public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug middle-end/36253]  New: Caller-save stack slot may not have proper alignment
@ 2008-05-17  0:18 hjl dot tools at gmail dot com
  2008-05-17  2:18 ` [Bug middle-end/36253] " hjl dot tools at gmail dot com
                   ` (5 more replies)
  0 siblings, 6 replies; 7+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-05-17  0:18 UTC (permalink / raw)
  To: gcc-bugs

setup_save_areas calls assign_stack_local to allocate a stack
slot for a hard register in the widest mode. assign_stack_local
has

  if (align == 0)
    {    
      tree type;

      if (mode == BLKmode)
        alignment = BIGGEST_ALIGNMENT;
      else 
        alignment = GET_MODE_ALIGNMENT (mode);

      /* Allow the target to (possibly) increase the alignment of this
         stack slot.  */
      type = lang_hooks.types.type_for_mode (mode, 0);
      if (type)
        alignment = LOCAL_ALIGNMENT (type, alignment);

      alignment /= BITS_PER_UNIT;
    }    

For x87, the widest mode is XF, which is 12byte aligned at 4 byte.
So a 12byte stack slot aligned at 4 byte is allocated for an x87
register.  For x87, DF is 8 byte aligned at 8 byte. If it turns out
later that we need to save/restore DF instead XF, we load/store the DF
register at 4 byte instead of 8 byte. It won't give us the best x87
performance. One fix is change

     if (type)
        alignment = LOCAL_ALIGNMENT (type, alignment);

to

    alignment = LOCAL_ALIGNMENT (type, mode, alignment);

If type is NULL, it is used for stack slot. We can return proper
alignment based on mode.


-- 
           Summary: Caller-save stack slot may not have proper alignment
           Product: gcc
           Version: 4.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hjl dot tools at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36253


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug middle-end/36253] Caller-save stack slot may not have proper alignment
  2008-05-17  0:18 [Bug middle-end/36253] New: Caller-save stack slot may not have proper alignment hjl dot tools at gmail dot com
@ 2008-05-17  2:18 ` hjl dot tools at gmail dot com
  2008-05-17 15:19 ` hjl dot tools at gmail dot com
                   ` (4 subsequent siblings)
  5 siblings, 0 replies; 7+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-05-17  2:18 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #1 from hjl dot tools at gmail dot com  2008-05-17 02:17 -------
With this patch to verify alignment of caller-save slot:

Index: caller-save.c
===================================================================
--- caller-save.c       (revision 2603)
+++ caller-save.c       (working copy)
@@ -704,6 +704,12 @@ insert_restore (struct insn_chain *chain
     mem = adjust_address (mem, save_mode[regno], 0);
   else
     mem = copy_rtx (mem);
+
+  /* Verify that the alignment of spill space is equal to or greater
+     than required.  */
+  if (GET_MODE_ALIGNMENT (GET_MODE (mem)) > MEM_ALIGN (mem))
+    gcc_unreachable ();
+
   pat = gen_rtx_SET (VOIDmode,
                     gen_rtx_REG (GET_MODE (mem),
                                  regno), mem);
@@ -776,6 +782,12 @@ insert_save (struct insn_chain *chain, i
     mem = adjust_address (mem, save_mode[regno], 0);
   else
     mem = copy_rtx (mem);
+
+  /* Verify that the alignment of spill space is equal to or greater
+     than required.  */
+  if (GET_MODE_ALIGNMENT (GET_MODE (mem)) > MEM_ALIGN (mem))
+    gcc_unreachable ();
+
   pat = gen_rtx_SET (VOIDmode, mem,
                     gen_rtx_REG (GET_MODE (mem),
                                  regno));
Index: function.c
===================================================================
--- function.c  (revision 2603)
+++ function.c  (working copy)
@@ -341,7 +341,7 @@ assign_stack_local (enum machine_mode mo
 {
   rtx x, addr;
   int bigend_correction = 0;
-  unsigned int alignment;
+  unsigned int alignment, alignment_in_bits;
   int frame_off, frame_alignment, frame_phase;

   if (align == 0)
@@ -378,8 +378,10 @@ assign_stack_local (enum machine_mode mo
   if (alignment * BITS_PER_UNIT > PREFERRED_STACK_BOUNDARY)
     alignment = PREFERRED_STACK_BOUNDARY / BITS_PER_UNIT;

