linux.git/kernel/trace/Kconfig, branch v6.18.21

Merge tag 'trace-v6.17' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace

2025-08-01T17:29:36+00:00

Pull tracing updates from Steven Rostedt:

- Deprecate auto-mounting tracefs to /sys/kernel/debug/tracing

When tracefs was first introduced back in 2014, the directory
/sys/kernel/tracing was added and is the designated location to mount
tracefs. To keep backward compatibility, tracefs was auto-mounted in
/sys/kernel/debug/tracing as well.

All distros now mount tracefs on /sys/kernel/tracing. Having it seen
in two different locations has lead to various issues and
inconsistencies.

The VFS folks have to also maintain debugfs_create_automount() for
this single user.

It's been over 10 years. Tooling and scripts should start replacing
the debugfs location with the tracefs one. The reason tracefs was
created in the first place was to allow access to the tracing
facilities without the need to configure debugfs into the kernel.
Using tracefs should now be more robust.

A new config is created: CONFIG_TRACEFS_AUTOMOUNT_DEPRECATED which is
default y, so that the kernel is still built with the automount. This
config allows those that want to remove the automount from debugfs to
do so.

When tracefs is accessed from /sys/kernel/debug/tracing, the
following printk is triggerd:

pr_warn("NOTICE: Automounting of tracing to debugfs is deprecated and will be removed in 2030\n");

This gives users another 5 years to fix their scripts.

- Use queue_rcu_work() instead of call_rcu() for freeing event filters

The number of filters to be free can be many depending on the number
of events within an event system. Freeing them from softirq context
can potentially cause undesired latency. Use the RCU workqueue to
free them instead.

- Remove pointless memory barriers in latency code

Memory barriers were added to some of the latency code a long time
ago with the idea of "making them visible", but that's not what
memory barriers are for. They are to synchronize access between
different variables. There was no synchronization here making them
pointless.

- Remove "__attribute__()" from the type field of event format

When LLVM is used to compile the kernel with CONFIG_DEBUG_INFO_BTF=y
and PAHOLE_HAS_BTF_TAG=y, some of the format fields get expanded with
the following:

field:const char * filename; offset:24; size:8; signed:0;

Turns into:

field:const char __attribute__((btf_type_tag("user"))) * filename; offset:24; size:8; signed:0;

This confuses parsers. Add code to strip these tags from the strings.

- Add eprobe config option CONFIG_EPROBE_EVENTS

Eprobes were added back in 5.15 but were only enabled when another
probe was enabled (kprobe, fprobe, uprobe, etc). The eprobes had no
config option of their own. Add one as they should be a separate
entity.

It's default y to keep with the old kernels but still has
dependencies on TRACING and HAVE_REGS_AND_STACK_ACCESS_API.

- Add eprobe documentation

When eprobes were added back in 5.15 no documentation was added to
describe them. This needs to be rectified.

- Replace open coded cpumask_next_wrap() in move_to_next_cpu()

- Have preemptirq_delay_run() use off-stack CPU mask

- Remove obsolete comment about pelt_cfs event

DECLARE_TRACE() appends "_tp" to trace events now, but the comment
above pelt_cfs still mentioned appending it manually.

- Remove EVENT_FILE_FL_SOFT_MODE flag

The SOFT_MODE flag was required when the soft enabling and disabling
of trace events was first introduced. But there was a bug with this
approach as it only worked for a single instance. When multiple users
required soft disabling and disabling the code was changed to have a
ref count. The SOFT_MODE flag is now set iff the ref count is non
zero. This is redundant and just reading the ref count is good
enough.

- Fix typo in comment

* tag 'trace-v6.17' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
Documentation: tracing: Add documentation about eprobes
tracing: Have eprobes have their own config option
tracing: Remove "__attribute__()" from the type field of event format
tracing: Deprecate auto-mounting tracefs in debugfs
tracing: Fix comment in trace_module_remove_events()
tracing: Remove EVENT_FILE_FL_SOFT_MODE flag
tracing: Remove pointless memory barriers
tracing/sched: Remove obsolete comment on suffixes
kernel: trace: preemptirq_delay_test: use offstack cpu mask
tracing: Use queue_rcu_work() to free filters
tracing: Replace opencoded cpumask_next_wrap() in move_to_next_cpu()

Merge tag 'trace-unused-v6.17' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace

2025-07-30T23:41:58+00:00

Pull tracepoint cleanup from Steven Rostedt:
 "Remove or hide unused tracepoints

  Tracepoints take up memory (around 5K per tracepoint) even when they
  are unused. Changes are being made to detect when a tracepoint is
  defined but unused and a warning is shown at build. But those changes
  are not yet ready for inclusion.

   - Fix some of the unused tracepoints that it detected

     Some tracepoints were removed and others were hidden by config
     settings to match the config settings of where they are
     instantiated. Some tracepoints were moved into architecture
     specific code as only one architecture used them.

   - Call the ftrace_test_filter tracepoint in an unreachable if
     statement

     The ftrace_test_filter tracepoint which is defined when ftrace
     selftests are configured and is used to test the filter logic, but
     the tracepoint is not actually called. It is put into an if
     statement to not have it get compiled out, but also not warn for
     not being used"

* tag 'trace-unused-v6.17' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  tracing: sched: Hide numa events under CONFIG_NUMA_BALANCING
  powerpc/thp: tracing: Hide hugepage events under CONFIG_PPC_BOOK3S_64
  tracing: Call trace_ftrace_test_filter() for the event
  tracing: arm: arm64: Hide trace events ipi_raise, ipi_entry and ipi_exit
  binder: Remove unused binder lock events
  PM: tracing: Hide power_domain_target event under ARCH_OMAP2PLUS
  PM: tracing: Hide device_pm_callback events under PM_SLEEP
  PM: tracing: Hide psci_domain_idle events under ARM_PSCI_CPUIDLE
  PM: cpufreq: powernv/tracing: Move powernv_throttle trace event
  alarmtimer: Hide alarmtimer_suspend event when RTC_CLASS is not configured
  tracing, AER: Hide PCIe AER event when PCIEAER is not configured

tracing: Have eprobes have their own config option

2025-07-30T14:38:43+00:00

Eprobes were added in 5.15 and were selected whenever any of the other
probe events were selected. If kprobe events were enabled (which it is by
default if kprobes are enabled) it would enable eprobe events as well. The
same for uprobes and fprobes.

Have eprobes have its own config and it gets enabled by default if tracing
is enabled.

Link: https://lore.kernel.org/all/20250729102636.b7cce553e7cc263722b12365@kernel.org/

Cc: Mark Rutland 
Cc: Mathieu Desnoyers 
Cc: Andrew Morton 
Cc: Namhyung Kim 
Cc: Jonathan Corbet 
Cc: Randy Dunlap 
Link: https://lore.kernel.org/20250730140945.360286733@kernel.org
Suggested-by: Masami Hiramatsu (Google) 
Acked-by: Masami Hiramatsu (Google) 
Signed-off-by: Steven Rostedt (Google)

tracing: arm: arm64: Hide trace events ipi_raise, ipi_entry and ipi_exit

2025-07-23T18:58:55+00:00

The ipi tracepoints are mostly generic, but the tracepoints ipi_raise,
ipi_entry and ipi_exit are only used by arm and arm64. This means these
trace events are wasting memory in all the other architectures that do not
use them.

Add CONFIG_HAVE_EXTRA_IPI_TRACEPOINTS and have arm and arm64 select it to
enable these trace events. The config makes it easy if other architectures
decide to trace these as well.

Cc: Masami Hiramatsu 
Cc: Mathieu Desnoyers 
Cc: Will Deacon 
Cc: Russell King 
Cc: Mark Rutland 
Cc: Thomas Gleixner 
Cc: Valentin Schneider 
Cc: Nicolas Pitre 
Cc: Peter Zijlstra 
Link: https://lore.kernel.org/20250722103714.64eba013@gandalf.local.home
Signed-off-by: Steven Rostedt (Google) 
Acked-by: Catalin Marinas

tracing: Deprecate auto-mounting tracefs in debugfs

2025-07-23T14:47:26+00:00

In January 2015, tracefs was created to allow access to the tracing
infrastructure without needing to compile in debugfs. When tracefs is
configured, the directory /sys/kernel/tracing will exist and tooling is
expected to use that path to access the tracing infrastructure.

