kumquat-buildroot/toolchain/gcc/3.4.0/820-no-mips-empic-relocs.patch

84 lines
3.3 KiB
Diff

From: cgd at broadcom dot com
To: gcc-patches at gcc dot gnu dot org
Cc: mark at codesourcery dot com
Date: 13 Jun 2004 22:51:30 -0700
Subject: [trunk + 3.4-branch RFA] don't use empic relocs for mips-linuxeh
This patch changes mips-linux to avoid using embedded-pic relocs for
its eh data. (Support for generating these for new code is removed in
current binutils srcs.)
Relating to this, previously, mips-linux and mips64-linux would use
different representations for their EH data (even for mips64-linux o32
abi), due to the mips64-linux n32/64 BFDs not supporting the
embedded-pic relocs. This was a bug.
For more explanation, see the thread of the URL quoted in the comment
in linux.h.
Tested the same w/ sources of about a week ago for c/c++ for
mips-linux (native) before/after. Also verified .o compatibility
before/after just to be sure.
I'd like this approved for the branch as well, so 3.4.1 will work
nicely w/ the next major binutils release.
thanks,
chris
2004-06-13 Chris Demetriou <cgd@broadcom.com>
* config/mips/linux.h (ASM_PREFERRED_EH_DATA_FORMAT): Redefine
to return DW_EH_PE_absptr.
* config/mips/linux64.h (ASM_PREFERRED_EH_DATA_FORMAT): Remove.
Index: config/mips/linux.h
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/mips/linux.h,v
retrieving revision 1.77
diff -u -p -r1.77 linux.h
--- gcc/gcc/config/mips/linux.h 19 Feb 2004 22:07:51 -0000 1.77
+++ gcc/gcc/config/mips/linux.h 14 Jun 2004 05:49:51 -0000
@@ -170,10 +170,11 @@ Boston, MA 02111-1307, USA. */
#undef FUNCTION_NAME_ALREADY_DECLARED
#define FUNCTION_NAME_ALREADY_DECLARED 1
-#define ASM_PREFERRED_EH_DATA_FORMAT(CODE, GLOBAL) \
- (flag_pic \
- ? ((GLOBAL) ? DW_EH_PE_indirect : 0) | DW_EH_PE_pcrel | DW_EH_PE_sdata4\
- : DW_EH_PE_absptr)
+/* If possible, we should attempt to use GP-relative relocs for this
+ (see <a href="http://sources.redhat.com/ml/binutils/2004-05/msg00227.html">http://sources.redhat.com/ml/binutils/2004-05/msg00227.html</a>).
+ However, until that is implement, this just uses standard, absolute
+ references. */
+#define ASM_PREFERRED_EH_DATA_FORMAT(CODE, GLOBAL) DW_EH_PE_absptr
/* The glibc _mcount stub will save $v0 for us. Don't mess with saving
it, since ASM_OUTPUT_REG_PUSH/ASM_OUTPUT_REG_POP do not work in the
Index: config/mips/linux64.h
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/mips/linux64.h,v
retrieving revision 1.7
diff -u -p -r1.7 linux64.h
--- gcc/gcc/config/mips/linux64.h 18 Feb 2004 15:05:10 -0000 1.7
+++ gcc/gcc/config/mips/linux64.h 14 Jun 2004 05:49:51 -0000
@@ -81,14 +81,6 @@ Boston, MA 02111-1307, USA. */
specification. The SGI/MIPS ABI defines it to be the same as PTR_SIZE. */
#define DWARF_OFFSET_SIZE PTR_SIZE
-#undef ASM_PREFERRED_EH_DATA_FORMAT
-#if 0 /* We can't use relative addressing modes on NEWABI :-( */
-#define ASM_PREFERRED_EH_DATA_FORMAT(CODE, GLOBAL) \
- (flag_pic \
- ? ((GLOBAL) ? DW_EH_PE_indirect : 0) | DW_EH_PE_sdata4 \
- : DW_EH_PE_absptr)
-#endif
-
/* GNU/Linux doesn't use the same floating-point format that IRIX uses
for long double. There's no need to override this here, since
ieee_quad_format is the default, but let's put this here to make