate
-mmad Use multiply accumulate
-mdouble-float Don't use single (32-bit) FP only
-msingle-float Use single (32-bit) FP only
-mel Use little-endian byte order
-meb Use big-endian byte order
-mno-embedded-data Don't use ROM instead of RAM
-membedded-data Use ROM instead of RAM
-mno-embedded-pic Don't use embedded PIC
-membedded-pic Use embedded PIC
-mno-long-calls Don't use indirect calls
-mlong-calls Use indirect calls
-mno-half-pic Don't use OSF PIC
-mhalf-pic Use OSF PIC
-mno-abicalls Don't use Irix PIC
-mabicalls Use Irix PIC
-mgp32 Use 32-bit general registers
-mgp64 Use 64-bit general registers
-mfp32 Use 32-bit FP registers
-mfp64 Use 64-bit FP registers
-mhard-float Use hardware floating point
-msoft-float Use software floating point
-mno-mips-tfile Don't use mips-tfile asm postpass
-mmips-tfile Use mips-tfile asm postpass
-mno-memcpy Optimize block moves
-mmemcpy Don't optimize block moves
-mno-stats Don't output compiler statistics
-mstats Output compiler statistics
-mno-gpopt Don't use GP relative sdata/sbss sections
-mno-gpOPT Don't use GP relative sdata/sbss sections
-mgpopt Use GP relative sdata/sbss sections
-mgpOPT Use GP relative sdata/sbss sections
-mno-rnames Don't use symbolic register names
-mrnames Use symbolic register names
-mgas Use GNU as
-mmips-as Use MIPS as
-mno-split-addresses Don't optimize lui/addiu address loads
-msplit-addresses Optimize lui/addiu address loads
-mlong32 Use 32-bit long type
-mlong64 Use 64-bit long type
-mint64 Use 64-bit int type
-mno-mips16 Don't use MIPS16 instructions
-mentry Use mips16 entry/exit psuedo ops
-mips Specify MIPS ISA
-mcpu= Specify CPU for scheduling purposes
There are undocumented target specific options as well.