ftrace: Make DYNAMIC_FTRACE always enabled for architectures that support it
Subsystems (19)
Steven Rostedt
maintainerRFC PATCH 1/1 ftrace: Make DYNAMIC_FTRACE always enabled for architectures that support it
From: Steven Rostedt <[email protected]>
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.
--- kernel/trace/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/trace/Kconfig +1 -1
@@ -275,7 +275,7 @@ config FUNCTION_TRACE_ARGS funcgraph-args (for the function graph tracer) config DYNAMIC_FTRACE - bool "enable/disable function tracing dynamically" + bool depends on FUNCTION_TRACER depends on HAVE_DYNAMIC_FTRACE default y
Signature · 1 line
2.47.2
Quoted 3 lines · Steven Rostedt
Remove the prompt to pick
DYNAMIC_FTRACEand simply enable it if the architecture supports it.
I don't disagree with removing pointless Kconfig options, but if we do this, we should do it properly.
Because now we have that entirely pointless "HAVE_DYNAMIC_FTRACE*" set of config options. Which with this patch would be just stupid and extra indirection that makes things harder to see.
IOW, with this patch we'd have
(a) architectures that support dynamic ftrace will do
select HAVE_DYNAMIC_FTRACE
select HAVE_DYNAMIC_FTRACE_WITH_REGS
(or whatever combination)
and then
(b) kernel/trace/Kconfig will just turn that into the properDYNAMIC_FTRACE_xyz macros
and when this no longer involves any questions, there is absolutely *zero* reason I can see for that pointless indirection through another config variables.
Yes, yes, there's still the "depends on FUNCTION_TRACER" that technically is a "reason" for that indirection, but it seems to be a reall ybad one.
IOW, I think that if we do this, we should just get rid of the "HAVE_DYNAMIC_FTRACE{_XYZ}" config variables entirely, and just make architectures say
select DYNAMIC_FTRACE if FUNCTION_TRACER
because the "HAVE_xyz" config variables seem to add no actual value, only confusion.
Or am I missing some reason for still having that extra config variable indirection?
Linus
Steven Rostedt
maintainerQuoted 11 lines · On Thu, 3 Jul 2025 09:57:37 -0700 Linus Torvalds…
IOW, I think that if we do this, we should just get rid of the "HAVE_DYNAMIC_FTRACE{_XYZ}" config variables entirely, and just make architectures say
select
DYNAMIC_FTRACEifFUNCTION_TRACERbecause the "
HAVE_xyz" config variables seem to add no actual value, only confusion.Or am I missing some reason for still having that extra config variable indirection?
I always thought the "HAVE_" configs was a way for architectures to state that it supports something but doesn't necessarily enable it. Whereas the not "HAVE_" configs are user selectable.
If I go and make all the architectures have:
select DYNAMIC_FTRACE if FUNCTION_TRACER
it seems to be duplicating the work, where all the architectures need to know the dependencies. To me, that belongs in the generic configs.
The FUNCTION_TRACER config is user selectable, and DYNMIC_FTRACE gets set when the arch supports dynamic ftrace and the user enables function tracing.
Now, I probably could go and try to clean up some of the HAVE_FTRACE_* configs. Ftrace has a lot of code that is tightly coupled to architecture specific code because it relies on trampolines written in assembly. Throughout the years, I added new features and optimizations that required architecture fixes. As a lot of the architectures support ftrace but I have no idea how to update them, I implemented the change in x86 and added a HAVE_FTRACE_* config so that other architectures could have time to update to the new feature too and the old way still works.
I should look to see what hasn't been ported to every architecture that supports ftrace and remove the configs that are now universal.
-- Steve
Steven Rostedt
maintainerQuoted 2 lines · On Thu, 3 Jul 2025 15:26:43 -0400 Steven Rostedt <[email protected]> wrote:
I should look to see what hasn't been ported to every architecture that supports ftrace and remove the configs that are now universal.
