[Bug rtl-optimization/34999] New: Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[Bug rtl-optimization/34999] New: Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[eres at il dot ibm dot com]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1721 bytes --]

I run built-in-setjmp.c testcase (taken from gcc.c-torture/execute dir);
first with -fprofile-generate -freorder-blocks-and-partition
and then with -fprofile-use -freorder-blocks-and-partition
under x86_64 and ppc64-linux with r131883 mainline and I get the ICE:

vn/build/build/spu/gcc/gcc/testsuite/gcc/built-in-setjmp.c -fprofile-use
-freorder-blocks-and-partition
../gcc/gcc/testsuite/gcc.c-torture/execute/built-in-setjmp.c: In function
âmainâ:
../gcc/gcc/testsuite/gcc.c-torture/execute/built-in-setjmp.c:39: internal
compiler error: in add_labels_and_missing_jumps, at bb-reorder.c:1304
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

The code for fixing up fallthru edges that cross between hot and cold sections
is in add_labels_and_missing_jumps () and fix_up_fall_thru_edges () functions.
The first function deals with fallthru edges with single successor and the
later
deals with fallthru edges with more than one successor.

It seems that the cause to this fail is because both of the functions do not
take into account the fact that a basic-block with a crossing edge can end with
a call insn.  I am working to fix that.


-- 
           Summary: Fallthru crossing edges in
                    partition_hot_cold_basic_blocks are not been fixed when
                    the section ends with call insn
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: eres at il dot ibm dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[eres at il dot ibm dot com]

------- Comment #1 from eres at il dot ibm dot com  2008-01-28 15:22 -------
Created an attachment (id=15037)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15037&action=view)
the testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[eres at il dot ibm dot com]

------- Comment #2 from eres at il dot ibm dot com  2008-01-29 16:07 -------
Created an attachment (id=15049)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15049&action=view)
A patch I am currently testing


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[revitale at gcc dot gnu dot org]

------- Comment #3 from revitale at gcc dot gnu dot org  2008-02-27 13:28 -------
Subject: Bug 34999

Author: revitale
Date: Wed Feb 27 13:27:56 2008
New Revision: 132711

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=132711
Log:
Fix PR rtl-optimization/34999

Added:
    trunk/gcc/testsuite/gcc.dg/tree-prof/pr34999.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/bb-reorder.c
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ubizjak at gmail dot com]

------- Comment #4 from ubizjak at gmail dot com  2008-02-28 09:47 -------
Test case fails on x86_64:

Running target unix
FAIL: gcc.dg/tree-prof/pr34999.c compilation,  -fprofile-use -D_PROFILE_USE
UNRESOLVED: gcc.dg/tree-prof/pr34999.c execution,    -fprofile-use
-D_PROFILE_USE

http://gcc.gnu.org/ml/gcc-testresults/2008-02/msg01856.html


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2008-02-28 09:47:54
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[eres at il dot ibm dot com]

------- Comment #5 from eres at il dot ibm dot com  2008-02-28 13:00 -------
I can not reproduce this fail on my local x86_64-unknown-linux-gnu machine:

PASS: gcc.dg/tree-prof/pr34999.c compilation,  -fprofile-use -D_PROFILE_USE
PASS: gcc.dg/tree-prof/pr34999.c execution,    -fprofile-use -D_PROFILE_USE

I appreciate any info on this ICE so I could try to fix it.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ubizjak at gmail dot com]

------- Comment #6 from ubizjak at gmail dot com  2008-02-28 13:26 -------
(In reply to comment #5)
> I can not reproduce this fail on my local x86_64-unknown-linux-gnu machine:
> 
> PASS: gcc.dg/tree-prof/pr34999.c compilation,  -fprofile-use -D_PROFILE_USE
> PASS: gcc.dg/tree-prof/pr34999.c execution,    -fprofile-use -D_PROFILE_USE
> 
> I appreciate any info on this ICE so I could try to fix it.

Sometimes --enable-checking=release triggers bugs that are hidden by other
--enable-checking settings.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[eres at il dot ibm dot com]

------- Comment #7 from eres at il dot ibm dot com  2008-02-28 13:49 -------
(In reply to comment #6)
> (In reply to comment #5)
> > I appreciate any info on this ICE so I could try to fix it.
> Sometimes --enable-checking=release triggers bugs that are hidden by other
> --enable-checking settings.

