[Bug target/10567] -fno-delayed-branch no honored when -mflat

[Bug target/10567] -fno-delayed-branch no honored when -mflat

[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=10567



------- Additional Comments From dhazeghi@yahoo.com  2003-05-26 08:38 -------
Hello,

is this problem still present in gcc 3.3 and mainline? Thanks,

Dara



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

[Bug target/10567] -fno-delayed-branch no honored when -mflat

[pinskia@physics.uc.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=10567


pinskia@physics.uc.edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


------- Additional Comments From pinskia@physics.uc.edu  2003-05-26 14:09 -------
See Dara's question.



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

[Bug target/10567] -fno-delayed-branch no honored when -mflat

[pinskia 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=10567


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2003-08-24 02:35:42
               date|                            |


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-24 02:35 -------
I can confirm this on the mainline (20030714).

[Bug target/10567] -fno-delayed-branch no honored when -mflat

[pinskia 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=10567



------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-24 02:41 -------
It is also in the mainline (20030823).

[Bug target/10567] -fno-delayed-branch no honored when -mflat

[pinskia 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=10567



------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-24 02:54 -------
David could you test a patch for me as I do not have access to a sparc?
Index: sparc.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/sparc/sparc.c,v
retrieving revision 1.256
diff -u -p -r1.256 sparc.c
--- sparc.c     16 Aug 2003 23:59:56 -0000      1.256
+++ sparc.c     24 Aug 2003 02:51:18 -0000
@@ -7223,11 +7223,13 @@ sparc_flat_function_epilogue (FILE *file
            fprintf (file, "\tset\t" HOST_WIDE_INT_PRINT_DEC ", %s\n",
                     size, t1_str);
        }
-
-      if (current_function_returns_struct)
-       fprintf (file, "\tjmp\t%%o7+12\n");
-      else
-       fprintf (file, "\tretl\n");
+      if (flag_delayed_branch)
+        {
+          if (current_function_returns_struct)
+           fprintf (file, "\tjmp\t%%o7+12\n");
+          else
+           fprintf (file, "\tretl\n");
+        }
 
       /* If the only register saved is the return address, we need a
         nop, unless we have an instruction to put into it.  Otherwise
@@ -7247,8 +7249,16 @@ sparc_flat_function_epilogue (FILE *file
       else if (size > 0)
        fprintf (file, "\tadd\t%s, %d, %s\n", sp_str, (int) size, sp_str);
 
-      else
+      else if (flag_delayed_branch)
        fprintf (file, "\tnop\n");
+     
+      if (!flag_delayed_branch)
+        {
+          if (current_function_returns_struct)
+            fprintf (file, "\tjmp\t%%o7+12\n\tnop\n");
+          else
+            fprintf (file, "\tretl\n\tnop\n");
+        }
     }
 
   /* Reset state info for each function.  */

[Bug target/10567] -fno-delayed-branch not honored in function epilogues

[ebotcazou 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=10567


ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|-fno-delayed-branch no      |-fno-delayed-branch not
                   |honored when -mflat         |honored in function
                   |                            |epilogues


------- Additional Comments From ebotcazou at gcc dot gnu dot org  2003-09-08 09:07 -------
It turns out that -fno-delayed-branch is not honored in the non-flat case
either, so Andrew's patch is not sufficient.

[Bug target/10567] -fno-delayed-branch not honored in function epilogues

[ebotcazou 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=10567


ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |ebotcazou at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


------- Additional Comments From ebotcazou at gcc dot gnu dot org  2003-09-08 09:08 -------
I'm going to complete Andrew's patch and test it.

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou 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=10567


ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|-fno-delayed-branch not     |-fno-delayed-branch not
                   |honored in function         |honored in back-end
                   |epilogues                   |


------- Additional Comments From ebotcazou at gcc dot gnu dot org  2003-09-09 14:19 -------
-fno-delayed-branch is not honored either by the following two peepholes:

;; Now peepholes to do a call followed by a jump.

(define_peephole
  [(parallel [(set (match_operand 0 "" "")
		   (call (mem:SI (match_operand:SI 1 "call_operand_address" "ps"))
			 (match_operand 2 "" "")))
	      (clobber (reg:SI 15))])
   (set (pc) (label_ref (match_operand 3 "" "")))]
  "short_branch (INSN_UID (insn), INSN_UID (operands[3]))
   && (USING_SJLJ_EXCEPTIONS || ! can_throw_internal (ins1))
   && sparc_cpu != PROCESSOR_ULTRASPARC
   && sparc_cpu != PROCESSOR_ULTRASPARC3"
  "call\t%a1, %2\n\tadd\t%%o7, (%l3-.-4), %%o7")