HAVE_FTRACE_MCUONT_RECORD looks like a candidate for removal:
$ git grep -e HAVE_FTRACE_MCOUNT -e HAVE_FUNCTION_TRACER arch arch/arm/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD if !XIP_KERNEL arch/arm/Kconfig: select HAVE_FUNCTION_TRACER if !XIP_KERNEL arch/arm64/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/arm64/Kconfig: select HAVE_FUNCTION_TRACER arch/csky/Kconfig: select HAVE_FUNCTION_TRACER arch/csky/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/loongarch/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/loongarch/Kconfig: select HAVE_FUNCTION_TRACER arch/microblaze/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/microblaze/Kconfig: select HAVE_FUNCTION_TRACER arch/mips/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/mips/Kconfig: select HAVE_FUNCTION_TRACER arch/parisc/Kconfig: select HAVE_FUNCTION_TRACER arch/parisc/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD if HAVE_DYNAMIC_FTRACE arch/powerpc/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/powerpc/Kconfig: select HAVE_FUNCTION_TRACER if !COMPILE_TEST && (PPC64 || (PPC32 && CC_IS_GCC)) arch/riscv/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD if !XIP_KERNEL arch/riscv/Kconfig: select HAVE_FUNCTION_TRACER if !XIP_KERNEL arch/s390/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/s390/Kconfig: select HAVE_FUNCTION_TRACER arch/sh/Kconfig: select HAVE_FUNCTION_TRACER arch/sh/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/sparc/Kconfig: select HAVE_FUNCTION_TRACER arch/sparc/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/x86/Kconfig: select HAVE_FTRACE_MCOUNT_RECORD arch/x86/Kconfig: select HAVE_FUNCTION_TRACER arch/xtensa/Kconfig: select HAVE_FUNCTION_TRACER
HAVE_FTRACE_MCOUNT_RECORD only means something if HAVE_FUNCTION_TRACER is set. Arm and riscv have the same dependencies for both.
It is also only meaningful for HAVE_DYNAMIC_FTRACE, so parisc is fine.
But it appears that xtensa is the only one that doesn't select it. But it is also the only arch that doesn't support DYNAMIC_FTRACE either so that's fine. The HAVE_FTRACE_MCOUNT_RECORD can be replaced by HAVE_DYNAMIC_FTRACE, as MCOUNT_RECORD is what is used to create the sections needed for DYNAMIC_FTRACE.
In fact the config has:
config FTRACE_MCOUNT_RECORD
def_bool y
depends on DYNAMIC_FTRACE
depends on HAVE_FTRACE_MCOUNT_RECORD
Now this config is only used for adding a section in vmlinux.lds. Thus, we can remove the MCOUNT_RECORD configs and just use CONFIG_DYNAMIC_FTRACE instead.
Something like below. I haven tested it. I would need to compile every arch. But I think this would work.
-- Steve
Documentation/trace/ftrace-design.rst +4 -8
@@ -238,19 +238,15 @@ You need very few things to get the syscalls tracing in an arch. - Tag this arch as HAVE_SYSCALL_TRACEPOINTS. -HAVE_FTRACE_MCOUNT_RECORD -------------------------- +HAVE_DYNAMIC_FTRACE +------------------- See scripts/recordmcount.pl for more info. Just fill in the arch-specific details for how to locate the addresses of mcount call sites via objdump. This option doesn't make much sense without also implementing dynamic ftrace. - -HAVE_DYNAMIC_FTRACE -------------------- - -You will first need HAVE_FTRACE_MCOUNT_RECORD and HAVE_FUNCTION_TRACER, so -scroll your reader back up if you got over eager. +You will first need HAVE_FUNCTION_TRACER, so scroll your reader back up if you +got over eager. Once those are out of the way, you will need to implement: - asm/ftrace.h:
arch/arm/Kconfig +0 -1
@@ -107,7 +107,6 @@ config ARM select HAVE_EFFICIENT_UNALIGNED_ACCESS if (CPU_V6 || CPU_V6K || CPU_V7) && MMU select HAVE_EXIT_THREAD select HAVE_GUP_FAST if ARM_LPAE - select HAVE_FTRACE_MCOUNT_RECORD if !XIP_KERNEL select HAVE_FUNCTION_ERROR_INJECTION select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACER if !XIP_KERNEL
arch/arm64/Kconfig +0 -1
