Skip to contentSkip to conversation

gcc-runtime: Hacks for libgfortran with gcc-4.8

RFC PATCH In review Richard Purdie 2 participants 8 messages Sep 5, 2013 – Sep 6, 2013

lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 gcc-runtime: Hacks for libgfortran with gcc-4.8

These are the hacks I needed to make libgfortran build. This is ugly, no argument from me. We could probably get better results if we patch configure and libtool to stop doing nasty things. I've probably taken this as far as I'd want to though, not being a particular fan of fortran...

Khem: Any thoughts on this?

---

meta/recipes-devtools/gcc/gcc-runtime.inc +18 -0
index 2599760..395623f 100644
@@ -18,6 +18,9 @@ RUNTIMETARGET = "libssp libstdc++-v3 libgomp"
 # libmudflap
 # libgfortran
 
+DEPENDS_append = " chrpath-replacement-native"
+EXTRANATIVEPATH += "chrpath-native"
+
 do_configure () {
 	export CXX="${CXX} -nostdinc++ -nostdlib++"
 	mtarget=`echo ${MULTIMACH_TARGET_SYS} | sed -e s#-${SDKPKGSUFFIX}##`
@@ -30,6 +33,11 @@ do_configure () {
 		cd ${B}/$target/$d/
 		chmod a+x ${S}/$d/configure
 		${S}/$d/configure ${CONFIGUREOPTS} ${EXTRA_OECONF}
+		# Ugly hack, libgfortran configure looks for ../libquadmath/libquadmath.la 
+		# so we need to compile it before configure
+		if [ "$d" = "libquadmath" ]; then
+			oe_runmake MULTIBUILDTOP=${B}/$target/$d/
+		fi
 	done
 }
 
@@ -38,6 +46,16 @@ do_compile () {
 	for d in libgcc ${RUNTIMETARGET}; do
 		cd ${B}/$target/$d/
 		oe_runmake MULTIBUILDTOP=${B}/$target/$d/
+		if [ "$d" = "libgfortran" ]; then
+			# libtool needs libdir to match the final installation directory which configure
+			# sets from output from this command (e.g. both set to /usr/lib/../lib
+			# It also adds bogus RPATHS which we have to delete
+			fulllibdir=`$CC -print-multi-os-directory`
+			if [ $fulllibdir != "." ]; then
+				sed -i -e "s#relink_command=.*#relink_command=#" ${B}/$target/$d/libgfortran.la
+				chrpath -d `readlink -f ${B}/$target/$d/.libs/libgfortran.so`
+			fi
+		fi
 	done
 }
 
#1 lore raw
Quoted 25 lines · Richard Purdie

These are the hacks I needed to make libgfortran build. This is ugly, no argument from me. We could probably get better results if we patch configure and libtool to stop doing nasty things. I've probably taken this as far as I'd want to though, not being a particular fan of fortran...

Khem: Any thoughts on this?

---

meta/recipes-devtools/gcc/gcc-runtime.inc +3 -0
index 2599760..395623f 100644
@@ -18,6 +18,9 @@ RUNTIMETARGET = "libssp libstdc++-v3 libgomp"
 # libmudflap
 # libgfortran
 
+DEPENDS_append = " chrpath-replacement-native"
+EXTRANATIVEPATH += "chrpath-native"
+
 do_configure () {
 	export CXX="${CXX} -nostdinc++ -nostdlib++"
 	mtarget=`echo ${MULTIMACH_TARGET_SYS} | sed -e s#-${SDKPKGSUFFIX}##`

meta/recipes-devtools/gcc/gcc-runtime.inc do_configure () { · line 36

Quoted 4 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +1 -0
@@ -30,6 +33,11 @@ do_configure () {
 		cd ${B}/$target/$d/
 		chmod a+x ${S}/$d/configure
 		${S}/$d/configure ${CONFIGUREOPTS} ${EXTRA_OECONF}
+		# Ugly hack, libgfortran configure looks for ../libquadmath/libquadmath.la

Maybe we should explicitly --enable-libquadmath in gcc-cross when fortran is asked for in RUNTIMETARGETS might avoid some of below.