-  if (crtl->stack_alignment_needed < alignment * BITS_PER_UNIT)
-    crtl->stack_alignment_needed = alignment * BITS_PER_UNIT;
+  alignment_in_bits = alignment * BITS_PER_UNIT;
+
+  if (crtl->stack_alignment_needed < alignment_in_bits)
+    crtl->stack_alignment_needed = alignment_in_bits;

   /* Calculate how many bytes the start of local variables is off from
      stack alignment.  */
@@ -432,6 +434,7 @@ assign_stack_local (enum machine_mode mo
     frame_offset += size;

   x = gen_rtx_MEM (mode, addr);
+  set_mem_align (x, alignment_in_bits);
   MEM_NOTRAP_P (x) = 1;

   stack_slot_list

On Linux/Intel64, I got

gnu/java/awt/GradientPaintContext.java: In class
'gnu.java.awt.GradientPaintContext':
gnu/java/awt/GradientPaintContext.java: In method
'gnu.java.awt.GradientPaintContext.getRaster(int,int,int,int)':
In file included from gnu/java/awt/EventModifier.java:105,
                 from gnu/java/awt/EmbeddedWindow.java:137,
                 from gnu/java/awt/ComponentReshapeEvent.java:84,
                 from gnu/java/awt/ComponentDataBlitOp.java:154,
                 from gnu/java/awt/ClasspathToolkit.java:231,
                 from gnu/java/awt/ClasspathGraphicsEnvironment.java:65,
                 from gnu/java/awt/Buffers.java:222,
                 from
/export/gnu/src/gcc-expand/gcc/libjava/classpath/gnu/java/awt/BitwiseXORComposite.java:292,
                 from
/export/gnu/src/gcc-expand/gcc/libjava/classpath/gnu/java/awt/BitwiseXORComposite.java:202,
                 from gnu/java/awt/BitwiseXORComposite.java:130,
                 from gnu/java/awt/BitMaskExtent.java:77,
                 from
/export/gnu/src/gcc-expand/gcc/libjava/classpath/gnu/java/awt/AWTUtilities.java:253,
                 from
/export/gnu/src/gcc-expand/gcc/libjava/classpath/gnu/java/awt/AWTUtilities.java:296,
                 from gnu/java/awt/AWTUtilities.java:895,
                 from <built-in>:63:
gnu/java/awt/GradientPaintContext.java:153: internal compiler error: in
insert_save, at caller-save.c:789
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
make[6]: *** [gnu/java/awt.lo] Error 1
make[6]: *** Waiting for unfinished jobs....

when I was building 32bit libjava.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36253


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug middle-end/36253] Caller-save stack slot may not have proper alignment
  2008-05-17  0:18 [Bug middle-end/36253] New: Caller-save stack slot may not have proper alignment hjl dot tools at gmail dot com
  2008-05-17  2:18 ` [Bug middle-end/36253] " hjl dot tools at gmail dot com
@ 2008-05-17 15:19 ` hjl dot tools at gmail dot com
  2008-05-17 23:51 ` hjl dot tools at gmail dot com
                   ` (3 subsequent siblings)
  5 siblings, 0 replies; 7+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-05-17 15:19 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #2 from hjl dot tools at gmail dot com  2008-05-17 15:18 -------
A patch is posted at

http://gcc.gnu.org/ml/gcc-patches/2008-05/msg01021.html


-- 

hjl dot tools at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://gcc.gnu.org/ml/gcc-
                   |                            |patches/2008-
                   |                            |05/msg01021.html
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36253


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug middle-end/36253] Caller-save stack slot may not have proper alignment
  2008-05-17  0:18 [Bug middle-end/36253] New: Caller-save stack slot may not have proper alignment hjl dot tools at gmail dot com
  2008-05-17  2:18 ` [Bug middle-end/36253] " hjl dot tools at gmail dot com
  2008-05-17 15:19 ` hjl dot tools at gmail dot com
@ 2008-05-17 23:51 ` hjl dot tools at gmail dot com
  2008-05-17 23:59 ` hjl at gcc dot gnu dot org
                   ` (2 subsequent siblings)
  5 siblings, 0 replies; 7+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-05-17 23:51 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #3 from hjl dot tools at gmail dot com  2008-05-17 23:50 -------
An updated patch is posted at

http://gcc.gnu.org/ml/gcc-patches/2008-05/msg01029.html


-- 

hjl dot tools at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |Joey dot ye at intel dot
                   |                            |com, xuepeng dot guo at
                   |                            |intel dot com
                URL|http://gcc.gnu.org/ml/gcc-  |http://gcc.gnu.org/ml/gcc-
                   |patches/2008-               |patches/2008-
                   |05/msg01021.html            |05/msg01029.html


