[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[dhazeghi@yahoo.com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407


dhazeghi@yahoo.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|dhazeghi@yahoo.com          |
  GCC build triplet|                            |ia64-unknown-linux
   GCC host triplet|                            |ia64-unknown-linux
 GCC target triplet|                            |ia64-unknown-linux


------- Additional Comments From dhazeghi@yahoo.com  2003-06-04 00:10 -------
Confirmed still present with gcc 3.3 branch and mainline (20030602).



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[dhazeghi at yahoo dot com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407


dhazeghi at yahoo dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2003-05-08 00:00:00         |2003-07-09 19:13:34
               date|                            |


------- Additional Comments From dhazeghi at yahoo dot com  2003-07-09 19:13 -------
Still present on mainline and branch (20030708).

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kobras at debian dot org


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-11 14:35 -------
*** Bug 11499 has been marked as a duplicate of this bug. ***

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[kobras at debian dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407



------- Additional Comments From kobras at debian dot org  2003-07-14 09:49 -------
Sorry for the duplicate--shouldn't have included the line number of the abort() when looking for 
existing reports...

Some additional data points: The ICE only happens when the test function PointInTri() takes five or 
more arguments of type Point2 and the fifth or higher argument appears in an assignment. I 
wasn't able to reproduce the ICE as soon as I alter the definition of Point2 to contain a different 
number of float variables. gcc won't abort either when the type is changed from float to double or 
int, or when the assignment is replaced by an equivalent memcpy().

(My testcase was extracted from the OpenDX sources as well, by the way. Not much of a surprise--
it seems to be pretty hard to trigger this bug.)

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|critical                    |normal

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[schwab at suse dot de]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407


schwab at suse dot de changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |3.4

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[rakdver at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407



------- Additional Comments From rakdver at gcc dot gnu dot org  2003-10-13 13:00 -------
I posted a fix for this several month ago:

http://gcc.gnu.org/ml/gcc-patches/2003-03/msg01222.html

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[roger at eyesopen dot com]

------- Additional Comments From roger at eyesopen dot com  2003-11-16 22:16 -------
I've confirmed that this has been fixed on mainline on ia64-unknown-linux-gnu.
Janis' simple testcase which ICEs with gcc 3.2.3, now compiles fine with 3.4.
I believe it was fixed by the following patch:

2003-08-27  Richard Sandiford  <rsandifo@redhat.com>

        * calls.c (initialize_argument_information): If an argument has no
        stack space associated with it, and BLOCK_REG_PADDING is defined,
        use it to decide at which end the argument should be padded.
        * function.c (assign_parms): Allocate BLKmode stack slots.
        * config/mips/mips-protos.h (mips_pad_arg_upward): Declare.
        (mips_pad_reg_upward): Declare.
        ...

where the change to assign_parms in function.c, looks identical to Zdenek's
proposed solution to the problem (as referenced in the bugzilla PR) posted
at http://gcc.gnu.org/ml/gcc-patches/2003-03/msg01222.html


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[schwab at suse dot de]

------- Additional Comments From schwab at suse dot de  2003-11-23 19:33 -------
The ICE is still reproducable as of 20031123 when compiling with
-fno-unit-at-a-time (or just remove static from the testcase).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2003-07-29 15:23:47         |2003-11-23 19:33:39
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[schwab at suse dot de]

------- Additional Comments From schwab at suse dot de  2003-11-23 19:35 -------
"Reconfirm bug" does not reopen the bug?

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |NEW


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[wilson at specifixinc dot com]

------- Additional Comments From wilson at specifixinc dot com  2003-12-01 09:22 -------
Subject: Re:  [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711
 with very simple code

schwab at suse dot de wrote:
> ------- Additional Comments From schwab at suse dot de  2003-11-23 19:33 -------
> The ICE is still reproducable as of 20031123 when compiling with
> -fno-unit-at-a-time (or just remove static from the testcase).

There is a critical difference between Zdenek's and Richard's patches. 
Richard's patch puts a PUT_MODE operation after the assign_stack_local 
call, which changes the stack slot back to DImode after it is allocated. 
  Passing BLKmode to assign_stack_local forces it to use 
BIGGEST_ALIGNMENT, which is all Richard was trying to do.

Deleting the PUT_MODE line gets us Zdenek's patch, which solves the 
problem.  However, it isn't clear if this is the right solution.

