[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[sthalik at misaki dot pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

Stanisław Halik <sthalik at misaki dot pl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sthalik at misaki dot pl

--- Comment #1 from Stanisław Halik <sthalik at misaki dot pl> ---
Created attachment 31033
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31033&action=edit
build command

hand-made mkspec could be wrong, i.e. PEBKAC
>From gcc-bugs-return-432118-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Oct 18 11:27:08 2013
Return-Path: <gcc-bugs-return-432118-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 23047 invoked by alias); 18 Oct 2013 11:27:08 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 23019 invoked by uid 48); 18 Oct 2013 11:27:05 -0000
From: "sthalik at misaki dot pl" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug objc++/58783] New: fails to cross-compile Qt 5:dev in obj-c++
Date: Fri, 18 Oct 2013 11:27:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: new
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: objc++
X-Bugzilla-Version: 4.8.2
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: sthalik at misaki dot pl
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: bug_id short_desc product version bug_status bug_severity priority component assigned_to reporter cc
Message-ID: <bug-58783-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-10/txt/msg01262.txt.bz2
Content-length: 576

http://gcc.gnu.org/bugzilla/show_bug.cgi?idX783

            Bug ID: 58783
           Summary: fails to cross-compile Qt 5:dev in obj-c++
           Product: gcc
           Version: 4.8.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: objc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: sthalik at misaki dot pl
                CC: iains at gcc dot gnu.org

This is host=x86_64-pc-linux-gnu, target=i686-apple-darwin10

Compiling Qt fails, attached are preprocessed source and faling command.

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[sthalik at misaki dot pl]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="UTF-8", Size: 3487 bytes --]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

--- Comment #2 from Stanisław Halik <sthalik at misaki dot pl> ---
Created attachment 31034
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31034&action=edit
FWIW, qt5 mkspec
>From gcc-bugs-return-432121-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Oct 18 11:34:04 2013
Return-Path: <gcc-bugs-return-432121-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 28789 invoked by alias); 18 Oct 2013 11:34:04 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 28773 invoked by uid 48); 18 Oct 2013 11:34:01 -0000
From: "yvan.roux at linaro dot org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/58784] New: [ARM] LRA legitimate address issue with misalign neon_store
Date: Fri, 18 Oct 2013 11:34:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: new
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 4.9.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: yvan.roux at linaro dot org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: bug_id short_desc product version bug_status bug_severity priority component assigned_to reporter attachments.created
Message-ID: <bug-58784-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-10/txt/msg01265.txt.bz2
Content-length: 1430

http://gcc.gnu.org/bugzilla/show_bug.cgi?idX784

            Bug ID: 58784
           Summary: [ARM] LRA legitimate address issue with misalign
                    neon_store
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: yvan.roux at linaro dot org

Created attachment 31035
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id1035&actioníit
testcase

The attached testcase failed on ARM with LRA enabled during the legitimate
address checking of the movmisaligndi_neon_store insn. Notice that this issue
is the root cause of all the LRA testsuite regressions (in gcc/g++, libstdc++
and fortran).

command line to reproduce:

cc1 -quiet unalign-1.c -march=armv7-a -mfloat-abi=hard -mfpu=neon -marm -O2
-mlra

error on insn:

(insn 12 7 14 2 (set (mem:DI (plus:SI (plus:SI (mult:SI (reg:SI 0 r0 [
sequenceId ])
                        (const_int 8 [0x8]))
                    (reg/f:SI 102 sfp))
                (const_int -3 [0xfffffffffffffffd])) [2 *_4+0 S8 A8])
        (unspec:DI [
                (reg:DI 117 [ D.4956 ])
            ] UNSPEC_MISALIGNED_ACCESS)) unalign-1.c:13 819
{*movmisaligndi_neon_store}
     (expr_list:REG_DEAD (reg:DI 117 [ D.4956 ])
        (expr_list:REG_DEAD (reg:SI 0 r0 [ sequenceId ])
            (nil))))

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[sthalik at misaki dot pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

