diff mbox series

[v9,01/16] tracing: move function tracer options to Kconfig

Message ID 20201211184633.3213045-2-samitolvanen@google.com (mailing list archive)
State New, archived
Headers show
Series Add support for Clang LTO | expand

Commit Message

Sami Tolvanen Dec. 11, 2020, 6:46 p.m. UTC 
Move function tracer options to Kconfig to make it easier to add
new methods for generating __mcount_loc, and to make the options
available also when building kernel modules.

Note that FTRACE_MCOUNT_USE_* options are updated on rebuild and
therefore, work even if the .config was generated in a different
environment.

Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
---
 Makefile               | 20 ++++++++------------
 kernel/trace/Kconfig   | 16 ++++++++++++++++
 scripts/Makefile.build |  6 ++----
 3 files changed, 26 insertions(+), 16 deletions(-)

Comments

Guenter Roeck Feb. 24, 2021, 8:17 p.m. UTC | #1 
On Fri, Dec 11, 2020 at 10:46:18AM -0800, Sami Tolvanen wrote:
> Move function tracer options to Kconfig to make it easier to add
> new methods for generating __mcount_loc, and to make the options
> available also when building kernel modules.
> 
> Note that FTRACE_MCOUNT_USE_* options are updated on rebuild and
> therefore, work even if the .config was generated in a different
> environment.
> 
> Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
> Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>

With this patch in place, parisc:allmodconfig no longer builds.

Error log:
Arch parisc is not supported with CONFIG_FTRACE_MCOUNT_RECORD at scripts/recordmcount.pl line 405.
make[2]: *** [scripts/mod/empty.o] Error 2

Due to this problem, CONFIG_FTRACE_MCOUNT_RECORD can no longer be
enabled in parisc builds. Since that is auto-selected by DYNAMIC_FTRACE,
DYNAMIC_FTRACE can no longer be enabled, and with it everything that
depends on it.

Bisect log attached.

Guenter

---
# bad: [414eece95b98b209cef0f49cfcac108fd00b8ced] Merge tag 'clang-lto-v5.12-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux
# good: [b12b47249688915e987a9a2a393b522f86f6b7ab] Merge tag 'powerpc-5.12-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
git bisect start '414eece95b98' 'b12b47249688'
# bad: [f6e1e1d1e149802ed4062fa514c2d184d30aacdf] Merge tag 'gfs2-for-5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2
git bisect bad f6e1e1d1e149802ed4062fa514c2d184d30aacdf
# bad: [79db4d2293eba2ce6265a341bedf6caecad5eeb3] Merge tag 'clang-lto-v5.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux
git bisect bad 79db4d2293eba2ce6265a341bedf6caecad5eeb3
# good: [9d5032f97e9e0655e8c507ab1f43237e31520b00] dt-bindings: mediatek: mt8192: Fix dt_binding_check warning
git bisect good 9d5032f97e9e0655e8c507ab1f43237e31520b00
# good: [f81f213850ca84b3d5e59e17d17acb2ecfc24076] Merge tag 'for-linus-5.12-1' of git://github.com/cminyard/linux-ipmi
git bisect good f81f213850ca84b3d5e59e17d17acb2ecfc24076
# bad: [112b6a8e038d793d016e330f53acb9383ac504b3] arm64: allow LTO to be selected
git bisect bad 112b6a8e038d793d016e330f53acb9383ac504b3
# bad: [3578ad11f3fba07e64c26d8db68cfd3dde28c59e] init: lto: fix PREL32 relocations
git bisect bad 3578ad11f3fba07e64c26d8db68cfd3dde28c59e
# bad: [22d429e75f24d114d99223389d6ba7047e952e32] kbuild: lto: limit inlining
git bisect bad 22d429e75f24d114d99223389d6ba7047e952e32
# bad: [dc5723b02e523b2c4a68667f7e28c65018f7202f] kbuild: add support for Clang LTO
git bisect bad dc5723b02e523b2c4a68667f7e28c65018f7202f
# bad: [3b15cdc15956673ba1551d79bceae471436ac6a9] tracing: move function tracer options to Kconfig
git bisect bad 3b15cdc15956673ba1551d79bceae471436ac6a9
# first bad commit: [3b15cdc15956673ba1551d79bceae471436ac6a9] tracing: move function tracer options to Kconfig
Kees Cook Feb. 24, 2021, 8:38 p.m. UTC | #2 
On Wed, Feb 24, 2021 at 12:17:23PM -0800, Guenter Roeck wrote:
> On Fri, Dec 11, 2020 at 10:46:18AM -0800, Sami Tolvanen wrote:
> > Move function tracer options to Kconfig to make it easier to add
> > new methods for generating __mcount_loc, and to make the options
> > available also when building kernel modules.
> > 
> > Note that FTRACE_MCOUNT_USE_* options are updated on rebuild and
> > therefore, work even if the .config was generated in a different
> > environment.
> > 
> > Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
> > Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
> 
> With this patch in place, parisc:allmodconfig no longer builds.
> 
> Error log:
> Arch parisc is not supported with CONFIG_FTRACE_MCOUNT_RECORD at scripts/recordmcount.pl line 405.
> make[2]: *** [scripts/mod/empty.o] Error 2
> 
> Due to this problem, CONFIG_FTRACE_MCOUNT_RECORD can no longer be
> enabled in parisc builds. Since that is auto-selected by DYNAMIC_FTRACE,
> DYNAMIC_FTRACE can no longer be enabled, and with it everything that
> depends on it.