Quoted 12 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +4 -0
+		# so we need to compile it before configure
+		if [ "$d" = "libquadmath" ]; then
+			oe_runmake MULTIBUILDTOP=${B}/$target/$d/
+		fi
 	done
 }
 
@@ -38,6 +46,16 @@ do_compile () {
 	for d in libgcc ${RUNTIMETARGET}; do
 		cd ${B}/$target/$d/
 		oe_runmake MULTIBUILDTOP=${B}/$target/$d/

meta/recipes-devtools/gcc/gcc-runtime.inc do_compile () { · line 58

Quoted 10 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +10 -0
+		if [ "$d" = "libgfortran" ]; then
+			# libtool needs libdir to match the final installation directory which configure
+			# sets from output from this command (e.g. both set to /usr/lib/../lib
+			# It also adds bogus RPATHS which we have to delete
+			fulllibdir=`$CC -print-multi-os-directory`
+			if [ $fulllibdir != "." ]; then
+				sed -i -e "s#relink_command=.*#relink_command=#" ${B}/$target/$d/libgfortran.la
+				chrpath -d `readlink -f ${B}/$target/$d/.libs/libgfortran.so`
+			fi
+		fi

hmm remind me but I think we use unmodified libtool that comes with gcc IIRC. if we used libtool-cross then this could be fixed there

Quoted 6 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +0 -0
 	done
 }
 
#2 lore raw
Quoted 20 lines · Khem Raj
Quoted 18 lines · Richard Purdie

These are the hacks I needed to make libgfortran build. This is ugly, no argument from me. We could probably get better results if we patch configure and libtool to stop doing nasty things. I've probably taken this as far as I'd want to though, not being a particular fan of fortran...

Khem: Any thoughts on this?

---

meta/recipes-devtools/gcc/gcc-runtime.inc +0 -0
index 2599760..395623f 100644
@@ -18,6 +18,9 @@ RUNTIMETARGET = "libssp libstdc++-v3 libgomp"
 # libmudflap
 # libgfortran
 
Quoted 14 lines

meta/recipes-devtools/gcc/gcc-runtime.inc do_configure () { · line 36

Quoted 11 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +4 -0
+DEPENDS_append = " chrpath-replacement-native"
+EXTRANATIVEPATH += "chrpath-native"
+
 do_configure () {
 	export CXX="${CXX} -nostdinc++ -nostdlib++"
 	mtarget=`echo ${MULTIMACH_TARGET_SYS} | sed -e s#-${SDKPKGSUFFIX}##`
@@ -30,6 +33,11 @@ do_configure () {
 		cd ${B}/$target/$d/
 		chmod a+x ${S}/$d/configure
 		${S}/$d/configure ${CONFIGUREOPTS} ${EXTRA_OECONF}
+		# Ugly hack, libgfortran configure looks for ../libquadmath/libquadmath.la

Maybe we should explicitly --enable-libquadmath in gcc-cross when fortran is asked for in RUNTIMETARGETS might avoid some of below.

That would mean the gcc-cross recipe has to package it. We've basically now agreed and changed the code so all the packaging doesn't happen in -cross packages since it was always problematic.

FWIW I also tried disabling quadmath but that caused different build failures.

Quoted 17 lines
Quoted 17 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +9 -0
+		# so we need to compile it before configure
+		if [ "$d" = "libquadmath" ]; then
+			oe_runmake MULTIBUILDTOP=${B}/$target/$d/
+		fi
 	done
 }
 
@@ -38,6 +46,16 @@ do_compile () {
 	for d in libgcc ${RUNTIMETARGET}; do
 		cd ${B}/$target/$d/
 		oe_runmake MULTIBUILDTOP=${B}/$target/$d/
+		if [ "$d" = "libgfortran" ]; then
+			# libtool needs libdir to match the final installation directory which configure
+			# sets from output from this command (e.g. both set to /usr/lib/../lib
+			# It also adds bogus RPATHS which we have to delete
+			fulllibdir=`$CC -print-multi-os-directory`
Quoted 8 lines

meta/recipes-devtools/gcc/gcc-runtime.inc do_compile () { · line 58

Quoted 5 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +5 -0
+			if [ $fulllibdir != "." ]; then
+				sed -i -e "s#relink_command=.*#relink_command=#" ${B}/$target/$d/libgfortran.la
+				chrpath -d `readlink -f ${B}/$target/$d/.libs/libgfortran.so`
+			fi
+		fi

hmm remind me but I think we use unmodified libtool that comes with gcc IIRC. if we used libtool-cross then this could be fixed there