Thanks, unfortunately I still can not reproduce the fail.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ubizjak at gmail dot com]

------- Comment #8 from ubizjak at gmail dot com  2008-02-28 18:34 -------
(In reply to comment #7)

> Thanks, unfortunately I still can not reproduce the fail.

Probably you need newer binutils:

GNU ld (GNU Binutils) 2.18

Executing on host: /home/uros/gcc-build/gcc/xgcc -B/home/uros/gcc-build/gcc/
/home/uros/gcc-svn/trunk/gcc/testsuite/gcc.dg/tree-prof/pr34999.c   -O2
-freorder-blocks-and-partition -fprofile-use -D_PROFILE_USE -fno-show-column 
-lm   -o /home/uros/gcc-build/gcc/testsuite/gcc/pr34999.x02    (timeout = 300)
/usr/local/x86_64-unknown-linux-gnu/bin/ld: error in
/tmp/cceUz9XT.o(.eh_frame); no .eh_frame_hdr table will be created.^M
output is:
/usr/local/x86_64-unknown-linux-gnu/bin/ld: error in
/tmp/cceUz9XT.o(.eh_frame); no .eh_frame_hdr table will be created.^M

FAIL: gcc.dg/tree-prof/pr34999.c compilation,  -fprofile-use -D_PROFILE_USE
UNRESOLVED: gcc.dg/tree-prof/pr34999.c execution,    -fprofile-use
-D_PROFILE_USE

HJ, is this a binutils bug or gcc bug?


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hjl dot tools at gmail dot
                   |                            |com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ubizjak at gmail dot com]

------- Comment #9 from ubizjak at gmail dot com  2008-02-28 19:12 -------
Created an attachment (id=15241)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15241&action=view)
asm that crashes linker

gcc pr34999.s 
/usr/local/lib/gcc/x86_64-unknown-linux-gnu/4.4.0/../../../../x86_64-unknown-linux-gnu/bin/ld:
error in /tmp/ccuXalsV.o(.eh_frame); no .eh_frame_hdr table will be created.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[eres at il dot ibm dot com]

------- Comment #10 from eres at il dot ibm dot com  2008-02-29 05:23 -------
(In reply to comment #8)
> (In reply to comment #7)
> > Thanks, unfortunately I still can not reproduce the fail.
> Probably you need newer binutils:
> GNU ld (GNU Binutils) 2.18

Yes, using a newer binutils the fail is reproducible. thanks


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[hjl dot tools at gmail dot com]

------- Comment #11 from hjl dot tools at gmail dot com  2008-02-29 05:54 -------
I don't think it is a linker bug. This bug looks very similar
to PR 34249. Uros, you fixed PR 34249. Maybe there is another
similar problem elsewhere in dwarf2.out.c.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ubizjak at gmail dot com]

------- Comment #12 from ubizjak at gmail dot com  2008-02-29 06:07 -------
(In reply to comment #11)
> I don't think it is a linker bug. This bug looks very similar
> to PR 34249. Uros, you fixed PR 34249. Maybe there is another
> similar problem elsewhere in dwarf2.out.c.

The problem is in this part of the frame:

.LASFDE3:
        .long   .LASFDE3-.Lframe1
        .long   .LHOTB2
        .long   .LHOTE2-.LHOTB2
        .long   .LCOLDB2
>>>	.long	.LCOLDE2-.LCOLDB2
        .uleb128 0x0
        .byte   0x4
        .long   .LCFI2-.LFB3
        .byte   0xe
        .uleb128 0x10
        .byte   0x86
        .uleb128 0x2
        .byte   0x4
        .long   .LCFI3-.LCFI2
        .byte   0xd
        .uleb128 0x6
        .align 8
.LEFDE3:

If the marked line is changed to .long  .LCOLDB2-.LCOLDB2 (or ..E2 - ..E2) the
compilation succeeds. However, looking through asm, it looks correct to me.


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ubizjak at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ubizjak at gmail dot com]

------- Comment #13 from ubizjak at gmail dot com  2008-02-29 06:32 -------
(In reply to comment #12)

> If the marked line is changed to .long  .LCOLDB2-.LCOLDB2 (or ..E2 - ..E2) the
> compilation succeeds. However, looking through asm, it looks correct to me.

That is, the frame and label positions look correct to me. The problem in PR
34249 was malformed frame, which is not the case here.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ubizjak at gmail dot com]

------- Comment #14 from ubizjak at gmail dot com  2008-02-29 06:47 -------
CC Jakub, perhaps he can give us a clue what goes wrong here.