Ew. Any idea why this didn't show up while it was in linux-next?
Sami Tolvanen Feb. 24, 2021, 8:54 p.m. UTC | #3 
On Wed, Feb 24, 2021 at 12:38 PM Kees Cook <keescook@chromium.org> wrote:
>
> On Wed, Feb 24, 2021 at 12:17:23PM -0800, Guenter Roeck wrote:
> > On Fri, Dec 11, 2020 at 10:46:18AM -0800, Sami Tolvanen wrote:
> > > Move function tracer options to Kconfig to make it easier to add
> > > new methods for generating __mcount_loc, and to make the options
> > > available also when building kernel modules.
> > >
> > > Note that FTRACE_MCOUNT_USE_* options are updated on rebuild and
> > > therefore, work even if the .config was generated in a different
> > > environment.
> > >
> > > Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
> > > Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
> >
> > With this patch in place, parisc:allmodconfig no longer builds.
> >
> > Error log:
> > Arch parisc is not supported with CONFIG_FTRACE_MCOUNT_RECORD at scripts/recordmcount.pl line 405.
> > make[2]: *** [scripts/mod/empty.o] Error 2
> >
> > Due to this problem, CONFIG_FTRACE_MCOUNT_RECORD can no longer be
> > enabled in parisc builds. Since that is auto-selected by DYNAMIC_FTRACE,
> > DYNAMIC_FTRACE can no longer be enabled, and with it everything that
> > depends on it.
>
> Ew. Any idea why this didn't show up while it was in linux-next?

Does anyone build parisc allmodconfig from -next?

parisc seems to always use -fpatchable-function-entry with dynamic
ftrace, so we just need to select
FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY to stop it from defaulting
to recordmcount:

diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig
index ecef9aff9d72..9ee806f68123 100644
--- a/arch/parisc/Kconfig
+++ b/arch/parisc/Kconfig
@@ -60,6 +60,7 @@ config PARISC
        select HAVE_KPROBES
        select HAVE_KRETPROBES
        select HAVE_DYNAMIC_FTRACE if
$(cc-option,-fpatchable-function-entry=1,1)
+       select FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY if HAVE_DYNAMIC_FTRACE
        select HAVE_FTRACE_MCOUNT_RECORD if HAVE_DYNAMIC_FTRACE
        select HAVE_KPROBES_ON_FTRACE
        select HAVE_DYNAMIC_FTRACE_WITH_REGS

I'll send a proper patch shortly.

Sami
Guenter Roeck Feb. 24, 2021, 10:28 p.m. UTC | #4 
On Wed, Feb 24, 2021 at 12:38:54PM -0800, Kees Cook wrote:
> On Wed, Feb 24, 2021 at 12:17:23PM -0800, Guenter Roeck wrote:
> > On Fri, Dec 11, 2020 at 10:46:18AM -0800, Sami Tolvanen wrote:
> > > Move function tracer options to Kconfig to make it easier to add
> > > new methods for generating __mcount_loc, and to make the options
> > > available also when building kernel modules.
> > > 
> > > Note that FTRACE_MCOUNT_USE_* options are updated on rebuild and
> > > therefore, work even if the .config was generated in a different
> > > environment.
> > > 
> > > Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
> > > Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
> > 
> > With this patch in place, parisc:allmodconfig no longer builds.
> > 
> > Error log:
> > Arch parisc is not supported with CONFIG_FTRACE_MCOUNT_RECORD at scripts/recordmcount.pl line 405.
> > make[2]: *** [scripts/mod/empty.o] Error 2
> > 
> > Due to this problem, CONFIG_FTRACE_MCOUNT_RECORD can no longer be
> > enabled in parisc builds. Since that is auto-selected by DYNAMIC_FTRACE,
> > DYNAMIC_FTRACE can no longer be enabled, and with it everything that
> > depends on it.
> 
> Ew. Any idea why this didn't show up while it was in linux-next?
> 

It did, I just wasn't able to bisect it there.

