configure: fix --enable-fuzzing linker failures

With --enable-fuzzing, QEMU_CFLAGS include -fsanitize=fuzzer-no-link.
This should allow us to build non-fuzzer binaries using objects
instrumented for fuzzing. However, to do that, we also need to link with
-fsanitize=fuzzer-no-link. We were not doing that.

Reported-by: Li Qiang <liq3ea@163.com>,
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <20210221174510.22542-1-alxndr@bu.edu>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

1 files changed, 10 insertions, 0 deletions

diff --git a/configure b/configure
index a79b3746d4..19f2b88589 100755
--- a/configure
+++ b/configure
@@ -6096,7 +6096,17 @@ if test "$fuzzing" = "yes" ; then
   # If LIB_FUZZING_ENGINE is set, assume we are running on OSS-Fuzz, and the
   # needed CFLAGS have already been provided
   if test -z "${LIB_FUZZING_ENGINE+xxx}" ; then
+    # Add CFLAGS to tell clang to add fuzzer-related instrumentation to all the
+    # compiled code.
     QEMU_CFLAGS="$QEMU_CFLAGS -fsanitize=fuzzer-no-link"
+    # To build non-fuzzer binaries with --enable-fuzzing, link everything with
+    # fsanitize=fuzzer-no-link. Otherwise, the linker will be unable to bind
+    # the fuzzer-related callbacks added by instrumentation.
+    QEMU_LDFLAGS="$QEMU_LDFLAGS -fsanitize=fuzzer-no-link"
+    # For the actual fuzzer binaries, we need to link against the libfuzzer
+    # library. Provide the flags for doing this in FUZZ_EXE_LDFLAGS. The meson
+    # rule for the fuzzer adds these to the link_args. They need to be
+    # configurable, to support OSS-Fuzz
     FUZZ_EXE_LDFLAGS="-fsanitize=fuzzer"
   else
     FUZZ_EXE_LDFLAGS="$LIB_FUZZING_ENGINE"