[Bug rtl-optimization/96475] New: direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[Bug rtl-optimization/96475] New: direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[npiggin at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

            Bug ID: 96475
           Summary: direct threaded interpreter with computed gotos
                    generates suboptimal dispatch loop
           Product: gcc
           Version: 10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: npiggin at gmail dot com
                CC: segher at gcc dot gnu.org
  Target Milestone: ---
            Target: powerpc64le-linux-gnu

Created attachment 48999
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=48999&action=edit
test case

The attached test case code generation with -O2 for run_program_goto generates
a central indirect branch dispatch to handlers that branch back to the central
dispatcher.

Direct threaded code with indirect branches between handlers is faster on a
POWER9 when there are no branch mispredictions due to fewer branches, and it
should generally do better with branch prediction when there is an indirect
branch from each handler.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[mikpelinux at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

Mikael Pettersson <mikpelinux at gmail dot com> changed:

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

--- Comment #1 from Mikael Pettersson <mikpelinux at gmail dot com> ---
You might want to use the -fno-tree-copyrename option.  The Erlang/OTP VM,
which is a directly-threaded interpreter, uses it on all targets since gcc-4.3.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #2 from Segher Boessenkool <segher at gcc dot gnu.org> ---
That option was removed in GCC 6, already.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[mikpelinux at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #3 from Mikael Pettersson <mikpelinux at gmail dot com> ---
(In reply to Segher Boessenkool from comment #2)
> That option was removed in GCC 6, already.

Oh.  Do you know if they added a replacement or just dropped the functionality
altogether?

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #4 from Segher Boessenkool <segher at gcc dot gnu.org> ---
https://gcc.gnu.org/g:1f9ceff11132

-ftree-coalesce-vars it sounds like?  But that isn't a 1-1 replacement
probably (or, what was the point of this change otherwise?)

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[mikpelinux at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #5 from Mikael Pettersson <mikpelinux at gmail dot com> ---
I can't reproduce on {ppc64le, ppc64, aarch64, m68k, x86_64}-linux-gnu, they
all inline the dispatcher at the end of each handler, w/ gcc-10.2.0 -O2. 
However, for sparc64 gcc-10.2.0 -O2 does end each handler with a branch to a
shared dispatcher (which is not what we want).

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #6 from Segher Boessenkool <segher at gcc dot gnu.org> ---
I have a patch (since yesterday; just no time to send it :-/ )

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Segher Boessenkool <segher@gcc.gnu.org>:

https://gcc.gnu.org/g:69ca5f3a988266da8905aef9cf22aa02807e0471

commit r11-3078-g69ca5f3a988266da8905aef9cf22aa02807e0471
Author: Segher Boessenkool <segher@kernel.crashing.org>
Date:   Fri Aug 7 01:31:38 2020 +0000

    bb-reorder: Remove a misfiring micro-optimization (PR96475)

    When the compgotos pass copies the tail of blocks ending in an indirect
    jump, there is a micro-optimization to not copy the last one, since the
    original block will then just be deleted.  This does not work properly
    if cleanup_cfg does not merge all pairs of blocks we expect it to.  It
    also does not work if that last block can be merged into multiple
    predecessors.

    2020-09-09  Segher Boessenkool  <segher@kernel.crashing.org>

            PR rtl-optimization/96475
            * bb-reorder.c (maybe_duplicate_computed_goto): Remove
single_pred_p
            micro-optimization.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

Segher Boessenkool <segher at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2020-09-09
             Status|UNCONFIRMED                 |NEW

--- Comment #8 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Fixed on trunk.  Needs backports.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

Christophe Lyon <clyon at gcc dot gnu.org> changed:

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

--- Comment #9 from Christophe Lyon <clyon at gcc dot gnu.org> ---
(In reply to Segher Boessenkool from comment #8)
> Fixed on trunk.  Needs backports.

