netdev - Re: [PATCH v3 bpf] selftests/bpf: do not ignore clang failures

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAEf4Bzb6mY-F-wUNNimS+hMSRbJetTKXNcGDQbsJXhXDywA+tg@mail.gmail.com>
Date:   Thu, 11 Jul 2019 07:55:34 -0700
From:   Andrii Nakryiko <andrii.nakryiko@...il.com>
To:     Ilya Leoshkevich <iii@...ux.ibm.com>
Cc:     bpf <bpf@...r.kernel.org>, Networking <netdev@...r.kernel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Song Liu <liu.song.a23@...il.com>
Subject: Re: [PATCH v3 bpf] selftests/bpf: do not ignore clang failures

On Thu, Jul 11, 2019 at 2:14 AM Ilya Leoshkevich <iii@...ux.ibm.com> wrote:
>
> When compiling an eBPF prog fails, make still returns 0, because
> failing clang command's output is piped to llc and therefore its
> exit status is ignored.
>
> When clang fails, pipe the string "clang failed" to llc. This will make
> llc fail with an informative error message. This solution was chosen
> over using pipefail, having separate targets or getting rid of llc
> invocation due to its simplicity.
>
> In addition, pull Kbuild.include in order to get .DELETE_ON_ERROR target,

In your original patch you explicitly declared .DELETE_ON_ERROR, but
in this one you just include Kbuild.include.
Is it enough to just include that file to get desired behavior or your
forgot to add .DELETE_ON_ERROR?

> which would cause partial .o files to be removed.
>
> Signed-off-by: Ilya Leoshkevich <iii@...ux.ibm.com>
> ---

Thanks!

Acked-by: Andrii Nakryiko <andriin@...com>

> v1->v2: use intermediate targets instead of pipefail
> v2->v3: pipe "clang failed" instead of using intermediate targets
>
> tools/testing/selftests/bpf/Makefile | 13 +++++++------
>  1 file changed, 7 insertions(+), 6 deletions(-)
>
> diff --git a/tools/testing/selftests/bpf/Makefile b/tools/testing/selftests/bpf/Makefile
> index e36356e2377e..e375f399b7a6 100644
> --- a/tools/testing/selftests/bpf/Makefile
> +++ b/tools/testing/selftests/bpf/Makefile
> @@ -1,4 +1,5 @@
>  # SPDX-License-Identifier: GPL-2.0
> +include ../../../../scripts/Kbuild.include
>
>  LIBDIR := ../../../lib
>  BPFDIR := $(LIBDIR)/bpf
> @@ -185,8 +186,8 @@ $(ALU32_BUILD_DIR)/test_progs_32: prog_tests/*.c
>
>  $(ALU32_BUILD_DIR)/%.o: progs/%.c $(ALU32_BUILD_DIR) \
>                                         $(ALU32_BUILD_DIR)/test_progs_32
> -       $(CLANG) $(CLANG_FLAGS) \
> -                -O2 -target bpf -emit-llvm -c $< -o - |      \
> +       ($(CLANG) $(CLANG_FLAGS) -O2 -target bpf -emit-llvm -c $< -o - || \
> +               echo "clang failed") | \
>         $(LLC) -march=bpf -mattr=+alu32 -mcpu=$(CPU) $(LLC_FLAGS) \
>                 -filetype=obj -o $@
>  ifeq ($(DWARF2BTF),y)
> @@ -197,16 +198,16 @@ endif
>  # Have one program compiled without "-target bpf" to test whether libbpf loads
>  # it successfully
>  $(OUTPUT)/test_xdp.o: progs/test_xdp.c
> -       $(CLANG) $(CLANG_FLAGS) \
> -               -O2 -emit-llvm -c $< -o - | \
> +       ($(CLANG) $(CLANG_FLAGS) -O2 -emit-llvm -c $< -o - || \
> +               echo "clang failed") | \
>         $(LLC) -march=bpf -mcpu=$(CPU) $(LLC_FLAGS) -filetype=obj -o $@
>  ifeq ($(DWARF2BTF),y)
>         $(BTF_PAHOLE) -J $@
>  endif
>
>  $(OUTPUT)/%.o: progs/%.c
> -       $(CLANG) $(CLANG_FLAGS) \
> -                -O2 -target bpf -emit-llvm -c $< -o - |      \
> +       ($(CLANG) $(CLANG_FLAGS) -O2 -target bpf -emit-llvm -c $< -o - || \
> +               echo "clang failed") | \
>         $(LLC) -march=bpf -mcpu=$(CPU) $(LLC_FLAGS) -filetype=obj -o $@
>  ifeq ($(DWARF2BTF),y)
>         $(BTF_PAHOLE) -J $@
> --
> 2.21.0
>