[Bug other/59927] New: [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[Bug other/59927] New: [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[doko at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

            Bug ID: 59927
           Summary: [4.9 Regression] ICE sorry, unimplemented: ms_abi
                    attribute requires -maccumulate-outgoing-args or
                    subtarget optimization implying it
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: other
          Assignee: unassigned at gcc dot gnu.org
          Reporter: doko at gcc dot gnu.org

Created attachment 31939
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31939&action=edit
preprocessed source

seen with trunk 20140122, works with the 4.8 branch, lowering to -O0 works
around the issue.

$ gcc -c -g -O2 fdi.i 
fdi.c: In function 'fdi_decomp':
fdi.c:2005:33: sorry, unimplemented: ms_abi attribute requires
-maccumulate-outgoing-args or subtarget optimization implying it
fdi.c:2005: confused by earlier errors, bailing out

[Bug other/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-01-23
                 CC|                            |mpolacek at gcc dot gnu.org
   Target Milestone|---                         |4.9.0
     Ever confirmed|0                           |1

--- Comment #1 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
With checking -g -O2 isn't needed:

fdi.c: In function ‘fdi_Ziphuft_build’:
fdi.c:1046:12: sorry, unimplemented: ms_abi attribute requires
-maccumulate-outgoing-args or subtarget optimization implying it
fdi.c:1046:12: internal compiler error: Segmentation fault
0xb7a3df crash_signal
    /home/marek/src/gcc/gcc/toplev.c:337
0x14cdc97 add_file_to_cache_tab
    /home/marek/src/gcc/gcc/input.c:296
0x14cdc97 lookup_or_add_file_to_cache_tab
    /home/marek/src/gcc/gcc/input.c:330
0x14cdc97 location_get_source_line(expanded_location, int*)
    /home/marek/src/gcc/gcc/input.c:704
0x14c89d4 diagnostic_show_locus(diagnostic_context*, diagnostic_info const*)
    /home/marek/src/gcc/gcc/diagnostic.c:304
0x14c921c diagnostic_report_diagnostic(diagnostic_context*, diagnostic_info*)
    /home/marek/src/gcc/gcc/diagnostic.c:800
0x14c9efb sorry(char const*, ...)
    /home/marek/src/gcc/gcc/diagnostic.c:1095
0xe3a394 init_cumulative_args(ix86_args*, tree_node*, rtx_def*, tree_node*,
int)
    /home/marek/src/gcc/gcc/config/i386/i386.c:6098
0x76a509 expand_call(tree_node*, rtx_def*, int)
    /home/marek/src/gcc/gcc/calls.c:2504
0x882562 expand_expr_real_1(tree_node*, rtx_def*, machine_mode,
expand_modifier, rtx_def**, bool)
    /home/marek/src/gcc/gcc/expr.c:10330
0x8912b0 store_expr(tree_node*, rtx_def*, int, bool)
    /home/marek/src/gcc/gcc/expr.c:5328
0x893731 expand_assignment(tree_node*, tree_node*, bool)
    /home/marek/src/gcc/gcc/expr.c:5114
0x78062b expand_call_stmt
    /home/marek/src/gcc/gcc/cfgexpand.c:2257
0x78062b expand_gimple_stmt_1
    /home/marek/src/gcc/gcc/cfgexpand.c:3157
0x78062b expand_gimple_stmt
    /home/marek/src/gcc/gcc/cfgexpand.c:3309
0x781b02 expand_gimple_basic_block
    /home/marek/src/gcc/gcc/cfgexpand.c:5149
0x785ee3 gimple_expand_cfg
    /home/marek/src/gcc/gcc/cfgexpand.c:5715
0x785ee3 execute
    /home/marek/src/gcc/gcc/cfgexpand.c:5935
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.
>From gcc-bugs-return-441374-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Thu Jan 23 20:52:39 2014
Return-Path: <gcc-bugs-return-441374-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 9275 invoked by alias); 23 Jan 2014 20:52:38 -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 9247 invoked by uid 48); 23 Jan 2014 20:52:35 -0000
From: "igtrnt at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/59928] New: Possible incorrect warning "comparison is always false due to limited range of data type"
Date: Thu, 23 Jan 2014 20:52:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: new
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.8.2
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: igtrnt at gmail dot com
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
Message-ID: <bug-59928-4@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: 2014-01/txt/msg02516.txt.bz2
Content-length: 853

http://gcc.gnu.org/bugzilla/show_bug.cgi?idY928

            Bug ID: 59928
           Summary: Possible incorrect warning "comparison is always false
                    due to limited range of data type"
           Product: gcc
           Version: 4.8.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: igtrnt at gmail dot com

