arm-linux-gcc.1

gcc编译器 arm-2.95.3.tar.gz

.Sp
Level 3 includes extra information, such as all the macro definitions
present in the program.  Some debuggers support macro expansion when
you use `\|\c
.B \-g3\c
\&\|'.
.TP
.B \-p
Generate extra code to write profile information suitable for the
analysis program \c
.B prof\c
\&.
.TP
.B \-pg
Generate extra code to write profile information suitable for the
analysis program \c
.B gprof\c
\&.
.TP
.B \-a
Generate extra code to write profile information for basic blocks,
which will record the number of times each basic block is executed.
This data could be analyzed by a program like \c
.B tcov\c
\&.  Note,
however, that the format of the data is not what \c
.B tcov\c
\& expects.
Eventually GNU \c
.B gprof\c
\& should be extended to process this data.
.TP
.B \-ax
Generate extra code to read basic block profiling parameters from 
file `bb.in' and write profiling results to file `bb.out'.
`bb.in' contains a list of functions. Whenever a function on the list
is entered, profiling is turned on. When the outmost function is left,
profiling is turned off. If a function name is prefixed with `-'
the function is excluded from profiling. If a function name is not
unique it can be disambiguated by writing
`/path/filename.d:functionname'. `bb.out' will list some available
filenames.
Four function names have a special meaning:
`__bb_jumps__' will cause jump frequencies to be written to `bb.out'.
`__bb_trace__' will cause the sequence of basic blocks to be piped 
into `gzip' and written to file `bbtrace.gz'.
`__bb_hidecall__' will cause call instructions to be excluded from
the trace.
`__bb_showret__' will cause return instructions to be included in
the trace.
.TP
.BI "\-d" "letters"
Says to make debugging dumps during compilation at times specified by
.I letters\c
\&.  This is used for debugging the compiler.  The file names
for most of the dumps are made by appending a word to the source file
name (e.g.  `\|\c
.B foo.c.rtl\c
\&\|' or `\|\c
.B foo.c.jump\c
\&\|').
.TP
.B \-dM
Dump all macro definitions, at the end of preprocessing, and write no
output.
.TP
.B \-dN
Dump all macro names, at the end of preprocessing.
.TP
.B \-dD
Dump all macro definitions, at the end of preprocessing, in addition to
normal output.
.TP
.B \-dy
Dump debugging information during parsing, to standard error.
.TP
.B \-dr
Dump after RTL generation, to `\|\c
.I file\c
.B \&.rtl\c
\&\|'.
.TP
.B \-dx
Just generate RTL for a function instead of compiling it.  Usually used
with `\|\c
.B r\c
\&\|'.
.TP
.B \-dj
Dump after first jump optimization, to `\|\c
.I file\c
.B \&.jump\c
\&\|'.
.TP
.B \-ds
Dump after CSE (including the jump optimization that sometimes
follows CSE), to `\|\c
.I file\c
.B \&.cse\c
\&\|'.
.TP
.B \-dL
Dump after loop optimization, to `\|\c
.I file\c
.B \&.loop\c
\&\|'.
.TP
.B \-dt
Dump after the second CSE pass (including the jump optimization that
sometimes follows CSE), to `\|\c
.I file\c
.B \&.cse2\c
\&\|'.
.TP
.B \-df
Dump after flow analysis, to `\|\c
.I file\c
.B \&.flow\c
\&\|'.
.TP
.B \-dc
Dump after instruction combination, to `\|\c
.I file\c
.B \&.combine\c
\&\|'.
.TP
.B \-dS
Dump after the first instruction scheduling pass, to
`\|\c
.I file\c
.B \&.sched\c
\&\|'.
.TP
.B \-dl
Dump after local register allocation, to `\|\c
.I file\c
.B \&.lreg\c
\&\|'.
.TP
.B \-dg
Dump after global register allocation, to `\|\c
.I file\c
.B \&.greg\c
\&\|'.
.TP
.B \-dR
Dump after the second instruction scheduling pass, to
`\|\c
.I file\c
.B \&.sched2\c
\&\|'.
