summaryrefslogtreecommitdiff
path: root/patches/gcc/3.3
diff options
context:
space:
mode:
Diffstat (limited to 'patches/gcc/3.3')
-rw-r--r--patches/gcc/3.3/compat.exp.patch37
-rw-r--r--patches/gcc/3.3/config.sub.patch72
-rw-r--r--patches/gcc/3.3/gcc-3.2.3-g++.exp.patch98
-rw-r--r--patches/gcc/3.3/gcc-3.3-libstdc++-v3-dg.exp.patch54
-rw-r--r--patches/gcc/3.3/sh-pic-set_fpscr.patch60
-rw-r--r--patches/gcc/3.3/sh-predef-gnu_source.patch15
-rw-r--r--patches/gcc/3.3/sh-spec.patch45
-rw-r--r--patches/gcc/3.3/sh4-kaz-workaround.patch145
-rw-r--r--patches/gcc/3.3/sh4-no-fix-protos.patch12
-rw-r--r--patches/gcc/3.3/sh4-pthread.patch42
10 files changed, 0 insertions, 580 deletions
diff --git a/patches/gcc/3.3/compat.exp.patch b/patches/gcc/3.3/compat.exp.patch
deleted file mode 100644
index 763b8a1..0000000
--- a/patches/gcc/3.3/compat.exp.patch
+++ /dev/null
@@ -1,37 +0,0 @@
-Fixes the following problem when testing a cross-compiler:
-
-============================================
-Testing g++.dg/compat/abi/bitfield1,
-...
-Invoking the compiler as /crosstool-0.12/build/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/build-gcc/gcc/testsuite/../g++ ...
-/crosstool-0.12/build/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/gcc-3.3/gcc/testsuite/g++.dg/compat/abi/bitfield1_main.C
-pid is 16028 -16028
-compiler exited with status 1
-output is:
-sh: error while loading shared libraries: /opt/cegl-2.0/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/powerpc-750-linux-gnu/./lib/libdl.so.2: ELF file data encoding not little-endian
-/crosstool-0.12/build/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/build-gcc/gcc/testsuite/../g++: error while loading shared libraries: /opt/cegl-2.0/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/powerpc-750-linux-gnu/./lib/libc.so.6: ELF file data encoding not little-endian
-
-FAIL: g++.dg/compat/abi/bitfield1 main_tst.o compile
-============================================
-
---- gcc-3.3/gcc/testsuite/g++.dg/compat/compat.exp.old Thu Oct 17 17:22:57 2002
-+++ gcc-3.3/gcc/testsuite/g++.dg/compat/compat.exp Mon Jul 14 10:59:19 2003
-@@ -42,11 +42,13 @@
- proc compat-fix-library-path { } {
- global ld_library_path
-
-- # See comments in lib/g++.exp for why this is needed.
-- setenv LD_LIBRARY_PATH $ld_library_path
-- setenv SHLIB_PATH $ld_library_path
-- setenv LD_LIBRARYN32_PATH $ld_library_path
-- setenv LD_LIBRARY64_PATH $ld_library_path
-+ if {![is_remote target]} {
-+ # See comments in lib/g++.exp for why this is needed.
-+ setenv LD_LIBRARY_PATH $ld_library_path
-+ setenv SHLIB_PATH $ld_library_path
-+ setenv LD_LIBRARYN32_PATH $ld_library_path
-+ setenv LD_LIBRARY64_PATH $ld_library_path
-+ }
- }
-
- #
diff --git a/patches/gcc/3.3/config.sub.patch b/patches/gcc/3.3/config.sub.patch
deleted file mode 100644
index 928f911..0000000
--- a/patches/gcc/3.3/config.sub.patch
+++ /dev/null
@@ -1,72 +0,0 @@
-When configuring a s390->s390 or cris->cris crosscompiler
-(ok, I haven't hit this yet, but one of these days I'll get me an account
-on an s390, and then I'll need this patch :-), you'll get the
-following error:
-
-+ /build/s390-unknown-linux-gnu/gcc-3.3-glibc-2.3.2/gcc-3.3/configure
---target=s390-unknown-linux-gnu --host=s390-host_unknown-linux-gnu
---prefix=/result/s390-unknown-linux-gnu/gcc-3.3-glibc-2.3.2
---enable-languages=c
---with-local-prefix=/result/s390-unknown-linux-gnu/gcc-3.3-glibc-2.3.2/s390-
-unknown-linux-gnu --enable-threads=no --without-headers --with-newlib
---disable-shared
-...
