[Buildroot] [Bug 10091] New: gcc7.1 does not build with graphite support due to old isl

[Buildroot] [Bug 10091] New: gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

            Bug ID: 10091
           Summary: gcc7.1 does not build with graphite support due to old
                    isl
           Product: buildroot
           Version: unspecified
          Hardware: All
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P5
         Component: Other
          Assignee: unassigned at buildroot.uclibc.org
          Reporter: ivan.nazarenko at gmail.com
                CC: buildroot at uclibc.org
  Target Milestone: ---

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

Ivan <ivan.nazarenko@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |minor

--- Comment #1 from Ivan <ivan.nazarenko@gmail.com> ---
graphite support in gcc7.1 demands isl 0.15 or newer.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #2 from Thomas Petazzoni <thomas.petazzoni@free-electrons.com> ---
Thanks for your bug report!

Do you know which versions of isl are acceptable for gcc 4.9, 5.x, 6.x and 7.x
? Indeed, Buildroot support all those versions of gcc.

In addition, isl was originally not updated to 0.15 for the following reasons
(from package/isl/isl.mk):

# The latest 0.15 version is not yet compatible with cloog 0.18.3, so
# bumping isl is not possible until a new version of cloog is
# published.
ISL_VERSION = 0.14.1

So we would also need to look at which version of cloog is acceptable for the
different versions of gcc/isl. Is this something you can help us with?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #3 from Andrey Yurovsky <yurovsky@gmail.com> ---
I just tried bumping ISL and was able to build GCC7 with graphite (for ARMv5 in
this case), I'm trying to better understand the dependencies here but perhaps
ISL 0.18 doesn't need cloog or can work with the older version.

diff --git a/package/isl/isl.hash b/package/isl/isl.hash
index e2c14ad14..22941de42 100644
--- a/package/isl/isl.hash
+++ b/package/isl/isl.hash
@@ -1,2 +1,4 @@
 # Locally computed
 sha256 8882c9e36549fc757efa267706a9af733bb8d7fe3905cbfde43e17a89eea4675       
isl-0.14.1.tar.xz
+sha256 0f35051cc030b87c673ac1f187de40e386a1482a0cfdf2c552dd6031b307ddc4 
isl-0.18.tar.xz
+
diff --git a/package/isl/isl.mk b/package/isl/isl.mk
index 74de125b6..85d013723 100644
--- a/package/isl/isl.mk
+++ b/package/isl/isl.mk
@@ -7,7 +7,7 @@
 # The latest 0.15 version is not yet compatible with cloog 0.18.3, so
 # bumping isl is not possible until a new version of cloog is
 # published.
-ISL_VERSION = 0.14.1
+ISL_VERSION = 0.18
 ISL_SOURCE = isl-$(ISL_VERSION).tar.xz
 ISL_SITE = http://isl.gforge.inria.fr
 ISL_LICENSE = MIT

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #4 from Ivan <ivan.nazarenko@gmail.com> ---
Build of gcc 7.1, with graphite support enabled, failed in the "checking isl >=
0.15" pass.

Disabling graphite allowed gcc build.

See https://gcc.gnu.org/install/prerequisites.html

While 6.x, 5.x and 4.9 still builds with isl 0.14, the gcc webpage specify isl
0.15 as the current minimum.

Verifying the ChangeLog-2016, 2015, etc files shows that in 2014 the dependency
of cloog was removed. This is true for the last gcc5.x, 6.x and 7.x.

But gcc4.x still depends on cloog.

So the options seems to be either:

- keep cloog, with corresponding old isl, and gcc4.x users have graphite, or
- removes cloog, new isl, and gcc7.1 users have graphite.

I am not aware if isl 0.15 would disrupt graphite in 5,x or 6.x. Improbable,
though, since 0.15 is from 2015 and the last gcc5.x from 2016.

I feel myself bit overwhelmed by builroot, and do not think I would be able to
experiment with anything beyond the trivial "make menuconfig". I do not event
know bash. But I would help where I can.