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36253


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug middle-end/36253] Caller-save stack slot may not have proper alignment
  2008-05-17  0:18 [Bug middle-end/36253] New: Caller-save stack slot may not have proper alignment hjl dot tools at gmail dot com
                   ` (2 preceding siblings ...)
  2008-05-17 23:51 ` hjl dot tools at gmail dot com
@ 2008-05-17 23:59 ` hjl at gcc dot gnu dot org
  2008-05-26 14:01 ` hjl at gcc dot gnu dot org
  2008-05-26 14:29 ` hjl dot tools at gmail dot com
  5 siblings, 0 replies; 7+ messages in thread
From: hjl at gcc dot gnu dot org @ 2008-05-17 23:59 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #4 from hjl at gcc dot gnu dot org  2008-05-17 23:58 -------
Subject: Bug 36253

Author: hjl
Date: Sat May 17 23:57:59 2008
New Revision: 135486

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=135486
Log:
2008-05-17  H.J. Lu  <hongjiu.lu@intel.com>

        PR middle-end/36253
        * caller-save.c (insert_restore): Verify alignment of spill
        space.
        (insert_save): Likewise.

        * cfgexpand.c (LOCAL_ALIGNMENT): Removed.
        (get_decl_align_unit): Pass VOIDmode to LOCAL_ALIGNMENT.

        * defaults.h (LOCAL_ALIGNMENT): New. Provide default.

        * function.c (LOCAL_ALIGNMENT): Removed.
        (get_stack_local_alignment): New.
        (assign_stack_local): Use it.  Set alignment on stack slot.
        (assign_stack_temp_for_type): Use get_stack_local_alignment.

        * config/bfin/bfin.c (bfin_local_alignment): Updated.
        * config/bfin/bfin.h (LOCAL_ALIGNMENT): Likewise.
        * config/bfin/bfin-protos.h (bfin_local_alignment): Likewise.
        * config/i386/i386.h (LOCAL_ALIGNMENT): Likewise.
        * config/i386/i386-protos.h (ix86_local_alignment): Likewise.
        * config/mips/mips.h (LOCAL_ALIGNMENT): Likewise.
        * config/mmix/mmix.c (mmix_local_alignment): Likewise.
        * config/mmix/mmix.h (LOCAL_ALIGNMENT): Likewise.
        * config/mmix/mmix-protos.h (mmix_local_alignment): Likewise.
        * config/rs6000/rs6000.h (LOCAL_ALIGNMENT): Likewise.
        * config/score/score.h (LOCAL_ALIGNMENT): Likewise.
        * config/sh/sh.h (LOCAL_ALIGNMENT): Likewise.
        * config/sparc/sparc.h (LOCAL_ALIGNMENT): Likewise.
        * config/spu/spu.h (LOCAL_ALIGNMENT): Likewise.

        * config/i386/i386.c (ix86_local_alignment): Handle caller-save
        stack slot in XFmode.

        * config/rs6000/rs6000.h (LOCAL_ALIGNMENT_1): New.
        * config/sh/sh.h (LOCAL_ALIGNMENT_1): Likewise.

        * doc/tm.texi (LOCAL_ALIGNMENT): Add mode.