-- 

ubizjak at gmail dot com changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[jakub at gcc dot gnu dot org]

------- Comment #15 from jakub at gcc dot gnu dot org  2008-02-29 08:56 -------
.LSFDE3:
        .long   .LEFDE3-.LASFDE3        # FDE Length
.LASFDE3:
        .long   .LASFDE3-.Lframe1       # FDE CIE offset
        .long   .LHOTB2 # FDE initial location
        .long   .LHOTE2-.LHOTB2 # FDE address range
        .long   .LCOLDB2        # FDE initial location
        .long   .LCOLDE2-.LCOLDB2       # FDE address range
        .uleb128 0x0    # Augmentation size
        .byte   0x4     # DW_CFA_advance_loc4
        .long   .LCFI2-.LFB3
        .byte   0xe     # DW_CFA_def_cfa_offset
        .uleb128 0x10
        .byte   0x86    # DW_CFA_offset, column 0x6
        .uleb128 0x2
        .byte   0x4     # DW_CFA_advance_loc4
        .long   .LCFI3-.LCFI2
        .byte   0xd     # DW_CFA_def_cfa_register
        .uleb128 0x6
        .align 8
.LEFDE3:

is not a valid .eh_frame FDE, the
2005-03-31  Caroline Tice  <ctice@apple.com>

        (output_call_frame_info): Add test to see if function switches text
        sections in the middle; if so, use appropriate extra hot and cold
        section labels to compute size deltas for the hot and cold sections.

change was totally broken.  There is no such thing as a two ranges FDE in
Dwarf2/3.  FDE starts with FDE length, CIE offset, FDE initial location, FDE
address range, depending on CIE optionally with augmentation size and
augmentation and then CFA instructions.  So, to describe .text/.text.unlikely
split function, you need two FDEs, one describing the .text section, the other
describing .text.unlikely section.  When the code jumps in between the
sections, you need to insert CFA instructions that reflect what really changed.
The easiest would be to use .cfi_* assembler directives that recentish gas
supports and emitting them inline in the code, rather than creating separate
.eh_frame.  But I believe we need some more help from gas - particularly more
powerful .cfi_escape and some easy way to just save current state resp. restore
it.  That way you could save the state at the jump location and restore it at
the jump target and let gas compute what CFA instructions are needed.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[uweigand at gcc dot gnu dot org]

------- Comment #16 from uweigand at gcc dot gnu dot org  2008-03-04 14:51 -------
Hi Jakub,

we need the same changes in both .eh_frame and .dwarf_frame;
does the gas .cfi_ support both sections?

I'm wondering how "save & restore" should work across two
different FDEs -- in the new FDE, we'd have to emit the full
set of CFA instructions to get to the "base-line" state ...


-- 

uweigand at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |uweigand at de dot ibm dot
                   |                            |com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[jakub at gcc dot gnu dot org]

------- Comment #17 from jakub at gcc dot gnu dot org  2008-03-04 15:10 -------
.cfi_* doesn't create .dwarf_frame, but perhaps it could be taught to do that
optionally (some flag on .cfi_startproc that would switch on additional
creation of .dwarf_frame).
The plan with save/restore was that you could do say
.text
.cfi_startproc
.cfi_personality 3, __gxx_personality_v0
.cfi_lsda 3, .LLSDA2
some code plus .cfi_* instructions for it
jmp 1f
.cfi_save .L0123
.section .text.unlikely
.cfi_startproc
.cfi_personality 3, __gxx_personality_v0
.cfi_lsda .LLSDA2
.cfi_restore .L0123
1:
some code plus .cfi_* instructions for it
.cfi_endproc
.previous
.cfi_endproc

where that .cfi_save would add a list of .cfi_* instructions from
.cfi_startproc
till that .cfi_save into .L0123 queue, then .cfi_restore would just add those
queued .cfi directives all at the same location (with some optimizations, e.g.
.cfi_* instructions that clearly cancel themselves would be optimized out, any
location advances of course taken out, etc.).
You can of course do that in dwarf2out.c, but we eventually want to use inline
.cfi_* instructions for other reasons - e.g. so that inline asm code can add
its  unwind info.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[danglin at gcc dot gnu dot org]