This is causing ICEs when compiling glibc (2.29) for
aarch64-none-linux-gnu and 
arm-none-linux-gnueabihf --with-mode=thumb --with-cpu=cortex-a15
--with-fpu=neon-vfpv4

0x82a054 vec<basic_block_def*, va_gc, vl_embed>::operator[](unsigned int)
        /tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/vec.h:880
0x82a054 pre_and_rev_post_order_compute_fn(function*, int*, int*, bool)
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/cfganal.c:1036
0x82a0b7 pre_and_rev_post_order_compute(int*, int*, bool)
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/cfganal.c:1050
0x7db017 init_alias_analysis()
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/alias.c:3392
0x180c862 sched_init()
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/haifa-sched.c:7326
0x180e1ed haifa_sched_init()
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/haifa-sched.c:7363
0xcd64dc schedule_insns
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3514
0xcd6da0 schedule_insns
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3930
0xcd6da0 rest_of_handle_sched_fusion
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3757
0xcd6da0 execute
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3932
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.
make[2]: *** [../o-iterator.mk:9:
/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/glibc-1/stdio-common/vfprintf-internal.o]
Error 1

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #10 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Please show the full message?  (It starts with "internal compiler error".)

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #11 from Christophe Lyon <clyon at gcc dot gnu.org> ---
Ha sorry I missed the beginning (interleaved in the log with other commands):
during RTL pass: sched_fusion
vfprintf-internal.c: In function '__vfprintf_internal':
vfprintf-internal.c:1702:1: internal compiler error: in operator[], at
vec.h:880
 1702 | }
      | ^
0x82a054 vec<basic_block_def*, va_gc, vl_embed>::operator[](unsigned int)
        /tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/vec.h:880
0x82a054 pre_and_rev_post_order_compute_fn(function*, int*, int*, bool)
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/cfganal.c:1036
0x82a0b7 pre_and_rev_post_order_compute(int*, int*, bool)
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/cfganal.c:1050
0x7db017 init_alias_analysis()
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/alias.c:3392
0x180c862 sched_init()
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/haifa-sched.c:7326
0x180e1ed haifa_sched_init()
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/haifa-sched.c:7363
0xcd64dc schedule_insns
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3514
0xcd6da0 schedule_insns
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3930
0xcd6da0 rest_of_handle_sched_fusion
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3757
0xcd6da0 execute
       
/tmp/4800342_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/sched-rgn.c:3932


I can prepare the preprocessed source tomorrow if needed.


The full compilation command:
/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/tools/bin/aarch64-none-linux-gnu-gcc
vfprintf-internal.c -c -std=gnu11 -fgnu89-inline  -g -O2 -Wall -Wwrite-strings
-Wundef -fmerge-all-const
ants -frounding-math -fno-stack-protector -Wstrict-prototypes
-Wold-style-definition -fmath-errno      -ftls-model=initial-exec     
-I../include -I/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-
gccsrc/obj-aarch64-none-linux-gnu/glibc-1/stdio-common 
-I/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/glibc-1
 -I../sysdeps/unix/sysv/linux/aarch64  -I../sysd
eps/aarch64/nptl  -I../sysdeps/unix/sysv/linux/generic 
-I../sysdeps/unix/sysv/linux/wordsize-64  -I../sysdeps/unix/sysv/linux/include
-I../sysdeps/unix/sysv/linux  -I../sysdeps/nptl  -I../sysdeps/
pthread  -I../sysdeps/gnu  -I../sysdeps/unix/inet  -I../sysdeps/unix/sysv 
-I../sysdeps/unix  -I../sysdeps/posix  -I../sysdeps/aarch64/fpu 
-I../sysdeps/aarch64/multiarch  -I../sysdeps/aarch64  -I.
./sysdeps/wordsize-64  -I../sysdeps/ieee754/ldbl-128 
-I../sysdeps/ieee754/dbl-64/wordsize-64  -I../sysdeps/ieee754/dbl-64 
-I../sysdeps/ieee754/flt-32  -I../sysdeps/ieee754  -I../sysdeps/generic  
-I.. -I../libio -I. -nostdinc -isystem
/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/tools/lib/gcc/aarch64-none-linux-gnu/11.0.0/include
-isystem /tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/
gcc-fsf-gccsrc/tools/lib/gcc/aarch64-none-linux-gnu/11.0.0/include-fixed
-isystem
/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/sysroot-aarch64-none-linux-gnu/usr/include
 -D_LIBC_REENTRA
NT -include
/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/glibc-1/libc-modules.h
-DMODULE_NAME=libc -include ../include/libc-symbols.h       -DTOP_NAMESPACE=gli
bc -D_IO_MTSAFE_IO -o
/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/glibc-1/stdio-common/vfprintf-internal.o
-MD -MP -MF /tmp/4800342_6.tmpdir/aci-gcc-fsf/build
s/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/glibc-1/stdio-common/vfprintf-internal.o.dt
-MT
/tmp/4800342_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/glibc-1/stdio-common/v
fprintf-internal.o

