[Bug ld/12721] New: (as or ld) produce a bad jump on mips eabi
http://sourceware.org/bugzilla/show_bug.cgi?id=12721
Summary: (as or ld) produce a bad jump on mips eabi
Product: binutils
Version: 2.18
Status: NEW
Severity: normal
Priority: P2
Component: ld
AssignedTo: unassig...@sources.redhat.com
ReportedBy: wes...@terpstra.ca
I've been bitten by a miscompilation of 'goto' on the mips target. I have
isolated it into a single test-case referenced below. I believe this bug to be
at least 6 years old and it is still present in version 2.18.1~cvs20080103-7
(as used in debian/sid).
The problem boils down to these assembler constructs which SHOULD behave the
same as I understand it:
sw $3,%lo(nextFun)($2)
- lw $1,%got($L894)($28)
+ lui $1,%hi($L894)
nop
addiu $1,$1,%lo($L894)
jr $1
The '%got' variant which is used by default on EABI will lead to a segmentation
fault.
I attached a tarball with the assembler to be (mis)compiled, the original C
file, and a library sufficient to link the final program.
If you want to blame gcc (though i think the assembler it generated is valid)
here is how to rebuild the assembler:
gcc -std=gnu99 -O0 -g -w -fno-strict-aliasing mlyacc.6.preprocessed.c -S -o
mlyacc.6.preprocessed.s
To link the two alternatives:
gcc -o ok mlyacc.6.preprocessed.ok.s link.a -lgmp -lm
gcc -o bad mlyacc.6.preprocessed.bad.s link.a -lgmp -lm
... the gmp version used here is 5.0.1+dfsg-7 from debian sid.
... the glibc version is 2.11.2-11 (again from debian sid)
To observe the bad jump:
gdb ./bad
break mlyacc.6.preprocessed.c:2915
run
Breakpoint 1, Chunk6 () at mlyacc.6.preprocessed.c:2915
2915 *(Word8*)(stackTop + (60)) = (Word8)(Word8)0x0;
(gdb) s
2916 *(CPointer*)(stackTop + (56)) = 25;
(gdb)
2917 do { if (0) fprintf (stderr, "%s:%d: Push (%d)\n", "mlyacc.6.c",
912, 60); stackTop += (60); } while (0);
(gdb)
2919 cont.nextChunk = (void*)Chunk0;
(gdb)
2920 nextFun = 2736; }
(gdb)
2922 goto leaveChunk; /* !!! This jump goes the wrong way !!! */
(gdb)
Program received signal SIGSEGV, Segmentation fault.
0x2aabead0 in _dl_check_caller (caller=0x0, mask=0) at dl-caller.c:39
39 dl-caller.c: No such file or directory.
in dl-caller.c
If you run the same sequence on 'gdb ./ok' the program will proceed past the
bad jump and die at the next bad jump. (Yes, in this file there are multiple
jumps which get miscompiled. mlyacc.6.preprocessed.ok.s only has the first
jump corrected.
--
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.
___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/12633] Crash with dlltool geneated import *.libs in MSVC++ Release target (Debug OK)
http://sourceware.org/bugzilla/show_bug.cgi?id=12633 --- Comment #2 from Martin Storsjö 2011-05-02 11:26:59 UTC --- Created attachment 5700 --> http://sourceware.org/bugzilla/attachment.cgi?id=5700 Minimal test case Here's an attached quite minimal testcase for this issue. The makefile builds a tiny DLL using mingw/gcc, and generates an import library, both using --out-implib with ld and using dlltool. Then the makefile builds two test exes that links to the created import libraries - both exes run just fine on windows. The batch script msvc.bat builds the same test exes using the MSVC command line tools, both with and without the /OPT:REF flag, against both of the created import libraries. The ones built with /OPT:NOREF work fine, while the ones built with /OPT:REF fail. For comparison, it also creates an import library from the .def file using lib.exe, and links against it. For the exes built against this import library, both versions work. -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/12720] [regression] ar d segfaults
http://sourceware.org/bugzilla/show_bug.cgi?id=12720 --- Comment #3 from cvs-commit at gcc dot gnu.org 2011-05-02 12:47:50 UTC --- CVSROOT:/cvs/src Module name:src Changes by:h...@sourceware.org2011-05-02 12:47:47 Modified files: binutils/testsuite: ChangeLog binutils/testsuite/binutils-all: ar.exp Log message: Add testcases for "ar -d" and "ar -m". 2011-05-02 H.J. Lu PR binutils/12720 * binutils-all/ar.exp (delete_an_element): New. (move_an_element): Likewise. Run delete_an_element and move_an_element. Patches: http://sourceware.org/cgi-bin/cvsweb.cgi/src/binutils/testsuite/ChangeLog.diff?cvsroot=src&r1=1.234&r2=1.235 http://sourceware.org/cgi-bin/cvsweb.cgi/src/binutils/testsuite/binutils-all/ar.exp.diff?cvsroot=src&r1=1.15&r2=1.16 -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug ld/12721] (as or ld) produce a bad jump on mips eabi
http://sourceware.org/bugzilla/show_bug.cgi?id=12721 --- Comment #1 from Wesley W. Terpstra 2011-05-02 12:58:47 UTC --- Created attachment 5701 --> http://sourceware.org/bugzilla/attachment.cgi?id=5701 Testcase for bad linking -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