-*** --with-headers is only supported when cross compiling
-
-This error pops up only when you're using Daniel Jacobowitz's technique
-of slightly changing the target and host tuples to make them different
-enough to force gcc's build system to not try to pull in system libraries
-or headers. This technique is needed e.g. to build an x86 -> x86
-cross-compiler.
-(The LFS developers ran into the same bug that prompted me to use
-this technique; they point people who run into it to
-http://linuxfromscratch.org/~greg/pure_lfs.txt, which is a different
-way of avoiding this problem. I think the tuple tweak is the way to go, though.)
-
-config-patches@gnu.org rejected this patch, on the grounds that there
-is only one vendor of each of those two architectures, so the
-canonicalization is by definition correct. When I pointed out the
-difficulty this causes for people building s390 -> s390 or
-cris -> cris compilers that are incompatible with the system
-libraries and thus must be built like cross-compilers, he grumped and said
-"autoconf should let you specify a cross-compiler in some other way than
-comparing tuple strings".
-
-
-
---- gcc-3.3/config.sub.old Sun Jun 8 20:38:47 2003
-+++ gcc-3.3/config.sub Sun Jun 8 20:40:34 2003
-@@ -433,9 +433,12 @@
- crds | unos)
- basic_machine=m68k-crds
- ;;
-- cris | cris-* | etrax*)
-+ cris | etrax*)
- basic_machine=cris-axis
- ;;
-+ cris-*)
-+ basic_machine=cris-`echo $basic_machine | sed 's/^[^-]*-//'`
-+ ;;
- da30 | da30-*)
- basic_machine=m68k-da30
- ;;
-@@ -820,11 +823,17 @@
- rtpc | rtpc-*)
- basic_machine=romp-ibm
- ;;
-- s390 | s390-*)
-+ s390)
- basic_machine=s390-ibm
- ;;
-- s390x | s390x-*)
-+ s390-*)
-+ basic_machine=s390-`echo $basic_machine | sed 's/^[^-]*-//'`
-+ ;;
-+ s390x)
- basic_machine=s390x-ibm
-+ ;;
-+ s390x-*)
-+ basic_machine=s390x-`echo $basic_machine | sed 's/^[^-]*-//'`
- ;;
- sa29200)
- basic_machine=a29k-amd
diff --git a/patches/gcc/3.3/gcc-3.2.3-g++.exp.patch b/patches/gcc/3.3/gcc-3.2.3-g++.exp.patch
deleted file mode 100644
index 0de46fc..0000000
--- a/patches/gcc/3.3/gcc-3.2.3-g++.exp.patch
+++ /dev/null
@@ -1,98 +0,0 @@
-g++ testsuite fixes for cross-compilers, Dan Kegel, Ixia Communications, 12 July 2003
-
-The first hunk fixes the error
-
-/testsuite_flags: /testsuite_flags: No such file or directory
- while executing
-"exec sh ${odir_v3}/testsuite_flags --build-includes"
- (procedure "g++_include_flags" line 21)
- invoked from within
-"g++_include_flags [get_multilibs] "
- (procedure "g++_init" line 63)
- invoked from within
-"${tool}_init $test_file_name"
- (procedure "runtest" line 19)
- invoked from within
-"runtest $test_name"
- ("foreach" body line 42)
- invoked from within
-...
-make[1]: [check-g++] Error 1 (ignored)
-
-The fix isn't especially pretty, but it worked for me, and can't hurt the
-more common native compiler case. Maybe someone who knows the code better
-can come up with a better fix.
-
-The second hunk fixes the error
-
-sh: error while loading shared libraries: /opt/cegl-2.0/powerpc-405-linux-gnu/gcc-3.2.3-glibc-2.2.5/powerpc-405-linux-gnu/./lib/libdl.so.2: ELF file data encoding not little-endian
-
-when trying to compile g++ testcases (!); setting up
-the shared library environment when running crosstests of g++
-should either be done by a special board file, or by
-setting up a remote chroot environment (see http://kegel.com/crosstool),
-not by blithely setting LD_LIBRARY_PATH on the local system.