------- Comment #18 from danglin at gcc dot gnu dot org  2008-03-16 16:27 -------
Executing on host: /test/gnu/gcc/objdir/gcc/xgcc -B/test/gnu/gcc/objdir/gcc/
/te
st/gnu/gcc/gcc/gcc/testsuite/gcc.dg/tree-prof/pr34999.c   -O2
-freorder-blocks-a
nd-partition -fprofile-use -D_PROFILE_USE -fno-show-column  -lm   -o
/test/gnu/g
cc/objdir/gcc/testsuite/gcc/pr34999.x02    (timeout = 300)
ld: An unknown relocation type 8 was encountered when applying the relocation
fo
r symbol ".text.unlikely" at offset 0x15c in section index 1 of file
/var/tmp//c
cFqB2Ph.o
1 errors.
collect2: ld returned 1 exit status
compiler exited with status 1
output is:
ld: An unknown relocation type 8 was encountered when applying the relocation
fo
r symbol ".text.unlikely" at offset 0x15c in section index 1 of file
/var/tmp//c
cFqB2Ph.o
1 errors.
collect2: ld returned 1 exit status

FAIL: gcc.dg/tree-prof/pr34999.c compilation,  -fprofile-use -D_PROFILE_USE

I believe the linker is complaining about the R_PARISC_PCREL12F relocation
used in this branch.

 15c:   9f f3 20 00     cmpb,*= r19,r31,164 <main+0x134>

Thought this hot and cold stuff was disabled on hppa as it breaks a lot
of stuff.


-- 

danglin at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |danglin at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[eres at il dot ibm dot com]

------- Comment #19 from eres at il dot ibm dot com  2008-04-08 11:07 -------
> The easiest would be to use .cfi_* assembler directives that recentish gas
> supports and emitting them inline in the code, rather than creating separate
> .eh_frame.  

I apologize ahead if I am totally wrong about it but 
on targets with limited local store when an overlay technique is used it might
be prefered to place the unwind info in a separate section that will always be
in the memory rather then inlining it using .cfi_* assembler.  


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[jakub at gcc dot gnu dot org]

------- Comment #20 from jakub at gcc dot gnu dot org  2008-04-08 11:13 -------
Please read the documentation about .cfi_* directives.  They construct an
.eh_frame section for you.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[pinskia at gcc dot gnu dot org]

------- Comment #21 from pinskia at gcc dot gnu dot org  2008-12-29 02:35 -------
*** Bug 37058 has been marked as a duplicate of this bug. ***


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rguenth at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[ghazi at gcc dot gnu dot org]

------- Comment #22 from ghazi at gcc dot gnu dot org  2009-03-27 16:37 -------
I still see failures in bb-reorg.c and pr34999.c on x86_64-unknown-linux-gnu:
http://gcc.gnu.org/ml/gcc-testresults/2009-03/msg02744.html

The error message is slightly different than what others have posted here.  I'm
using "GNU assembler 2.17 Debian GNU/Linux" on the GCC compile farm:

FAIL: gcc.dg/tree-prof/bb-reorg.c compilation,  -fprofile-use -D_PROFILE_USE
/tmp/cc87gzwS.s: Assembler messages:
/tmp/cc87gzwS.s:116: Error: can't resolve `.text.unlikely' {.text.unlikely
section} - `L0^A' {.text section}
compiler exited with status 1

FAIL: gcc.dg/tree-prof/pr34999.c compilation,  -fprofile-use -D_PROFILE_USE
/tmp/ccDOht2P.s: Assembler messages:
/tmp/ccDOht2P.s:109: Error: can't resolve `.text.unlikely' {.text.unlikely
section} - `L0^A' {.text section}
compiler exited with status 1

These are the last C testsuite failures I get on this box.  Perhaps we can
xfail them when we branch if it won't be fixed in time for 4.4?


-- 

ghazi at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ghazi at gcc dot gnu dot org
   Last reconfirmed|2008-02-28 09:47:54         |2009-03-27 16:37:36
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Fallthru crossing edges in partition_hot_cold_basic_blocks are not been fixed when the section ends with call insn

[steven at gcc dot gnu dot org]

------- Comment #23 from steven at gcc dot gnu dot org  2009-03-27 17:27 -------
The checkin of the broken patch mentioned in comment #15 is here:
http://gcc.gnu.org/ml/gcc-cvs/2005-03/msg01521.html

Poor ctice also didn't manage to properly check in a ChangeLog entry, but oh
well.


-- 

steven at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |steven at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Incorrect FDE entries with hot/cold code section splitting (partition_hot_cold_basic_blocks)

[jakub at gcc dot gnu dot org]

------- Comment #24 from jakub at gcc dot gnu dot org  2009-07-24 15:37 -------
I have a patch.