To allow backward compatibility, when debugfs is mounted, it would
automount tracefs in its "tracing" directory so that tooling that had hard
coded /sys/kernel/debug/tracing would still work.

It has been over 10 years since the new interface was introduced, and all
tooling should now be using it. Start the process of deprecating the old
path so that it doesn't need to be maintained anymore.

A new config is added to allow distributions to disable automounting of
tracefs on debugfs.

If /sys/kernel/debug/tracing is accessed, a pr_warn() will trigger stating:

"NOTICE: Automounting of tracing to debugfs is deprecated and will be removed in 2030"

Expect to remove this feature in 5 years (2030).

Cc: 
Cc: Masami Hiramatsu 
Cc: Mathieu Desnoyers 
Cc: Mark Rutland 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: Sebastian Andrzej Siewior 
Cc: Namhyung Kim 
Cc: Linus Torvalds 
Cc: Andrew Morton 
Cc: Greg Kroah-Hartman 
Cc: Al Viro 
Cc: Christian Brauner 
Cc: Jan Kara 
Cc: Arnaldo Carvalho de Melo 
Cc: Frederic Weisbecker 
Cc: Jiri Olsa 
Cc: Ian Rogers 
Link: https://lore.kernel.org/20250722170806.40c068c6@gandalf.local.home
Signed-off-by: Steven Rostedt (Google)

tracing: Remove redundant config HAVE_FTRACE_MCOUNT_RECORD

2025-07-23T00:15:56+00:00

Ftrace is tightly coupled with architecture specific code because it
requires the use of trampolines written in assembly. This means that when
a new feature or optimization is made, it must be done for all
architectures. To simplify the approach, CONFIG_HAVE_FTRACE_* configs are
added to denote which architecture has the new enhancement so that other
architectures can still function until they too have been updated.

The CONFIG_HAVE_FTRACE_MCOUNT was added to help simplify the
DYNAMIC_FTRACE work, but now every architecture that implements
DYNAMIC_FTRACE also has HAVE_FTRACE_MCOUNT set too, making it redundant
with the HAVE_DYNAMIC_FTRACE.

Remove the HAVE_FTRACE_MCOUNT config and use DYNAMIC_FTRACE directly where
applicable.

Link: https://lore.kernel.org/all/20250703154916.48e3ada7@gandalf.local.home/

Cc: Masami Hiramatsu 
Cc: Mathieu Desnoyers 
Cc: Mark Rutland 
Cc: Linus Torvalds 
Link: https://lore.kernel.org/20250704104838.27a18690@gandalf.local.home
Signed-off-by: Steven Rostedt (Google)

ftrace: Make DYNAMIC_FTRACE always enabled for architectures that support it

2025-07-23T00:13:16+00:00

ftrace has two flavors:

 1) static: Where every function always calls the ftrace trampoline

 2) dynamic: Where each function has nops that can be changed on demand to
             jump to the ftrace trampoline when needed.

The static flavor has very high performance overhead and was only created
to make it easier for architectures to implement the dynamic flavor. An
architecture developer can first implement the static ftrace to make sure
the trampolines work before working on the more complicated dynamic aspect
of ftrace. Once the architecture can support dynamic ftrace, there's no
reason to continue to support the static flavor. In fact, the static
flavor tends to bitrot and bugs start to appear in them.

Remove the prompt to pick DYNAMIC_FTRACE and simply enable it if the
architecture supports it.