An alternative would be to fix the IA-64 function_arg function to use 
the argument mode instead of DImode for a single-register HFA.  This 
seems a bit more reasonable.  The following patch is untested.
Index: config/ia64/ia64.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/ia64/ia64.c,v
retrieving revision 1.259
diff -p -r1.259 ia64.c
*** config/ia64/ia64.c	29 Oct 2003 14:23:45 -0000	1.259
--- config/ia64/ia64.c	1 Dec 2003 09:16:54 -0000
*************** ia64_function_arg (CUMULATIVE_ARGS *cum,
*** 3527,3535 ****
  		      ? 1 : GET_MODE_SIZE (gr_mode) / UNITS_PER_WORD;
  	}
  
!       /* If we ended up using just one location, just return that one loc.  */
        if (i == 1)
! 	return XEXP (loc[0], 0);
        else
  	return gen_rtx_PARALLEL (mode, gen_rtvec_v (i, loc));
      }
--- 3527,3536 ----
  		      ? 1 : GET_MODE_SIZE (gr_mode) / UNITS_PER_WORD;
  	}
  
!       /* If we ended up using just one location, just return that one loc, but
! 	 change the mode back to the argument mode.  */
        if (i == 1)
! 	return gen_rtx_REG (mode, REGNO (XEXP (loc[0], 0)));
        else
  	return gen_rtx_PARALLEL (mode, gen_rtvec_v (i, loc));
      }


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[schwab at suse dot de]

------- Additional Comments From schwab at suse dot de  2003-12-01 12:36 -------
I can confirm that this fixes the bug. 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2003-12-02 01:12 -------
Subject: Bug 8407

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	wilson@gcc.gnu.org	2003-12-02 01:12:11

Modified files:
	gcc            : ChangeLog 
	gcc/config/ia64: ia64.c 

Log message:
	PR target/8407
	* config/ia64/ia64.c (ia64_function_arg): For single-reg HFA, call
	gen_rtx_REG to create new reg with argument mode.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.1885&r2=2.1886
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/ia64/ia64.c.diff?cvsroot=gcc&r1=1.259&r2=1.260



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[wilson at gcc dot gnu dot org]

------- Additional Comments From wilson at gcc dot gnu dot org  2003-12-02 01:34 -------
Patch checked in.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[gdr at integrable-solutions dot net]

------- Additional Comments From gdr at integrable-solutions dot net  2003-12-02 02:46 -------
Subject: Re:  [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

"wilson at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org> writes:

| Patch checked in.

Jim, do you think such a patch could be safely applied to 3_3-branch?

Gaby
 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[wilson at specifixinc dot com]

------- Additional Comments From wilson at specifixinc dot com  2003-12-02 02:59 -------
Subject: Re:  [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711
 with very simple code

gdr at integrable-solutions dot net wrote:
> Jim, do you think such a patch could be safely applied to 3_3-branch?

Yes.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2003-12-02 11:51 -------
Subject: Bug 8407

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	hammer-3_3-branch
Changes by:	schwab@gcc.gnu.org	2003-12-02 11:51:44

Modified files:
	gcc            : ChangeLog.hammer function.c 
	gcc/config/ia64: ia64.c 

Log message:
	Merge from mainline
	
	2003-12-01  James E Wilson  <wilson@specifixinc.com>
	
	PR target/8407
	* config/ia64/ia64.c (ia64_function_arg): For single-reg HFA, call
	gen_rtx_REG to create new reg with argument mode.
	
	Merge from mainline
	
	2003-08-27  Richard Sandiford  <rsandifo@redhat.com>
	
	* function.c (assign_parms): Allocate BLKmode stack slots.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.hammer.diff?cvsroot=gcc&only_with_tag=hammer-3_3-branch&r1=1.1.2.230&r2=1.1.2.231
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/function.c.diff?cvsroot=gcc&only_with_tag=hammer-3_3-branch&r1=1.387.2.15&r2=1.387.2.16
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/ia64/ia64.c.diff?cvsroot=gcc&only_with_tag=hammer-3_3-branch&r1=1.195.2.11&r2=1.195.2.12



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[schwab at suse dot de]

------- Additional Comments From schwab at suse dot de  2003-12-02 12:39 -------
> > Jim, do you think such a patch could be safely applied to 3_3-branch? 
>  
> Yes. 
 
But AFAIK it needs to be combined with the patch referenced in comment #11 to 
have effect. 
 
 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[wilson at specifixinc dot com]

------- Additional Comments From wilson at specifixinc dot com  2003-12-03 01:04 -------
Subject: Re:  [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711
 with very simple code

schwab at suse dot de wrote:
> But AFAIK it needs to be combined with the patch referenced in comment #11 to 
> have effect. 

No, we don't need Richard Sandiford's patch.  The comments in #11 were 
based on a misunderstanding of what Richard's patch did.

I just tested a 3.3 cross compiler with my patch hacked in, and it fixed 
the testcase.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407

[Bug target/8407] [IA-64] ICE in simplify_gen_subreg, at simplify-rtx.c:2711 with very simple code

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2003-12-03 19:58 -------
Subject: Bug 8407

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_3-branch
Changes by:	wilson@gcc.gnu.org	2003-12-03 19:58:39

Modified files:
	gcc            : ChangeLog 
	gcc/config/ia64: ia64.c 

Log message:
	PR target/8407
	* config/ia64/ia64.c (ia64_function_arg): For single-reg HFA, call
	gen_rtx_REG to create new reg with argument mode.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.16114.2.819&r2=1.16114.2.820
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/ia64/ia64.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.198.2.15&r2=1.198.2.16



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=8407