[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <194f38f2dc7d521375e5a660baaf1be31536be9a.camel@gmail.com>
Date: Mon, 16 Jan 2023 14:59:56 +0200
From: Eduard Zingerman <eddyz87@...il.com>
To: Quentin Monnet <quentin@...valent.com>,
Peter Foley <pefoley2@...oley.com>,
Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Andrii Nakryiko <andrii@...nel.org>,
Martin KaFai Lau <martin.lau@...ux.dev>,
Song Liu <song@...nel.org>, Yonghong Song <yhs@...com>,
John Fastabend <john.fastabend@...il.com>,
KP Singh <kpsingh@...nel.org>,
Stanislav Fomichev <sdf@...gle.com>,
Hao Luo <haoluo@...gle.com>, Jiri Olsa <jolsa@...nel.org>,
Nathan Chancellor <nathan@...nel.org>,
Nick Desaulniers <ndesaulniers@...gle.com>,
Tom Rix <trix@...hat.com>
Cc: bpf@...r.kernel.org, linux-kernel@...r.kernel.org,
llvm@...ts.linux.dev
Subject: Re: [PATCH] tools: bpf: Disable stack protector
On Mon, 2023-01-16 at 10:30 +0000, Quentin Monnet wrote:
> 2023-01-14 18:00 UTC-0500 ~ Peter Foley <pefoley2@...oley.com>
> > Avoid build errors on distros that force the stack protector on by
> > default.
> > e.g.
> > CLANG /home/peter/linux/work/tools/bpf/bpftool/pid_iter.bpf.o
> > skeleton/pid_iter.bpf.c:53:5: error: A call to built-in function '__stack_chk_fail' is not supported.
> > int iter(struct bpf_iter__task_file *ctx)
> > ^
> > 1 error generated.
> >
> > Signed-off-by: Peter Foley <pefoley2@...oley.com>
> > ---
> > tools/bpf/bpftool/Makefile | 1 +
> > tools/bpf/runqslower/Makefile | 5 +++--
> > 2 files changed, 4 insertions(+), 2 deletions(-)
> >
> > diff --git a/tools/bpf/bpftool/Makefile b/tools/bpf/bpftool/Makefile
> > index f610e184ce02a..36ac0002e386f 100644
> > --- a/tools/bpf/bpftool/Makefile
> > +++ b/tools/bpf/bpftool/Makefile
> > @@ -215,6 +215,7 @@ $(OUTPUT)%.bpf.o: skeleton/%.bpf.c $(OUTPUT)vmlinux.h $(LIBBPF_BOOTSTRAP)
> > -I$(or $(OUTPUT),.) \
> > -I$(srctree)/tools/include/uapi/ \
> > -I$(LIBBPF_BOOTSTRAP_INCLUDE) \
> > + -fno-stack-protector \
> > -g -O2 -Wall -target bpf -c $< -o $@
> > $(Q)$(LLVM_STRIP) -g $@
> >
>
> For bpftool, a similar patch was already submitted and merged to the
> bpf-next tree last Friday: 878625e1c7a1 ("bpftool: Always disable stack
> protection for BPF objects").
>
> > diff --git a/tools/bpf/runqslower/Makefile b/tools/bpf/runqslower/Makefile
> > index 8b3d87b82b7a2..f7313cc966a04 100644
> > --- a/tools/bpf/runqslower/Makefile
> > +++ b/tools/bpf/runqslower/Makefile
> > @@ -60,8 +60,9 @@ $(OUTPUT)/%.skel.h: $(OUTPUT)/%.bpf.o | $(BPFTOOL)
> > $(QUIET_GEN)$(BPFTOOL) gen skeleton $< > $@
> >
> > $(OUTPUT)/%.bpf.o: %.bpf.c $(BPFOBJ) | $(OUTPUT)
> > - $(QUIET_GEN)$(CLANG) -g -O2 -target bpf $(INCLUDES) \
> > - -c $(filter %.c,$^) -o $@ && \
> > + $(QUIET_GEN)$(CLANG) -g -O2 -target bpf $(INCLUDES) \
> > + -fno-stack-protector \
> > + -c $(filter %.c,$^) -o $@ && \
> > $(LLVM_STRIP) -g $@
> >
> > $(OUTPUT)/%.o: %.c | $(OUTPUT)
>
> This one looks good, thanks!
>
> I note a few more places in the repository where we compile to BPF using
> clang. Given that there have been patches to add -fno-stack-protector at
> several locations already, have you checked if any of these also need
> the flag, by any chance, so we could fix this once and for all?
>
> $ git grep -l 'target bpf ' | egrep -v '(Documentation|bpftool)'
> kernel/bpf/preload/iterators/Makefile
> samples/bpf/Makefile
> samples/bpf/test_lwt_bpf.sh
> tools/bpf/runqslower/Makefile
> tools/build/feature/Makefile
> tools/perf/Makefile.perf
> tools/perf/util/llvm-utils.c
> tools/testing/selftests/bpf/Makefile
> tools/testing/selftests/net/bpf/Makefile
> tools/testing/selftests/tc-testing/Makefile
A bit tangential, but since BPF LLVM backend does not support the
stack protector (should it?) there is also an option to adjust LLVM
to avoid this instrumentation, WDYT?
Powered by blists - more mailing lists