Guenter
Guenter Roeck Feb. 24, 2021, 10:29 p.m. UTC | #5 
On Wed, Feb 24, 2021 at 12:54:15PM -0800, Sami Tolvanen wrote:
> On Wed, Feb 24, 2021 at 12:38 PM Kees Cook <keescook@chromium.org> wrote:
> >
> > On Wed, Feb 24, 2021 at 12:17:23PM -0800, Guenter Roeck wrote:
> > > On Fri, Dec 11, 2020 at 10:46:18AM -0800, Sami Tolvanen wrote:
> > > > Move function tracer options to Kconfig to make it easier to add
> > > > new methods for generating __mcount_loc, and to make the options
> > > > available also when building kernel modules.
> > > >
> > > > Note that FTRACE_MCOUNT_USE_* options are updated on rebuild and
> > > > therefore, work even if the .config was generated in a different
> > > > environment.
> > > >
> > > > Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
> > > > Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
> > >
> > > With this patch in place, parisc:allmodconfig no longer builds.
> > >
> > > Error log:
> > > Arch parisc is not supported with CONFIG_FTRACE_MCOUNT_RECORD at scripts/recordmcount.pl line 405.
> > > make[2]: *** [scripts/mod/empty.o] Error 2
> > >
> > > Due to this problem, CONFIG_FTRACE_MCOUNT_RECORD can no longer be
> > > enabled in parisc builds. Since that is auto-selected by DYNAMIC_FTRACE,
> > > DYNAMIC_FTRACE can no longer be enabled, and with it everything that
> > > depends on it.
> >
> > Ew. Any idea why this didn't show up while it was in linux-next?
> 
> Does anyone build parisc allmodconfig from -next?
> 

https://kerneltests.org/builders/next-parisc-next

Guenter

> parisc seems to always use -fpatchable-function-entry with dynamic
> ftrace, so we just need to select
> FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY to stop it from defaulting
> to recordmcount:
> 
> diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig
> index ecef9aff9d72..9ee806f68123 100644
> --- a/arch/parisc/Kconfig
> +++ b/arch/parisc/Kconfig
> @@ -60,6 +60,7 @@ config PARISC
>         select HAVE_KPROBES
>         select HAVE_KRETPROBES
>         select HAVE_DYNAMIC_FTRACE if
> $(cc-option,-fpatchable-function-entry=1,1)
> +       select FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY if HAVE_DYNAMIC_FTRACE
>         select HAVE_FTRACE_MCOUNT_RECORD if HAVE_DYNAMIC_FTRACE
>         select HAVE_KPROBES_ON_FTRACE
>         select HAVE_DYNAMIC_FTRACE_WITH_REGS
> 
> I'll send a proper patch shortly.
> 
> Sami
Kees Cook Feb. 24, 2021, 10:42 p.m. UTC | #6 
On Wed, Feb 24, 2021 at 02:28:07PM -0800, Guenter Roeck wrote:
> On Wed, Feb 24, 2021 at 12:38:54PM -0800, Kees Cook wrote:
> > On Wed, Feb 24, 2021 at 12:17:23PM -0800, Guenter Roeck wrote:
> > > On Fri, Dec 11, 2020 at 10:46:18AM -0800, Sami Tolvanen wrote:
> > > > Move function tracer options to Kconfig to make it easier to add
> > > > new methods for generating __mcount_loc, and to make the options
> > > > available also when building kernel modules.
> > > > 
> > > > Note that FTRACE_MCOUNT_USE_* options are updated on rebuild and
> > > > therefore, work even if the .config was generated in a different
> > > > environment.
> > > > 
> > > > Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
> > > > Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
> > > 
> > > With this patch in place, parisc:allmodconfig no longer builds.
> > > 
> > > Error log:
> > > Arch parisc is not supported with CONFIG_FTRACE_MCOUNT_RECORD at scripts/recordmcount.pl line 405.
> > > make[2]: *** [scripts/mod/empty.o] Error 2
> > > 
> > > Due to this problem, CONFIG_FTRACE_MCOUNT_RECORD can no longer be
> > > enabled in parisc builds. Since that is auto-selected by DYNAMIC_FTRACE,
> > > DYNAMIC_FTRACE can no longer be enabled, and with it everything that
> > > depends on it.
> > 
> > Ew. Any idea why this didn't show up while it was in linux-next?
> > 
> 
> It did, I just wasn't able to bisect it there.