We do use libtool from gcc, yes. To do otherwise safely, you have to reautoconf gcc which "is painful". Mixing two libtool versions usually gives stacks of errors so our only option would be to patch libtool in gcc to apply our patches. Even then, I'm not sure they'd cope with the way quadmath is linked :/.

This is why I ended up hacking things in the metadata since it seemed the least worse solution...

Cheers,

Richard

#3 lore raw
Quoted 36 lines · Richard Purdie
Quoted 20 lines · Khem Raj
Quoted 18 lines · Richard Purdie

These are the hacks I needed to make libgfortran build. This is ugly, no argument from me. We could probably get better results if we patch configure and libtool to stop doing nasty things. I've probably taken this as far as I'd want to though, not being a particular fan of fortran...

Khem: Any thoughts on this?

---

meta/recipes-devtools/gcc/gcc-runtime.inc +0 -0
index 2599760..395623f 100644
@@ -18,6 +18,9 @@ RUNTIMETARGET = "libssp libstdc++-v3 libgomp"
 # libmudflap
 # libgfortran
 
Quoted 14 lines

meta/recipes-devtools/gcc/gcc-runtime.inc do_configure () { · line 36

Quoted 11 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +4 -0
+DEPENDS_append = " chrpath-replacement-native"
+EXTRANATIVEPATH += "chrpath-native"
+
 do_configure () {
 	export CXX="${CXX} -nostdinc++ -nostdlib++"
 	mtarget=`echo ${MULTIMACH_TARGET_SYS} | sed -e s#-${SDKPKGSUFFIX}##`
@@ -30,6 +33,11 @@ do_configure () {
 		cd ${B}/$target/$d/
 		chmod a+x ${S}/$d/configure
 		${S}/$d/configure ${CONFIGUREOPTS} ${EXTRA_OECONF}
+		# Ugly hack, libgfortran configure looks for ../libquadmath/libquadmath.la

Maybe we should explicitly --enable-libquadmath in gcc-cross when fortran is asked for in RUNTIMETARGETS might avoid some of below.

Quoted 6 lines

That would mean the gcc-cross recipe has to package it. We've basically now agreed and changed the code so all the packaging doesn't happen in -cross packages since it was always problematic.

FWIW I also tried disabling quadmath but that caused different build failures.

But we stash the build artifacts from gcc-cross that then we reuse to build gcc-runtime so I am hoping that it will do the configuration bits right probably and we dont have to do libtool surgery.

Quoted 27 lines
Quoted 18 lines
Quoted 18 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +9 -0
+		# so we need to compile it before configure
+		if [ "$d" = "libquadmath" ]; then
+			oe_runmake MULTIBUILDTOP=${B}/$target/$d/
+		fi
 	done
 }
 
@@ -38,6 +46,16 @@ do_compile () {
 	for d in libgcc ${RUNTIMETARGET}; do
 		cd ${B}/$target/$d/
 		oe_runmake MULTIBUILDTOP=${B}/$target/$d/
+		if [ "$d" = "libgfortran" ]; then
+			# libtool needs libdir to match the final installation directory which configure
+			# sets from output from this command (e.g. both set to /usr/lib/../lib
+			# It also adds bogus RPATHS which we have to delete
+			fulllibdir=`$CC -print-multi-os-directory`
Quoted 8 lines

meta/recipes-devtools/gcc/gcc-runtime.inc do_compile () { · line 58

Quoted 5 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +5 -0
+			if [ $fulllibdir != "." ]; then
+				sed -i -e "s#relink_command=.*#relink_command=#" ${B}/$target/$d/libgfortran.la
+				chrpath -d `readlink -f ${B}/$target/$d/.libs/libgfortran.so`
+			fi
+		fi

hmm remind me but I think we use unmodified libtool that comes with gcc IIRC. if we used libtool-cross then this could be fixed there

Quoted 8 lines

We do use libtool from gcc, yes. To do otherwise safely, you have to reautoconf gcc which "is painful". Mixing two libtool versions usually gives stacks of errors so our only option would be to patch libtool in gcc to apply our patches. Even then, I'm not sure they'd cope with the way quadmath is linked :/.