The code below produces the following warning:

comparison is always false due to limited range of data type [-Wtype-limits]


CODE:

    bool b;
    if ( static_cast<unsigned int>(b) > static_cast<unsigned int>(1) )
    {
        // DO SMTH.
    }

In case boolean variable is uninitialized, comparison may be true because
integer value of `b` may contain junk and not neceserally `0` or `1`.

[Bug other/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org,
                   |                            |ktietz at gcc dot gnu.org

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
   0x14cdc97 add_file_to_cache_tab
           /home/marek/src/gcc/gcc/input.c:296

That's unrelated segfault,
http://gcc.gnu.org/ml/gcc-patches/2014-01/msg01514.html

Supposedly ms_abi attribute, if accumulate_outgoing_args isn't set, could just
implicitly add target ("accumulate-outgoing-args")?  Or what is the reason why
ms_abi doesn't support it?

[Bug other/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[ktietz at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

--- Comment #3 from Kai Tietz <ktietz at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #2)
>    0x14cdc97 add_file_to_cache_tab
>            /home/marek/src/gcc/gcc/input.c:296
> 
> That's unrelated segfault,
> http://gcc.gnu.org/ml/gcc-patches/2014-01/msg01514.html
> 
> Supposedly ms_abi attribute, if accumulate_outgoing_args isn't set, could
> just implicitly add target ("accumulate-outgoing-args")?  Or what is the
> reason why ms_abi doesn't support it?

That might be a way.  ACCUMULATE_OUTGOING_ARGS has to be set for ABI do
resevered stack-area.  On windows 64-bit we have additionally the problem that
drap-register can't be expressed in SEH-block (unwind information), and so to
turn it off is violating the ABI.

[Bug other/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Reduced testcase:

typedef unsigned int (__attribute__ ((ms_abi)) *F) (long);

void
foo (F f)
{
  f (0);
}

or even:

extern void bar (int) __attribute__ ((ms_abi));
void
foo (void)
{
  bar (0);
}


Why do you need ACCUMULATE_OUTGOING_ARGS in the non-ms_abi caller of a ms_abi
function though?  There is no SEH involved there... I see the subsequent ICE is
about the stack pointer delta not being sufficiently aligned, but it is unclear
if it is because of i386.c bug in the ms_abi support, or calls.c bug.
Changing ACCUMULATE_OUTGOING_ARGS on the fly if we notice that the current
function calls or might call a ms_abi function would be both weird and not sure
where the target would have a chance to see the calls to decide that prior to
expansion (and, changing that while already in expansion is too late).

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |i?86-*-*
           Priority|P3                          |P1
          Component|other                       |target

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[rth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

Richard Henderson <rth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rth at gcc dot gnu.org

--- Comment #5 from Richard Henderson <rth at gcc dot gnu.org> ---
There's definitely a calls.c bug.

We allocate 24 bytes of stack at line 2973, to make up the 32-byte preferred
alignment from the 8 bytes that we "pushed".  Except that we shouldn't pushing
any bytes at all.

Then we allocate 32 bytes of stack at line 3111 to allocate all the space for
the arguments passed in registers.

Then we ICE because (24+32) % 32 != 0.

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[rth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

Richard Henderson <rth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rth at gcc dot gnu.org

--- Comment #6 from Richard Henderson <rth at gcc dot gnu.org> ---
Created attachment 32080
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32080&action=edit
testing patch

Here's a patch that seems altogether too straightforward.

It suppresses the incorrect first adjustment mentioned in the previous
comment while retaining the reg_parm_stack_space adjust.

I also tested the obvious 5 argument call and see

    subq    $8, %rsp
    pushq    $4
    subq    $32, %rsp
    movl    $3, %r9d
    movl    $2, %r8d
    movl    $1, %edx
    movl    $0, %ecx
    call    baz

which seems inefficient, but is also correct.

I'll attempt to test this at some point this weekend, but I'm not
currently in a position to build on x86_64 windows.  Kai, can you
fire something off?

