[Bug inline-asm/32178] New: RAX and EAX are used as two different registers

[Bug inline-asm/32178] New: RAX and EAX are used as two different registers

[henne at nachtwindheim dot de]

First here is my gcc -v output:
Using built-in specs.
Target: x86_64-suse-linux
Configured with: ../configure --enable-threads=posix --prefix=/usr
--with-local-prefix=/usr/local --infodir=/usr/share/info
--mandir=/usr/share/man --libdir=/usr/lib64 --libexecdir=/usr/lib64
--enable-languages=c,c++,objc,fortran,obj-c++,java,ada
--enable-checking=release --with-gxx-include-dir=/usr/include/c++/4.1.2
--enable-ssp --disable-libssp --disable-libgcj --with-slibdir=/lib64
--with-system-zlib --enable-shared --enable-__cxa_atexit
--enable-libstdcxx-allocator=new --program-suffix=-4.1
--enable-version-specific-runtime-libs --without-system-libunwind
--with-cpu=generic --host=x86_64-suse-linux
Thread model: posix
gcc version 4.1.2 20061115 (prerelease) (SUSE Linux)

I have tracked down a problem while compiling a larger program and make a
compilable code snippet which reproduces the problem (SEGFAULT):

/*snippet.c*/

typedef volatile long fastlock_t;

static inline int fastlock_trylock(fastlock_t *l)
{

        int lacquired = 0;

        /* note: no 'lock' prefix even on SMP since xchg is always atomic */
        asm volatile("movl  $1,%0 \n\t" \
                "xchgl %0,%1 \n\t" \
                "xorl  $1,%0"
                : "=r"(lacquired)
                : "m"(*l)
                : "memory");

        if (lacquired)
                return 1;

        return 0;
}

int main(void)
{
        fastlock_t* test;

        fastlock_trylock(test);
        return 0;
}

The compilation worked without warnigs.
The precompiled file looked like this:

# 1 "snippet.c"
# 1 "<built-in>"
# 1 "<command line>"
# 1 "snippet.c"
typedef volatile long fastlock_t;

static inline int fastlock_trylock(fastlock_t *l)
{

 int lacquired = 0;


 asm volatile("movl  $1,%0 \n\t" "xchgl %0,%1 \n\t" "xorl  $1,%0"


  : "=r"(lacquired)
  : "m"(*l)
  : "memory");

 if (lacquired)
  return 1;

 return 0;
}

int main(void)
{
 fastlock_t* test;

 fastlock_trylock(test);
 return 0;
}

Here until now everything looks fine,
but when generating the assembly the error happens.
Here is the assembly of fastlock_trylock:

fastlock_trylock:
.LFB2:
        pushq   %rbp
.LCFI3:
        movq    %rsp, %rbp
.LCFI4:
        movq    %rdi, -24(%rbp)
        movl    $0, -4(%rbp)
        movq    -24(%rbp), %rax
#APP
        movl  $1,%eax 
        xchgl %eax,(%rax) 
        xorl  $1,%eax
#NO_APP
        movl    %eax, -4(%rbp)
        cmpl    $0, -4(%rbp)
        je      .L4
        movl    $1, -28(%rbp)
        jmp     .L6
.L4:
        movl    $0, -28(%rbp)
.L6:
        movl    -28(%rbp), %eax
        leave
        ret

In rax is *l and eax is used for the value, so the pointer is overwritten and
the SEGFAULT happens while accessing the pointer with xchgl.
I'm not familiar with gcc inline asm but I belive that the code of snippet.c
should be allright.

Greets Henne


-- 
           Summary: RAX and EAX are used as two different registers
           Product: gcc
           Version: 4.1.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: inline-asm
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: henne at nachtwindheim dot de
 GCC build triplet: gcc version 4.1.2 20061115 (prerelease) (SUSE Linux)
  GCC host triplet: x86_64-suse-linux
GCC target triplet: x86_64-suse-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32178

[Bug inline-asm/32178] RAX and EAX are used as two different registers

[henne at nachtwindheim dot de]

------- Comment #1 from henne at nachtwindheim dot de  2007-06-01 11:38 -------
Created an attachment (id=13645)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13645&action=view)
The eample snippet.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32178

[Bug inline-asm/32178] RAX and EAX are used as two different registers

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2007-06-01 14:14 -------
Your inline-asm is broken, you early clobber the output so you need to mark as
such like:

        asm volatile("movl  $1,%0 \n\t" \
                "xchgl %0,%1 \n\t" \
                "xorl  $1,%0"
                : "=&r"(lacquired)
                : "m"(*l)
                : "memory");


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32178