Although the name "ftrace" is used to describe the kernel tracing
infrastructure, in actuality it is really just the function hook
infrastructure. The TRACE_EVENT system is more generic and is used
by tracing, perf and lttng as well as others.
This gets confusing at times, as it is not consistent in the naming.
The kernel should only use "ftrace" when describing the function hook
infrastructure. For trace_events and recording, it really should just
use "trace".
This only changes the internal infrastructure code and to make sure
such a large change has no impact, I tested it by merging this branch
into linux-next and building an allmodconfig. There were no build
failures or warnings due to this change set.
I plan on doing further testing and hopefully get this ready for 4.2.
When my testing is complete, I'll be pushing this into linux-next.
-- Steve
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git
ftrace/rename-ftrace
Head SHA1: d0ee8f4a1f5f3df933cdc9a80f5dea13e2fb5022
Steven Rostedt (Red Hat) (19):
tracing: Rename trace/ftrace.h to trace/trace_events.h
tracing: Move the perf code out of trace_event.h
tracing: Rename ftrace_event.h to trace_events.h
tracing: Rename ftrace_print_*() functions ta trace_print_*()
tracing: Rename (un)register_ftrace_event() to (un)register_trace_event()
tracing: Rename ftrace_event_file to trace_event_file
tracing: Rename ftrace_event_{call,class} to trace_event_{call,class}
tracing: Rename ftrace_event_buffer to trace_event_buffer.
tracing: Rename ftrace_output functions to trace_output
tracing: Rename FTRACE_MAX_EVENT to TRACE_EVENT_TYPE_MAX
tracing: Rename ftrace_event_name() to trace_event_name()
tracing: Rename struct ftrace_subsystem_dir to trace_subsystem_dir
tracing: Rename FTRACE_EVENT_FL_* flags to EVENT_FILE_FL_*
tracing: Rename ftrace_trigger_soft_disabled() to trace_trigger_soft_disabled()
tracing: Rename ftrace_raw_##call event structures to trace_event_raw_##call
tracing: Rename ftrace_data_offset_##call to trace_event_data_offset_##call
tracing: Rename ftrace_event_type_funcs_##call to trace_event_type_funcs_##call
tracing: Rename ftrace_define_fields_##call() to trace_event_define_fields_##call()
tracing: Rename ftrace_get_offsets_##call() to trace_event_get_offsets_##call()
----
arch/x86/kvm/mmutrace.h | 2 +-
arch/x86/kvm/svm.c | 2 +-
arch/x86/kvm/vmx.c | 2 +-
include/linux/ftrace_event.h | 616 ---------------
include/linux/module.h | 2 +-
include/linux/perf_event.h | 2 +-
include/linux/syscalls.h | 12 +-
include/linux/trace_events.h | 616 +++++++++++++++
include/trace/define_trace.h | 3 +-
include/trace/events/power.h | 2 +-
include/trace/ftrace.h | 859 ---------------------
include/trace/perf.h | 350 +++++++++
include/trace/syscall.h | 6 +-
include/trace/trace_events.h | 508 ++++++++++++
kernel/events/core.c | 2 +-
kernel/module.c | 2 +-
kernel/rcu/tiny.c | 2 +-
kernel/rcu/tree.c | 2 +-
kernel/trace/blktrace.c | 4 +-
kernel/trace/ring_buffer.c | 2 +-
kernel/trace/trace.c | 22 +-
kernel/trace/trace.h | 42 +-
kernel/trace/trace_branch.c | 4 +-
kernel/trace/trace_event_perf.c | 20 +-
kernel/trace/trace_events.c | 304 ++++----
kernel/trace/trace_events_filter.c | 86 +--
kernel/trace/trace_events_trigger.c | 70 +-
kernel/trace/trace_export.c | 10 +-
kernel/trace/trace_functions_graph.c | 8 +-
kernel/trace/trace_kprobe.c | 70 +-
kernel/trace/trace_mmiotrace.c | 4 +-
kernel/trace/trace_output.c | 78 +-
kernel/trace/trace_output.h | 2 +-
kernel/trace/trace_probe.h | 8 +-
kernel/trace/trace_sched_wakeup.c | 4 +-
kernel/trace/trace_syscalls.c | 72 +-
kernel/trace/trace_uprobe.c | 46 +-
mm/debug.c | 2 +-
.../perf/util/scripting-engines/trace-event-perl.c | 4 +-
.../util/scripting-engines/trace-event-python.c | 4 +-
40 files changed, 1928 insertions(+), 1928 deletions(-)
Hi Steven,
On 05/14/2015 05:13 PM, Steven Rostedt wrote:
> Although the name "ftrace" is used to describe the kernel tracing
> infrastructure, in actuality it is really just the function hook
> infrastructure. The TRACE_EVENT system is more generic and is used
> by tracing, perf and lttng as well as others.
>
> This gets confusing at times, as it is not consistent in the naming.
> The kernel should only use "ftrace" when describing the function hook
> infrastructure. For trace_events and recording, it really should just
> use "trace".
>
> This only changes the internal infrastructure code and to make sure
> such a large change has no impact, I tested it by merging this branch
> into linux-next and building an allmodconfig. There were no build
> failures or warnings due to this change set.
>
> I plan on doing further testing and hopefully get this ready for 4.2.
> When my testing is complete, I'll be pushing this into linux-next.
I compiled each patch individually and did some smoke testing. No
problems found so far. I don't know if that already classifies for the
Tested-by tag, though feel free adding it.
cheers,
daniel
On Fri, 22 May 2015 14:28:57 +0200
Daniel Wagner <[email protected]> wrote:
> I compiled each patch individually and did some smoke testing. No
> problems found so far. I don't know if that already classifies for the
> Tested-by tag, though feel free adding it.
Thanks, but it has already been added to linux-next, which means no
more rebasing for adding tags :-/
One of my tests (done with ktest) that I do before adding anything
to linux-next is to test each patch individually to make sure it doesn't
add any new warnings or errors, and boots, and also runs a smoke test.
-- Steve