Modified:
    branches/stack/gcc/ChangeLog.stackalign
    branches/stack/gcc/caller-save.c
    branches/stack/gcc/cfgexpand.c
    branches/stack/gcc/config/bfin/bfin-protos.h
    branches/stack/gcc/config/bfin/bfin.c
    branches/stack/gcc/config/bfin/bfin.h
    branches/stack/gcc/config/i386/i386-protos.h
    branches/stack/gcc/config/i386/i386.c
    branches/stack/gcc/config/i386/i386.h
    branches/stack/gcc/config/mips/mips.h
    branches/stack/gcc/config/mmix/mmix-protos.h
    branches/stack/gcc/config/mmix/mmix.c
    branches/stack/gcc/config/mmix/mmix.h
    branches/stack/gcc/config/rs6000/rs6000.h
    branches/stack/gcc/config/score/score.h
    branches/stack/gcc/config/sh/sh.h
    branches/stack/gcc/config/sparc/sparc.h
    branches/stack/gcc/config/spu/spu.h
    branches/stack/gcc/defaults.h
    branches/stack/gcc/doc/tm.texi
    branches/stack/gcc/function.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36253


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug middle-end/36253] Caller-save stack slot may not have proper alignment
  2008-05-17  0:18 [Bug middle-end/36253] New: Caller-save stack slot may not have proper alignment hjl dot tools at gmail dot com
                   ` (3 preceding siblings ...)
  2008-05-17 23:59 ` hjl at gcc dot gnu dot org
@ 2008-05-26 14:01 ` hjl at gcc dot gnu dot org
  2008-05-26 14:29 ` hjl dot tools at gmail dot com
  5 siblings, 0 replies; 7+ messages in thread
From: hjl at gcc dot gnu dot org @ 2008-05-26 14:01 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #5 from hjl at gcc dot gnu dot org  2008-05-26 14:00 -------
Subject: Bug 36253

Author: hjl
Date: Mon May 26 13:59:56 2008
New Revision: 135927

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=135927
Log:
2008-05-26  H.J. Lu  <hongjiu.lu@intel.com>

        PR middle-end/36253
        * caller-save.c (insert_restore): Verify alignment of spill
        space.
        (insert_save): Likewise.

        * cfgexpand.c (LOCAL_ALIGNMENT): Removed.

        * defaults.h (LOCAL_ALIGNMENT): New. Provide default.
        (STACK_SLOT_ALIGNMENT): Likewise.

        * function.c (LOCAL_ALIGNMENT): Removed.
        (get_stack_local_alignment): New.
        (assign_stack_local): Use it.  Set alignment on stack slot.
        (assign_stack_temp_for_type): Use get_stack_local_alignment.

        * config/i386/i386.h (LOCAL_ALIGNMENT): Updated.
        (STACK_SLOT_ALIGNMENT): New.

        * config/i386/i386.c (ix86_local_alignment): Handle caller-save
        stack slot in XFmode.

        * doc/tm.texi (STACK_SLOT_ALIGNMENT): New.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/caller-save.c
    trunk/gcc/cfgexpand.c
    trunk/gcc/config/i386/i386-protos.h
    trunk/gcc/config/i386/i386.c
    trunk/gcc/config/i386/i386.h
    trunk/gcc/defaults.h
    trunk/gcc/doc/tm.texi
    trunk/gcc/function.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36253


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug middle-end/36253] Caller-save stack slot may not have proper alignment
  2008-05-17  0:18 [Bug middle-end/36253] New: Caller-save stack slot may not have proper alignment hjl dot tools at gmail dot com
                   ` (4 preceding siblings ...)
  2008-05-26 14:01 ` hjl at gcc dot gnu dot org
@ 2008-05-26 14:29 ` hjl dot tools at gmail dot com
  5 siblings, 0 replies; 7+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-05-26 14:29 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #6 from hjl dot tools at gmail dot com  2008-05-26 14:28 -------
Fixed by

http://gcc.gnu.org/ml/gcc-patches/2008-05/msg01615.html


-- 

hjl dot tools at gmail dot com changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36253


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2008-05-26 14:29 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-05-17  0:18 [Bug middle-end/36253] New: Caller-save stack slot may not have proper alignment hjl dot tools at gmail dot com
2008-05-17  2:18 ` [Bug middle-end/36253] " hjl dot tools at gmail dot com
2008-05-17 15:19 ` hjl dot tools at gmail dot com
2008-05-17 23:51 ` hjl dot tools at gmail dot com
2008-05-17 23:59 ` hjl at gcc dot gnu dot org
2008-05-26 14:01 ` hjl at gcc dot gnu dot org
2008-05-26 14:29 ` hjl dot tools at gmail dot com

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).