netdev - Re: [PATCH bpf-next] selftests/bpf: fix building bpf selftests statically

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: <CAJD7tkaxHKUrb44-J5Zxm6x1KH-XnPVXDgZyu4+Rc094D7MEhA@mail.gmail.com>
Date:   Fri, 13 May 2022 16:05:03 -0700
From:   Yosry Ahmed <yosryahmed@...gle.com>
To:     Andrii Nakryiko <andrii.nakryiko@...il.com>
Cc:     Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        John Fastabend <john.fastabend@...il.com>,
        KP Singh <kpsingh@...nel.org>, Hao Luo <haoluo@...gle.com>,
        Networking <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH bpf-next] selftests/bpf: fix building bpf selftests statically

On Fri, May 13, 2022 at 3:57 PM Andrii Nakryiko
<andrii.nakryiko@...il.com> wrote:
>
> On Thu, May 12, 2022 at 5:41 PM Yosry Ahmed <yosryahmed@...gle.com> wrote:
> >
> > bpf selftests can no longer be built with CFLAGS=-static with
> > liburandom_read.so and its dependent target.
> >
> > Filter out -static for liburandom_read.so and its dependent target.
> >
> > Signed-off-by: Yosry Ahmed <yosryahmed@...gle.com>
> > ---
> >  tools/testing/selftests/bpf/Makefile | 9 ++++++---
> >  1 file changed, 6 insertions(+), 3 deletions(-)
> >
> > diff --git a/tools/testing/selftests/bpf/Makefile b/tools/testing/selftests/bpf/Makefile
> > index 6bbc03161544..4eaefc187d5b 100644
> > --- a/tools/testing/selftests/bpf/Makefile
> > +++ b/tools/testing/selftests/bpf/Makefile
> > @@ -168,14 +168,17 @@ $(OUTPUT)/%:%.c
> >         $(call msg,BINARY,,$@)
> >         $(Q)$(LINK.c) $^ $(LDLIBS) -o $@
> >
> > +# If the tests are being built statically, exclude dynamic libraries defined
> > +# in this Makefile and their dependencies.
> > +DYNAMIC_CFLAGS := $(filter-out -static,$(CFLAGS))
>
> I don't particularly like yet another CFLAGS global variable, but also
> you are not filtering out -static from LDFLAGS, which would be
> problematic if you try to do
>
> make SAN_FLAGS=-static
>
> which otherwise would work (and is probably better than overriding all of CFLAGS
>
> How about something like this

Yeah this looks good, thanks for pointing out the problem with LDFLAGS.

>
> diff --git a/tools/testing/selftests/bpf/Makefile
> b/tools/testing/selftests/bpf/Makefile
> index 6bbc03161544..2e8eddf240af 100644
> --- a/tools/testing/selftests/bpf/Makefile
> +++ b/tools/testing/selftests/bpf/Makefile
> @@ -170,11 +170,11 @@ $(OUTPUT)/%:%.c
>
>  $(OUTPUT)/liburandom_read.so: urandom_read_lib1.c urandom_read_lib2.c
>         $(call msg,LIB,,$@)
> -       $(Q)$(CC) $(CFLAGS) -fPIC $(LDFLAGS) $^ $(LDLIBS) --shared -o $@
> +       $(Q)$(CC) $(filter-out -static,$(CFLAGS) $(LDFLAGS)) $^
> $(LDLIBS) -fPIC -shared -o $@
>
>  $(OUTPUT)/urandom_read: urandom_read.c urandom_read_aux.c
> $(OUTPUT)/liburandom_read.so
>         $(call msg,BINARY,,$@)
> -       $(Q)$(CC) $(CFLAGS) $(LDFLAGS) $(filter %.c,$^)                        \
> +       $(Q)$(CC) $(filter-out -static,$(CFLAGS) $(LDFLAGS)) $(filter %.c,$^)  \
>                   liburandom_read.so $(LDLIBS)                                 \
>                   -Wl,-rpath=. -Wl,--build-id=sha1 -o $@
>
> ?
>
>
> But I also have a question, this leaves urandom_read relying on
> system-wide shared libraries, isn't that still a problem for you? Or
> you intend to just ignore urandom_read-related tests?
>

I wasn't running those tests, and I thought having most tests compile
statically is better than nothing. Maybe this can be fixed by defining
a static target for liburandom_read? I am honestly not sure, I have
little experience with Makefiles/compilation.

>
> $ ldd urandom_read
>         linux-vdso.so.1 (0x00007ffd0d5e5000)
>         liburandom_read.so (0x00007fc7f7d76000)
>         libelf.so.1 => /lib64/libelf.so.1 (0x00007fc7f7937000)
>         libz.so.1 => /lib64/libz.so.1 (0x00007fc7f7720000)
>         librt.so.1 => /lib64/librt.so.1 (0x00007fc7f7518000)
>         libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fc7f72f8000)
>         libc.so.6 => /lib64/libc.so.6 (0x00007fc7f6f33000)
>         /lib64/ld-linux-x86-64.so.2 (0x00007fc7f7b50000)
>
>
>
>
> >  $(OUTPUT)/liburandom_read.so: urandom_read_lib1.c urandom_read_lib2.c
> >         $(call msg,LIB,,$@)
> > -       $(Q)$(CC) $(CFLAGS) -fPIC $(LDFLAGS) $^ $(LDLIBS) --shared -o $@
> > +       $(Q)$(CC) $(DYNAMIC_CFLAGS) -fPIC $(LDFLAGS) $^ $(LDLIBS) --shared -o $@
> >
> >  $(OUTPUT)/urandom_read: urandom_read.c urandom_read_aux.c $(OUTPUT)/liburandom_read.so
> >         $(call msg,BINARY,,$@)
> > -       $(Q)$(CC) $(CFLAGS) $(LDFLAGS) $(filter %.c,$^)                        \
> > -                 liburandom_read.so $(LDLIBS)                                 \
> > +       $(Q)$(CC) $(DYNAMIC_CFLAGS) $(LDFLAGS) $(filter %.c,$^)                 \
> > +                 liburandom_read.so $(LDLIBS)                                  \
> >                   -Wl,-rpath=. -Wl,--build-id=sha1 -o $@
> >
> >  $(OUTPUT)/bpf_testmod.ko: $(VMLINUX_BTF) $(wildcard bpf_testmod/Makefile bpf_testmod/*.[ch])
> > --
> > 2.36.0.550.gb090851708-goog
> >