qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] fuzz: Add support for custom fuzzing library


From: Paolo Bonzini
Subject: Re: [PATCH] fuzz: Add support for custom fuzzing library
Date: Tue, 1 Sep 2020 21:00:48 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.11.0

On 01/09/20 20:18, Alexander Bulekov wrote:
> ---
>  configure                    | 12 ++++++++++--
>  meson.build                  |  6 +++++-
>  tests/qtest/fuzz/meson.build |  5 ++---
>  3 files changed, 17 insertions(+), 6 deletions(-)
> 
> 
> Hi Paolo,
> Here I'm trying to specify the linker-script with
> add_project_link_arguments. How I'm testing this:
> 
> $ CC=clang-10 CXX=clang++-10 ../configure --enable-fuzzing
> $ make V=1 "-j$(nproc)" qemu-fuzz-i386
> 
> clang++-10  -o qemu-fuzz-i386 
> qemu-fuzz-i386.p/tests_qtest_fuzz_qtest_wrappers.c.o \
> ... libblock.fa chardev/libchardev.fa \
> -Wl,--start-group tests/qtest/libqos/libqos.a -Wl,--no-whole-archive \
> -Wl,-T,/home/alxndr/Development/qemu/tests/qtest/fuzz/fork_fuzz.ld \
> ... \
> -Wl,-rpath-link,/home/alxndr/Development/qemu/build/ -lstdc++ -Wl,--end-group
> 
> Maybe if I can get the oss-fuzz LIB_FUZZING_ENGINE
> (/usr/lib/libFuzzingEngine.a) into the --start-group, that could also
> solve the issue... I'll take another look at exactly what the oss-fuzz
> build container does.

Actually your patch is pretty close, just by hacking

diff --git a/tests/qtest/libqos/meson.build b/tests/qtest/libqos/meson.build
index 19931b9248..2bc46c5a84 100644
--- a/tests/qtest/libqos/meson.build
+++ b/tests/qtest/libqos/meson.build
@@ -52,6 +52,6 @@ libqos = static_library('qos',
         'arm-xilinx-zynq-a9-machine.c',
         'ppc64_pseries-machine.c',
         'x86_64_pc-machine.c',
-), build_by_default: false)
+), name_suffix: 'fa', build_by_default: false)

 qos = declare_dependency(link_whole: libqos)

I can get it to work.  Better find a way to fix it in Meson though,
because relying on the ".a" suffix is very brittle.

Paolo




reply via email to

[Prev in Thread] Current Thread [Next in Thread]