Commit 4736718a by Matthias Klose Committed by Anthony Green

Improve libffi comments

From-SVN: r162936
parent 3ac097a5
...@@ -39,6 +39,11 @@ ...@@ -39,6 +39,11 @@
* libffi/src/x86/unix64.S (.eh_frame) * libffi/src/x86/unix64.S (.eh_frame)
[HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type. [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
2010-03-14 Matthias Klose <doko@ubuntu.com>
* src/x86/ffi64.c: Fix typo in comment.
* src/x86/ffi.c: Use /* ... */ comment style.
2010-02-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE> 2010-02-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
* doc/libffi.texi (The Closure API): Fix typo. * doc/libffi.texi (The Closure API): Fix typo.
......
...@@ -585,10 +585,10 @@ ffi_prep_raw_closure_loc (ffi_raw_closure* closure, ...@@ -585,10 +585,10 @@ ffi_prep_raw_closure_loc (ffi_raw_closure* closure,
return FFI_BAD_ABI; return FFI_BAD_ABI;
} }
// we currently don't support certain kinds of arguments for raw /* we currently don't support certain kinds of arguments for raw
// closures. This should be implemented by a separate assembly language closures. This should be implemented by a separate assembly
// routine, since it would require argument processing, something we language routine, since it would require argument processing,
// don't do now for performance. something we don't do now for performance. */
for (i = cif->nargs-1; i >= 0; i--) for (i = cif->nargs-1; i >= 0; i--)
{ {
......
...@@ -50,9 +50,10 @@ extern void ffi_call_unix64 (void *args, unsigned long bytes, unsigned flags, ...@@ -50,9 +50,10 @@ extern void ffi_call_unix64 (void *args, unsigned long bytes, unsigned flags,
gcc/config/i386/i386.c. Do *not* change one without the other. */ gcc/config/i386/i386.c. Do *not* change one without the other. */
/* Register class used for passing given 64bit part of the argument. /* Register class used for passing given 64bit part of the argument.
These represent classes as documented by the PS ABI, with the exception These represent classes as documented by the PS ABI, with the
of SSESF, SSEDF classes, that are basically SSE class, just gcc will exception of SSESF, SSEDF classes, that are basically SSE class,
use SF or DFmode move instead of DImode to avoid reformating penalties. just gcc will use SF or DFmode move instead of DImode to avoid
reformatting penalties.
Similary we play games with INTEGERSI_CLASS to use cheaper SImode moves Similary we play games with INTEGERSI_CLASS to use cheaper SImode moves
whenever possible (upper half does contain padding). */ whenever possible (upper half does contain padding). */
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment