[Bug rtl-optimization/64539] New: [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[Bug rtl-optimization/64539] New: [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

            Bug ID: 64539
           Summary: [5 regression] cprop_hardreg does not respect clobbers
                    in C_I_F_U
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vries at gcc dot gnu.org

As discussed in PR64154 comment 2 and 3, the clobbers in
CALL_INSN_FUNCTION_USAGE are in general not registered by
copyprop_hardreg_forward_1.

F.i. in this thumb1 call insn there's a clobber of the ip reg, that is ignored
by cprop_hardreg:
...
(call_insn 141 292 142 13 (parallel [
            (call (mem:SI (symbol_ref:SI ("f2") [flags 0x3]  <function_decl
0x7f8182689100 f2>) [0 f2 S4 A32])
                (const_int 0 [0]))
            (use (const_int 0 [0]))
            (clobber (reg:SI 14 lr))
        ])
/myssd/terguo01/toolchain-build/GCC32RM-424/src/gcc/gcc/testsuite/gcc.dg/vshift-3.c:119
770 {*call_insn}
     (expr_list:REG_CALL_DECL (symbol_ref:SI ("f2") [flags 0x3]  <function_decl
0x7f8182689100 f2>)
        (expr_list:REG_EH_REGION (const_int 0 [0])
            (nil)))
    (expr_list (clobber (reg:SI 12 ip))
        (nil)))
...

I can't rule out that this is a generic bug, so marking tentatively as 5.0
regression.

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

--- Comment #1 from vries at gcc dot gnu.org ---
In order to fix this, I'll probably first have to understand the PR57003 issue
and fix.

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

--- Comment #2 from vries at gcc dot gnu.org ---
At first glance, this seems appropriate:
...
diff --git a/gcc/regcprop.c b/gcc/regcprop.c
index 8c4f564..b42a4b7 100644
--- a/gcc/regcprop.c
+++ b/gcc/regcprop.c
@@ -801,6 +801,18 @@ copyprop_hardreg_forward_1 (basic_block bb, struct
value_data *vd)
         I wouldn't think this were true for regular insns, but
         scan_rtx treats them like that...  */
       note_stores (PATTERN (insn), kill_clobbered_value, vd);
+      if (CALL_P (insn))
+       {
+         rtx exp;
+         for (exp = CALL_INSN_FUNCTION_USAGE (insn);
+              exp;
+              exp = XEXP (exp, 1))
+           {
+             rtx x = XEXP (exp, 0);
+             if (GET_CODE (x) == CLOBBER)
+               kill_value (SET_DEST (x), vd);
+           }
+       }

       /* Kill all auto-incremented values.  */
       /* ??? REG_INC is useless, since stack pushes aren't done that way.  */
...

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

--- Comment #3 from vries at gcc dot gnu.org ---
Created attachment 34411
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34411&action=edit
Tentative patch

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

vries at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |terry.guo at arm dot com

--- Comment #4 from vries at gcc dot gnu.org ---
Hi Terry,

I've bootstrapped and reg-tested the patch from comment 3 on x86_64, no issues
found.

Can you try the patch, and see if it fixes the problem that you observed in
PR64154 comment 2 ?

Thanks,
- Tom

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[terry.guo at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

--- Comment #5 from Terry Guo <terry.guo at arm dot com> ---
Thanks and it did fixed the issue I observed.

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

vries at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch, wrong-code

--- Comment #6 from vries at gcc dot gnu.org ---
https://gcc.gnu.org/ml/gcc-patches/2015-01/msg00477.html

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |5.0

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

--- Comment #7 from vries at gcc dot gnu.org ---
Author: vries
Date: Fri Jan  9 18:54:20 2015
New Revision: 219400

URL: https://gcc.gnu.org/viewcvs?rev=219400&root=gcc&view=rev
Log:
Handle CALL_INSN_FUNCTION_USAGE clobbers in regcprop.c

2015-01-09  Tom de Vries  <tom@codesourcery.com>

    PR rtl-optimization/64539
    * regcprop.c (kill_clobbered_values): Factor out of ...
    (copyprop_hardreg_forward_1): ... here.  Use kill_clobbered_values
    instead of note_stores with kill_clobbered_value.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/regcprop.c

[Bug rtl-optimization/64539] [5 regression] cprop_hardreg does not respect clobbers in C_I_F_U

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64539

vries at gcc dot gnu.org changed:

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

--- Comment #8 from vries at gcc dot gnu.org ---
Patch committed, marking resolved-fixed.