While I am completely bewildered by all this buildroot stuff (I am not even
familiar with bash), I though it would be nice to report failed builds. But do
not feel

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #5 from Ivan <ivan.nazarenko@gmail.com> ---
Anyway it seems that gcc7.1 is still a bit buggy. I could not finish the build
of qt5base with it - and the issue was not that "crc" of the other bug report.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #6 from Peter Seiderer <ps.report@gmx.net> ---
Mind to share your .config file and the (last several lines showing the error)
of your build log?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #7 from Ivan <ivan.nazarenko@gmail.com> ---
Here are they: The error:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/bin/arm-buildroot-linux-musleabihf-g++
-c -pipe -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Os
--sysroot=/home/ivan/buildroot-git/buildroot/buildroot/output/host/arm-buildroot-linux-musleabihf/sysroot
-O3 -O3 -std=c++1z -fvisibility=hidden -fvisibility-inlines-hidden -Wall -W
-Wvla -Wdate-time -Wshift-overflow=2 -Wduplicated-cond -D_REENTRANT -fPIC
-DQT_NO_USING_NAMESPACE -DQT_NO_FOREACH -DQT_HAVE_POLL -DQT_HAVE_PPOLL
-DQT_BUILD_CORE_LIB -DQT_BUILDING_QT -DQT_NO_CAST_TO_ASCII
-DQT_ASCII_CAST_WARNINGS -DQT_MOC_COMPAT -DQT_USE_QSTRINGBUILDER
-DQT_DEPRECATED_WARNINGS -DQT_DISABLE_DEPRECATED_BEFORE=0x050000
-D_LARGEFILE64_SOURCE -D_LARGEFILE_SOURCE -DQT_NO_DEBUG -I. -Iglobal
-I../3rdparty/harfbuzz/src -I../3rdparty/md5 -I../3rdparty/md4
-I../3rdparty/sha3 -I../3rdparty/double-conversion/include
-I../3rdparty/double-conversion/include/double-conversion -I../3rdparty/forkfd
-I../../include -I../../include/QtCore -I../../include/QtCore/5.8.0
-I../../include/QtCore/5.8.0/QtCore -I.moc
-I/home/ivan/buildroot-git/buildroot/buildroot/output/host/arm-buildroot-linux-musleabihf/sysroot/usr/include/glib-2.0
-I/home/ivan/buildroot-git/buildroot/buildroot/output/host/arm-buildroot-linux-musleabihf/sysroot/usr/lib/glib-2.0/include
-I/home/ivan/buildroot-git/buildroot/buildroot/output/host/arm-buildroot-linux-musleabihf/sysroot/usr/include
-I../../mkspecs/devices/linux-buildroot-g++ -o .obj/qreadwritelock.o
thread/qreadwritelock.cpp
In file included from
../../include/QtCore/5.8.0/QtCore/private/../../../../../src/corelib/tools/qsimd_p.h:293:0,
                 from ../../include/QtCore/5.8.0/QtCore/private/qsimd_p.h:1,
                 from global/qlogging.cpp:57:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_ldc(unsigned int, unsigned int, const void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:48:49:
error: invalid conversion from ?const void*? to ?const int*? [-fpermissive]
   return __builtin_arm_ldc (__coproc, __CRd, __p);
                                                 ^
<built-in>: note:   initializing argument 3 of ?void __builtin_arm_ldc(unsigned
int, unsigned int, const int*)?
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_ldcl(unsigned int, unsigned int, const void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:55:50:
error: invalid conversion from ?const void*? to ?const int*? [-fpermissive]
   return __builtin_arm_ldcl (__coproc, __CRd, __p);
                                                  ^
<built-in>: note:   initializing argument 3 of ?void
__builtin_arm_ldcl(unsigned int, unsigned int, const int*)?
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_stc(unsigned int, unsigned int, void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:62:49:
error: invalid conversion from ?void*? to ?int*? [-fpermissive]
