[PATCHv2,2/8] module/ftrace: handle patchable-function-entry

Message ID 20191029165832.33606-3-mark.rutland@arm.com
State New
Headers show
Series
  • arm64: ftrace cleanup + FTRACE_WITH_REGS
Related show

Commit Message

Mark Rutland Oct. 29, 2019, 4:58 p.m.
When using patchable-function-entry, the compiler will record the
callsites into a section named "__patchable_function_entries" rather
than "__mcount_loc". Let's abstract this difference behind a new
FTRACE_CALLSITE_SECTION, so that architectures don't have to handle this
explicitly (e.g. with custom module linker scripts).

As parisc currently handles this explicitly, it is fixed up accordingly,
with its custom linker script removed. Since FTRACE_CALLSITE_SECTION is
only defined when DYNAMIC_FTRACE is selected, the parisc module loading
code is updated to only use the definition in that case. When
DYNAMIC_FTRACE is not selected, modules shouldn't have this section, so
this removes some redundant work in that case.

I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,
and verified that the section made it into the .ko files for modules.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>

Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

Cc: Helge Deller <deller@gmx.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>
Cc: Jessica Yu <jeyu@kernel.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Sven Schnelle <svens@stackframe.org>
Cc: linux-parisc@vger.kernel.org
---
 arch/parisc/Makefile          |  1 -
 arch/parisc/kernel/module.c   | 10 +++++++---
 arch/parisc/kernel/module.lds |  7 -------
 include/linux/ftrace.h        |  5 +++++
 kernel/module.c               |  2 +-
 5 files changed, 13 insertions(+), 12 deletions(-)
 delete mode 100644 arch/parisc/kernel/module.lds

-- 
2.11.0

Comments

Torsten Duwe Oct. 30, 2019, 3:03 p.m. | #1
On Tue, Oct 29, 2019 at 04:58:26PM +0000, Mark Rutland wrote:
> When using patchable-function-entry, the compiler will record the

> callsites into a section named "__patchable_function_entries" rather

> than "__mcount_loc". Let's abstract this difference behind a new

> FTRACE_CALLSITE_SECTION, so that architectures don't have to handle this

> explicitly (e.g. with custom module linker scripts).

> 

> As parisc currently handles this explicitly, it is fixed up accordingly,

> with its custom linker script removed. Since FTRACE_CALLSITE_SECTION is

> only defined when DYNAMIC_FTRACE is selected, the parisc module loading

> code is updated to only use the definition in that case. When

> DYNAMIC_FTRACE is not selected, modules shouldn't have this section, so

> this removes some redundant work in that case.

> 

> I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,

> and verified that the section made it into the .ko files for modules.


This is because of remaining #ifdeffery in include/asm-generic/vmlinux.lds.h:

#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY
#define MCOUNT_REC()    . = ALIGN(8);                           \
                        __start_mcount_loc = .;                 \
                        KEEP(*(__patchable_function_entries))   \
                        __stop_mcount_loc = .;
#else
#define MCOUNT_REC()    . = ALIGN(8);                           \
                        __start_mcount_loc = .;                 \
                        KEEP(*(__mcount_loc))                   \
                        __stop_mcount_loc = .;
#endif

Maybe you want to tackle that as well? I suggest to have at least one
FTRACE_CALLSITE_SECTION definition without double quotes. Alternatively, my
earlier solution just kept both sections, in case either one or both are
present.

                        KEEP(*(__patchable_function_entries))   \
                        KEEP(*(__mcount_loc))                   \

	Torsten
Mark Rutland Oct. 31, 2019, 9:02 a.m. | #2
On Wed, Oct 30, 2019 at 04:03:02PM +0100, Torsten Duwe wrote:
> On Tue, Oct 29, 2019 at 04:58:26PM +0000, Mark Rutland wrote:

> > When using patchable-function-entry, the compiler will record the

> > callsites into a section named "__patchable_function_entries" rather

> > than "__mcount_loc". Let's abstract this difference behind a new

> > FTRACE_CALLSITE_SECTION, so that architectures don't have to handle this

> > explicitly (e.g. with custom module linker scripts).

