[Bug other/13221] New: libffi's closure couldn't pass sequence of char and/or short arguments.

[Bug other/13221] New: libffi's closure couldn't pass sequence of char and/or short arguments.

[hos at tamanegi dot org]

In 32bit x86 platform, libffi's closure couldn't pass sequence of char
and/or short type arguments.
Minimum test program is append below.

This probrem maybe fix by following patch below, because passing
arguments are always alignment 4bytes (Also if argument size was less
than 4bytes).

--

/* test program */

#include <ffi.h>

void test_func(ffi_cif *cif, void *rval, void **avals, void *data)
{
    printf("%d, %d\n", *(char *)avals[0], *(char *)avals[1]);
}

main()
{
    ffi_cif cif;
    ffi_closure closure;
    ffi_type *args[2];

    args[0] = &ffi_type_schar;
    args[1] = &ffi_type_schar;

    ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 2, &ffi_type_void, args);
    ffi_prep_closure(&closure, &cif, test_func, NULL);

    ((void(*)(char, char))(&closure))(1, 2);
}

--

Index: ffi.c
===================================================================
RCS file: /cvsroot/gcc/gcc/libffi/src/x86/ffi.c,v
retrieving revision 1.10
diff -u -r1.10 ffi.c
--- ffi.c	21 Mar 2003 13:43:20 -0000	1.10
+++ ffi.c	28 Nov 2003 17:25:13 -0000
@@ -301,8 +301,8 @@
       size_t z;
 
       /* Align if necessary */
-      if (((*p_arg)->alignment - 1) & (unsigned) argp) {
-	argp = (char *) ALIGN(argp, (*p_arg)->alignment);
+      if ((sizeof(int) - 1) & (unsigned) argp) {
+	argp = (char *) ALIGN(argp, sizeof(int));
       }
 
       z = (*p_arg)->size;

-- 
           Summary: libffi's closure couldn't pass sequence of char and/or
                    short arguments.
           Product: gcc
           Version: 3.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: other
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hos at tamanegi dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-cygwin
  GCC host triplet: i686-pc-cygwin
GCC target triplet: i686-pc-cygwin


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2003-11-28 19:03 -------
Can you sent the patch to gcc-patches@?
Andreas what can you say about this patch and test case?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andreast at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2003-11-28 19:03:17
               date|                            |
   Target Milestone|---                         |3.4


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[andreast at gcc dot gnu dot org]

------- Additional Comments From andreast at gcc dot gnu dot org  2003-11-28 19:12 -------
I'll check the type of test case. I think I already added something like this.
But I could be wrong. If not I'll add it.
Thanks.

Regarding the patch, I hope an x86 expert can speak up. I don't own such
architectures :)

 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[andreast at gcc dot gnu dot org]

------- Additional Comments From andreast at gcc dot gnu dot org  2003-11-30 07:01 -------
I added some testcases for this bug:

http://gcc.gnu.org/ml/gcc-patches/2003-11/msg02305.html

With this patch from Hosaka Yuji the failures go away:

http://gcc.gnu.org/ml/gcc-patches/2003-11/msg02311.html

He reported success on the above testcases.
Native configuration is i686-pc-linux-gnu

=== libffi Summary ===

# of expected passes		182
# of unsupported tests		2

So hope to get some feedback from a x86 responsible.

Nevertheless, I'll commit the testcases soon as part of the testsuite.

I also would like to commit the patch if it is ok. But here I need ok for approval.






-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|andreast at gcc dot gnu dot |
                   |org                         |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch
   Last reconfirmed|2003-11-28 19:03:17         |2003-11-30 08:26:13
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2003-12-01 07:23 -------
Subject: Bug 13221

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	andreast@gcc.gnu.org	2003-12-01 07:23:28

Modified files:
	libffi         : ChangeLog 
	libffi/testsuite/libffi.special: unwindtest.cc 
Added files:
	libffi/testsuite/libffi.call: cls_multi_sshort.c 
	                              cls_multi_sshortchar.c 
	                              cls_multi_uchar.c 
	                              cls_multi_schar.c 
	                              cls_multi_ushortchar.c 
	                              cls_multi_ushort.c 

Log message:
	2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
	
	PR other/13221
	* testsuite/libffi.call/cls_multi_sshort.c: New test case.
	* testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
	* testsuite/libffi.call/cls_multi_uchar.c: Likewise.
	* testsuite/libffi.call/cls_multi_schar.c: Likewise.
	* testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
	* testsuite/libffi.call/cls_multi_ushort.c: Likewise.
	
	* testsuite/libffi.special/unwindtest.cc: Cosmetics.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/ChangeLog.diff?cvsroot=gcc&r1=1.147&r2=1.148
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/testsuite/libffi.special/unwindtest.cc.diff?cvsroot=gcc&r1=1.6&r2=1.7
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/testsuite/libffi.call/cls_multi_sshort.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/testsuite/libffi.call/cls_multi_sshortchar.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/testsuite/libffi.call/cls_multi_uchar.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/testsuite/libffi.call/cls_multi_schar.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/testsuite/libffi.call/cls_multi_ushortchar.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/testsuite/libffi.call/cls_multi_ushort.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[zack at gcc dot gnu dot org]

------- Additional Comments From zack at gcc dot gnu dot org  2003-12-02 01:21 -------
patch observed to have been checked in

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2003-12-02 01:22 -------
That was only the testsuite patch :).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[zack at gcc dot gnu dot org]

------- Additional Comments From zack at gcc dot gnu dot org  2003-12-02 01:24 -------
Why is (*p_arg)->alignment wrong?

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[zack at gcc dot gnu dot org]

------- Additional Comments From zack at gcc dot gnu dot org  2003-12-02 05:23 -------
... after some discussion on IRC it was determined that (*p_arg)->alignment is
generic, and so the patch from Hosaka Yuji is correct.  Please check that in and
close the bug report.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2003-12-02 05:39 -------
Subject: Bug 13221

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	andreast@gcc.gnu.org	2003-12-02 05:39:07

Modified files:
	libffi         : ChangeLog 
	libffi/src/x86 : ffi.c 

Log message:
	2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
	
	PR other/13221
	* src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
	Align arguments to 32 bits.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/ChangeLog.diff?cvsroot=gcc&r1=1.148&r2=1.149
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libffi/src/x86/ffi.c.diff?cvsroot=gcc&r1=1.10&r2=1.11



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221

[Bug other/13221] libffi's closure couldn't pass sequence of char and/or short arguments.

[andreast at gcc dot gnu dot org]

------- Additional Comments From andreast at gcc dot gnu dot org  2003-12-02 05:43 -------
Done so.

Andreas

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13221