   return __builtin_arm_stc (__coproc, __CRd, __p);
                                                 ^
<built-in>: note:   initializing argument 3 of ?void __builtin_arm_stc(unsigned
int, unsigned int, int*)?
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_stcl(unsigned int, unsigned int, void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:69:50:
error: invalid conversion from ?void*? to ?int*? [-fpermissive]
   return __builtin_arm_stcl (__coproc, __CRd, __p);
                                                  ^
<built-in>: note:   initializing argument 3 of ?void
__builtin_arm_stcl(unsigned int, unsigned int, int*)?
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_ldc2(unsigned int, unsigned int, const void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:100:50:
error: invalid conversion from ?const void*? to ?const int*? [-fpermissive]
   return __builtin_arm_ldc2 (__coproc, __CRd, __p);
                                                  ^
<built-in>: note:   initializing argument 3 of ?void
__builtin_arm_ldc2(unsigned int, unsigned int, const int*)?
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_ldc2l(unsigned int, unsigned int, const void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:107:51:
error: invalid conversion from ?const void*? to ?const int*? [-fpermissive]
   return __builtin_arm_ldc2l (__coproc, __CRd, __p);
                                                   ^
<built-in>: note:   initializing argument 3 of ?void
__builtin_arm_ldc2l(unsigned int, unsigned int, const int*)?
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_stc2(unsigned int, unsigned int, void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:114:50:
error: invalid conversion from ?void*? to ?int*? [-fpermissive]
   return __builtin_arm_stc2 (__coproc, __CRd, __p);
                                                  ^
<built-in>: note:   initializing argument 3 of ?void
__builtin_arm_stc2(unsigned int, unsigned int, int*)?
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_stc2l(unsigned int, unsigned int, void*)?:
/home/ivan/buildroot-git/buildroot/buildroot/output/host/lib/gcc/arm-buildroot-linux-musleabihf/7.1.0/include/arm_acle.h:121:51:
error: invalid conversion from ?void*? to ?int*? [-fpermissive]
   return __builtin_arm_stc2l (__coproc, __CRd, __p);
                                                   ^
<built-in>: note:   initializing argument 3 of ?void
__builtin_arm_stc2l(unsigned int, unsigned int, int*)?
g++ -Wl,--gc-sections -Wl,-O1 -o ../../../bin/uic .obj/customwidgetsinfo.o
.obj/databaseinfo.o .obj/driver.o .obj/treewalker.o .obj/ui4.o .obj/validator.o
.obj/cppextractimages.o .obj/cppwritedeclaration.o .obj/cppwriteicondata.o
.obj/cppwriteicondeclaration.o .obj/cppwriteiconinitialization.o
.obj/cppwriteincludes.o .obj/cppwriteinitialization.o .obj/main.o .obj/uic.o  
-L/home/ivan/buildroot-git/buildroot/buildroot/output/build/qt5base-5.8.0/lib
-lQt5Bootstrap -lpthread 
Makefile:11089: recipe for target '.obj/qlogging.o' failed
make[4]: *** [.obj/qlogging.o] Error 1
make[4]: *** Waiting for unfinished jobs....
Makefile:163: recipe for target 'sub-corelib-make_first' failed
make[3]: *** [sub-corelib-make_first] Error 2
Makefile:46: recipe for target 'sub-src-make_first' failed
make[2]: *** [sub-src-make_first] Error 2
package/pkg-generic.mk:227: recipe for target
'/home/ivan/buildroot-git/buildroot/buildroot/output/build/qt5base-5.8.0/.stamp_built'
failed
make[1]: ***
[/home/ivan/buildroot-git/buildroot/buildroot/output/build/qt5base-5.8.0/.stamp_built]
Error 2
Makefile:79: recipe for target '_all' failed
make: *** [_all] Error 2