> > 

> > As parisc currently handles this explicitly, it is fixed up accordingly,

> > with its custom linker script removed. Since FTRACE_CALLSITE_SECTION is

> > only defined when DYNAMIC_FTRACE is selected, the parisc module loading

> > code is updated to only use the definition in that case. When

> > DYNAMIC_FTRACE is not selected, modules shouldn't have this section, so

> > this removes some redundant work in that case.

> > 

> > I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,

> > and verified that the section made it into the .ko files for modules.

> 

> This is because of remaining #ifdeffery in include/asm-generic/vmlinux.lds.h:

> 

> #ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY

> #define MCOUNT_REC()    . = ALIGN(8);                           \

>                         __start_mcount_loc = .;                 \

>                         KEEP(*(__patchable_function_entries))   \

>                         __stop_mcount_loc = .;

> #else

> #define MCOUNT_REC()    . = ALIGN(8);                           \

>                         __start_mcount_loc = .;                 \

>                         KEEP(*(__mcount_loc))                   \

>                         __stop_mcount_loc = .;

> #endif


For modules we use a combination of scripts/module-common.lds and an
architecture's own module.lds, not vmlinux.lds.h. So I don't think the above is
relevant for modules.

For modules the kernel's ELF loader looks for the ELF ection, not the
__{start,stop}_mcount_loc symbols that we use for the main kernel image.

FWIW, when building a module, I see the following linker operations:

| [mark@blommer:~/src/linux]% toolchain korg gcc-8.1.0-nolibc make V=1 ARCH=arm64 CROSS_COMPILE=aarch64-linux- arch/arm64/crypto/sha512-ce.ko | grep aarch64-linux-ld
|   aarch64-linux-ld  -EL  -maarch64elf    -r -o arch/arm64/crypto/sha512-ce.o arch/arm64/crypto/sha512-ce-glue.o arch/arm64/crypto/sha512-ce-core.o
|   aarch64-linux-ld -r  -EL  -maarch64elf  --build-id  -T ./scripts/module-common.lds -T ./arch/arm64/kernel/module.lds -o arch/arm64/crypto/sha512-ce.ko arch/arm64/crypto/sha512-ce.o arch/arm64/crypto/sha512-ce.mod.o;  true

> Maybe you want to tackle that as well? I suggest to have at least one

> FTRACE_CALLSITE_SECTION definition without double quotes. Alternatively, my

> earlier solution just kept both sections, in case either one or both are

> present.

> 

>                         KEEP(*(__patchable_function_entries))   \

>                         KEEP(*(__mcount_loc))                   \


I agree that the CC_USING_PATCHABLE_FUNCTION_ENTRY ifdeffery could be
simplified, and that it would be nice to consistently use
FTRACE_CALLSITE_SECTION if we can. However, the generic linker script doesn't
include anything, and I don't see a good location for that to live.

What I could do is add an explicit comment:

/*
 * 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
 * FTRACE_CALLSITE_SECTION. We capture all of them here to avoid header
 * dependencies.
 */
#define MCOUNT_REC()						\
			. = ALIGN(8);				\
			__start_mcount_loc = .;			\
			KEEP(*(__patchable_function_entries))   \
			KEEP(*(__mcount_loc))                   \
			__stop_mcount_loc = .;			

... which should make the dependency clear. Does that sound good to you?

Thanks,
Mark.
Torsten Duwe Oct. 31, 2019, 11:42 a.m. | #3
On Thu, Oct 31, 2019 at 09:02:32AM +0000, Mark Rutland wrote:
> On Wed, Oct 30, 2019 at 04:03:02PM +0100, Torsten Duwe wrote:

> > On Tue, Oct 29, 2019 at 04:58:26PM +0000, Mark Rutland wrote:

> > > 

> > > I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,

> > > and verified that the section made it into the .ko files for modules.

