diff mbox series

[1/2] PR target/113652, Fix failed bootstrap on ppc unrecognized opcode: `lfiwzx' with -mcpu=7450

Message ID Zpip3CReP0nxz1t8@cowardly-lion.the-meissners.org
State New
Headers show
Series PR target/113652, Fix failed bootstrap on ppc unrecognized opcode: `lfiwzx' with -mcpu=7450 | expand

Commit Message

Michael Meissner July 18, 2024, 5:36 a.m. UTC 
Currently, we add -mvsx when building the float128 support in libgcc.  This
allows us to build the float128 support on a big endian system where the
default cpu is power4.  While the libgcc support can be built, given there is
no glibc support for float128 available.

However, adding -mvsx and building the libgcc float128 support causes problems
if you set the default cpu to something like a 7540, which does not have VSX
support.  The assembler complains that when the code does a ".machine 7450", you
cannot use VSX instructions.

With these patches, the float128 libgcc support is only built if the default
compiler has VSX support.  If somebody wanted to enable the glibc support for
big endian, they would need to set the base cpu to power8 to enable building the
libgcc float128 libraries.

The second patch in this series fixes up the GCC testsuite so that float128
tests are only run if the compiler enables float128 by default.

I built little endian compilers and there were no regressions.

I built big endian compilers with the --with-cpu=power5 configure option, and I
verified that none of the float128 support functions are built.

I also built big endian compilers on a power9 with the --with-cpu=native
configure option, and I verified that the float128 support functions were
built, since the default compiler used the VSX instruction set.

I verified that on both sets of big endian builds, that all of the float128
tests were skipped, since there is no support for float128 in glibc and the GCC
compiler does not enable float128 on those systems.

Can I check these patches into the trunk assuming the original bugzilla author
says they fix the problem?

2024-07-18 Michael Meissner  <meissner@linux.ibm.com>

libgcc/

	PR target/113652
	* config.host (powerpc*-*-linux*): Do not add t-float128-hw or
	t-float128-p10-hw if the default compiler does not support float128.
	* config/rs6000/t-float128 (FP128_CFLAGS_SW): Do not add -mvsx when
	building the basic float128 support.
	* config/rs6000/t-float128-hw (FP128_CFLAGS_HW): Likewise.
	* config/rs6000/t-float128-p10-hw (FP128_3_1_CFLAGS_HW): Likewise.
	* configure.ac (powerpc*-*-linux*): Do not add -mvsx when testing
	whether to build the float128 support.
	* configure: Regenerate.
---
 libgcc/config.host                     | 12 ++++++------
 libgcc/config/rs6000/t-float128        |  8 +++++++-
 libgcc/config/rs6000/t-float128-hw     |  3 +--
 libgcc/config/rs6000/t-float128-p10-hw |  3 +--
 libgcc/configure                       |  8 +++++++-
 libgcc/configure.ac                    |  8 +++++++-
 6 files changed, 29 insertions(+), 13 deletions(-)
diff mbox series

Patch

diff --git a/libgcc/config.host b/libgcc/config.host
index 9fae51d4ce7..261b08859a4 100644
--- a/libgcc/config.host
+++ b/libgcc/config.host
@@ -1292,14 +1292,14 @@  powerpc*-*-linux*)
 
 	if test $libgcc_cv_powerpc_float128 = yes; then
 		tmake_file="${tmake_file} rs6000/t-float128"
-	fi
 
-	if test $libgcc_cv_powerpc_float128_hw = yes; then
-		tmake_file="${tmake_file} rs6000/t-float128-hw"
-	fi
+		if test $libgcc_cv_powerpc_float128_hw = yes; then
+			tmake_file="${tmake_file} rs6000/t-float128-hw"
 
-	if test $libgcc_cv_powerpc_3_1_float128_hw = yes; then
-		tmake_file="${tmake_file} rs6000/t-float128-p10-hw"
+			if test $libgcc_cv_powerpc_3_1_float128_hw = yes; then
+				tmake_file="${tmake_file} rs6000/t-float128-p10-hw"
+			fi
+		fi
 	fi
 
 	extra_parts="$extra_parts ecrti.o ecrtn.o ncrti.o ncrtn.o"
diff --git a/libgcc/config/rs6000/t-float128 b/libgcc/config/rs6000/t-float128
index b09b5664af0..93e78adcd62 100644
--- a/libgcc/config/rs6000/t-float128
+++ b/libgcc/config/rs6000/t-float128
@@ -74,7 +74,13 @@  fp128_includes		= $(srcdir)/soft-fp/double.h \
 			  $(srcdir)/soft-fp/soft-fp.h
 