-- 

jakub at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |jakub at gcc dot gnu dot org
                   |dot org                     |
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2009-03-27 16:37:36         |2009-07-24 15:37:04
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Incorrect FDE entries with hot/cold code section splitting (partition_hot_cold_basic_blocks)

[jakub at gcc dot gnu dot org]

------- Comment #25 from jakub at gcc dot gnu dot org  2009-07-24 23:30 -------
Subject: Bug 34999

Author: jakub
Date: Fri Jul 24 23:30:39 2009
New Revision: 150069

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=150069
Log:
        PR rtl-optimization/34999
        * dwarf2out.c (struct dw_fde_struct): Add dw_fde_switch_cfi
        and dw_fde_switched_cold_to_hot fields.
        (output_cfi_p): New function.
        (output_call_frame_info): If fde->dw_fde_switched_sections,
        output 2 FDEs instead of one with corrupted header.
        (dwarf2out_do_cfi_startproc): New function.
        (dwarf2out_begin_prologue): Use it.  Initialize fde->dw_fde_switch_cfi
        and fde->dw_fde_switched_cold_to_hot.
        (dwarf2out_switch_text_section): Compute
        fde->dw_fde_switched_cold_to_hot.  Switch to new text section here.
        If dwarf2out_do_cfi_asm, emit .cfi_endproc before it and call
        dwarf2out_do_cfi_startproc plus emit again currently active CFI insns.
        Otherwise, compute fde->dw_fde_switch_cfi.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/dwarf2out.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Incorrect FDE entries with hot/cold code section splitting (partition_hot_cold_basic_blocks)

[ghazi at gcc dot gnu dot org]

------- Comment #26 from ghazi at gcc dot gnu dot org  2009-08-06 19:54 -------
The patch fixed the bb-reorg.c and pr34999.c testsuite failures on my x86_64
box on mainline.  However I still see the failures on the 4.4 branch.

Jakub - Is your patch suitable for 4.4?
If so, will you please backport it?

Thanks.


-- 

ghazi at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2009-07-24 15:37:04         |2009-08-06 19:54:19
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Incorrect FDE entries with hot/cold code section splitting (partition_hot_cold_basic_blocks)

[ghazi at gcc dot gnu dot org]

------- Comment #27 from ghazi at gcc dot gnu dot org  2009-09-07 01:04 -------
(In reply to comment #25)
> Subject: Bug 34999
> Author: jakub
> Date: Fri Jul 24 23:30:39 2009
> New Revision: 150069
> URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=150069
> Log:
>         PR rtl-optimization/34999
>         * dwarf2out.c (struct dw_fde_struct): Add dw_fde_switch_cfi
>         and dw_fde_switched_cold_to_hot fields.
>         (output_cfi_p): New function.
>         (output_call_frame_info): If fde->dw_fde_switched_sections,
>         output 2 FDEs instead of one with corrupted header.
>         (dwarf2out_do_cfi_startproc): New function.
>         (dwarf2out_begin_prologue): Use it.  Initialize fde->dw_fde_switch_cfi
>         and fde->dw_fde_switched_cold_to_hot.
>         (dwarf2out_switch_text_section): Compute
>         fde->dw_fde_switched_cold_to_hot.  Switch to new text section here.
>         If dwarf2out_do_cfi_asm, emit .cfi_endproc before it and call
>         dwarf2out_do_cfi_startproc plus emit again currently active CFI insns.
>         Otherwise, compute fde->dw_fde_switch_cfi.
> Modified:
>     trunk/gcc/ChangeLog
>     trunk/gcc/dwarf2out.c

Is this patch suitable for 4.4.x?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Incorrect FDE entries with hot/cold code section splitting (partition_hot_cold_basic_blocks)

[ghazi at gcc dot gnu dot org]

------- Comment #28 from ghazi at gcc dot gnu dot org  2009-12-01 15:14 -------
Backport to 4.4?  Or close this and PR41501?


-- 

ghazi at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999

[Bug rtl-optimization/34999] Incorrect FDE entries with hot/cold code section splitting (partition_hot_cold_basic_blocks)

[rguenth at gcc dot gnu dot org]

------- Comment #29 from rguenth at gcc dot gnu dot org  2009-12-01 15:16 -------
The CFI stuff isn't on the 4.4 branch, so WONTFIX there.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34999