https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100289
--- Comment #13 from rudi at heitbaum dot com --- I have looked further into this; and here are the locations of sys/mman.h, this include comes from the glibc package. gcc builds in package #80, whilst glibc builds in package #82. The dependancies - given the way the gcc11 configure script works (and must be pulling from /usr/include/x86_64-linux-gnu/sys/mman.h) is a recursive dependency. Here are the packages in question. gcc:bootstrap does not have a glibc dependency. == gcc package == PKG_NAME="gcc" PKG_DEPENDS_BOOTSTRAP="ccache:host autoconf:host binutils:host gmp:host mpfr:host mpc:host zstd:host" PKG_DEPENDS_TARGET="toolchain" PKG_DEPENDS_HOST="ccache:host autoconf:host binutils:host gmp:host mpfr:host mpc:host zstd:host glibc" PKG_DEPENDS_INIT="toolchain" == glibc package == PKG_NAME="glibc" PKG_DEPENDS_TARGET="ccache:host autotools:host linux:host gcc:bootstrap pigz:host Python3:host" PKG_DEPENDS_INIT="glibc" == here is the actual creation of mman.h (using gcc10.3.0) == grep sys/mman.h 82.log /usr/bin/install -c -m 644 sys/mman.h /storage/home-rudi/LibreELEC.tv/build.LibreELEC-Generic.x86_64-10.0-devel/.sysroot/glibc.target/usr/include/sys/mman.h /usr/bin/install -c -m 644 sys/mman.h /storage/home-rudi/LibreELEC.tv/build.LibreELEC-Generic.x86_64-10.0-devel/install_pkg/glibc-2.33/usr/include/sys/mman.h == the find after a full build == find . -name mman.h | grey sys/ ./build.LibreELEC-Generic.x86_64-10.0-devel/install_pkg/glibc-2.33/usr/include/sys/mman.h ./build.LibreELEC-Generic.x86_64-10.0-devel/toolchain/x86_64-libreelec-linux-gnu/sysroot/usr/include/sys/mman.h ./build.LibreELEC-Generic.x86_64-10.0-devel/build/glibc-2.33/misc/sys/mman.h ./build.LibreELEC-Generic.x86_64-10.0-devel/build/glibc-2.33/include/sys/mman.h ./build.LibreELEC-Generic.x86_64-10.0-devel/build/gcc-10.3.0/fixincludes/tests/base/sys/mman.h ./build.LibreELEC-Generic.x86_64-10.0-devel/build/xorg-intel-gpu-tools-1.26/lib/stubs/syscalls/sys/mman.h
