[Bug target/116592] New: illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[Bug target/116592] New: illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[shuizhuyuanluo at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

            Bug ID: 116592
           Summary: illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3
                    when compiling for risc-v xtheadvector
           Product: gcc
           Version: 15.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: shuizhuyuanluo at gmail dot com
  Target Milestone: ---

$ ./riscv64-unknown-linux-gnu-g++ -c test3.c -o test3.o
-march=rv64gc_zfh_xtheadvector -O2
/tmp/ccQPGUNc.s: Assembler messages:
/tmp/ccQPGUNc.s:86: Error: illegal operands `th.vsetvli zero,0,e32,m8'


-O1 compiles fine
-O2 and -O3 gives the above error




$ ./riscv64-unknown-linux-gnu-g++ -v
Using built-in specs.
COLLECT_GCC=./riscv64-unknown-linux-gnu-g++
COLLECT_LTO_WRAPPER=/data/action/osd/riscv/libexec/gcc/riscv64-unknown-linux-gnu/15.0.0/lto-wrapper
Target: riscv64-unknown-linux-gnu
Configured with: /data/action/osd/riscv-gnu-toolchain/gcc/configure
--target=riscv64-unknown-linux-gnu --prefix=/data/action/osd/riscv
--with-sysroot=/data/action/osd/riscv/sysroot --with-pkgversion=g155da081706
--with-system-zlib --enable-shared --enable-tls
--enable-languages=c,c++,fortran --disable-libmudflap --disable-libssp
--disable-libquadmath --disable-libsanitizer --disable-nls --disable-bootstrap
--src=.././gcc --disable-default-pie --disable-multilib --with-abi=lp64d
--with-arch=rv64imafdc --with-tune=rocket --with-isa-spec=20191213
'CFLAGS_FOR_TARGET=-O2    -mcmodel=medlow' 'CXXFLAGS_FOR_TARGET=-O2   
-mcmodel=medlow'
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 15.0.0 20240829 (experimental) (g155da081706) 



binutils master branch

commit 5cca20f614694ba6af3c10ca5bb93c32cdb9522b (HEAD -> master, origin/master,
origin/HEAD)
Author: GDB Administrator <gdbadmin@sourceware.org>
Date:   Thu Aug 29 00:00:14 2024 +0000

    Automatic date update in version.in



------------- soucecode -------------


#include <math.h>
#include <riscv_vector.h>

static vfloat32m8_t atan2_ps(vfloat32m8_t a, vfloat32m8_t b, size_t vl)
{
    float tmpx[vl];
    float tmpy[vl];
    __riscv_vse32_v_f32m8(tmpx, a, vl);
    __riscv_vse32_v_f32m8(tmpy, b, vl);
    for (size_t i = 0; i < vl; i++)
    {
        tmpx[i] = atan2(tmpx[i], tmpy[i]);
    }
    return __riscv_vle32_v_f32m8(tmpx, vl);
}

void my_atan2(const float* x, const float* y, float* out, int size)
{
    int n = size;
    while (n > 0)
    {
        size_t vl = __riscv_vsetvl_e32m8(n);
        vfloat32m8_t _x = __riscv_vle32_v_f32m8(x, vl);
        vfloat32m8_t _y = __riscv_vle32_v_f32m8(y, vl);
        vfloat32m8_t _out = atan2_ps(_x, _y, vl);
        __riscv_vse32_v_f32m8(out, _out, vl);
        n -= vl;
        x += vl;
        y += vl;
        out += vl;
    }
}

[Bug target/116592] illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I think it is time to remove xtheadvector support it is just broken.

[Bug target/116592] illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[kito at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

Kito Cheng <kito at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |cmuellner at gcc dot gnu.org,
                   |                            |kito at gcc dot gnu.org

--- Comment #2 from Kito Cheng <kito at gcc dot gnu.org> ---
Hi Christoph:

Not sure if Vrull is still care theadvector?

[Bug target/116592] illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[cooper.qu at linux dot alibaba.com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

Cooper Qu <cooper.qu at linux dot alibaba.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |cooper.qu at linux dot alibaba.com

--- Comment #3 from Cooper Qu <cooper.qu at linux dot alibaba.com> ---
Hi,

We are currently maintaining xtheadvector. Some bugs, including this one, have
already been fixed downstream, and we will submit them gradually.

[Bug target/116592] illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[jinma at linux dot alibaba.com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

Jin Ma <jinma at linux dot alibaba.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jinma at linux dot alibaba.com

--- Comment #4 from Jin Ma <jinma at linux dot alibaba.com> ---
Hi, all

I have noticed this bug recently and have fixed it, and there are several other
related patches being sorted out. I will try to push it upstream recently. In
fact, we have been using it and polishing it as much as possible.