(define_peephole
  [(parallel [(call (mem:SI (match_operand:SI 0 "call_operand_address" "ps"))
		    (match_operand 1 "" ""))
	      (clobber (reg:SI 15))])
   (set (pc) (label_ref (match_operand 2 "" "")))]
  "short_branch (INSN_UID (insn), INSN_UID (operands[2]))
   && (USING_SJLJ_EXCEPTIONS || ! can_throw_internal (ins1))
   && sparc_cpu != PROCESSOR_ULTRASPARC
   && sparc_cpu != PROCESSOR_ULTRASPARC3"
  "call\t%a0, %1\n\tadd\t%%o7, (%l2-.-4), %%o7")

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou 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=10567


ebotcazou at gcc dot gnu dot org changed:

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

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2003-11-26 07:48 -------
David, I have a patch but it doesn't fix all the cases (IIRC sibcalls and I'm
not sure this is fixable).  It also aims to prettify the assembly emitted for
delay slots, i.e. always indent like in

   jmp %i7+8
    restore

while, currently, there is a mix of indentation and non-indentation.

Would you mind if I re-target this bug for 3.5?


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  GCC build triplet|sparc-unknown-linux-gnu     |sparc*-*-*
   GCC host triplet|sparc-unknown-linux-gnu     |sparc*-*-*
 GCC target triplet|sparc-unknown-linux-gnu     |sparc*-*-*


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[davem at gcc dot gnu dot org]

------- Additional Comments From davem at gcc dot gnu dot org  2003-11-26 08:37 -------
The sibcall thing is easy to fix, just turn sibcalls off in the
sparc option handling fixup code if -fno-delayed-branch has been
specified.

No problem retargeting this to 3.5

I also agree with your delay slot indentation consolidation.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2003-11-26 08:47 -------
Agreed.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.4                         |3.5


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
   Last reconfirmed|2003-11-21 18:38:58         |2004-02-26 01:45:02
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.5.0                       |---


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2004-07-03 08:57 -------
David, just to let you know that I've not forgotten this PR.

I have a patch that almost works, except that there is a last problem:
sparc_output_mi_thunk unconditionally uses a sibcall, which is forbidden with
-fno-delayed-branches.  Trying to find a workaround to this...


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[davem at gcc dot gnu dot org]

------- Additional Comments From davem at gcc dot gnu dot org  2004-07-06 20:55 -------
It may not be pretty, but I believe it solves the bug correctly.
I think it is worth applying.  If we come up with a more clever
way to do this stuff, then we'll do so.  Meanwhile the bug
will be fixed, there will be no regressions, and the ugliness
will be confined to the sparc backend.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-07-07 21:33 -------
Subject: Bug 10567

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	ebotcazou@gcc.gnu.org	2004-07-07 21:33:09

Modified files:
	gcc            : ChangeLog 
	gcc/config/sparc: sparc.c sparc.md 

Log message:
	PR target/10567
	* config/sparc/sparc.c (load_pic_register): Honor flag_delayed_branch.
	(output_return): Likewise.
	(output_sibcall): Abort if !flag_delayed_branch.
	(sparc_function_ok_for_sibcall): Return 0 if !flag_delayed_branch.
	(emit_and_preserve): New function.
	(sparc_output_mi_thunk): Use it.  Honor flag_delayed_branch.  Emit an
	indirect jump to the thunked-to function if !flag_delayed_branch.
	* config/sparc/sparc.md (delayed_branch): New attribute.
	(load_pcrel_sym): Honor flag_delayed_branch.  Use above
	attribute to compute the length of the insn.
	(goto_handler_and_restore): Likewise.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.4349&r2=2.4350
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/sparc/sparc.c.diff?cvsroot=gcc&r1=1.311&r2=1.312
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/sparc/sparc.md.diff?cvsroot=gcc&r1=1.206&r2=1.207



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-07-07 22:21 -------
So this is fixed and be closed, right?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2004-07-07 22:37 -------
I want to add a link to the gcc-patches message as usual.  I sent it more than
one  hour ago but it seems to be blocked somewhere.  Don't worry, I'll close the
bug after the message has arrived.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2004-07-08 06:25 -------
(From update of attachment 6695)
The patch uploaded in the audit trail is not quite correct, the C++ thunk
support is broken.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #6695 is|0                           |1
           obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2004-07-08 06:26 -------
See http://gcc.gnu.org/ml/gcc-patches/2004-07/msg00699.html

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567

[Bug target/10567] -fno-delayed-branch not honored in back-end

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-07-08 15:25 -------
Subject: Bug 10567

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	ebotcazou@gcc.gnu.org	2004-07-08 15:25:05

Modified files:
	gcc            : ChangeLog 
	gcc/config/sparc: sparc.md 

Log message:
	PR target/10567
	* config/sparc/sparc.md (update_return): Honor flag_delayed_branch.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.4379&r2=2.4380
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/sparc/sparc.md.diff?cvsroot=gcc&r1=1.207&r2=1.208



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10567