@@ -223,7 +223,6 @@ config ARM64 select HAVE_EFFICIENT_UNALIGNED_ACCESS select HAVE_GUP_FAST select HAVE_FTRACE_GRAPH_FUNC - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_TRACER select HAVE_FUNCTION_ERROR_INJECTION select HAVE_FUNCTION_GRAPH_FREGS
arch/csky/Kconfig +0 -1
@@ -80,7 +80,6 @@ config CSKY select HAVE_FUNCTION_TRACER select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_ERROR_INJECTION - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_KERNEL_GZIP select HAVE_KERNEL_LZO select HAVE_KERNEL_LZMA
arch/loongarch/Kconfig +0 -1
@@ -144,7 +144,6 @@ config LOONGARCH select HAVE_EXIT_THREAD select HAVE_GUP_FAST select HAVE_FTRACE_GRAPH_FUNC - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_ARG_ACCESS_API select HAVE_FUNCTION_ERROR_INJECTION select HAVE_FUNCTION_GRAPH_FREGS
arch/microblaze/Kconfig +0 -1
@@ -28,7 +28,6 @@ config MICROBLAZE select HAVE_DEBUG_KMEMLEAK select HAVE_DMA_CONTIGUOUS select HAVE_DYNAMIC_FTRACE - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACER select HAVE_PAGE_SIZE_4KB
arch/mips/Kconfig +0 -1
@@ -73,7 +73,6 @@ config MIPS select HAVE_EBPF_JIT if !CPU_MICROMIPS select HAVE_EXIT_THREAD select HAVE_GUP_FAST - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACER select HAVE_GCC_PLUGINS
arch/parisc/Kconfig +0 -1
@@ -81,7 +81,6 @@ config PARISC select HAVE_KPROBES select HAVE_KRETPROBES select HAVE_DYNAMIC_FTRACE if $(cc-option,-fpatchable-function-entry=1,1) - select HAVE_FTRACE_MCOUNT_RECORD if HAVE_DYNAMIC_FTRACE select FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY if DYNAMIC_FTRACE select HAVE_KPROBES_ON_FTRACE select HAVE_DYNAMIC_FTRACE_WITH_REGS
arch/powerpc/Kconfig +0 -1
@@ -246,7 +246,6 @@ config PPC select HAVE_EFFICIENT_UNALIGNED_ACCESS select HAVE_GUP_FAST select HAVE_FTRACE_GRAPH_FUNC - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_ARG_ACCESS_API select HAVE_FUNCTION_DESCRIPTORS if PPC64_ELF_ABI_V1 select HAVE_FUNCTION_ERROR_INJECTION
arch/riscv/Kconfig +0 -1
@@ -158,7 +158,6 @@ config RISCV select HAVE_DYNAMIC_FTRACE_WITH_CALL_OPS if (DYNAMIC_FTRACE_WITH_ARGS && !CFI_CLANG) select HAVE_DYNAMIC_FTRACE_WITH_ARGS if HAVE_DYNAMIC_FTRACE select HAVE_FTRACE_GRAPH_FUNC - select HAVE_FTRACE_MCOUNT_RECORD if !XIP_KERNEL select HAVE_FUNCTION_GRAPH_TRACER if HAVE_DYNAMIC_FTRACE_WITH_ARGS select HAVE_FUNCTION_GRAPH_FREGS select HAVE_FUNCTION_TRACER if !XIP_KERNEL
arch/s390/Kconfig +0 -1
@@ -199,7 +199,6 @@ config S390 select HAVE_GUP_FAST select HAVE_FENTRY select HAVE_FTRACE_GRAPH_FUNC - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_ARG_ACCESS_API select HAVE_FUNCTION_ERROR_INJECTION select HAVE_FUNCTION_GRAPH_FREGS
arch/sh/Kconfig +0 -1
@@ -40,7 +40,6 @@ config SUPERH select HAVE_GUP_FAST if MMU select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACER - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_HW_BREAKPOINT select HAVE_IOREMAP_PROT if MMU && !X2TLB select HAVE_KERNEL_BZIP2
arch/sparc/Kconfig +0 -1
@@ -78,7 +78,6 @@ config SPARC64 select MMU_GATHER_NO_FLUSH_CACHE select HAVE_ARCH_TRANSPARENT_HUGEPAGE select HAVE_DYNAMIC_FTRACE - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_PAGE_SIZE_8KB select HAVE_SYSCALL_TRACEPOINTS select HAVE_CONTEXT_TRACKING_USER
arch/x86/Kconfig +0 -1
@@ -244,7 +244,6 @@ config X86 select HAVE_GUP_FAST select HAVE_FENTRY if X86_64 || DYNAMIC_FTRACE select HAVE_FTRACE_GRAPH_FUNC if HAVE_FUNCTION_GRAPH_TRACER - select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_GRAPH_FREGS if HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_GRAPH_TRACER if X86_32 || (X86_64 && DYNAMIC_FTRACE) select HAVE_FUNCTION_TRACER
include/asm-generic/vmlinux.lds.h +1 -1
@@ -167,7 +167,7 @@ defined(CONFIG_AUTOFDO_CLANG) || defined(CONFIG_PROPELLER_CLANG)#define FTRACE_STUB_HACK #endif -#ifdef CONFIG_FTRACE_MCOUNT_RECORD +#ifdef CONFIG_DYNAMIC_FTRACE /* * The ftrace call sites are logged to a section whose name depends on the * compiler option used. A given kernel image will only use one, AKA