BR
Jin

[Bug target/116592] illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[jinma at linux dot alibaba.com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

--- Comment #5 from Jin Ma <jinma at linux dot alibaba.com> ---
(In reply to nihui from comment #0)
> $ ./riscv64-unknown-linux-gnu-g++ -c test3.c -o test3.o
> -march=rv64gc_zfh_xtheadvector -O2
> /tmp/ccQPGUNc.s: Assembler messages:
> /tmp/ccQPGUNc.s:86: Error: illegal operands `th.vsetvli zero,0,e32,m8'
> 
> 
> -O1 compiles fine
> -O2 and -O3 gives the above error
> 
> 
> 
> 
> $ ./riscv64-unknown-linux-gnu-g++ -v
> Using built-in specs.
> COLLECT_GCC=./riscv64-unknown-linux-gnu-g++
> COLLECT_LTO_WRAPPER=/data/action/osd/riscv/libexec/gcc/riscv64-unknown-linux-
> gnu/15.0.0/lto-wrapper
> Target: riscv64-unknown-linux-gnu
> Configured with: /data/action/osd/riscv-gnu-toolchain/gcc/configure
> --target=riscv64-unknown-linux-gnu --prefix=/data/action/osd/riscv
> --with-sysroot=/data/action/osd/riscv/sysroot --with-pkgversion=g155da081706
> --with-system-zlib --enable-shared --enable-tls
> --enable-languages=c,c++,fortran --disable-libmudflap --disable-libssp
> --disable-libquadmath --disable-libsanitizer --disable-nls
> --disable-bootstrap --src=.././gcc --disable-default-pie --disable-multilib
> --with-abi=lp64d --with-arch=rv64imafdc --with-tune=rocket
> --with-isa-spec=20191213 'CFLAGS_FOR_TARGET=-O2    -mcmodel=medlow'
> 'CXXFLAGS_FOR_TARGET=-O2    -mcmodel=medlow'
> Thread model: posix
> Supported LTO compression algorithms: zlib
> gcc version 15.0.0 20240829 (experimental) (g155da081706) 
> 
> 
> 
> binutils master branch
> 
> commit 5cca20f614694ba6af3c10ca5bb93c32cdb9522b (HEAD -> master,
> origin/master, origin/HEAD)
> Author: GDB Administrator <gdbadmin@sourceware.org>
> Date:   Thu Aug 29 00:00:14 2024 +0000
> 
>     Automatic date update in version.in
> 
> 
> 
> ------------- soucecode -------------
> 
> 
> #include <math.h>
> #include <riscv_vector.h>
> 
> static vfloat32m8_t atan2_ps(vfloat32m8_t a, vfloat32m8_t b, size_t vl)
> {
>     float tmpx[vl];
>     float tmpy[vl];
>     __riscv_vse32_v_f32m8(tmpx, a, vl);
>     __riscv_vse32_v_f32m8(tmpy, b, vl);
>     for (size_t i = 0; i < vl; i++)
>     {
>         tmpx[i] = atan2(tmpx[i], tmpy[i]);
>     }
>     return __riscv_vle32_v_f32m8(tmpx, vl);
> }
> 
> void my_atan2(const float* x, const float* y, float* out, int size)
> {
>     int n = size;
>     while (n > 0)
>     {
>         size_t vl = __riscv_vsetvl_e32m8(n);
>         vfloat32m8_t _x = __riscv_vle32_v_f32m8(x, vl);
>         vfloat32m8_t _y = __riscv_vle32_v_f32m8(y, vl);
>         vfloat32m8_t _out = atan2_ps(_x, _y, vl);
>         __riscv_vse32_v_f32m8(out, _out, vl);
>         n -= vl;
>         x += vl;
>         y += vl;
>         out += vl;
>     }
> }


I tried to fix this issue, see the patch below link:

https://gcc.gnu.org/pipermail/gcc-patches/2024-September/662463.html

BR
Jin

[Bug target/116592] illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

--- Comment #6 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jeff Law <law@gcc.gnu.org>:

https://gcc.gnu.org/g:d620499b3a24f14cfb98529640584e63d7eca149

commit r15-3527-gd620499b3a24f14cfb98529640584e63d7eca149
Author: Jin Ma <jinma@linux.alibaba.com>
Date:   Sat Sep 7 10:29:02 2024 -0600

    [PATCH v4] [target/116592] RISC-V: Fix illegal operands "th.vsetvli
zero,0,e32,m8" for XTheadVector

    Since the THeadVector vsetvli does not support vl as an immediate, we
    need to convert 0 to zero when outputting asm.

            PR target/116592

    gcc/ChangeLog:

            * config/riscv/thead.cc (th_asm_output_opcode): Change '0' to
            "zero"

    gcc/testsuite/ChangeLog:

            * gcc.target/riscv/rvv/xtheadvector/pr116592.c: New test.

[Bug target/116592] illegal operands th.vsetvli zero,0,e32,m8 with -O2 -O3 when compiling for risc-v xtheadvector

[law at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116592

Jeffrey A. Law <law at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
                 CC|                            |law at gcc dot gnu.org
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #7 from Jeffrey A. Law <law at gcc dot gnu.org> ---
Fixed on the trunk.