-
---- gcc-3.2.3/gcc/testsuite/lib/g++.exp.old Fri Jul 11 15:42:47 2003
-+++ gcc-3.2.3/gcc/testsuite/lib/g++.exp Sat Jul 12 12:57:07 2003
-@@ -72,6 +72,8 @@
- #
- proc g++_include_flags { paths } {
- global srcdir
-+ global objdir
-+ global target_triplet
- global HAVE_LIBSTDCXX_V3
- global TESTING_IN_BUILD_TREE
-
-@@ -90,6 +92,20 @@
-
- if { ${HAVE_LIBSTDCXX_V3} } {
- set odir_v3 [lookfor_file ${gccpath} libstdc++-v3]
-+ if { $odir_v3 == "" } {
-+ verbose "g++_include_flags: couldn't find libstdc++-v3 on first try, now looking in build directory $objdir"
-+ # first assume no multilibs
-+ set odir_v3 [lookfor_file ${objdir} "$target_triplet/libstdc++-v3"]
-+ }
-+ if { $odir_v3 == "" } {
-+ verbose "g++_include_flags: couldn't find libstdc++-v3 on second try, trying multilib"
-+ # assume multilib only one level deep
-+ set multisub [file tail $gccpath]
-+ set odir_v3 [lookfor_file ${objdir} "$target_triplet/$multisub/libstdc++-v3"]
-+ }
-+ if { $odir_v3 == "" } {
-+ error "Can't find libstdc++-v3"
-+ }
- append flags [exec sh ${odir_v3}/testsuite_flags --build-includes]
- } else {
- set odir_v2 [lookfor_file ${gccpath} libstdc++]
-@@ -192,16 +192,20 @@
- }
- }
-
-- # On IRIX 6, we have to set variables akin to LD_LIBRARY_PATH, but
-- # called LD_LIBRARYN32_PATH (for the N32 ABI) and LD_LIBRARY64_PATH
-- # (for the 64-bit ABI). The right way to do this would be to modify
-- # unix.exp -- but that's not an option since it's part of DejaGNU
-- # proper, so we do it here. We really only need to do
-- # this on IRIX, but it shouldn't hurt to do it anywhere else.
-- setenv LD_LIBRARY_PATH $ld_library_path
-- setenv SHLIB_PATH $ld_library_path
-- setenv LD_LIBRARYN32_PATH $ld_library_path
-- setenv LD_LIBRARY64_PATH $ld_library_path
-+ if {![is_remote target]} {
-+ # On IRIX 6, we have to set variables akin to LD_LIBRARY_PATH, but
-+ # called LD_LIBRARYN32_PATH (for the N32 ABI) and LD_LIBRARY64_PATH
-+ # (for the 64-bit ABI). The right way to do this would be to modify
-+ # unix.exp -- but that's not an option since it's part of DejaGNU
-+ # proper, so we do it here. We really only need to do
-+ # this on IRIX, but it shouldn't hurt to do it anywhere else.
-+
-+ # Doing this causes us to be unable to run cross-compilers.
-+ setenv LD_LIBRARY_PATH $ld_library_path
-+ setenv SHLIB_PATH $ld_library_path
-+ setenv LD_LIBRARYN32_PATH $ld_library_path
-+ setenv LD_LIBRARY64_PATH $ld_library_path
-+ }
-
- return "$flags"
- }
diff --git a/patches/gcc/3.3/gcc-3.3-libstdc++-v3-dg.exp.patch b/patches/gcc/3.3/gcc-3.3-libstdc++-v3-dg.exp.patch
deleted file mode 100644
index aafc130..0000000
--- a/patches/gcc/3.3/gcc-3.3-libstdc++-v3-dg.exp.patch
+++ /dev/null
@@ -1,54 +0,0 @@
-Without this patch, the command
-
-RUNTESTFLAGS=--target=powerpc-750-linux-gnu -v -v -v -v make check-target-libstdc++-v3
-
-fails in two ways:
-1. the -L option meant to locate the testsuite directory is incorrect, and
-2. the wrong compiler is invoked, causing all sorts of havoc, not least of which
-is the native compiler is invoked when we really wanted to invoke the cross-compiler
-we just built.
-
-Here's an example log of the problem in action. Every testcase fails, this shows just one:
--------------
-Invoking the compiler as g++ -ggdb3 -DDEBUG_ASSERT -I/home3/dank/wk/ixos/dap_interim/linux/3rdParty/crosstool/crosstool-0.12/build/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/gcc-3.3/libstdc++-v3/testsuite /home3/dank/wk/ixos/dap_interim/linux/3rdParty/crosstool/crosstool-0.12/build/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/gcc-3.3/libstdc++-v3/testsuite/20_util/binders.cc -g -O2 -DDEBUG_ASSERT -L/testsuite -lv3test -lm -o ./binders.exe
-compiler exited with status 1
-output is:
-/usr/lib/gcc-lib/i386-linux/2.95.4/../../../../include/g++-3/stl_function.h: In method `void binder2nd<mem_fun1_ref_t<void,Elem,int> >::operator ()(const Elem &) const':^M
-/usr/lib/gcc-lib/i386-linux/2.95.4/../../../../include/g++-3/stl_algo.h:83: instantiated from `for_each<Elem *, binder2nd<mem_fun1_ref_t<void,Elem,int> > >(Elem *, Elem *, binder2nd<mem_fun1_ref_t<void,Elem,int> >)'^M
-/home3/dank/wk/ixos/dap_interim/linux/3rdParty/crosstool/crosstool-0.12/build/powerpc-750-linux-gnu/gcc-3.3-glibc-2.2.5/gcc-3.3/libstdc++-v3/testsuite/20_util/binders.cc:43: instantiated from here^M
-/usr/lib/gcc-lib/i386-linux/2.95.4/../../../../include/g++-3/stl_function.h:224: conversion from `const Elem' to `Elem &' discards qualifiers^M
-/usr/lib/gcc-lib/i386-linux/2.95.4/../../../../include/g++-3/stl_function.h:622: in passing argument 1 of `mem_fun1_ref_t<void,Elem,int>::operator ()(Elem &, int) const'^M
-...
-FAIL: 20_util/binders.cc (test for excess errors)
-WARNING: 20_util/binders.cc compilation failed to produce executable
--------------
-
-And here's the patch. I'm not happy with it, and it probably gets some cases wrong,
-but it seems to work for the common native case and for my cross-compiler case.
-
---- gcc-3.3/libstdc++-v3/testsuite/lib/libstdc++-v3-dg.exp.old Sun Jul 13 10:42:01 2003
-+++ gcc-3.3/libstdc++-v3/testsuite/lib/libstdc++-v3-dg.exp Sun Jul 13 11:39:54 2003
-@@ -46,8 +46,23 @@
- global gluefile wrap_flags
- global ld_library_path
- global tool_root_dir
-+ global target_triplet
-
- set blddir [lookfor_file [get_multilibs] libstdc++-v3]
-+ if { $blddir == "" } {
-+ set multilibs [get_multilibs]
-+ # FIXME: assume multilib only one level deep
-+ set multisub [file tail $multilibs]
-+ verbose "libstdc++-v3-init: couldn't find libstdc++-v3 in $multilibs, trying $objdir"
-+ set blddir [lookfor_file ${objdir} "$target_triplet/$multisub/libstdc++-v3"]
-+ }
-+ if { $blddir == "" } {
-+ verbose "libstdc++-v3-init: couldn't find libstdc++-v3, trying $objdir without multilibs"
-+ set blddir [lookfor_file ${objdir} "$target_triplet/libstdc++-v3"]
-+ }
-+ if { $blddir == "" } {
-+ error "Can't find libstdc++-v3"
-+ }
-
- # By default, we assume we want to run program images.
- global dg-do-what-default
diff --git a/patches/gcc/3.3/sh-pic-set_fpscr.patch b/patches/gcc/3.3/sh-pic-set_fpscr.patch
deleted file mode 100644
index 00cfb4c..0000000
--- a/patches/gcc/3.3/sh-pic-set_fpscr.patch
+++ /dev/null
@@ -1,60 +0,0 @@
-Should fix
-make[2]: *** [/gcc-3.3-glibc-2.3.2/build-glibc/elf/check-textrel.out] Error 1
-in glibc's "make tests".