include/linux/ftrace.h +1 -1
@@ -1108,7 +1108,7 @@ static __always_inline unsigned long get_lock_parent_ip(void)# define trace_preempt_off (a0, a1) do { } while (0)#endif -#ifdef CONFIG_FTRACE_MCOUNT_RECORD +#ifdef CONFIG_DYNAMIC_FTRACE extern void ftrace_init (void );#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY #define FTRACE_CALLSITE_SECTION "__patchable_function_entries"
include/linux/kernel.h +3 -3
@@ -373,9 +373,9 @@ ftrace_vprintk(const char *fmt, va_list ap)static inline void ftrace_dump (enum ftrace_dump_mode oops_dump_mode) { }#endif /* CONFIG_TRACING */ -/* Rebuild everything on CONFIG_FTRACE_MCOUNT_RECORD */ -#ifdef CONFIG_FTRACE_MCOUNT_RECORD -# define REBUILD_DUE_TO_FTRACE_MCOUNT_RECORD +/* Rebuild everything on CONFIG_DYNAMIC_FTRACE */ +#ifdef CONFIG_DYNAMIC_FTRACE +# define REBUILD_DUE_TO_DYNAMIC_FTRACE #endif /* Permissions on a sysfs file: you didn't miss the 0 prefix did you? */
include/linux/module.h +1 -1
@@ -539,7 +539,7 @@ struct module {struct trace_eval_map **trace_evals;unsigned int num_trace_evals;#endif -#ifdef CONFIG_FTRACE_MCOUNT_RECORD +#ifdef CONFIG_DYNAMIC_FTRACE unsigned int num_ftrace_callsites;unsigned long *ftrace_callsites;#endif
include/linux/unwind_deferred.h +2 -2
@@ -28,8 +28,8 @@ void unwind_deferred_cancel(struct unwind_work *work); static __always_inline void unwind_reset_info(void) { -if (unlikely (current->unwind_info.id)) - current->unwind_info.id =0 ; +if (unlikely (current->unwind_info.id.id)) + current->unwind_info.id.id =0 ;/* * As unwind_user_faultable() can be called directly and * depends on nr_entries being cleared on exit to user,
kernel/module/main.c +1 -1
@@ -2639,7 +2639,7 @@ static int find_module_sections(struct module *mod, struct load_info *info)sizeof (*mod->trace_bprintk_fmt_start), &mod->num_trace_bprintk_fmt);#endif -#ifdef CONFIG_FTRACE_MCOUNT_RECORD +#ifdef CONFIG_DYNAMIC_FTRACE /* sechdrs[0].sh_size is always zero */ mod->ftrace_callsites = section_objs(info, FTRACE_CALLSITE_SECTION, sizeof(*mod->ftrace_callsites),
kernel/trace/Kconfig +4 -14
@@ -74,11 +74,6 @@ config HAVE_DYNAMIC_FTRACE_NO_PATCHABLE If the architecture generates __patchable_function_entries sections but does not want them included in the ftrace locations. -config HAVE_FTRACE_MCOUNT_RECORD - bool - help - See Documentation/trace/ftrace-design.rst - config HAVE_SYSCALL_TRACEPOINTS bool help
@@ -803,27 +798,22 @@ config BPF_KPROBE_OVERRIDE Allows BPF to override the execution of a probed function and set a different return value. This is used for error injection. -config FTRACE_MCOUNT_RECORD - def_bool y - depends on DYNAMIC_FTRACE - depends on HAVE_FTRACE_MCOUNT_RECORD - config FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY bool - depends on FTRACE_MCOUNT_RECORD + depends on DYNAMIC_FTRACE config FTRACE_MCOUNT_USE_CC def_bool y depends on $(cc-option,-mrecord-mcount) depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY - depends on FTRACE_MCOUNT_RECORD + depends on DYNAMIC_FTRACE config FTRACE_MCOUNT_USE_OBJTOOL def_bool y depends on HAVE_OBJTOOL_MCOUNT depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY depends on !FTRACE_MCOUNT_USE_CC - depends on FTRACE_MCOUNT_RECORD + depends on DYNAMIC_FTRACE select OBJTOOL config FTRACE_MCOUNT_USE_RECORDMCOUNT
@@ -831,7 +821,7 @@ config FTRACE_MCOUNT_USE_RECORDMCOUNT depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY depends on !FTRACE_MCOUNT_USE_CC depends on !FTRACE_MCOUNT_USE_OBJTOOL - depends on FTRACE_MCOUNT_RECORD + depends on DYNAMIC_FTRACE config TRACING_MAP bool
kernel/trace/ftrace.c +0 -4
@@ -1042,10 +1042,6 @@ static struct ftrace_ops *removed_ops;*/ static bool update_all_ops; -#ifndef CONFIG_FTRACE_MCOUNT_RECORD -# error Dynamic ftrace depends on MCOUNT_RECORD -#endif -struct ftrace_func_probe {struct ftrace_probe_ops *probe_ops;struct ftrace_ops ops;