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #12 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Thanks.  Unfortunately it isn't saying much more (well, during which
pass this happened, pretty important ;-) )

> I can prepare the preprocessed source tomorrow if needed.

Thanks!  That will make reproducing the problem much simpler.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #13 from Christophe Lyon <clyon at gcc dot gnu.org> ---
Created attachment 49206
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49206&action=edit
glibc testcase for aarch64 ICE

The vfprintf-internal.i preprocessed source from glibc-2.29 creates an ICE on
aarch64 with:
cc1 -fpreprocessed vfprintf-internal.i -O2 -quiet
during RTL pass: sched_fusion
vfprintf-internal.c: In function '__vfprintf_internal':
vfprintf-internal.c:1702:1: internal compiler error: in operator[], at
vec.h:880
 1702 | }
      | ^
0x850282 vec<basic_block_def*, va_gc, vl_embed>::operator[](unsigned int)
        /home/christophe.lyon/src/GCC/sources/gcc-fsf-git/trunk/gcc/vec.h:880
0x850282 pre_and_rev_post_order_compute_fn(function*, int*, int*, bool)
       
/home/christophe.lyon/src/GCC/sources/gcc-fsf-git/trunk/gcc/cfganal.c:1036
0x850337 pre_and_rev_post_order_compute(int*, int*, bool)
[...]

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #14 from Christophe Lyon <clyon at gcc dot gnu.org> ---
I that's easier for you, there are similar ICEs while running the GCC testsuite
for aarch64-elf (newlib build works):

