| Message ID | 20210116182133.2286884-2-qais.yousef@arm.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | BPF |
| Headers | show |
| Series | Allow attaching to bare tracepoints | expand |
| Context | Check | Description |
|---|---|---|
| netdev/cover_letter | success | Link |
| netdev/fixes_present | success | Link |
| netdev/patch_count | success | Link |
| netdev/tree_selection | success | Clearly marked for bpf-next |
| netdev/subject_prefix | success | Link |
| netdev/cc_maintainers | warning | 7 maintainers not CCed: corbet@lwn.net linux-doc@vger.kernel.org kpsingh@kernel.org kafai@fb.com songliubraving@fb.com john.fastabend@gmail.com mingo@redhat.com |
| netdev/source_inline | success | Was 0 now: 0 |
| netdev/verify_signedoff | success | Link |
| netdev/module_param | success | Was 0 now: 0 |
| netdev/build_32bit | success | Errors and warnings before: 1109 this patch: 1109 |
| netdev/kdoc | success | Errors and warnings before: 0 this patch: 0 |
| netdev/verify_fixes | success | Link |
| netdev/checkpatch | fail | ERROR: Macros with complex values should be enclosed in parentheses |
| netdev/build_allmodconfig_warn | success | Errors and warnings before: 1011 this patch: 1011 |
| netdev/header_inline | success | Link |
| netdev/stable | success | Stable not CCed |
On 1/16/21 10:21 AM, Qais Yousef wrote: > 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> Acked-by: Yonghong Song <yhs@fb.com>
diff --git a/Documentation/bpf/bpf_design_QA.rst b/Documentation/bpf/bpf_design_QA.rst index 2df7b067ab93..0e15f9b05c9d 100644 --- a/Documentation/bpf/bpf_design_QA.rst +++ b/Documentation/bpf/bpf_design_QA.rst @@ -208,6 +208,12 @@ data structures and compile with kernel internal headers. Both of these kernel internals are subject to change and can break with newer kernels such that the program needs to be adapted accordingly. +Q: Are tracepoints part of the stable ABI? +------------------------------------------ +A: NO. Tracepoints are tied to internal implementation details hence they are +subject to change and can break with newer kernels. BPF programs need to change +accordingly when this happens. + Q: How much stack space a BPF program uses? ------------------------------------------- A: Currently all program types are limited to 512 bytes of stack diff --git a/include/trace/bpf_probe.h b/include/trace/bpf_probe.h index cd74bffed5c6..a23be89119aa 100644 --- a/include/trace/bpf_probe.h +++ b/include/trace/bpf_probe.h @@ -55,8 +55,7 @@ /* tracepoints with more than 12 arguments will hit build error */ #define CAST_TO_U64(...) CONCATENATE(__CAST, COUNT_ARGS(__VA_ARGS__))(__VA_ARGS__) -#undef DECLARE_EVENT_CLASS -#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ +#define __BPF_DECLARE_TRACE(call, proto, args) \ static notrace void \ __bpf_trace_##call(void *__data, proto) \ { \ @@ -64,6 +63,10 @@ __bpf_trace_##call(void *__data, proto) \ CONCATENATE(bpf_trace_run, COUNT_ARGS(args))(prog, CAST_TO_U64(args)); \ } +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ + __BPF_DECLARE_TRACE(call, PARAMS(proto), PARAMS(args)) + /* * This part is compiled out, it is only here as a build time check * to make sure that if the tracepoint handling changes, the @@ -111,6 +114,11 @@ __DEFINE_EVENT(template, call, PARAMS(proto), PARAMS(args), size) #define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) +#undef DECLARE_TRACE +#define DECLARE_TRACE(call, proto, args) \ + __BPF_DECLARE_TRACE(call, PARAMS(proto), PARAMS(args)) \ + __DEFINE_EVENT(call, call, PARAMS(proto), PARAMS(args), 0) + #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) #undef DEFINE_EVENT_WRITABLE
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> --- Documentation/bpf/bpf_design_QA.rst | 6 ++++++ include/trace/bpf_probe.h | 12 ++++++++++-- 2 files changed, 16 insertions(+), 2 deletions(-)