[Bug target/44707] operand requires impossible reload

[Bug target/44707] operand requires impossible reload

[bergner at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

--- Comment #7 from Peter Bergner <bergner at gcc dot gnu.org> 2011-03-09 19:20:47 UTC ---
Author: bergner
Date: Wed Mar  9 19:20:42 2011
New Revision: 170829

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=170829
Log:
gcc/
    Backport from mainline
    2010-07-02  Ulrich Weigand  <Ulrich.Weigand@de.ibm.com>

    PR target/44707
    * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Recognize
    (lo_sum (high ...) ...) patterns generated by earlier passes.

gcc/testsuite/
    Backport from mainline
    2010-07-02  Ulrich Weigand  <Ulrich.Weigand@de.ibm.com>

    PR target/44707
    * gcc.c-torture/compile/pr44707.c: New test.

Added:
    branches/ibm/gcc-4_5-branch/gcc/testsuite/gcc.c-torture/compile/pr44707.c
Modified:
    branches/ibm/gcc-4_5-branch/gcc/ChangeLog.ibm
    branches/ibm/gcc-4_5-branch/gcc/config/rs6000/rs6000.c
    branches/ibm/gcc-4_5-branch/gcc/testsuite/ChangeLog.ibm

[Bug target/44707] operand requires impossible reload

[fang at csl dot cornell.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

David Fang <fang at csl dot cornell.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fang at csl dot cornell.edu

--- Comment #8 from David Fang <fang at csl dot cornell.edu> 2011-07-10 17:34:25 UTC ---
Hi, I see this test failing on powerpc-darwin8 (I know, not a critical
platform).
http://gcc.gnu.org/ml/gcc-testresults/2011-07/msg01092.html
What information can I provide from my test runs?

[Bug target/44707] operand requires impossible reload

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

Jack Howarth <howarth at nitro dot med.uc.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |howarth at nitro dot
                   |                            |med.uc.edu

--- Comment #9 from Jack Howarth <howarth at nitro dot med.uc.edu> 2011-07-11 03:03:55 UTC ---
On powerpc-apple-darwin9, this fails under gcc-4.6.1 as...

Executing on host: /sw/src/fink.build/gcc46-4.6.1-1000/darwin_objdir/gcc/xgcc
-B/sw/src/fink.build/gcc46-4.6.1-1000/darwin_objdir/gcc/   -O1  -w -c  -m32 -o
pr44707.o
/sw/src/fink.build/gcc46-4.6.1-1000/gcc-4.6.1/gcc/testsuite/gcc.c-torture/compile/pr44707.c
   (timeout = 300)
/var/tmp//ccZ2tK1u.s:15:non-relocatable subtraction expression, "_w" minus
"L00000000001$pb"^M
/var/tmp//ccZ2tK1u.s:15:symbol: "_w" can't be undefined in a subtraction
expression^M
/var/tmp//ccZ2tK1u.s:14:non-relocatable subtraction expression, "_w" minus
"L00000000001$pb"^M
/var/tmp//ccZ2tK1u.s:14:symbol: "_w" can't be undefined in a subtraction
expression^M
/var/tmp//ccZ2tK1u.s:13:non-relocatable subtraction expression, "_v" minus
"L00000000001$pb"^M
/var/tmp//ccZ2tK1u.s:13:symbol: "_v" can't be undefined in a subtraction
expression^M
/var/tmp//ccZ2tK1u.s:12:non-relocatable subtraction expression, "_v" minus
"L00000000001$pb"^M
/var/tmp//ccZ2tK1u.s:12:symbol: "_v" can't be undefined in a subtraction
expression^M
compiler exited with status 1

which may be an assembler issue remembering that darwin is using the rather
old...

Apple Inc version cctools-698.1~1, GNU assembler version 1.38

[Bug target/44707] operand requires impossible reload

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

--- Comment #10 from Jack Howarth <howarth at nitro dot med.uc.edu> 2011-07-11 03:07:34 UTC ---
Created attachment 24734
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=24734
preprocessed source for pr44707.c from powerpc-apple-darwin9

[Bug target/44707] operand requires impossible reload

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

--- Comment #11 from Jack Howarth <howarth at nitro dot med.uc.edu> 2011-07-11 03:08:07 UTC ---
Created attachment 24735
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=24735
assembly for pr44707.c from powerpc-apple-darwin9

[Bug target/44707] operand requires impossible reload

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

--- Comment #12 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-07-11 09:44:00 UTC ---
> Hi, I see this test failing on powerpc-darwin8 (I know, not a critical platform).
> http://gcc.gnu.org/ml/gcc-testresults/2011-07/msg01092.html
> What information can I provide from my test runs?

This is pr45233.

[Bug target/44707] operand requires impossible reload

[ebotcazou at gcc dot gnu dot org]

------- Comment #6 from ebotcazou at gcc dot gnu dot org  2010-07-25 21:46 -------
Subject: Bug 44707

Author: ebotcazou
Date: Sun Jul 25 21:46:32 2010
New Revision: 162521

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=162521
Log:
        PR target/44707
        * config/sparc/sparc-protos.h (sparc_legitimize_reload_address): New.
        * config/sparc/sparc.c: Include reload.h.
        (legitimize_tls_address): Rename into...
        (sparc_legitimize_tls_address): ...this.
        (legitimize_pic_address): Rename into...
        (sparc_legitimize_pic_address): ...this.
        (sparc_expand_move): Adjust to above renaming.
        (sparc_tls_referenced_p): Likewise.
        (sparc_legitimize_tls_address): Likewise.
        (sparc_legitimize_pic_address): Likewise.
        (sparc_legitimize_address): Likewise.
        (sparc_output_mi_thunk): Likewise.
        (sparc_legitimize_reload_address): New global function.  Recognize
        (lo_sum (high ...) ...) patterns generated by earlier passes.
        * config/sparc/sparc.h (LEGITIMIZE_RELOAD_ADDRESS): Use above function.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/sparc/sparc-protos.h
    trunk/gcc/config/sparc/sparc.c
    trunk/gcc/config/sparc/sparc.h


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

[Bug target/44707] operand requires impossible reload

[uweigand at gcc dot gnu dot org]

------- Comment #5 from uweigand at gcc dot gnu dot org  2010-07-02 11:50 -------
Fixed.


-- 

uweigand at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

[Bug target/44707] operand requires impossible reload

[uweigand at gcc dot gnu dot org]

------- Comment #4 from uweigand at gcc dot gnu dot org  2010-07-02 11:48 -------
Subject: Bug 44707

Author: uweigand
Date: Fri Jul  2 11:48:30 2010
New Revision: 161703

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=161703
Log:
ChangeLog:

        PR target/44707
        * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Recognize
        (lo_sum (high ...) ...) patterns generated by earlier passes.

testsuite/ChangeLog:

        PR target/44707
        * gcc.c-torture/compile/pr44707.c: New test.

Added:
    trunk/gcc/testsuite/gcc.c-torture/compile/pr44707.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/rs6000/rs6000.c
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

[Bug target/44707] operand requires impossible reload

[uweigand at gcc dot gnu dot org]

------- Comment #3 from uweigand at gcc dot gnu dot org  2010-07-01 19:14 -------
Patch posted here:
http://gcc.gnu.org/ml/gcc-patches/2010-07/msg00082.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

[Bug target/44707] operand requires impossible reload

[uweigand at gcc dot gnu dot org]

-- 

uweigand at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |uweigand at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2010-06-29 16:56:47         |2010-07-01 19:07:33
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

[Bug target/44707] operand requires impossible reload

[uweigand at gcc dot gnu dot org]

------- Comment #2 from uweigand at gcc dot gnu dot org  2010-06-29 17:03 -------
Created an attachment (id=21041)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21041&action=view)
Recognize (lo_sum (high ...) ...) in rs6000_legitimize_reload_address

> It seems to me that simply extending rs6000_legitimate_reload_address to handle
> this case as well should fix the bug.

And indeed, this (otherwise untested) patch fixes the bug for me.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707

[Bug target/44707] operand requires impossible reload

[uweigand at gcc dot gnu dot org]

------- Comment #1 from uweigand at gcc dot gnu dot org  2010-06-29 16:56 -------
I agree, this looks like a longstanding bug in
rs6000_legitimize_reload_address.

What happens here is that find_reloads is called on this insn:

(insn 15 8 18 2 pr44707.c:13 (asm_operands/v ("/* %0 %1 %2 %3 %4 */") ("") 0 [
            (mem/s/c:SI (symbol_ref:SI ("v") [flags 0xc0] <var_decl 0xf6f90fc0
v>) [3 v.a+0 S4 A32])
            (mem/c/i:SI (symbol_ref:SI ("w") [flags 0xc4] <var_decl 0xf6f91020
w>) [3 w+0 S4 A32])
            (mem/s/c:SI (const:SI (plus:SI (symbol_ref:SI ("v") [flags 0xc0]
<var_decl 0xf6f90fc0 v>)
                        (const_int 4 [0x4]))) [3 v.b+0 S4 A32])
            (mem/s/c:SI (const:SI (plus:SI (symbol_ref:SI ("v") [flags 0xc0]
<var_decl 0xf6f90fc0 v>)
                        (const_int 8 [0x8]))) [3 v.c+0 S4 A32])
            (mem/s/c:SI (const:SI (plus:SI (symbol_ref:SI ("v") [flags 0xc0]
<var_decl 0xf6f90fc0 v>)
                        (const_int 12 [0xc]))) [3 v.d+0 S4 A32])
        ]
         [
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
        ]
         [] pr44707.c:14) -1 (nil))

rs6000_find_reloads_address notices that it can rewrite
  (symbol_ref:SI "v")
to
  (lo_sum:SI (high:SI (symbol_ref:SI "v")) (symbol_ref:SI "v"))
(and place a reload on the (high:SI) subexpression) and does so.  This change
remains in the insn, and when in the next iteration find_reloads is called
again, the insn now looks like:

(insn 15 8 18 2 pr44707.c:13 (asm_operands/v ("/* %0 %1 %2 %3 %4 */") ("") 0 [
            (mem/s/c:SI (lo_sum:SI (high:SI (symbol_ref:SI ("v") [flags 0xc0]
<var_decl 0xf6f90fc0 v>))
                    (symbol_ref:SI ("v") [flags 0xc0] <var_decl 0xf6f90fc0 v>))
[3 v.a+0 S4 A32])
            (mem/c/i:SI (lo_sum:SI (high:SI (symbol_ref:SI ("w") [flags 0xc4]
<var_decl 0xf6f91020 w>))
                    (symbol_ref:SI ("w") [flags 0xc4] <var_decl 0xf6f91020 w>))
[3 w+0 S4 A32])
            (mem/s/c:SI (const:SI (plus:SI (symbol_ref:SI ("v") [flags 0xc0]
<var_decl 0xf6f90fc0 v>)
                        (const_int 4 [0x4]))) [3 v.b+0 S4 A32])
            (mem/s/c:SI (const:SI (plus:SI (symbol_ref:SI ("v") [flags 0xc0]
<var_decl 0xf6f90fc0 v>)
                        (const_int 8 [0x8]))) [3 v.c+0 S4 A32])
            (mem/s/c:SI (const:SI (plus:SI (symbol_ref:SI ("v") [flags 0xc0]
<var_decl 0xf6f90fc0 v>)
                        (const_int 12 [0xc]))) [3 v.d+0 S4 A32])
        ]
         [
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
            (asm_input:SI ("nro") (null):0)
        ]
         [] pr44707.c:14) -1 (nil))

However, this expression is now no longer recognized by
rs6000_legitimize_reload_address, and therefore no reload on (high:SI) is
pushed.

Thus, when the reload is finally processed, a reload insn like this is
generated:

(insn 26 8 27 2 pr44707.c:13 (set (reg:SI 10 10)
        (lo_sum:SI (high:SI (symbol_ref:SI ("v") [flags 0xc0] <var_decl
0xf6f90fc0 v>))
            (symbol_ref:SI ("v") [flags 0xc0] <var_decl 0xf6f90fc0 v>))) -1
(nil))

As this does not actually correspond to any valid pattern, an assertion is
triggered.

The underlying problem is that with the current reload setup, an implementation
of LEGITIMATE_RELOAD_ADDRESS must always recognize expressions it has itself
generated in an earlier call.  And indeed, that's what a comment in
rs6000_legitimize_reload_address says:

static rtx
rs6000_legitimize_reload_address (rtx x, enum machine_mode mode,
                                  int opnum, int type,
                                  int ind_levels ATTRIBUTE_UNUSED, int *win)
{
  bool reg_offset_p = reg_offset_addressing_ok_p (mode);

  /* We must recognize output that we have already generated ourselves.  */
  if (GET_CODE (x) == PLUS
      && GET_CODE (XEXP (x, 0)) == PLUS
      && GET_CODE (XEXP (XEXP (x, 0), 0)) == REG
      && GET_CODE (XEXP (XEXP (x, 0), 1)) == CONST_INT
      && GET_CODE (XEXP (x, 1)) == CONST_INT)


However, this recognizes only certain types of such output, and in particular
not the one that shows up in this test case.

It seems to me that simply extending rs6000_legitimate_reload_address to handle
this case as well should fix the bug.


-- 

uweigand at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2010-06-29 16:56:47
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44707