  Executed from: gcc.c-torture/compile/compile.exp
    gcc.c-torture/compile/950613-1.c   -O2  (internal compiler error)
    gcc.c-torture/compile/950613-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/compile/950613-1.c   -O3 -g  (internal compiler error)
    gcc.c-torture/compile/pr25224.c   -O2  (internal compiler error)
    gcc.c-torture/compile/pr25224.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/compile/pr25224.c   -O3 -fomit-frame-pointer -funroll-loops
-fpeel-loops -ftracer -finline-functions  (internal compiler error)
    gcc.c-torture/compile/pr25224.c   -O3 -g  (internal compiler error)
    gcc.c-torture/compile/pr27863.c   -O2  (internal compiler error)
    gcc.c-torture/compile/pr27863.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/compile/pr27863.c   -O3 -g  (internal compiler error)
    gcc.c-torture/compile/pr28489.c   -O2  (internal compiler error)
    gcc.c-torture/compile/pr28489.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/compile/pr28489.c   -O3 -g  (internal compiler error)
    gcc.c-torture/compile/pr29128.c   -O2  (internal compiler error)
    gcc.c-torture/compile/pr29128.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/compile/pr29128.c   -O3 -g  (internal compiler error)
    gcc.c-torture/compile/pr46107.c   -O2  (internal compiler error)
    gcc.c-torture/compile/pr46107.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/compile/pr46107.c   -O3 -g  (internal compiler error)
  Executed from: gcc.c-torture/execute/execute.exp
    gcc.c-torture/execute/20040302-1.c   -O2  (internal compiler error)
    gcc.c-torture/execute/20040302-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/execute/20040302-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  (internal compiler error)
    gcc.c-torture/execute/20040302-1.c   -O3 -g  (internal compiler error)
    gcc.c-torture/execute/20071210-1.c   -O2  (internal compiler error)
    gcc.c-torture/execute/20071210-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/execute/20071210-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  (internal compiler error)
    gcc.c-torture/execute/20071210-1.c   -O3 -g  (internal compiler error)
    gcc.c-torture/execute/920302-1.c   -O2  (internal compiler error)
    gcc.c-torture/execute/920302-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/execute/920302-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  (internal compiler error)
    gcc.c-torture/execute/920302-1.c   -O3 -fomit-frame-pointer -funroll-loops
-fpeel-loops -ftracer -finline-functions  (internal compiler error)
    gcc.c-torture/execute/920302-1.c   -O3 -g  (internal compiler error)
    gcc.c-torture/execute/comp-goto-1.c   -O2  (internal compiler error)
    gcc.c-torture/execute/comp-goto-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.c-torture/execute/comp-goto-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  (internal compiler error)
    gcc.c-torture/execute/comp-goto-1.c   -O3 -fomit-frame-pointer
-funroll-loops -fpeel-loops -ftracer -finline-functions  (internal compiler
error)
    gcc.c-torture/execute/comp-goto-1.c   -O3 -g  (internal compiler error)
  Executed from: gcc.dg/torture/dg-torture.exp
    gcc.dg/torture/pr53695.c   -O2  (internal compiler error)
    gcc.dg/torture/pr53695.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error)
    gcc.dg/torture/pr53695.c   -O3 -g  (internal compiler error)

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #15 from Alex Coplan <acoplan at gcc dot gnu.org> ---
> there are similar ICEs while running the GCC testsuite for aarch64-elf

Confirmed, seeing these ICEs when running the testsuite on
aarch64-none-linux-gnu.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[dinuxbg at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

Dimitar Dimitrov <dinuxbg at gmail dot com> changed:

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

--- Comment #16 from Dimitar Dimitrov <dinuxbg at gmail dot com> ---
Patch is also causing ICE in a few test cases on pru-elf target:

during RTL pass: ce3
/home/dinux/projects/pru/testbot-workspace/gcc/gcc/testsuite/gcc.c-torture/compile/pr46107.c:
In function 'bar':
/home/dinux/projects/pru/testbot-workspace/gcc/gcc/testsuite/gcc.c-torture/compile/pr46107.c:18:1:
internal compiler error: in calc_dfs_tree, at dominance.c:458
0x598a77 calc_dfs_tree
        /home/dinux/projects/pru/testbot-workspace/gcc/gcc/dominance.c:458
0x96f509 calculate_dominance_info(cdi_direction)
        /home/dinux/projects/pru/testbot-workspace/gcc/gcc/dominance.c:734
0x90d292 flow_loops_find(loops*)
        /home/dinux/projects/pru/testbot-workspace/gcc/gcc/cfgloop.c:431
0xbb39ce loop_optimizer_init(unsigned int)
        /home/dinux/projects/pru/testbot-workspace/gcc/gcc/loop-init.c:93
0x1453d90 if_convert
        /home/dinux/projects/pru/testbot-workspace/gcc/gcc/ifcvt.c:5382
0x14562fd execute
        /home/dinux/projects/pru/testbot-workspace/gcc/gcc/ifcvt.c:5566

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

Arseny Solokha <asolokha at gmx dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asolokha at gmx dot com

--- Comment #17 from Arseny Solokha <asolokha at gmx dot com> ---
(In reply to Dimitar Dimitrov from comment #16)
> Patch is also causing ICE in a few test cases on pru-elf target:

It causes ICEs on practically any target, though I believe it should have been
reported separately rather than in this PR.