-
--------------
-
-Message-Id: <200307300255.h6U2tB906928@r-rr.iij4u.or.jp>
-To: dank@kegel.com
-Cc: linux-sh@m17n.org
-Subject: Re: Writing PIC version of __udivsi3_i4? (was: Re: "make tests" fails
- with "libc.so: text relocations used" in glibc-2.3.2 for sh4)
-In-Reply-To: Your message of "Wed, 30 Jul 2003 11:20:49 +0900"
- <200307300212.h6U2CH901209@r-rr.iij4u.or.jp>
-References: <200307300212.h6U2CH901209@r-rr.iij4u.or.jp>
-Date: Wed, 30 Jul 2003 12:03:01 +0900
-From: kaz Kojima <kkojima@rr.iij4u.or.jp>
-
-> I'll try to write the PIC version of it,
-
-Dan, does the attached patch work for you?
-
-[dank: original patch deleted; here's same thing rediffed against gcc-3.3 instead of HEAD]
-
---- gcc-3.3/gcc/config/sh/lib1funcs.asm.old Tue Jul 29 21:37:29 2003
-+++ gcc-3.3/gcc/config/sh/lib1funcs.asm Tue Jul 29 21:40:58 2003
-@@ -1792,7 +1792,17 @@
- .global GLOBAL(set_fpscr)
- GLOBAL(set_fpscr):
- lds r4,fpscr
-+#ifdef __PIC__
-+ mov.l r12,@-r15
-+ mova LOCAL(set_fpscr_L0),r0
-+ mov.l LOCAL(set_fpscr_L0),r12
-+ add r0,r12
-+ mov.l LOCAL(set_fpscr_L1),r0
-+ mov.l @(r0,r12),r1
-+ mov.l @r15+,r12
-+#else
- mov.l LOCAL(set_fpscr_L1),r1
-+#endif
- swap.w r4,r0
- or #24,r0
- #ifndef FMOVD_WORKS
-@@ -1820,8 +1830,16 @@
- mov.l r3,@(4,r1)
- #endif
- .align 2
-+#ifdef __PIC__
-+LOCAL(set_fpscr_L0):
-+ .long _GLOBAL_OFFSET_TABLE_
-+LOCAL(set_fpscr_L1):
-+ .long GLOBAL(fpscr_values@GOT)
-+#else
- LOCAL(set_fpscr_L1):
- .long GLOBAL(fpscr_values)
-+#endif
-+
- #ifdef __ELF__
- .comm GLOBAL(fpscr_values),8,4
- #else
diff --git a/patches/gcc/3.3/sh-predef-gnu_source.patch b/patches/gcc/3.3/sh-predef-gnu_source.patch
deleted file mode 100644
index dc94f4c..0000000
--- a/patches/gcc/3.3/sh-predef-gnu_source.patch
+++ /dev/null
@@ -1,15 +0,0 @@
-See http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11096
-and http://gcc.gnu.org/ml/gcc-patches/2003-06/msg00424.html
-
---- gcc-3.3/gcc/config/sh/linux.h.old Wed Jun 4 17:37:40 2003
-+++ gcc-3.3/gcc/config/sh/linux.h Wed Jun 4 17:38:16 2003
-@@ -53,5 +53,9 @@
- %{static:-static}"
-
-+/* The GNU C++ standard library requires that these macros be defined. */
-+#undef CPLUSPLUS_CPP_SPEC
-+#define CPLUSPLUS_CPP_SPEC "-D_GNU_SOURCE %(cpp)"
-+
- #undef LIB_SPEC
- #define LIB_SPEC \
- "%{shared: -lc} \
diff --git a/patches/gcc/3.3/sh-spec.patch b/patches/gcc/3.3/sh-spec.patch
deleted file mode 100644
index 6266c1f..0000000
--- a/patches/gcc/3.3/sh-spec.patch
+++ /dev/null
@@ -1,45 +0,0 @@
-Message-Id: <200308020452.h724q0n01509@r-rr.iij4u.or.jp>
-To: dank@kegel.com
-Cc: kkojima@rr.iij4u.or.jp
-Subject: Re: Writing PIC version of __udivsi3_i4?
