diff options
| author | Jan Hubicka <jh@suse.cz> | 2001-09-28 19:20:25 +0200 |
|---|---|---|
| committer | Jan Hubicka <hubicka@gcc.gnu.org> | 2001-09-28 17:20:25 +0000 |
| commit | ad91981212e1f9cf0c9b57a93f85e0a5cf1e096b (patch) | |
| tree | 12467869da5fe7bf490f2d11def08e5c06292f0f /gcc/doc/invoke.texi | |
| parent | 576fb787822fe6c80db9228fe52bec206c6be163 (diff) | |
| download | gcc-ad91981212e1f9cf0c9b57a93f85e0a5cf1e096b.zip gcc-ad91981212e1f9cf0c9b57a93f85e0a5cf1e096b.tar.gz gcc-ad91981212e1f9cf0c9b57a93f85e0a5cf1e096b.tar.bz2 | |
i386-protos.h (ix86_setup_incoming_varargs, [...]): Declare.
* i386-protos.h (ix86_setup_incoming_varargs, ix86_va_arg,
ix86_va_start, ix86_build_va_list): Declare.
* i386.c (ix86_setup_incoming_varargs, ix86_va_arg,
ix86_va_start, ix86_build_va_list): New global functions.
* i386.md (sse_prologue_save_insn): New insn.
(sse_prologue_save): New expander.
* i386.h (EXPAND_BUILTIN_VA_ARG, EXPAND_BUILTIN_VA_START,
BUILD_VA_LIST_TYPE, SETUP_INCOMING_VARARGS): New macros.
From-SVN: r45862
Diffstat (limited to 'gcc/doc/invoke.texi')
| -rw-r--r-- | gcc/doc/invoke.texi | 14 |
1 files changed, 13 insertions, 1 deletions
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi index d2b2afc..58c4d03 100644 --- a/gcc/doc/invoke.texi +++ b/gcc/doc/invoke.texi @@ -274,7 +274,7 @@ in the following sections. -fschedule-insns -fschedule-insns2 @gol -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol --funroll-all-loops -funroll-loops @gol +-funroll-all-loops -funroll-loops -fmessy-debugging @gol --param @var{name}=@var{value} -O -O0 -O1 -O2 -O3 -Os} @@ -3415,6 +3415,18 @@ Not all of the optimizations performed by GCC have @option{-f} options to control them. @table @gcctabopt +@item -fmessy-debugging +@opindex fmessy-debugging +Some optimizations can be stronger if compiler give up ability to generate +at least approximately usefull debug information for optimized programs. +These transformation are not done when @option{-g} switch is not specified, +as politics is to get the code with @option{-g} and without equivalent, +except for debugging section. + +This optimization enables such transformations. It is still compatible with +@option{-g}, but debugging resulting programs will be even higher challenge. + +@table @gcctabopt @item -fstrength-reduce @opindex fstrength-reduce Perform the optimizations of loop strength reduction and |