Link: https://lore.kernel.org/all/f7e12c6d-892e-4ca3-9ef0-fbb524d04a48@ghiti.fr/

Cc: Masami Hiramatsu 
Cc: Mathieu Desnoyers 
Cc: Mark Rutland 
Cc: Linus Torvalds 
Cc: Alexandre Ghiti 
Cc: ChenMiao 
Link: https://lore.kernel.org/20250703115222.2d7c8cd5@batman.local.home
Signed-off-by: Steven Rostedt (Google)

ftrace: Have tracing function args depend on PROBE_EVENTS_BTF_ARGS

2025-04-02T13:50:56+00:00

The option PROBE_EVENTS_BTF_ARGS enables the functions
btf_find_func_proto() and btf_get_func_param() which are used by the
function argument tracing code. The option FUNCTION_TRACE_ARGS was
dependent on the same configs that PROBE_EVENTS_BTF_ARGS was dependent on,
but it was also dependent on PROBE_EVENTS_BTF_ARGS. In fact, if
PROBE_EVENTS_BTF_ARGS is supported then FUNCTION_TRACE_ARGS is supported.

Just make FUNCTION_TRACE_ARGS depend on PROBE_EVENTS_BTF_ARGS.

Cc: Masami Hiramatsu 
Cc: Mathieu Desnoyers 
Cc: Mark Rutland 
Link: https://lore.kernel.org/20250401113601.17fa1129@gandalf.local.home
Fixes: 533c20b062d7c ("ftrace: Add print_function_args()")
Closes: https://lore.kernel.org/all/DB9PR08MB75820599801BAD118D123D7D93AD2@DB9PR08MB7582.eurprd08.prod.outlook.com/
Reported-by: Christian Loehle 
Tested-by: Christian Loehle 
Tested-by: Leon Romanovsky 
Signed-off-by: Steven Rostedt (Google)

ftrace: Add support for function argument to graph tracer

2025-03-04T16:27:23+00:00

Wire up the code to print function arguments in the function graph
tracer. This functionality can be enabled/disabled during runtime with
options/funcgraph-args.

Example usage:

6)              | dummy_xmit [dummy](skb = 0x8887c100, dev = 0x872ca000) {
6)              |   consume_skb(skb = 0x8887c100) {
6)              |     skb_release_head_state(skb = 0x8887c100) {
6)  0.178 us    |       sock_wfree(skb = 0x8887c100)
6)  0.627 us    |     }

Cc: Mark Rutland 
Cc: Mathieu Desnoyers 
Cc: Andrew Morton 
Cc: Paul Walmsley 
Cc: Palmer Dabbelt 
Cc: Albert Ou 
Cc: Guo Ren 
Cc: Donglin Peng 
Cc: Zheng Yejian 
Link: https://lore.kernel.org/20250227185822.810321199@goodmis.org
Reviewed-by: Masami Hiramatsu (Google) 
Co-developed-by: Steven Rostedt (Google) 
Signed-off-by: Sven Schnelle 
Signed-off-by: Steven Rostedt (Google)

ftrace: Add print_function_args()

2025-03-04T16:27:23+00:00

Add a function to decode argument types with the help of BTF. Will
be used to display arguments in the function and function graph
tracer.

It can only handle simply arguments and up to FTRACE_REGS_MAX_ARGS number
of arguments. When it hits a max, it will print ", ...":

   page_to_skb(vi=0xffff8d53842dc980, rq=0xffff8d53843a0800, page=0xfffffc2e04337c00, offset=6160, len=64, truesize=1536, ...)

And if it hits an argument that is not recognized, it will print the raw
value and the type of argument it is:

   make_vfsuid(idmap=0xffffffff87f99db8, fs_userns=0xffffffff87e543c0, kuid=0x0 (STRUCT))
   __pti_set_user_pgtbl(pgdp=0xffff8d5384ab47f8, pgd=0x110e74067 (STRUCT))

Cc: Mark Rutland 
Cc: Mathieu Desnoyers 
Cc: Andrew Morton 
Cc: Paul Walmsley 
Cc: Palmer Dabbelt 
Cc: Albert Ou 
Cc: Guo Ren 
Cc: Donglin Peng 
Cc: Zheng Yejian 
Link: https://lore.kernel.org/20250227185822.639418500@goodmis.org
Reviewed-by: Masami Hiramatsu (Google) 
Co-developed-by: Steven Rostedt (Google) 
Signed-off-by: Sven Schnelle 
Signed-off-by: Steven Rostedt (Google)