% gcc-11.0.0 -O1 -fexpensive-optimizations -w -c
gcc/testsuite/gcc.c-torture/execute/920302-1.c
during RTL pass: ce3
gcc/testsuite/gcc.c-torture/execute/920302-1.c: In function 'execute':
gcc/testsuite/gcc.c-torture/execute/920302-1.c:25:1: internal compiler error:
in calc_dfs_tree, at dominance.c:458
   25 | }
      | ^
0x618911 calc_dfs_tree
       
/var/tmp/portage/sys-devel/gcc-11.0.0_alpha20200913/work/gcc-11-20200913/gcc/dominance.c:458
0x98bec9 calculate_dominance_info(cdi_direction)
       
/var/tmp/portage/sys-devel/gcc-11.0.0_alpha20200913/work/gcc-11-20200913/gcc/dominance.c:734
0x921972 flow_loops_find(loops*)
       
/var/tmp/portage/sys-devel/gcc-11.0.0_alpha20200913/work/gcc-11-20200913/gcc/cfgloop.c:431
0xbd6eae loop_optimizer_init(unsigned int)
       
/var/tmp/portage/sys-devel/gcc-11.0.0_alpha20200913/work/gcc-11-20200913/gcc/loop-init.c:93
0x1761c73 if_convert
       
/var/tmp/portage/sys-devel/gcc-11.0.0_alpha20200913/work/gcc-11-20200913/gcc/ifcvt.c:5382
0x176494d execute
       
/var/tmp/portage/sys-devel/gcc-11.0.0_alpha20200913/work/gcc-11-20200913/gcc/ifcvt.c:5566

w/ gcc-11.0.0-alpha20200913 snapshot.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #18 from Segher Boessenkool <segher at gcc dot gnu.org> ---
I have a patch.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #19 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Created attachment 49215
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49215&action=edit
proposed patch for the ICEs

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #20 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Could you guys please test the attached patch?  Thanks in advance!

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #21 from Arseny Solokha <asolokha at gmx dot com> ---
(In reply to Segher Boessenkool from comment #20)
> Could you guys please test the attached patch?  Thanks in advance!

It fixed the ICE on x86_64 I've reported in comment 17. I didn't run a regtest,
though.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #22 from Christophe Lyon <clyon at gcc dot gnu.org> ---
Yes, that fixes the build for aarch64, thanks!

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[dinuxbg at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #23 from Dimitar Dimitrov <dinuxbg at gmail dot com> ---
(In reply to Segher Boessenkool from comment #20)
> Could you guys please test the attached patch?  Thanks in advance!

Yes, it fixed all new regressions for pru-elf. Thanks.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

--- Comment #24 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Segher Boessenkool <segher@gcc.gnu.org>:

https://gcc.gnu.org/g:bab581381681813ab3ccd007195777c1f816e656

commit r11-3205-gbab581381681813ab3ccd007195777c1f816e656
Author: Segher Boessenkool <segher@kernel.crashing.org>
Date:   Mon Sep 14 18:19:47 2020 +0000

    bb-reorder: Fix for ICEs caused by 69ca5f3a9882

    After the previous patch we are left with an unreachable BB.  This will
    ICE if either we have -fschedule-fusion, or we do not have peephole2.

    2020-09-15  Segher Boessenkool  <segher@kernel.crashing.org>

            PR rtl-optimization/96475
            * bb-reorder.c (duplicate_computed_gotos): If we did anything, run
            cleanup_cfg.

[Bug rtl-optimization/96475] direct threaded interpreter with computed gotos generates suboptimal dispatch loop

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96475

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED