projects / platform / upstream / bcc.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 7598249) | patch
Let perf_output be always per-cpu
authorBrenden Blanco <bblanco@plumgrid.com>
Fri, 6 Nov 2015 21:03:05 +0000 (13:03 -0800)
committerBrenden Blanco <bblanco@plumgrid.com>
Fri, 6 Nov 2015 21:03:05 +0000 (13:03 -0800)
commitaeca6bfc56e3408947bc40fdecc734d6d8b1a825
treedd0e9fa975a923f0bd5506fd9d736219e6f19c94tree | snapshot
parent75982493b74512dc89843fb87bb86b325506e407commit | diff
Let perf_output be always per-cpu

Incoporating feedback from bgregg:
Make per-cpu nature of output buffer be hidden and implicit. As such,
incoporate some rewriting into the definition of the PERF_OUTPUT.
Create two different macros to distinguish the perf_array (hardware
counters) from the perf_output (ring buffer) use cases.

Also, rename perf_output to perf_submit.

Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
examples/tracing/trace_perf_output.py diff | blob | history
src/cc/export/helpers.h diff | blob | history
src/cc/frontends/clang/b_frontend_action.cc diff | blob | history
src/python/bcc/__init__.py diff | blob | history
Domain: System / System Framework;