scripts/recordmcount.pl +1 -1
@@ -359,7 +359,7 @@ if ($arch eq "x86_64") { $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_CKCORE_PCREL_JSR_IMM26BY2\\s+_mcount\$"; $alignment = 2; } else { - die "Arch $arch is not supported with CONFIG_FTRACE_MCOUNT_RECORD"; + die "Arch $arch is not supported with CONFIG_DYNAMIC_FTRACE"; } my $text_found = 0;
Steven Rostedt
maintainerQuoted 5 lines · On Thu, 3 Jul 2025 15:49:16 -0400 Steven Rostedt <[email protected]> wrote:
include/linux/unwind_deferred.h +0
-0
include/linux/unwind_deferred.h void unwind_deferred_cancel(struct unwind_work *work); · line 35
Quoted 10 lines
include/linux/unwind_deferred.h +2
-2@@ -28,8 +28,8 @@ void unwind_deferred_cancel(struct unwind_work *work); static __always_inline void unwind_reset_info(void) { -if (unlikely (current->unwind_info.id)) - current->unwind_info.id =0 ; +if (unlikely (current->unwind_info.id.id)) + current->unwind_info.id.id =0 ;/* * As unwind_user_faultable() can be called directly and * depends on nr_entries being cleared on exit to user,
Bah, I did this while fixing the unwind patches.
Oops!
-- Steve
" part
Quoted 4 lines · Steven Rostedt
I always thought the "
HAVE_" configs was a way for architectures to state that it supports something but doesn't necessarily enable it. Whereas the not "HAVE_" configs are user selectable.
I don't actually think we have those kinds of rules in practice.
We have very random Kconfig things. A lot of things are "architecture just selects the feature directly", and the naming is a random collection of ARCH_HAS_abc, ARCH_USE_abc, ARCH_WANT_abc, HAVE_ARCH_abc and HAVE_abc.
And those are the patterns that _try_ to be some kind of pattern, and ignores all the "architecture just selects the feature directly".
Honestly, when we have *five* different syntaxes for "this architecture supports feature 'abc'", I'd argue that not only isn't there a standard way, all those different syntax forms are just confusing.
I get the strong feeling that the reason people have those different prefixes is literally to just group things together alphabetically - ie some people use "ARCH_HAS_xyz" just because that way they can put the "select" statement next to another ARCH_HAS_xyz that they also maintain.
And then in addition to those various "arch has", we have the "GENERIC_xyz" Kconfig entries that architectures also select.
I think those GENERIC_xyz Kconfig options may actually have more of a
pattern: that _tends_ to be about "I'm not implementing my own version, please just pick the generic version". That one is one of our better patterns, I think.
So the reason I dislike the HAVE_xyz pattern is exactly that there _isn't_ a pattern. When there are fifteen different patterns, it's not a pattern at all.
That said, maybe it's better to have one place that has that "if FUNCTION_TRACER, even if I despise the nonsensical "helper indirection" just because of the random naming.
Linus
Steven Rostedt
maintainerQuoted 7 lines · On Thu, 3 Jul 2025 13:58:17 -0700 Linus Torvalds…
So the reason I dislike the
HAVE_xyzpattern is exactly that there _isn't_ a pattern. When there are fifteen different patterns, it's not a pattern at all.That said, maybe it's better to have one place that has that "if
FUNCTION_TRACER, even if I despise the nonsensical "helper indirection" just because of the random naming.
At least with HAVE_FTRACE_* there is a pattern. The HAVE_* may not be consistent across other parts of the kernel, but it has been with ftrace.
As I have stated, ftrace is very tightly coupled with the architectures due to the assembly written trampolines. And having a simple way for the architectures to denote what it supports and what it does not makes the generic code much simpler to implement.
-- Steve