This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Bootstrap comparison failure on i686-linux
- To: gcc at gcc dot gnu dot org
- Subject: Bootstrap comparison failure on i686-linux
- From: Andreas Jaeger <aj at suse dot de>
- Date: 06 Jan 2001 20:17:33 +0100
Bootstrapping gcc CVS current (checked out around 18:00 UTC 6th
January 2001) with gcc 2.95.2 on i686-linux-gnu with glibc 2.2 gives a
comparsion failure on haifa-sched.o:
[...]
Bootstrap comparison failure!
haifa-sched.o differs
make[1]: *** [compare] Error 1
I've run objdump -xD on both haifa-sched.o files and diffed the
output. The difference is just register renaming. How can this
happen? Any ideas?
My configure command is:
/cvs/gcc/configure --prefix=/opt/gcc-2.97.test --enable-shared \
--enable-threads=posix \
--with-gnu-as --with-gnu-ld --disable-nls --with-system-zlib \
--enable-languages=c,objc,c++,f77
And I run "make bootstrap" - until I got the above failure. I didn't
pass any CFLAGS.
Andreas
--- h1 Sat Jan 6 20:10:38 2001
+++ h2 Sat Jan 6 20:10:35 2001
@@ -1,6 +1,6 @@
-haifa-sched.o: file format elf32-i386
-haifa-sched.o
+stage2/haifa-sched.o: file format elf32-i386
+stage2/haifa-sched.o
architecture: i386, flags 0x00000011:
HAS_RELOC, HAS_SYMS
start address 0x0000000000000000
@@ -2757,7 +2757,7 @@
1b4c: 8d 04 8d 00 00 00 00 lea 0x0(,%ecx,4),%eax
1b53: 29 c2 sub %eax,%edx
1b55: 83 f9 02 cmp $0x2,%ecx
- 1b58: 8d 5a 04 lea 0x4(%edx),%ebx
+ 1b58: 8d 72 04 lea 0x4(%edx),%esi
1b5b: 0f 84 34 05 00 00 je 2095 <schedule_block+0x705>
1b61: 83 f9 02 cmp $0x2,%ecx
1b64: 7e 14 jle 1b7a <schedule_block+0x1ea>
@@ -2765,7 +2765,7 @@
1b67: R_386_32 .text
1b6b: 6a 04 push $0x4
1b6d: 51 push %ecx
- 1b6e: 53 push %ebx
+ 1b6e: 56 push %esi
1b6f: e8 fc ff ff ff call 1b70 <schedule_block+0x1e0>
1b70: R_386_PC32 qsort
1b74: 8b 4d e4 mov 0xffffffe4(%ebp),%ecx
@@ -3225,27 +3225,27 @@
2090: e9 4b fa ff ff jmp 1ae0 <schedule_block+0x150>
2095: 8b 42 08 mov 0x8(%edx),%eax
2098: 31 ff xor %edi,%edi
- 209a: 89 de mov %ebx,%esi
+ 209a: 89 f3 mov %esi,%ebx
209c: 89 45 c4 mov %eax,0xffffffc4(%ebp)
209f: 90 nop
20a0: 85 ff test %edi,%edi
20a2: 78 21 js 20c5 <schedule_block+0x735>
20a4: 83 ec 08 sub $0x8,%esp
- 20a7: 8d 45 c4 lea 0xffffffc4(%ebp),%eax
- 20aa: 50 push %eax
- 20ab: 56 push %esi
+ 20a7: 8d 55 c4 lea 0xffffffc4(%ebp),%edx
+ 20aa: 52 push %edx
+ 20ab: 53 push %ebx
20ac: e8 9f e2 ff ff call 350 <rank_for_schedule>
20b1: 83 c4 10 add $0x10,%esp
20b4: 85 c0 test %eax,%eax
20b6: 78 0d js 20c5 <schedule_block+0x735>
- 20b8: 8b 04 bb mov (%ebx,%edi,4),%eax
- 20bb: 83 ee 04 sub $0x4,%esi
- 20be: 89 44 bb 04 mov %eax,0x4(%ebx,%edi,4)
+ 20b8: 8b 04 be mov (%esi,%edi,4),%eax
+ 20bb: 83 eb 04 sub $0x4,%ebx
+ 20be: 89 44 be 04 mov %eax,0x4(%esi,%edi,4)
20c2: 4f dec %edi
20c3: eb db jmp 20a0 <schedule_block+0x710>
20c5: 8b 45 c4 mov 0xffffffc4(%ebp),%eax
20c8: 8b 4d e4 mov 0xffffffe4(%ebp),%ecx
- 20cb: 89 44 bb 04 mov %eax,0x4(%ebx,%edi,4)
+ 20cb: 89 44 be 04 mov %eax,0x4(%esi,%edi,4)
20cf: e9 a6 fa ff ff jmp 1b7a <schedule_block+0x1ea>
20d4: a1 00 00 00 00 mov 0x0,%eax
20d5: R_386_32 sched_verbose
--
Andreas Jaeger
SuSE Labs aj@suse.de
private aj@arthur.inka.de
http://www.suse.de/~aj