-In-Reply-To: Your message of "Fri, 01 Aug 2003 21:15:27 -0700"
- <3F2B3ADF.6030206@kegel.com>
-References: <3F2B3ADF.6030206@kegel.com>
-Date: Sat, 02 Aug 2003 13:58:05 +0900
-From: kaz Kojima <kkojima@rr.iij4u.or.jp>
-
-Dan Kegel <dank@kegel.com> wrote:
-> It seems it might be from gcc's spec file:
->
-> *subtarget_link_spec:
-> %{shared:-shared} %{!static: %{rdynamic:-export-dynamic} %{!dynamic-linker:-dynamic-linker /lib/ld-linux.so.2} %{!rpath:-rpath /lib}} %{static:-static}
->
-> as the only place -rpath showed up was in the output of gcc -v as it
-> was linking libc.so.
->
-> Well, that's a clue, anyway! Guess I'll dig through the spec files
-> tomorrow to see if I can figure it out some more.
-
-I've grepped gcc/config/*/*.h and found only SH and old libc1 stuffs
-of i386/alpha include %{!rpath:-rpath ... in 3.3 release and the
-current CVS. And my sh-gcc's specs doesn't have this :-(
-It would be the Right Thing to remove this stuff simply. How about
-the gcc patch below?
-
-Regards,
- kaz
---
---- gcc-3.3/gcc/config/sh/linux.h.orig Sat Aug 2 13:20:57 2003
-+++ gcc-3.3/gcc/config/sh/linux.h Sat Aug 2 13:22:42 2003
-@@ -48,8 +48,7 @@ do { \
- "%{shared:-shared} \
- %{!static: \
- %{rdynamic:-export-dynamic} \
-- %{!dynamic-linker:-dynamic-linker /lib/ld-linux.so.2} \
-- %{!rpath:-rpath /lib}} \
-+ %{!dynamic-linker:-dynamic-linker /lib/ld-linux.so.2}} \
- %{static:-static}"
-
- /* The GNU C++ standard library requires that these macros be defined. */
-
-
diff --git a/patches/gcc/3.3/sh4-kaz-workaround.patch b/patches/gcc/3.3/sh4-kaz-workaround.patch
deleted file mode 100644
index 189b132..0000000
--- a/patches/gcc/3.3/sh4-kaz-workaround.patch
+++ /dev/null
@@ -1,145 +0,0 @@
-[lightly edited to fit my patch directory - dank]
-
-From: kaz Kojima <kkojima@rr.iij4u.or.jp>
-Date: Sat, 09 Aug 2003 09:46:21 +0900
-To: dank@kegel.com
-
-Hi Dan,
-
-I've come back from the vacation and looked glibc string test
-failures on sh4. This looks a gcc problem. gcc-3.3/3.4 doesn't
-compile these tests correctly. The attached testcase aborts on
-gcc-3.3/3.4 -O2 but exits normally gcc-3.2 and gcc-3.0.
-The option -O2 is not essential but it makes the testcase small.
-The failed string tests include the same pattern of the code with
-f=random to generate ramdom strings but they get strings with
-embedded NULL characters :-(
-
-...
-I've got a workaround below for this bug, though it might merely
-paper over the real bug. Anyway, I'd like to send a PR for this.
