trace: bpf: Allow bpf to attach to bare tracepoints
authorQais Yousef <qais.yousef@arm.com>
Tue, 19 Jan 2021 12:22:36 +0000 (12:22 +0000)
committerAlexei Starovoitov <ast@kernel.org>
Wed, 20 Jan 2021 22:13:59 +0000 (14:13 -0800)
commit6939f4ef16d48f2093f337162cfc041d0e30ed25
tree6ecd96bd4b33e542db6e479a76d26f013bc07fd2
parent86e6b4e993cf0c4dbe4c0ebfe052c89b9f9a2ade
trace: bpf: Allow bpf to attach to bare tracepoints

Some subsystems only have bare tracepoints (a tracepoint with no
associated trace event) to avoid the problem of trace events being an
ABI that can't be changed.

>From bpf presepective, bare tracepoints are what it calls
RAW_TRACEPOINT().

Since bpf assumed there's 1:1 mapping, it relied on hooking to
DEFINE_EVENT() macro to create bpf mapping of the tracepoints. Since
bare tracepoints use DECLARE_TRACE() to create the tracepoint, bpf had
no knowledge about their existence.

By teaching bpf_probe.h to parse DECLARE_TRACE() in a similar fashion to
DEFINE_EVENT(), bpf can find and attach to the new raw tracepoints.

Enabling that comes with the contract that changes to raw tracepoints
don't constitute a regression if they break existing bpf programs.
We need the ability to continue to morph and modify these raw
tracepoints without worrying about any ABI.

Update Documentation/bpf/bpf_design_QA.rst to document this contract.

Signed-off-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210119122237.2426878-2-qais.yousef@arm.com
Documentation/bpf/bpf_design_QA.rst
include/trace/bpf_probe.h