S390 Buildroot/gcc build issue

Alexander,

I am contacting you regarding a build issue I'm seeing with the S390 architecture in Buildroot.

With the following defconfig:

BR2_s390x=y
BR2_KERNEL_HEADERS_6_12=y
BR2_PACKAGE_GLIBC_KERNEL_COMPAT=y
BR2_BINUTILS_VERSION_2_46_X=y
BR2_BINUTILS_GPROFNG=y
BR2_GCC_VERSION_16_X=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_TOOLCHAIN_BUILDROOT_FORTRAN=y
BR2_GCC_ENABLE_OPENMP=y
BR2_PACKAGE_HOST_GDB=y
BR2_PACKAGE_HOST_GDB_TUI=y
BR2_PACKAGE_HOST_GDB_PYTHON3=y
BR2_GDB_VERSION_17=y
BR2_GNU_MIRROR="https://mirrors.kernel.org/gnu/"
BR2_PACKAGE_GDB=y

gcc 16.2.0 fails to build with:

../../../libgcc/config/s390/_dpd_sd_to_hf.c:27:25: error: decimal floating-point not supported for this target

   27 | HFtype __dpd_truncsdhf (_Decimal32);
      |                         ^~~~~~~~~~
../../../libgcc/config/s390/_dpd_sd_to_hf.c:31:16: error: decimal floating-point not supported for this target
   31 | force_convert (_Decimal32 x)
      |                ^~~~~~~~~~
../../../libgcc/config/s390/_dpd_hf_to_td.c:34:1: error: decimal floating-point not supported for this target
   34 | _Decimal128
      | ^~~~~~~~~~~
../../../libgcc/config/s390/_dpd_sd_to_hf.c:35:18: error: decimal floating-point not supported for this target
   35 | __dpd_truncsdhf (_Decimal32 x)
      |                  ^~~~~~~~~~

I haven't checked whether this is new with gcc 16.x or not, and I'm not familiar with floating point support on S390.

Are you familiar with this issue?

I see the same issue with gcc 16.1.0 in the autobuilders:

https://autobuild.buildroot.net/results/f88/f88565af0f653df831a7fa43d2507d167fdb9d36/build-end.log

But I don't seem to see any issue with gcc 15.x, so my guess is that this is a new issue with gcc 16.x.

the types _Decimal* require for the host compiler to be built with --enable-decimal-float.

package/gcc/gcc.mk:
-------------------

HOST_GCC_COMMON_CONF_OPTS = \
	--target=$(GNU_TARGET_NAME) \
	--with-sysroot=$(STAGING_DIR) \
	--enable-__cxa_atexit \
	--with-gnu-ld \
	--disable-libssp \
	--disable-multilib \
	--disable-decimal-float                 <----- !!!

Can we turn --disable-decimal-float into --enable-decimal-float on s390x arch ? It fixed the problem for me.

package/gcc/gcc.mk:
-------------------

ifeq ($(BR2_s390x),y)

HOST_GCC_COMMON_CONF_OPTS += \
	--with-long-double-128 \
	--enable-decimal-float                 <----- !!!

Hello,

On Wed, Aug 12, 2026 at 11:11:10AM +0200, Alexander Egorenkov wrote:

Quoted 2 lines

the types _Decimal* require for the host compiler to be built with --enable-decimal-float.

What do you call the "host compiler" in this context?

Are you talking about the native gcc compiler available on the host, or the cross-compiler built by Buildroot?

and for S390, it gets overridden.

See: it's the same variable that gets appended, so at the end we have:

HOST_GCC_COMMON_CONF_OPTS = \
	[...]
	--disable-decimal-float \
	[...]
	--enable-decimal-float

So on S390, the cross-compiler is built with --enable-decimal-float.

Hello,

Quoted 2 lines · Alexander Egorenkov

the types _Decimal* require for the host compiler to be built with --enable-decimal-float.

What do you call the "host compiler" in this context?

Are you talking about the native gcc compiler available on the host, or the cross-compiler built by Buildroot?

Quoted 4 lines

package/gcc/gcc.mk:
-------------------

Quoted 8 lines

HOST_GCC_COMMON_CONF_OPTS = \
	--target=$(GNU_TARGET_NAME) \
	--with-sysroot=$(STAGING_DIR) \
	--enable-__cxa_atexit \
	--with-gnu-ld \
	--disable-libssp \
	--disable-multilib \
	--disable-decimal-float                 <----- !!!

These are the common options...

Quoted 10 lines

Can we turn --disable-decimal-float into --enable-decimal-float on s390x arch ? It fixed the problem for me.

package/gcc/gcc.mk:
-------------------

ifeq ($(BR2_s390x),y)

HOST_GCC_COMMON_CONF_OPTS += \
	--with-long-double-128 \
	--enable-decimal-float                 <----- !!!

and for S390, it gets overridden.

See: it's the same variable that gets appended, so at the end we have:

HOST_GCC_COMMON_CONF_OPTS = \
	[...]
	--disable-decimal-float \
	[...]
	--enable-decimal-float

So on S390, the cross-compiler is built with --enable-decimal-float.

Best regards,

Thomas

Signature · 3 lines

Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com