-
-Regards,
- kaz
---
-int val = 0xff00;
-
-int f (void) { return val; }
-
-unsigned char a[1];
-
-void
-foo (void)
-{
- a[0] = f () & 255;
-
- if (!a[0])
- a[0] = f () & 255;
-
- if (!a[0])
- a[0] = 1 + (f () & 127);
-}
-
-int
-main (int argc, char **argv)
-{
- foo ();
- if (!a[0])
- abort ();
-
- return 0;
-}
-
---
-
-diff -u3prN ORIG/gcc/gcc/config/sh/sh.c LOCAL/gcc/gcc/config/sh/sh.c
---- gcc/gcc/config/sh/sh.c.old Fri Aug 8 18:39:02 2003
-+++ gcc/gcc/config/sh/sh.c Fri Aug 8 22:31:02 2003
-@@ -6657,6 +6657,19 @@ arith_reg_dest (op, mode)
- return arith_reg_operand (op, mode);
- }
-
-+/* Like above, but for SImode compare destinations: forbid paradoxical
-+ subregs, because it would get the combiner confused. */
-+int
-+arith_reg_cmp_dest (op, mode)
-+ rtx op;
-+ enum machine_mode mode;
-+{
-+ if (mode == SImode && GET_CODE (op) == SUBREG
-+ && GET_MODE_SIZE (GET_MODE (SUBREG_REG (op))) < 4)
-+ return 0;
-+ return arith_reg_operand (op, mode);
-+}
-+
- int
- int_gpr_dest (op, mode)
- rtx op;
-diff -u3prN ORIG/gcc/gcc/config/sh/sh.h LOCAL/gcc/gcc/config/sh/sh.h
---- gcc/gcc/config/sh/sh.h.old Fri Aug 8 18:39:02 2003
-+++ gcc/gcc/config/sh/sh.h Fri Aug 8 22:31:02 2003
-@@ -3365,6 +3365,7 @@ extern int rtx_equal_function_value_matt
- {"and_operand", {SUBREG, REG, CONST_INT}}, \
- {"any_register_operand", {SUBREG, REG}}, \
- {"arith_operand", {SUBREG, REG, CONST_INT}}, \
-+ {"arith_reg_cmp_dest", {SUBREG, REG}}, \
- {"arith_reg_dest", {SUBREG, REG}}, \
- {"arith_reg_operand", {SUBREG, REG}}, \
- {"arith_reg_or_0_operand", {SUBREG, REG, CONST_INT, CONST_VECTOR}}, \
-
---- gcc-3.3/gcc/config/sh/sh.md.orig Tue Apr 15 10:06:10 2003
-+++ gcc-3.3/gcc/config/sh/sh.md Sat Aug 9 22:31:13 2003
-@@ -616,7 +616,7 @@
-
- (define_insn ""
- [(set (reg:SI T_REG)
-- (eq:SI (and:SI (match_operand:SI 0 "arith_reg_operand" "z,r")
-+ (eq:SI (and:SI (match_operand:SI 0 "arith_reg_cmp_dest" "z,r")
- (match_operand:SI 1 "arith_operand" "L,r"))
- (const_int 0)))]
- "TARGET_SH1"
-@@ -631,7 +631,7 @@
-
- (define_insn "cmpeqsi_t"
- [(set (reg:SI T_REG)
-- (eq:SI (match_operand:SI 0 "arith_reg_operand" "r,z,r")
-+ (eq:SI (match_operand:SI 0 "arith_reg_cmp_dest" "r,z,r")
- (match_operand:SI 1 "arith_operand" "N,rI,r")))]
- "TARGET_SH1"
- "@
-@@ -642,7 +642,7 @@
-
- (define_insn "cmpgtsi_t"
- [(set (reg:SI T_REG)
-- (gt:SI (match_operand:SI 0 "arith_reg_operand" "r,r")
-+ (gt:SI (match_operand:SI 0 "arith_reg_cmp_dest" "r,r")
- (match_operand:SI 1 "arith_reg_or_0_operand" "r,N")))]
- "TARGET_SH1"
- "@
-@@ -652,7 +652,7 @@
-
- (define_insn "cmpgesi_t"
- [(set (reg:SI T_REG)
-- (ge:SI (match_operand:SI 0 "arith_reg_operand" "r,r")
-+ (ge:SI (match_operand:SI 0 "arith_reg_cmp_dest" "r,r")
- (match_operand:SI 1 "arith_reg_or_0_operand" "r,N")))]
- "TARGET_SH1"
- "@
-@@ -666,7 +666,7 @@
-
- (define_insn "cmpgeusi_t"
- [(set (reg:SI T_REG)
-- (geu:SI (match_operand:SI 0 "arith_reg_operand" "r")
-+ (geu:SI (match_operand:SI 0 "arith_reg_cmp_dest" "r")
- (match_operand:SI 1 "arith_reg_operand" "r")))]
- "TARGET_SH1"
- "cmp/hs %1,%0"
