[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20200908192719.GA1300@bpf-dev>
Date: Tue, 8 Sep 2020 16:27:20 -0300
From: Carlos Antonio Neira Bustos <cneirabustos@...il.com>
To: Andrii Nakryiko <andrii.nakryiko@...il.com>
Cc: Networking <netdev@...r.kernel.org>, Yonghong Song <yhs@...com>,
"Eric W. Biederman" <ebiederm@...ssion.com>,
Jesper Dangaard Brouer <brouer@...hat.com>,
bpf <bpf@...r.kernel.org>, Andrii Nakryiko <andriin@...com>
Subject: Re: [PATCH v1 bpf-next] bpf: new helper bpf_get_current_pcomm
On Wed, Aug 26, 2020 at 10:25:55PM -0700, Andrii Nakryiko wrote:
> On Wed, Aug 26, 2020 at 9:06 AM Carlos Neira <cneirabustos@...il.com> wrote:
> >
> > In multi-threaded applications bpf_get_current_comm is returning per-thread
> > names, this helper will return comm from real_parent.
> > This makes a difference for some Java applications, where get_current_comm is
> > returning per-thread names, but get_current_pcomm will return "java".
> >
>
> Why not bpf_probe_read_kernel_str(dst, 16, task->real_parent->comm)
> for fentry/fexit/tp_btf/etc BTF-aware BPF programs or
> BPF_CORE_READ_STR_INTO(dst, task, real_parent, comm) for any BPF
> program that has bpf_probe_read_kernel[_str]() (which is pretty much
> every BPF program nowadays, I think)?
>
> Yes, CONFIG_DEBUG_INFO_BTF=y Kconfig is a requirement, but it's a good
> idea to have that if you are using BPF anyways.
>
> > Signed-off-by: Carlos Neira <cneirabustos@...il.com>
> > ---
> > include/linux/bpf.h | 1 +
> > include/uapi/linux/bpf.h | 15 ++++-
> > kernel/bpf/core.c | 1 +
> > kernel/bpf/helpers.c | 28 +++++++++
> > kernel/trace/bpf_trace.c | 2 +
> > tools/include/uapi/linux/bpf.h | 15 ++++-
> > .../selftests/bpf/prog_tests/current_pcomm.c | 57 +++++++++++++++++++
> > .../selftests/bpf/progs/test_current_pcomm.c | 17 ++++++
> > 8 files changed, 134 insertions(+), 2 deletions(-)
> > create mode 100644 tools/testing/selftests/bpf/prog_tests/current_pcomm.c
> > create mode 100644 tools/testing/selftests/bpf/progs/test_current_pcomm.c
> >
>
> [...]
Thanks Andrii,
I'll use bpf_probe_read_kernel_str(dst, 16, task->real_parent->comm),
I was not aware of that.
Bests!.
Powered by blists - more mailing lists