--- Comment #3 from Stanisław Halik <sthalik at misaki dot pl> ---
Created attachment 31036
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31036&action=edit
preprocessed source
>From gcc-bugs-return-432122-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Oct 18 11:47:15 2013
Return-Path: <gcc-bugs-return-432122-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 4966 invoked by alias); 18 Oct 2013 11:47:15 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 4900 invoked by uid 55); 18 Oct 2013 11:47:12 -0000
From: "rguenther at suse dot de" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/56309] conditional moves instead of compare and branch result in almost 2x slower code
Date: Fri, 18 Oct 2013 11:47:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 4.7.2
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rguenther at suse dot de
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-56309-4-lDqOsXxnEP@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-56309-4@http.gcc.gnu.org/bugzilla/>
References: <bug-56309-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-10/txt/msg01266.txt.bz2
Content-length: 773

http://gcc.gnu.org/bugzilla/show_bug.cgi?idV309

--- Comment #31 from rguenther at suse dot de <rguenther at suse dot de> ---
On Fri, 18 Oct 2013, ubizjak at gmail dot com wrote:

> http://gcc.gnu.org/bugzilla/show_bug.cgi?idV309
>
> --- Comment #30 from Uro? Bizjak <ubizjak at gmail dot com> ---
> (In reply to Jakub Jelinek from comment #29)
>
> > On the by-ref-O3.ii testcase there is no change, but by-val-O3.ii seems to
> > be consistently faster, around 20% on AMD 8354 and 35% on Intel 2600.
>
> According to detailed analysis in Comment 26, perhaps a new tuning flag or
> target hook could be used for if-conversion of scalars. Also, I think that -Os
> can still use cmoves due to smaller code size.

But RTL if-conversion still will generate cmov, right?

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[sthalik at misaki dot pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

Stanisław Halik <sthalik at misaki dot pl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #31036|0                           |1
        is obsolete|                            |

--- Comment #4 from Stanisław Halik <sthalik at misaki dot pl> ---
Created attachment 31037
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31037&action=edit
.bz2 for Iain
>From gcc-bugs-return-432126-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Oct 18 11:57:55 2013
Return-Path: <gcc-bugs-return-432126-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 14160 invoked by alias); 18 Oct 2013 11:57:55 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 14096 invoked by uid 48); 18 Oct 2013 11:57:52 -0000
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/56309] conditional moves instead of compare and branch result in almost 2x slower code
Date: Fri, 18 Oct 2013 11:57:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 4.7.2
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jakub at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-56309-4-jfzCqExpEw@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-56309-4@http.gcc.gnu.org/bugzilla/>
References: <bug-56309-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-10/txt/msg01270.txt.bz2
Content-length: 1569

http://gcc.gnu.org/bugzilla/show_bug.cgi?idV309

--- Comment #32 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to rguenther@suse.de from comment #31)
> But RTL if-conversion still will generate cmov, right?

Sure.

The current if-conversion is performed without any cost model or anything
similar, say even for:
void
foo (int *a, int *b, int *c)
{
  int i;
  for (i = 0; i < 1024; i += 32)
    {
      int t1, t2;
#define V(N) \
      t1 = a[i + N]; t2 = b[i + N]; \
      a[i + N] = c[i + N / 8] ? t1 * (N + 1) / 3 + 21: t2 * (N + 3) / 17 + 9;
      V(0) V(1) V(2) V(3) V(4) V(5) V(6) V(7) V(8) V(9)
      V(10) V(11) V(12) V(13) V(14) V(15) V(16) V(17) V(18) V(19)
      V(20) V(21) V(22) V(23) V(24) V(25) V(26) V(27) V(28) V(29)
      V(30) V(31)
    }
}
it will just if-convert everything together, performing all the operations
unconditionally and doing lots of conditional moves.
If my patches are way to go, supposedly the vectorizer cost model should take
into account the non-if-converted loop cost instead of if-converted loop cost
and compare that to the cost of the vectorized if-converted loop, though with
multiple basic blocks in loop's body there won't be one cost, but perhaps one
can compute minimum and maximum and average cost or something similar (also
taking into account branch probabilities).
And, for targets where a tree if-conversion would be useful, after adding some
cost model and limiting how many statements we should be running
unconditionally, parts of the if-conversion framework could be used for a
different pass.

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[sthalik at misaki dot pl]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="UTF-8", Size: 3593 bytes --]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

--- Comment #5 from Stanisław Halik <sthalik at misaki dot pl> ---
Created attachment 31038
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31038&action=edit
platform info in .s
>From gcc-bugs-return-432128-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Oct 18 12:03:58 2013
Return-Path: <gcc-bugs-return-432128-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 19895 invoked by alias); 18 Oct 2013 12:03:58 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 19873 invoked by uid 48); 18 Oct 2013 12:03:55 -0000
From: "yvan.roux at linaro dot org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/58785] New: [ARM] LRA issue in Thumb mode with movhi
Date: Fri, 18 Oct 2013 12:03:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: new
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 4.9.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: yvan.roux at linaro dot org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: bug_id short_desc product version bug_status bug_severity priority component assigned_to reporter attachments.created
Message-ID: <bug-58785-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-10/txt/msg01272.txt.bz2
Content-length: 1549

http://gcc.gnu.org/bugzilla/show_bug.cgi?idX785

            Bug ID: 58785
           Summary: [ARM] LRA issue in Thumb mode with movhi
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: yvan.roux at linaro dot org

Created attachment 31039
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id1039&actioníit
testcase

The attached testcase failed on ARM with LRA enabled in Thumb mode (it is a
reduce testcase from the ICE which brakes bootstrap with Thumb).

Here LRA has to deal with thumb1_movhi_insn :

(insn 11 5 14 2 (set (reg:HI 0 r0)
        (const_int -1318 [0xfffffffffffffada])) ... {*thumb1_movhi_insn}
     (nil))

and creates new regs to do it :

11: r0:HI=r114:HI
    Inserting insn reload before:
18: r115:SI=0xfffffffffffffada
19: r114:HI=r115:SI#0
    REG_EQUAL 0xfffffffffffffada

 Choosing alt 6 in insn 18:  (0) l  (1) mi {*thumb1_movsi_insn}
 Creating newreg\x116 from oldreg\x115, assigning class LO_REGS to r116

18: r116:SI=0xfffffffffffffada
    Inserting insn reload after:
20: r115:SI=r116:SI

 Creating newreg\x117, assigning class LO_REGS to scratch r117

and during this move processing,
gen_thumb_movhi_clobber(rtx_def*,rtx_def*,rtx_def*) is called with a first rtx
parameter that doesn't satisfy the constraint of being a strict memory address
(it is a register).

command line to reproduce:

cc1 -quiet thumb2-lra.i -O2 -mthumb -mlra

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[iains at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

Iain Sandoe <iains at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-10-18
     Ever confirmed|0                           |1

--- Comment #6 from Iain Sandoe <iains at gcc dot gnu.org> ---
confirmed on a x86-64-darwin12 X i686-darwin10, but the test case could use
some reduction.

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[iains at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

Iain Sandoe <iains at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement

--- Comment #7 from Iain Sandoe <iains at gcc dot gnu.org> ---
at present, fast enumeration is only supported by GCC for Objective-C (both GNU
and NeXT/Apple runtimes).

we will need to investigate what is required to extend this to Objective-C++.

[Bug objc++/58783] fails to cross-compile Qt 5:dev in obj-c++

[egall at gwmail dot gwu.edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

Eric Gallager <egall at gwmail dot gwu.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |egall at gwmail dot gwu.edu

--- Comment #8 from Eric Gallager <egall at gwmail dot gwu.edu> ---
I recently ran into this with the Cocoa interface for DeSmuME, which also tries
to use fast enumeration from Objective-C++. So maybe this should be re-titled
to reflect the fact that it's a more general issue with fast enumeration in
Objective-C++, and not anything Qt5-specific.

[Bug objc++/58783] Fast enumeration is not supported for Objective-C++

[egallager at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58783

--- Comment #10 from Eric Gallager <egallager at gcc dot gnu.org> ---
Note that on the Standards page of the manual, it currently says "fast
enumeration (only for Objective-C)", so once this bug is fixed, that
parenthetical can be removed:
https://gcc.gnu.org/onlinedocs/gcc/Standards.html#Standards