-@@ -674,7 +674,7 @@
-
- (define_insn "cmpgtusi_t"
- [(set (reg:SI T_REG)
-- (gtu:SI (match_operand:SI 0 "arith_reg_operand" "r")
-+ (gtu:SI (match_operand:SI 0 "arith_reg_cmp_dest" "r")
- (match_operand:SI 1 "arith_reg_operand" "r")))]
- "TARGET_SH1"
- "cmp/hi %1,%0"
diff --git a/patches/gcc/3.3/sh4-no-fix-protos.patch b/patches/gcc/3.3/sh4-no-fix-protos.patch
deleted file mode 100644
index c3c478f..0000000
--- a/patches/gcc/3.3/sh4-no-fix-protos.patch
+++ /dev/null
@@ -1,12 +0,0 @@
-See http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10331
-and http://gcc.gnu.org/ml/gcc-patches/2003-06/msg00423.html
-
---- gcc-3.3/gcc/config/sh/t-linux.old Wed Jun 4 13:01:39 2003
-+++ gcc-3.3/gcc/config/sh/t-linux Wed Jun 4 13:02:12 2003
-@@ -1,3 +1,6 @@
-+# Don't run fixproto
-+STMP_FIXPROTO =
-+
- TARGET_LIBGCC2_CFLAGS = -fpic
- LIB1ASMFUNCS_CACHE = _ic_invalidate
-
diff --git a/patches/gcc/3.3/sh4-pthread.patch b/patches/gcc/3.3/sh4-pthread.patch
deleted file mode 100644
index a0c9233..0000000
--- a/patches/gcc/3.3/sh4-pthread.patch
+++ /dev/null
@@ -1,42 +0,0 @@
-Patch to fix following test case failure:
-
- === libstdc++-v3 tests ===
-FAIL: thread/pthread1.cc (test for excess errors)
-Excess errors:
-/opt/cegl-2.0/sh4-unknown-linux-gnu/gcc-ss-3_3-20030714-glibc-2.2.5/sh4-unknown-linux-gnu/bin/ld: cannot find -lthread
-collect2: ld returned 1 exit status
-
-Note that *any* program compiled with -pthread fails:
-
-/opt/cegl-2.0/sh4-unknown-linux-gnu/gcc-ss-3_3-20030714-glibc-2.2.5/bin/sh4-unknown-linux-gnu-gcc hello.c -pthread
-/opt/cegl-2.0/sh4-unknown-linux-gnu/gcc-ss-3_3-20030714-glibc-2.2.5/lib/gcc-lib/sh4-unknown-linux-gnu/3.3.1/../../../../sh4-unknown-linux-gnu/bin/ld: cannot find -lthread
-collect2: ld returned 1 exit status
-
-Compiling with -lpthread on the other hand works fine:
-/opt/cegl-2.0/sh4-unknown-linux-gnu/gcc-ss-3_3-20030714-glibc-2.2.5/bin/sh4-unknown-linux-gnu-gcc hello.c -lpthread
-
-So -pthread is broken in some trivial way; it should invoke -lpthread, not -lthread,
-at least when targeting Linux.
-
-http://www.sh-linux.org/rpm-2003/SRPMS/gcc-3.2.3-3.src.rpm contains a jumbo patch,
-gcc-20030210-sh-linux-1.patch, that includes a fix for this. Here's the
-appropriate hunk (brings in a fix for the documented -mieee option).
-No idea if this fix is completely right, but it works for me...
-- dank@kegel.com 20 Jul 2003
-
-Index: linux.h
-===================================================================
-RCS file: /cvsroot/gcc/gcc/gcc/config/sh/linux.h,v
-retrieving revision 1.9.20.1
-diff -u -d -u -r1.9.20.1 linux.h
---- gcc-ss-3_3-20030714/gcc/config/sh/linux.h.old 6 Jun 2003 02:30:59 -0000 1.9.20.1
-+++ gcc-ss-3_3-20030714/gcc/config/sh/linux.h 20 Jul 2003 23:36:50 -0000
-@@ -59,7 +59,7 @@
- #undef LIB_SPEC
- #define LIB_SPEC \
- "%{shared: -lc} \
-- %{!shared: %{pthread:-lthread} \
-+ %{!shared: %{mieee:-lieee} %{pthread:-lpthread} \
- %{profile:-lc_p} %{!profile: -lc}}"
-
- #undef STARTFILE_SPEC
generated by cgit v0.10.2-6-g49f6 at 2024-04-28 22:10:36 (GMT)