.TP
.B \-dJ
Dump after last jump optimization, to `\|\c
.I file\c
.B \&.jump2\c
\&\|'.
.TP
.B \-dd
Dump after delayed branch scheduling, to `\|\c
.I file\c
.B \&.dbr\c
\&\|'.
.TP
.B \-dk
Dump after conversion from registers to stack, to `\|\c
.I file\c
.B \&.stack\c
\&\|'.
.TP
.B \-da
Produce all the dumps listed above.
.TP
.B \-dm
Print statistics on memory usage, at the end of the run, to
standard error.
.TP
.B \-dp
Annotate the assembler output with a comment indicating which
pattern and alternative was used.
.TP
.B \-fpretend\-float
When running a cross-compiler, pretend that the target machine uses the
same floating point format as the host machine.  This causes incorrect
output of the actual floating constants, but the actual instruction
sequence will probably be the same as GNU CC would make when running on
the target machine.
.TP
.B \-save\-temps
Store the usual \*(lqtemporary\*(rq intermediate files permanently; place them
in the current directory and name them based on the source file.  Thus,
compiling `\|\c
.B foo.c\c
\&\|' with `\|\c
.B \-c \-save\-temps\c
\&\|' would produce files
`\|\c
.B foo.cpp\c
\&\|' and `\|\c
.B foo.s\c
\&\|', as well as `\|\c
.B foo.o\c
\&\|'.
.TP
.BI "\-print\-file\-name=" "library"
Print the full absolute name of the library file \|\c
.nh
.I library
.hy
\&\| that
would be used when linking\(em\&and do not do anything else.  With this
option, GNU CC does not compile or link anything; it just prints the
file name.
.TP
.B \-print\-libgcc\-file\-name
Same as `\|\c
.B \-print\-file\-name=libgcc.a\c
\&\|'.
.TP
.BI "\-print\-prog\-name=" "program"
Like `\|\c
.B \-print\-file\-name\c
\&\|', but searches for a program such as `\|\c
cpp\c
\&\|'.
.SH OPTIMIZATION OPTIONS
These options control various sorts of optimizations:
.TP
.B \-O
.TP
.B \-O1
Optimize.  Optimizing compilation takes somewhat more time, and a lot
more memory for a large function.
.Sp
Without `\|\c
.B \-O\c
\&\|', the compiler's goal is to reduce the cost of
compilation and to make debugging produce the expected results.
Statements are independent: if you stop the program with a breakpoint
between statements, you can then assign a new value to any variable or
change the program counter to any other statement in the function and
get exactly the results you would expect from the source code.
.Sp
Without `\|\c
.B \-O\c
\&\|', only variables declared \c
.B register\c
\& are
allocated in registers.  The resulting compiled code is a little worse
than produced by PCC without `\|\c
.B \-O\c
\&\|'.
.Sp
With `\|\c
.B \-O\c
\&\|', the compiler tries to reduce code size and execution
time.
.Sp
When you specify `\|\c
.B \-O\c
\&\|', the two options `\|\c
.B \-fthread\-jumps\c
\&\|' and `\|\c
.B \-fdefer\-pop\c
\&\|' are turned on.  On machines that have delay slots, the `\|\c
.B \-fdelayed\-branch\c
\&\|' option is turned on.  For those machines that can support debugging even
without a frame pointer, the `\|\c
.B \-fomit\-frame\-pointer\c
\&\|' option is turned on.  On some machines other flags may also be turned on.
.TP
.B \-O2
Optimize even more.  Nearly all supported optimizations that do not
involve a space-speed tradeoff are performed.  Loop unrolling and function
inlining are not done, for example.  As compared to
.B \-O\c
\&,
this option increases both compilation time and the performance of the
generated code.
.TP
.B \-O3
Optimize yet more. This turns on everything
.B \-O2
does, along with also turning on
.B \-finline\-functions.
.TP
.B \-O0
Do not optimize.