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #8 from Ivan <ivan.nazarenko@gmail.com> ---
Created attachment 7126
  --> https://bugs.busybox.net/attachment.cgi?id=7126&action=edit
config of a basic raspberrypi3 + gcc7 + qt + extras

and the config, attached. Note that if I just change gcc to 6.x, everything
builds.

And there is one more odd thing. Some hundreds of lines before this failure, I
noticed this:

g++ -c -pipe -ffunction-sections -O2 -fPIC -std=c++11 -fno-exceptions -Wall -W
-D_REENTRANT -DQT_VERSION_STR='"5.8.0"' -DQT_VERSION_MAJOR=5
-DQT_VERSION_MINOR=8 -DQT_VERSION_PATCH=0 -DQT_BOOTSTRAPPED
-DQT_NO_CAST_TO_ASCII -DQT_NO_FOREACH -DQT_NO_CAST_FROM_ASCII
-DQT_BUILD_BOOTSTRAP_LIB -DQT_BUILDING_QT -DQT_ASCII_CAST_WARNINGS
-DQT_MOC_COMPAT -DQT_USE_QSTRINGBUILDER -DQT_DEPRECATED_WARNINGS
-DQT_DISABLE_DEPRECATED_BEFORE=0x050000 -DQT_NO_EXCEPTIONS
-D_LARGEFILE64_SOURCE -D_LARGEFILE_SOURCE -DQT_NO_DEBUG -I.
-I../../3rdparty/zlib -I../../../include -I../../../include/QtCore
-I../../../include/QtCore/5.8.0 -I../../../include/QtCore/5.8.0/QtCore
-I../../../include/QtXml -I../../../include/QtXml/5.8.0
-I../../../include/QtXml/5.8.0/QtXml -I../../../mkspecs/linux-g++ -o
.obj/qvariant.o ../../corelib/kernel/qvariant.cpp
g++ -c -pipe -ffunction-sections -O2 -fPIC -std=c++11 -fno-exceptions -Wall -W
-D_REENTRANT -DQT_VERSION_STR='"5.8.0"' -DQT_VERSION_MAJOR=5
-DQT_VERSION_MINOR=8 -DQT_VERSION_PATCH=0 -DQT_BOOTSTRAPPED
-DQT_NO_CAST_TO_ASCII -DQT_NO_FOREACH -DQT_NO_CAST_FROM_ASCII
-DQT_BUILD_BOOTSTRAP_LIB -DQT_BUILDING_QT -DQT_ASCII_CAST_WARNINGS
-DQT_MOC_COMPAT -DQT_USE_QSTRINGBUILDER -DQT_DEPRECATED_WARNINGS
-DQT_DISABLE_DEPRECATED_BEFORE=0x050000 -DQT_NO_EXCEPTIONS
-D_LARGEFILE64_SOURCE -D_LARGEFILE_SOURCE -DQT_NO_DEBUG -I.
-I../../3rdparty/zlib -I../../../include -I../../../include/QtCore
-I../../../include/QtCore/5.8.0 -I../../../include/QtCore/5.8.0/QtCore
-I../../../include/QtXml -I../../../include/QtXml/5.8.0
-I../../../include/QtXml/5.8.0/QtXml -I../../../mkspecs/linux-g++ -o
.obj/qsystemerror.o ../../corelib/kernel/qsystemerror.cpp
g++ -c -pipe -ffunction-sections -O2 -fPIC -std=c++11 -fno-exceptions -Wall -W
-D_REENTRANT -DQT_VERSION_STR='"5.8.0"' -DQT_VERSION_MAJOR=5
-DQT_VERSION_MINOR=8 -DQT_VERSION_PATCH=0 -DQT_BOOTSTRAPPED
-DQT_NO_CAST_TO_ASCII -DQT_NO_FOREACH -DQT_NO_CAST_FROM_ASCII
-DQT_BUILD_BOOTSTRAP_LIB -DQT_BUILDING_QT -DQT_ASCII_CAST_WARNINGS
-DQT_MOC_COMPAT -DQT_USE_QSTRINGBUILDER -DQT_DEPRECATED_WARNINGS
-DQT_DISABLE_DEPRECATED_BEFORE=0x050000 -DQT_NO_EXCEPTIONS
-D_LARGEFILE64_SOURCE -D_LARGEFILE_SOURCE -DQT_NO_DEBUG -I.
-I../../3rdparty/zlib -I../../../include -I../../../include/QtCore
-I../../../include/QtCore/5.8.0 -I../../../include/QtCore/5.8.0/QtCore
-I../../../include/QtXml -I../../../include/QtXml/5.8.0
-I../../../include/QtXml/5.8.0/QtXml -I../../../mkspecs/linux-g++ -o
.obj/quuid.o ../../corelib/plugin/quuid.cpp
g++ -c -pipe -ffunction-sections -O2 -fPIC -std=c++11 -fno-exceptions -Wall -W
-D_REENTRANT -DQT_VERSION_STR='"5.8.0"' -DQT_VERSION_MAJOR=5
-DQT_VERSION_MINOR=8 -DQT_VERSION_PATCH=0 -DQT_BOOTSTRAPPED
-DQT_NO_CAST_TO_ASCII -DQT_NO_FOREACH -DQT_NO_CAST_FROM_ASCII
-DQT_BUILD_BOOTSTRAP_LIB -DQT_BUILDING_QT -DQT_ASCII_CAST_WARNINGS
-DQT_MOC_COMPAT -DQT_USE_QSTRINGBUILDER -DQT_DEPRECATED_WARNINGS
-DQT_DISABLE_DEPRECATED_BEFORE=0x050000 -DQT_NO_EXCEPTIONS
-D_LARGEFILE64_SOURCE -D_LARGEFILE_SOURCE -DQT_NO_DEBUG -I.
-I../../3rdparty/zlib -I../../../include -I../../../include/QtCore
-I../../../include/QtCore/5.8.0 -I../../../include/QtCore/5.8.0/QtCore
-I../../../include/QtXml -I../../../include/QtXml/5.8.0
-I../../../include/QtXml/5.8.0/QtXml -I../../../mkspecs/linux-g++ -o
.obj/qbitarray.o ../../corelib/tools/qbitarray.cpp