 # Build the emulator without ISA 3.0 hardware support.
-FP128_CFLAGS_SW		 = -Wno-type-limits -mvsx -mfloat128 \
+#
+# In the past we added -mvsx to build the float128 specific libraries with the
+# VSX instruction set.  This allowed the big endian GCC on server platforms to
+# build the float128 support.  However, is causes problems when other default
+# cpu targets are used such as the 7450.
+
+FP128_CFLAGS_SW		 = -Wno-type-limits -mfloat128 \
 			   -mno-float128-hardware -mno-gnu-attribute \
 			   -I$(srcdir)/soft-fp \
 			   -I$(srcdir)/config/rs6000 \
diff --git a/libgcc/config/rs6000/t-float128-hw b/libgcc/config/rs6000/t-float128-hw
index ed67b572580..82726c98b98 100644
--- a/libgcc/config/rs6000/t-float128-hw
+++ b/libgcc/config/rs6000/t-float128-hw
@@ -23,8 +23,7 @@  fp128_ifunc_obj		= $(fp128_ifunc_static_obj) $(fp128_ifunc_shared_obj)
 fp128_sed_hw		= -hw
 
 # Build the hardware support functions with appropriate hardware support
-FP128_CFLAGS_HW		 = -Wno-type-limits -mvsx -mfloat128 \
-			   -mcpu=power9 \
+FP128_CFLAGS_HW		 = -Wno-type-limits -mfloat128 -mcpu=power9 \
 			   -mfloat128-hardware -mno-gnu-attribute \
 			   -I$(srcdir)/soft-fp \
 			   -I$(srcdir)/config/rs6000 \
diff --git a/libgcc/config/rs6000/t-float128-p10-hw b/libgcc/config/rs6000/t-float128-p10-hw
index edaaee0e478..ee50d248ca1 100644
--- a/libgcc/config/rs6000/t-float128-p10-hw
+++ b/libgcc/config/rs6000/t-float128-p10-hw
@@ -13,8 +13,7 @@  fp128_3_1_hw_shared_obj	= $(addsuffix _s$(objext),$(fp128_3_1_hw_funcs))
 fp128_3_1_hw_obj	= $(fp128_3_1_hw_static_obj) $(fp128_3_1_hw_shared_obj)
 
 # Build the hardware support functions with appropriate hardware support
-FP128_3_1_CFLAGS_HW	 = -Wno-type-limits -mvsx -mfloat128 \
-			   -mcpu=power10 \
+FP128_3_1_CFLAGS_HW	 = -Wno-type-limits -mfloat128 -mcpu=power10 \
 			   -mfloat128-hardware -mno-gnu-attribute \
 			   -I$(srcdir)/soft-fp \
 			   -I$(srcdir)/config/rs6000 \
diff --git a/libgcc/configure b/libgcc/configure
index a69d314374a..635237a06c8 100755
--- a/libgcc/configure
+++ b/libgcc/configure
@@ -5184,9 +5184,15 @@  case ${host} in
 # check if we have VSX (ISA 2.06) support to build the software libraries, and
 # whether the assembler can handle xsaddqp for hardware support.  Also check if
 # a new glibc is being used so that __builtin_cpu_supports can be used.
+#
+# with the VSX instruction set.  This allowed the big endian GCC on server
+# platforms to build the float128 support.  However, is causes problems when
+# other default cpu targets are used such as the 7450.  Now
+# libgcc_cv_powerpc_float128 will fail if the default cpu cannot build the
+# float128 support.
 powerpc*-*-linux*)
   saved_CFLAGS="$CFLAGS"
-  CFLAGS="$CFLAGS -mabi=altivec -mvsx -mfloat128"
+  CFLAGS="$CFLAGS -mfloat128"
   { $as_echo "$as_me:${as_lineno-$LINENO}: checking for PowerPC ISA 2.06 to build __float128 libraries" >&5
 $as_echo_n "checking for PowerPC ISA 2.06 to build __float128 libraries... " >&6; }
 if ${libgcc_cv_powerpc_float128+:} false; then :
diff --git a/libgcc/configure.ac b/libgcc/configure.ac
index c2749fe0958..2a725a6f662 100644
--- a/libgcc/configure.ac
+++ b/libgcc/configure.ac
@@ -407,9 +407,15 @@  case ${host} in
 # check if we have VSX (ISA 2.06) support to build the software libraries, and
 # whether the assembler can handle xsaddqp for hardware support.  Also check if
 # a new glibc is being used so that __builtin_cpu_supports can be used.
+#
+# with the VSX instruction set.  This allowed the big endian GCC on server
+# platforms to build the float128 support.  However, is causes problems when
+# other default cpu targets are used such as the 7450.  Now
+# libgcc_cv_powerpc_float128 will fail if the default cpu cannot build the
+# float128 support.
 powerpc*-*-linux*)
   saved_CFLAGS="$CFLAGS"
-  CFLAGS="$CFLAGS -mabi=altivec -mvsx -mfloat128"
+  CFLAGS="$CFLAGS -mfloat128"
   AC_CACHE_CHECK([for PowerPC ISA 2.06 to build __float128 libraries],
 		 [libgcc_cv_powerpc_float128],
 		 [AC_COMPILE_IFELSE(