.Sp
If you use multiple
.B \-O
options, with or without level numbers, the last such option is the
one that is effective.
.PP
Options of the form `\|\c
.B \-f\c
.I flag\c
\&\c
\&\|' specify machine-independent
flags.  Most flags have both positive and negative forms; the negative
form of `\|\c
.B \-ffoo\c
\&\|' would be `\|\c
.B \-fno\-foo\c
\&\|'.  The following list shows
only one form\(em\&the one which is not the default.
You can figure out the other form by either removing `\|\c
.B no\-\c
\&\|' or
adding it.
.TP
.B \-ffloat\-store
Do not store floating point variables in registers.  This
prevents undesirable excess precision on machines such as the
68000 where the floating registers (of the 68881) keep more
precision than a \c
.B double\c
\& is supposed to have.
.Sp
For most programs, the excess precision does only good, but a few
programs rely on the precise definition of IEEE floating point.
Use `\|\c
.B \-ffloat\-store\c
\&\|' for such programs.
.TP
.B \-fmemoize\-lookups
.TP
.B \-fsave\-memoized
Use heuristics to compile faster (C++ only).  These heuristics are not
enabled by default, since they are only effective for certain input
files.  Other input files compile more slowly.
.Sp
The first time the compiler must build a call to a member function (or
reference to a data member), it must (1) determine whether the class
implements member functions of that name; (2) resolve which member
function to call (which involves figuring out what sorts of type
conversions need to be made); and (3) check the visibility of the member
function to the caller.  All of this adds up to slower compilation.
Normally, the second time a call is made to that member function (or
reference to that data member), it must go through the same lengthy
process again.  This means that code like this
.Sp
\&  cout << "This " << p << " has " << n << " legs.\en";
.Sp
makes six passes through all three steps.  By using a software cache,
a \*(lqhit\*(rq significantly reduces this cost.  Unfortunately, using the
cache introduces another layer of mechanisms which must be implemented,
and so incurs its own overhead.  `\|\c
.B \-fmemoize\-lookups\c
\&\|' enables
the software cache.
.Sp
Because access privileges (visibility) to members and member functions
may differ from one function context to the next,
.B g++
may need to flush the cache.  With the `\|\c
.B \-fmemoize\-lookups\c
\&\|' flag, the cache is flushed after every
function that is compiled.  The `\|\c
\-fsave\-memoized\c
\&\|' flag enables the same software cache, but when the compiler
determines that the context of the last function compiled would yield
the same access privileges of the next function to compile, it
preserves the cache.
This is most helpful when defining many member functions for the same
class: with the exception of member functions which are friends of
other classes, each member function has exactly the same access
privileges as every other, and the cache need not be flushed.
.TP
.B \-fno\-default\-inline
Don't make member functions inline by default merely because they are
defined inside the class scope (C++ only).
.TP
.B \-fno\-defer\-pop
Always pop the arguments to each function call as soon as that
function returns.  For machines which must pop arguments after a
function call, the compiler normally lets arguments accumulate on the
stack for several function calls and pops them all at once.
.TP
.B \-fforce\-mem
Force memory operands to be copied into registers before doing
arithmetic on them.  This may produce better code by making all
memory references potential common subexpressions.  When they are
not common subexpressions, instruction combination should
eliminate the separate register-load.  I am interested in hearing
about the difference this makes.
.TP
.B \-fforce\-addr
Force memory address constants to be copied into registers before
doing arithmetic on them.  This may produce better code just as
`\|\c
.B \-fforce\-mem\c
\&\|' may.  I am interested in hearing about the
difference this makes.
.TP
.B \-fomit\-frame\-pointer
Don't keep the frame pointer in a register for functions that
don't need one.  This avoids the instructions to save, set up and
restore frame pointers; it also makes an extra register available
in many functions.  \c
.I It also makes debugging impossible on most machines\c
\&.
.Sp
On some machines, such as the Vax, this flag has no effect, because
the standard calling sequence automatically handles the frame pointer
and nothing is saved by pretending it doesn't exist.  The
machine-description macro \c
.B FRAME_POINTER_REQUIRED\c
\& controls
whether a target machine supports this flag.
.TP
.B \-finline\-functions
Integrate all simple functions into their callers.  The compiler
heuristically decides which functions are simple enough to be worth
integrating in this way.
.Sp
If all calls to a given function are integrated, and the function is
declared \c
.B static\c
\&, then GCC normally does not output the function as
assembler code in its own right.
.TP
.B \-fcaller\-saves
Enable values to be allocated in registers that will be clobbered by
function calls, by emitting extra instructions to save and restore the
registers around such calls.  Such allocation is done only when it
seems to result in better code than would otherwise be produced.
.Sp
This option is enabled by default on certain machines, usually those
which have no call-preserved registers to use instead.
.TP
.B \-fkeep\-inline\-functions
Even if all calls to a given function are integrated, and the function
is declared \c
.B static\c
\&, nevertheless output a separate run-time
callable version of the function.
.TP
.B \-fno\-function\-cse
Do not put function addresses in registers; make each instruction that
calls a constant function contain the function's address explicitly.
.Sp
This option results in less efficient code, but some strange hacks
that alter the assembler output may be confused by the optimizations
performed when this option is not used.
.TP
.B \-fno\-peephole
Disable any machine-specific peephole optimizations.
.TP
.B \-ffast-math
This option allows GCC to violate some ANSI or IEEE rules/specifications
in the interest of optimizing code for speed.  For example, it allows
the compiler to assume arguments to the \c
.B sqrt\c
\& function are
non-negative numbers.
.Sp
This option should never be turned on by any `\|\c
.B \-O\c
\&\|' option since
it can result in incorrect output for programs which depend on
an exact implementation of IEEE or ANSI rules/specifications for
math functions.
.PP
The following options control specific optimizations.  The `\|\c
.B \-O2\c
\&\|'
option turns on all of these optimizations except `\|\c
.B \-funroll\-loops\c
\&\|'
and `\|\c
.B \-funroll\-all\-loops\c
\&\|'.
.PP
The `\|\c
.B \-O\c
\&\|' option usually turns on
the `\|\c
.B \-fthread\-jumps\c
\&\|' and `\|\c
.B \-fdelayed\-branch\c
\&\|' options, but
specific machines may change the default optimizations.
.PP
You can use the following flags in the rare cases when \*(lqfine-tuning\*(rq
of optimizations to be performed is desired.
.TP
.B \-fstrength\-reduce
Perform the optimizations of loop strength reduction and
elimination of iteration variables.
.TP
.B \-fthread\-jumps
Perform optimizations where we check to see if a jump branches to a
location where another comparison subsumed by the first is found.  If
so, the first branch is redirected to either the destination of the
second branch or a point immediately following it, depending on whether
the condition is known to be true or false.
.TP
.B \-funroll\-loops
Perform the optimization of loop unrolling.  This is only done for loops
whose number of iterations can be determined at compile time or run time.
.TP
.B \-funroll\-all\-loops
Perform the optimization of loop unrolling.  This is done for all loops.
This usually makes programs run more slowly.
.TP
.B \-fcse\-follow\-jumps
In common subexpression elimination, scan through jump instructions
when the target of the jump is not reached by any other path.  For
example, when CSE encounters an \c
.B if\c
\& statement with an
.B else\c
\& clause, CSE will follow the jump when the condition
tested is false.
.TP
.B \-fcse\-skip\-blocks
This is similar to `\|\c
.B \-fcse\-follow\-jumps\c
\&\|', but causes CSE to
follow jumps which conditionally skip over blocks.  When CSE
encounters a simple \c
.B if\c
\& statement with no else clause,
`\|\c
.B \-fcse\-skip\-blocks\c
\&\|' causes CSE to follow the jump around the
body of the \c
.B if\c
\&.
.TP
.B \-frerun\-cse\-after\-loop
Re-run common subexpression elimination after loop optimizations has been
performed.
.TP
.B \-felide\-constructors
El