The odd thing is why my system compiler is being called instead of the one
provided by buildroot?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #9 from Ivan <ivan.nazarenko@gmail.com> ---
but verifying against the output produced when using gcc6.x, I see it invoking
the same apparent system compiler, and yet everything works fine. So somehow it
is actually invoking the arm cross-compiler and I just do not understand how.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #10 from Peter Seiderer <ps.report@gmx.net> ---
This is the normal qt source build 'bootstrap', first build the qmake command
for the host system, then cross compile qt using the generated Makefiles
(generated by the host qmake command).

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #11 from Peter Seiderer <ps.report@gmx.net> ---
$ ./host/usr/bin/arm-buildroot-linux-gnueabihf-g++ --version
arm-buildroot-linux-gnueabihf-g++.br_real (Buildroot
2017.08-git-01151-g8303ff38ef) 7.1.0

Include arm_acle.h and compile as C:

$ cat test_arm_acle_include.c 
#include <arm_acle.h>

int main(int argc, char* argv[]) {
        return 0;
}

$ ./host/usr/bin/arm-buildroot-linux-gnueabihf-gcc test_arm_acle_include.c

O.k, no failure


Include arm_acle.h and compile as C++:

$ cat test_arm_acle_include.cpp 
#include <arm_acle.h>

int main(int argc, char* argv[]) {
        return 0;
}

$ ./host/usr/bin/arm-buildroot-linux-gnueabihf-g++ test_arm_acle_include.cpp 
In file included from test_arm_acle_include.cpp:1:0:
/home/seiderer/Work/zeiss/freescale-imx6/build_rpi3_002/build-002-gcc-7-001/host/lib/gcc/arm-buildroot-linux-gnueabihf/7.1.0/include/arm_acle.h:
In function ?void __arm_ldc(unsigned int, unsigned int, const void*)?:
/home/seiderer/Work/zeiss/freescale-imx6/build_rpi3_002/build-002-gcc-7-001/host/lib/gcc/arm-buildroot-linux-gnueabihf/7.1.0/include/arm_acle.h:48:49:
error: invalid conversion from ?const void*? to ?const int*? [-fpermissive]
   return __builtin_arm_ldc (__coproc, __CRd, __p);
[...]

$ ./host/usr/bin/arm-buildroot-linux-gnueabihf-g++ -fpermissive
test_arm_acle_include.cpp

O.k., no failure...