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[ktietz at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

--- Comment #7 from Kai Tietz <ktietz at gcc dot gnu.org> ---
(In reply to Richard Henderson from comment #6)
> I'll attempt to test this at some point this weekend, but I'm not
> currently in a position to build on x86_64 windows.  Kai, can you
> fire something off?


I will do a build over weekend for it.

One side-note I have about this patch.  Due this allow drap-register for x64,
we seem to need to enable eax_live for this target, too.

See in ix86_expand_prologue the lines
...
if (!TARGET_64BIT_MS_ABI)
  eax_live = ix86_eax_live_at_start_p ();
...

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[rth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

--- Comment #8 from Richard Henderson <rth at gcc dot gnu.org> ---
(In reply to Kai Tietz from comment #7)
> One side-note I have about this patch.  Due this allow drap-register for
> x64, we seem to need to enable eax_live for this target, too.
> 
> See in ix86_expand_prologue the lines
> ...
> if (!TARGET_64BIT_MS_ABI)
>   eax_live = ix86_eax_live_at_start_p ();
> ...


In theory it shouldn't matter.  Unlike the unix 64-bit abi, EAX is
never live for any w64 abi.  On the other hand, removing the condition
shouldn't hurt either, making the code less confusing.

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Author: jakub
Date: Tue Feb 11 09:06:37 2014
New Revision: 207686

URL: http://gcc.gnu.org/viewcvs?rev=207686&root=gcc&view=rev
Log:
    PR target/59927
    * gcc.target/i386/pr59927.c: New test.

Added:
    trunk/gcc/testsuite/gcc.target/i386/pr59927.c
Modified:
    trunk/gcc/testsuite/ChangeLog

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Seems to be fixed.

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[uros at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

--- Comment #12 from uros at gcc dot gnu.org ---
Author: uros
Date: Tue Feb 11 18:59:24 2014
New Revision: 207697

URL: http://gcc.gnu.org/viewcvs?rev=207697&root=gcc&view=rev
Log:
    PR target/59927
    Revert
    2013-12-15  Uros Bizjak  <ubizjak@gmail.com>

    PR testsuite/58630
    * gcc.target/i386/pr43662.c (dg-options):
    Add -maccumulate-outgoing-args.
    * gcc.target/i386/pr43869.c (dg-options): Ditto.
    * gcc.target/i386/pr57003.c (dg-options): Ditto.
    * gcc.target/i386/avx-vzeroupper-16.c (dg-options):
    Remove -mtune=generic and add -maccumulate-outgoing-args instead.
    * gcc.target/i386/avx-vzeroupper-17.c (dg-options): Ditto.
    * gcc.target/i386/avx-vzeroupper-18.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/func-1.c (dg-options):
    Add -maccumulate-outgoing-args.
    * gcc.target/x86_64/abi/callabi/func-2a.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/func-2b.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/func-indirect.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/func-indirect-2a.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/func-indirect-2b.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/leaf-1.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/leaf-2.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/pr38891.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/vaarg-1.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/vaarg-2.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/vaarg-3.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/vaarg-4a.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/vaarg-4b.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/vaarg-5a.c (dg-options): Ditto.
    * gcc.target/x86_64/abi/callabi/vaarg-5b.c (dg-options): Ditto.


Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gcc.target/i386/avx-vzeroupper-16.c
    trunk/gcc/testsuite/gcc.target/i386/avx-vzeroupper-17.c
    trunk/gcc/testsuite/gcc.target/i386/avx-vzeroupper-18.c
    trunk/gcc/testsuite/gcc.target/i386/pr43662.c
    trunk/gcc/testsuite/gcc.target/i386/pr43869.c
    trunk/gcc/testsuite/gcc.target/i386/pr57003.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/func-1.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/func-2a.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/func-2b.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/func-indirect-2a.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/func-indirect-2b.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/func-indirect.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/leaf-1.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/leaf-2.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/pr38891.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/vaarg-1.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/vaarg-2.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/vaarg-3.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/vaarg-4a.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/vaarg-4b.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/vaarg-5a.c
    trunk/gcc/testsuite/gcc.target/x86_64/abi/callabi/vaarg-5b.c

[Bug target/59927] [4.9 Regression] ICE sorry, unimplemented: ms_abi attribute requires -maccumulate-outgoing-args or subtarget optimization implying it

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59927

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |zsojka at seznam dot cz

--- Comment #13 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 51518 has been marked as a duplicate of this bug. ***