This is why I ended up hacking things in the metadata since it seemed the least worse solution…

yeah. I dont think we have much choice there. re-autoconfiguring gcc with different libtool is a project in itself. GCC does not recommend to reconfigure so thats fine.

Quoted 5 lines

Cheers,

Richard

#4 lore raw
Quoted 42 lines · Khem Raj
Quoted 40 lines · Richard Purdie
Quoted 20 lines · Khem Raj
Quoted 18 lines · Richard Purdie

These are the hacks I needed to make libgfortran build. This is ugly, no argument from me. We could probably get better results if we patch configure and libtool to stop doing nasty things. I've probably taken this as far as I'd want to though, not being a particular fan of fortran...

Khem: Any thoughts on this?

---

meta/recipes-devtools/gcc/gcc-runtime.inc +0 -0
index 2599760..395623f 100644
@@ -18,6 +18,9 @@ RUNTIMETARGET = "libssp libstdc++-v3 libgomp"
 # libmudflap
 # libgfortran
 
Quoted 14 lines

meta/recipes-devtools/gcc/gcc-runtime.inc do_configure () { · line 36

Quoted 11 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +4 -0
+DEPENDS_append = " chrpath-replacement-native"
+EXTRANATIVEPATH += "chrpath-native"
+
 do_configure () {
 	export CXX="${CXX} -nostdinc++ -nostdlib++"
 	mtarget=`echo ${MULTIMACH_TARGET_SYS} | sed -e s#-${SDKPKGSUFFIX}##`
@@ -30,6 +33,11 @@ do_configure () {
 		cd ${B}/$target/$d/
 		chmod a+x ${S}/$d/configure
 		${S}/$d/configure ${CONFIGUREOPTS} ${EXTRA_OECONF}
+		# Ugly hack, libgfortran configure looks for ../libquadmath/libquadmath.la

Maybe we should explicitly --enable-libquadmath in gcc-cross when fortran is asked for in RUNTIMETARGETS might avoid some of below.

That would mean the gcc-cross recipe has to package it. We've basically now agreed and changed the code so all the packaging doesn't happen in -cross packages since it was always problematic.

Quoted 5 lines
Quoted 2 lines

FWIW I also tried disabling quadmath but that caused different build failures.

But we stash the build artifacts from gcc-cross that then we reuse to build gcc-runtime so I am hoping that it will do the configuration bits right probably and we dont have to do libtool surgery.

It won't help, we'd have to install the libs/headers from quadmath into the target sysroot and we can't reasonably do that from the cross recipe.

Cheers,

Richard

#5 lore raw
Quoted 46 lines · Richard Purdie
Quoted 45 lines · Khem Raj
Quoted 42 lines · Richard Purdie
Quoted 20 lines · Khem Raj
Quoted 18 lines · Richard Purdie

These are the hacks I needed to make libgfortran build. This is ugly, no argument from me. We could probably get better results if we patch configure and libtool to stop doing nasty things. I've probably taken this as far as I'd want to though, not being a particular fan of fortran...

Khem: Any thoughts on this?

---

meta/recipes-devtools/gcc/gcc-runtime.inc +0 -0
index 2599760..395623f 100644
@@ -18,6 +18,9 @@ RUNTIMETARGET = "libssp libstdc++-v3 libgomp"
 # libmudflap
 # libgfortran
 