Quick fix: add -fpermissive for the qt compile in case of gcc-7.x

Long term fix: ask the gcc people to fix arm_acle.h for C++ compile

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #12 from Arnout Vandecappelle <arnout@mind.be> ---
(In reply to Ivan from comment #4)
> Verifying the ChangeLog-2016, 2015, etc files shows that in 2014 the dependency of cloog > was removed. This is true for the last gcc5.x, 6.x and 7.x.
> 
> But gcc4.x still depends on cloog.
> 
> So the options seems to be either:
> 
> - keep cloog, with corresponding old isl, and gcc4.x users have graphite, or
> - removes cloog, new isl, and gcc7.1 users have graphite.

I would remove cloog, bump isl, and disable graphite support for gcc 4.9. gcc
4.9 is going away at some point anyway.

BTW Ivan please clone this bug for the Qt5base issue, it has nothing to do with
graphite support.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #13 from Andrey Yurovsky <yurovsky@gmail.com> ---
Arnout, that seems like solid fix. Would you like me to submit a patch or are
you already working on that?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #14 from Ivan <ivan.nazarenko@gmail.com> ---
> BTW Ivan please clone this bug for the Qt5base issue, it has nothing to do with
> graphite support.

Reported it to the gcc people:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81497

Though I did not understand their comment. I would fix it simply type casting
the third parameter. But in this case I am sure I do not know what I am talking
about.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

--- Comment #15 from Peter Seiderer <ps.report@gmx.net> ---
(In reply to Ivan from comment #14)

Changing host/lib/gcc/arm-buildroot-linux-gnueabihf/7.1.0/include/arm_acle.h
the following way fixed the compile for me:

--- arm_acle.h_orig     2017-07-24 20:54:27.202341305 +0200
+++ arm_acle.h  2017-07-24 21:18:08.750558081 +0200
@@ -43,28 +43,28 @@

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_ldc (const unsigned int __coproc, const unsigned int __CRd,
-          const void * __p)
+          const int * __p)
 {
   return __builtin_arm_ldc (__coproc, __CRd, __p);
 }

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_ldcl (const unsigned int __coproc, const unsigned int __CRd,
-           const void * __p)
+           const int * __p)
 {
   return __builtin_arm_ldcl (__coproc, __CRd, __p);
 }

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_stc (const unsigned int __coproc, const unsigned int __CRd,
-          void * __p)
+          int * __p)
 {
   return __builtin_arm_stc (__coproc, __CRd, __p);
 }

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_stcl (const unsigned int __coproc, const unsigned int __CRd,
-           void * __p)
+           int * __p)
 {
   return __builtin_arm_stcl (__coproc, __CRd, __p);
 }
@@ -95,28 +95,28 @@

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_ldc2 (const unsigned int __coproc, const unsigned int __CRd,
-           const void * __p)
+           const int * __p)
 {
   return __builtin_arm_ldc2 (__coproc, __CRd, __p);
 }

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_ldc2l (const unsigned int __coproc, const unsigned int __CRd,
-            const void * __p)
+            const int * __p)
 {
   return __builtin_arm_ldc2l (__coproc, __CRd, __p);
 }

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_stc2 (const unsigned int __coproc, const unsigned int __CRd,
-           void * __p)
+           int * __p)
 {
   return __builtin_arm_stc2 (__coproc, __CRd, __p);
 }

 __extension__ static __inline void __attribute__ ((__always_inline__))
 __arm_stc2l (const unsigned int __coproc, const unsigned int __CRd,
-            void * __p)
+            int * __p)
 {
   return __builtin_arm_stc2l (__coproc, __CRd, __p);
 }

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Buildroot] [Bug 10091] gcc7.1 does not build with graphite support due to old isl

[bugzilla at busybox.net]

https://bugs.busybox.net/show_bug.cgi?id=10091

Thomas Petazzoni <thomas.petazzoni@free-electrons.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #16 from Thomas Petazzoni <thomas.petazzoni@free-electrons.com> ---
The graphite issue with gcc 7.x has been fixed in
https://git.buildroot.org/buildroot/commit/?id=d16b6bf9b9da2d58955bf1abaf15f25c5666b90a
and
https://git.buildroot.org/buildroot/commit/?id=94afc741a0cd837d9549c346babc5b8be68dea5a.
Please open a separate bug report for the Qt5 issue. Thanks!

-- 
You are receiving this mail because:
You are on the CC list for the bug.