[0/4] Refactor the logic for writing trace data in the file

Message

Tzvetomir Stoyanov (VMware) Oct. 8, 2021, 4:15 a.m. UTC

When a trace buffer data are written in the trace file, the buffer
option in the file metadata is updated with the file offset of the
tracing data. Hide this logic into the trace-cmd library.
Added new APIs:
 tracecmd_add_buffer_info()
 tracecmd_write_buffer_info()
Changed APIs:
 tracecmd_append_buffer_cpu_data()
Removed APIs:
 tracecmd_add_buffer_option()

Refactored the internal logic of tracecmd_write_cpu_data() API to be
suitable for upcoming trace file format changes and data compression.
The size and the offset of the trace data is saved in the file right
after the data is written. The old logic calculates the size and offset
in advance, but when the trace data is compressed it is hard to use
that approach.

This patch-set depends on "[PATCH 00/10] Refactor APIs for creating output handler",
should be applied on top of it:
 https://lore.kernel.org/linux-trace-devel/[e-mail]/

Tzvetomir Stoyanov (VMware) (4):
  trace-cmd library: Add internal database for trace buffers
  trace-cmd: Use the new trace buffers API
  trace-cmd library: Helper logic to update the trace buffer offset
  trace-cmd library: Refactor the logic for writing trace data in the
    file

 .../include/private/trace-cmd-private.h       |  10 +-
 lib/trace-cmd/include/trace-cmd-local.h       |  19 ++
 lib/trace-cmd/trace-output.c                  | 319 ++++++++++++------
 tracecmd/trace-listen.c                       |   2 +-
 tracecmd/trace-record.c                       |  18 +-
 5 files changed, 250 insertions(+), 118 deletions(-)