Ah-ha! Okay, thanks. Sorry it's been broken for so long! I've added
parisc to my local cross builder now.
diff mbox series

Patch

diff --git a/Makefile b/Makefile
index 9ec53d947628..94549264dd5c 100644
--- a/Makefile
+++ b/Makefile
@@ -851,12 +851,8 @@  KBUILD_CFLAGS += $(DEBUG_CFLAGS)
 export DEBUG_CFLAGS
 
 ifdef CONFIG_FUNCTION_TRACER
-ifdef CONFIG_FTRACE_MCOUNT_RECORD
-  # gcc 5 supports generating the mcount tables directly
-  ifeq ($(call cc-option-yn,-mrecord-mcount),y)
-    CC_FLAGS_FTRACE	+= -mrecord-mcount
-    export CC_USING_RECORD_MCOUNT := 1
-  endif
+ifdef CONFIG_FTRACE_MCOUNT_USE_CC
+  CC_FLAGS_FTRACE	+= -mrecord-mcount
   ifdef CONFIG_HAVE_NOP_MCOUNT
     ifeq ($(call cc-option-yn, -mnop-mcount),y)
       CC_FLAGS_FTRACE	+= -mnop-mcount
@@ -864,6 +860,12 @@  ifdef CONFIG_FTRACE_MCOUNT_RECORD
     endif
   endif
 endif
+ifdef CONFIG_FTRACE_MCOUNT_USE_RECORDMCOUNT
+  ifdef CONFIG_HAVE_C_RECORDMCOUNT
+    BUILD_C_RECORDMCOUNT := y
+    export BUILD_C_RECORDMCOUNT
+  endif
+endif
 ifdef CONFIG_HAVE_FENTRY
   ifeq ($(call cc-option-yn, -mfentry),y)
     CC_FLAGS_FTRACE	+= -mfentry
@@ -873,12 +875,6 @@  endif
 export CC_FLAGS_FTRACE
 KBUILD_CFLAGS	+= $(CC_FLAGS_FTRACE) $(CC_FLAGS_USING)
 KBUILD_AFLAGS	+= $(CC_FLAGS_USING)
-ifdef CONFIG_DYNAMIC_FTRACE
-	ifdef CONFIG_HAVE_C_RECORDMCOUNT
-		BUILD_C_RECORDMCOUNT := y
-		export BUILD_C_RECORDMCOUNT
-	endif
-endif
 endif
 
 # We trigger additional mismatches with less inlining
diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index e1bf5228fb69..616a198661d6 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig
@@ -595,6 +595,22 @@  config FTRACE_MCOUNT_RECORD
 	depends on DYNAMIC_FTRACE
 	depends on HAVE_FTRACE_MCOUNT_RECORD
 
+config FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY
+	bool
+	depends on FTRACE_MCOUNT_RECORD
+
+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
+
+config FTRACE_MCOUNT_USE_RECORDMCOUNT
+	def_bool y
+	depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY
+	depends on !FTRACE_MCOUNT_USE_CC
+	depends on FTRACE_MCOUNT_RECORD
+
 config TRACING_MAP
 	bool
 	depends on ARCH_HAVE_NMI_SAFE_CMPXCHG
diff --git a/scripts/Makefile.build b/scripts/Makefile.build
index 4c058f12dd73..22654a463ad8 100644
--- a/scripts/Makefile.build
+++ b/scripts/Makefile.build
@@ -178,8 +178,7 @@  cmd_modversions_c =								\
 	fi
 endif
 
-ifdef CONFIG_FTRACE_MCOUNT_RECORD
-ifndef CC_USING_RECORD_MCOUNT
+ifdef CONFIG_FTRACE_MCOUNT_USE_RECORDMCOUNT
 # compiler will not generate __mcount_loc use recordmcount or recordmcount.pl
 ifdef BUILD_C_RECORDMCOUNT
 ifeq ("$(origin RECORDMCOUNT_WARN)", "command line")
@@ -206,8 +205,7 @@  recordmcount_source := $(srctree)/scripts/recordmcount.pl
 endif # BUILD_C_RECORDMCOUNT
 cmd_record_mcount = $(if $(findstring $(strip $(CC_FLAGS_FTRACE)),$(_c_flags)),	\
 	$(sub_cmd_record_mcount))
-endif # CC_USING_RECORD_MCOUNT
-endif # CONFIG_FTRACE_MCOUNT_RECORD
+endif # CONFIG_FTRACE_MCOUNT_USE_RECORDMCOUNT
 
 ifdef CONFIG_STACK_VALIDATION
 ifneq ($(SKIP_STACK_VALIDATION),1)