diff options
author | Alexey Neyman <stilor@att.net> | 2017-01-26 21:50:18 (GMT) |
---|---|---|
committer | Alexey Neyman <stilor@att.net> | 2017-01-26 21:50:18 (GMT) |
commit | d7b043e360dbf261ddb97accf5c454d3cf88f92e (patch) | |
tree | 80adb198f805f8c2b8057dc12c03dd557a5062af /scripts/build/arch/mips.sh | |
parent | 675f7bfa4eb8415d5695dcda6111d9bd54419b0f (diff) |
Create <sgidefs.h> if MUSL doesn't create one
Document the status quo and work around for mips-musl.
Signed-off-by: Alexey Neyman <stilor@att.net>
Diffstat (limited to 'scripts/build/arch/mips.sh')
-rw-r--r-- | scripts/build/arch/mips.sh | 19 |
1 files changed, 19 insertions, 0 deletions
diff --git a/scripts/build/arch/mips.sh b/scripts/build/arch/mips.sh index 6097c89..f91a80a 100644 --- a/scripts/build/arch/mips.sh +++ b/scripts/build/arch/mips.sh @@ -69,3 +69,22 @@ CT_DoArchUClibcCflags() { esac done } + +CT_DoArchMUSLPostInstall() { + # GDB and MUSL maintainers seem to disagree on whether <sgidefs.h> + # is to be provided as a part of C library. GDB guys think it is + # a C library responsibility, while MUSL authors think GDB should + # not be using <sgidefs.h>. Neither side is willing to reach out + # to the other and negotiate the needed changes, and I don't want + # to play the middle man. Hence, provide our own wrapper for + # for <sgidefs.h> - the only solution short of telling MUSL users + # stop using it. This is why MUSL is experimental in ct-ng and + # will likely remain in that status. + # References: + # http://www.openwall.com/lists/musl/2017/01/26/2 + # https://sourceware.org/ml/gdb-patches/2017-01/msg00469.html + # https://www.sourceware.org/ml/libc-alpha/2004-11/msg00034.html + if [ ! -r "${CT_HEADERS_DIR}/sgidefs.h" ]; then + echo "#include <asm/sgidefs.h> // Redirected by ct-ng" > "${CT_HEADERS_DIR}/sgidefs.h" + fi +} |