Quoted 14 lines

meta/recipes-devtools/gcc/gcc-runtime.inc do_configure () { · line 36

Quoted 11 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +4 -0
+DEPENDS_append = " chrpath-replacement-native"
+EXTRANATIVEPATH += "chrpath-native"
+
 do_configure () {
 	export CXX="${CXX} -nostdinc++ -nostdlib++"
 	mtarget=`echo ${MULTIMACH_TARGET_SYS} | sed -e s#-${SDKPKGSUFFIX}##`
@@ -30,6 +33,11 @@ do_configure () {
 		cd ${B}/$target/$d/
 		chmod a+x ${S}/$d/configure
 		${S}/$d/configure ${CONFIGUREOPTS} ${EXTRA_OECONF}
+		# Ugly hack, libgfortran configure looks for ../libquadmath/libquadmath.la

Maybe we should explicitly --enable-libquadmath in gcc-cross when fortran is asked for in RUNTIMETARGETS might avoid some of below.

That would mean the gcc-cross recipe has to package it. We've basically now agreed and changed the code so all the packaging doesn't happen in -cross packages since it was always problematic.

FWIW I also tried disabling quadmath but that caused different build failures.

Quoted 6 lines
Quoted 2 lines

But we stash the build artifacts from gcc-cross that then we reuse to build gcc-runtime so I am hoping that it will do the configuration bits right probably and we dont have to do libtool surgery.

It won't help, we'd have to install the libs/headers from quadmath into the target sysroot and we can't reasonably do that from the cross recipe.

I was thinking just to have configured tree like we have for libgcc and others. full build of libquadmath and libgfortran of course happens with gcc-runtime phase.

Quoted 5 lines

Cheers,

Richard

#6 lore raw
Quoted 51 lines · Khem Raj
Quoted 49 lines · Richard Purdie
Quoted 47 lines · Khem Raj
Quoted 36 lines · Richard Purdie
Quoted 20 lines · Khem Raj
Quoted 18 lines · Richard Purdie

These are the hacks I needed to make libgfortran build. This is ugly, no argument from me. We could probably get better results if we patch configure and libtool to stop doing nasty things. I've probably taken this as far as I'd want to though, not being a particular fan of fortran...

Khem: Any thoughts on this?

---

meta/recipes-devtools/gcc/gcc-runtime.inc +0 -0
index 2599760..395623f 100644
@@ -18,6 +18,9 @@ RUNTIMETARGET = "libssp libstdc++-v3 libgomp"
 # libmudflap
 # libgfortran
 
Quoted 14 lines

meta/recipes-devtools/gcc/gcc-runtime.inc do_configure () { · line 36

Quoted 11 lines
meta/recipes-devtools/gcc/gcc-runtime.inc +4 -0
+DEPENDS_append = " chrpath-replacement-native"
+EXTRANATIVEPATH += "chrpath-native"
+
 do_configure () {
 	export CXX="${CXX} -nostdinc++ -nostdlib++"
 	mtarget=`echo ${MULTIMACH_TARGET_SYS} | sed -e s#-${SDKPKGSUFFIX}##`
@@ -30,6 +33,11 @@ do_configure () {
 		cd ${B}/$target/$d/
 		chmod a+x ${S}/$d/configure
 		${S}/$d/configure ${CONFIGUREOPTS} ${EXTRA_OECONF}
+		# Ugly hack, libgfortran configure looks for ../libquadmath/libquadmath.la

Maybe we should explicitly --enable-libquadmath in gcc-cross when fortran is asked for in RUNTIMETARGETS might avoid some of below.

Quoted 6 lines

That would mean the gcc-cross recipe has to package it. We've basically now agreed and changed the code so all the packaging doesn't happen in -cross packages since it was always problematic.

FWIW I also tried disabling quadmath but that caused different build failures.

But we stash the build artifacts from gcc-cross that then we reuse to build gcc-runtime so I am hoping that it will do the configuration bits right probably and we dont have to do libtool surgery.

Quoted 7 lines
Quoted 3 lines

It won't help, we'd have to install the libs/headers from quadmath into the target sysroot and we can't reasonably do that from the cross recipe.

I was thinking just to have configured tree like we have for libgcc and others. full build of libquadmath and libgfortran
of course happens with gcc-runtime phase.

I can't see how that will help solve this problem though?

Cheers,

Richard

#7 lore raw
Quoted 1 line · Richard Purdie

I can't see how that will help solve this problem though?

might help n getting the libtool hackery right.