perf callchain lbr: Make the leaf IP that of the sample
authorIan Rogers <irogers@google.com>
Thu, 5 Feb 2026 20:56:21 +0000 (12:56 -0800)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Fri, 6 Feb 2026 21:10:44 +0000 (18:10 -0300)
The current IP of a leaf function when reported from a perf record with
"--call-graph lbr" is the "to" field of the LBR branch stack record.

The sample for the event being recorded may be further into the function
and there may be inlining information associated with it.

Rather than use the branch stack "to" field in this case switch to the
callchain appending the sample->ip and thereby allowing the inline
information to show.

Before this change:
```
$ perf record --call-graph lbr perf test -w inlineloop
...
$ perf script --fields +srcline
...
perf-inlineloop  467586  4649.344493:     950905 cpu_core/cycles/P:
           55dfda2829c0 parent+0x0 (perf)
 inlineloop.c:31
           55dfda282a96 inlineloop+0x86 (perf)
 inlineloop.c:47
           55dfda236420 run_workload+0x59 (perf)
 builtin-test.c:715
           55dfda236b03 cmd_test+0x413 (perf)
 builtin-test.c:825
...
```

After this change:
```
$ perf record --call-graph lbr perf test -w inlineloop
...
$ perf script --fields +srcline
...
perf-inlineloop  529703 11878.680815:     950905 cpu_core/cycles/P:
            555ce86be9e6 leaf+0x26
  inlineloop.c:20 (inlined)
            555ce86be9e6 middle+0x26
  inlineloop.c:27 (inlined)
            555ce86be9e6 parent+0x26 (perf)
  inlineloop.c:32
            555ce86bea96 inlineloop+0x86 (perf)
  inlineloop.c:47
            555ce8672420 run_workload+0x59 (perf)
  builtin-test.c:715
            555ce8672b03 cmd_test+0x413 (perf)
  builtin-test.c:825
...
```

Reviewed-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Dmitriy Vyukov <dvyukov@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Krzysztof Ɓopatowski <krzysztof.m.lopatowski@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/util/machine.c

index 5b0f5a4..e76f8c8 100644 (file)
@@ -2423,8 +2423,14 @@ static int lbr_callchain_add_lbr_ip(struct thread *thread,
        }
 
        if (callee) {
-               /* Add LBR ip from first entries.to */
-               ip = entries[0].to;
+               /*
+                * Set the (first) leaf function's IP to sample->ip (the
+                * location of the sample) but if not recorded use entries.to
+                */
+               if (sample->ip)
+                       ip = sample->ip;
+               else
+                       ip = entries[0].to;
                flags = &entries[0].flags;
                *branch_from = entries[0].from;
                err = add_callchain_ip(thread, cursor, parent,
@@ -2477,8 +2483,14 @@ static int lbr_callchain_add_lbr_ip(struct thread *thread,
        }
 
        if (lbr_nr > 0) {
-               /* Add LBR ip from first entries.to */
-               ip = entries[0].to;
+               /*
+                * Set the (first) leaf function's IP to sample->ip (the
+                * location of the sample) but if not recorded use entries.to
+                */
+               if (sample->ip)
+                       ip = sample->ip;
+               else
+                       ip = entries[0].to;
                flags = &entries[0].flags;
                *branch_from = entries[0].from;
                err = add_callchain_ip(thread, cursor, parent,