> > 

> > This is because of remaining #ifdeffery in include/asm-generic/vmlinux.lds.h:

> > 

> > #ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY

> > #define MCOUNT_REC()    . = ALIGN(8);                           \

> >                         __start_mcount_loc = .;                 \

> >                         KEEP(*(__patchable_function_entries))   \

> >                         __stop_mcount_loc = .;

> > #else

> > #define MCOUNT_REC()    . = ALIGN(8);                           \

> >                         __start_mcount_loc = .;                 \

> >                         KEEP(*(__mcount_loc))                   \

> >                         __stop_mcount_loc = .;

> > #endif

> 

> For modules we use a combination of scripts/module-common.lds and an

> architecture's own module.lds, not vmlinux.lds.h. So I don't think the above is

> relevant for modules.


Sure, this is only loosely related,...

> I agree that the CC_USING_PATCHABLE_FUNCTION_ENTRY ifdeffery could be

> simplified, and that it would be nice to consistently use

> FTRACE_CALLSITE_SECTION if we can. However, the generic linker script doesn't

> include anything, and I don't see a good location for that to live.

> 

> What I could do is add an explicit comment:

> 

> /*

>  * 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

>  * FTRACE_CALLSITE_SECTION. We capture all of them here to avoid header

>  * dependencies.

>  */

> #define MCOUNT_REC()						\

> 			. = ALIGN(8);				\

> 			__start_mcount_loc = .;			\

> 			KEEP(*(__patchable_function_entries))   \

> 			KEEP(*(__mcount_loc))                   \

> 			__stop_mcount_loc = .;			

> 

> ... which should make the dependency clear. Does that sound good to you?


Beautiful. I just didn't want to miss the opportunity to have this cleaned
up as well, and deemed this patch "closest" because of the definition of 
FTRACE_CALLSITE_SECTION. Put it where you see it fit best.

Thanks,

	Torsten
Mark Rutland Oct. 31, 2019, 1 p.m. | #4
On Thu, Oct 31, 2019 at 12:42:23PM +0100, Torsten Duwe wrote:
> On Thu, Oct 31, 2019 at 09:02:32AM +0000, Mark Rutland wrote:

> > I agree that the CC_USING_PATCHABLE_FUNCTION_ENTRY ifdeffery could be

> > simplified, and that it would be nice to consistently use

> > FTRACE_CALLSITE_SECTION if we can. However, the generic linker script doesn't

> > include anything, and I don't see a good location for that to live.

> > 

> > What I could do is add an explicit comment:

> > 

> > /*

> >  * 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

> >  * FTRACE_CALLSITE_SECTION. We capture all of them here to avoid header

> >  * dependencies.

> >  */

> > #define MCOUNT_REC()						\

> > 			. = ALIGN(8);				\

> > 			__start_mcount_loc = .;			\

> > 			KEEP(*(__patchable_function_entries))   \

> > 			KEEP(*(__mcount_loc))                   \

> > 			__stop_mcount_loc = .;			

> > 

> > ... which should make the dependency clear. Does that sound good to you?

> 

> Beautiful. I just didn't want to miss the opportunity to have this cleaned

> up as well, and deemed this patch "closest" because of the definition of 

> FTRACE_CALLSITE_SECTION. Put it where you see it fit best.


Sure. I've folded the above into this patch, and pushed out an updated branch:

  https://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git/log/?h=arm64/ftrace-with-regs

Thanks,
Mark.
Steven Rostedt Nov. 4, 2019, 1:25 p.m. | #5
On Tue, 29 Oct 2019 16:58:26 +0000
Mark Rutland <mark.rutland@arm.com> wrote:

> When using patchable-function-entry, the compiler will record the

> callsites into a section named "__patchable_function_entries" rather

> than "__mcount_loc". Let's abstract this difference behind a new

> FTRACE_CALLSITE_SECTION, so that architectures don't have to handle this

> explicitly (e.g. with custom module linker scripts).

> 

> As parisc currently handles this explicitly, it is fixed up accordingly,

> with its custom linker script removed. Since FTRACE_CALLSITE_SECTION is

> only defined when DYNAMIC_FTRACE is selected, the parisc module loading

> code is updated to only use the definition in that case. When

> DYNAMIC_FTRACE is not selected, modules shouldn't have this section, so

> this removes some redundant work in that case.

> 

> I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,

> and verified that the section made it into the .ko files for modules.

> 

> Signed-off-by: Mark Rutland <mark.rutland@arm.com>

> Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

> Cc: Helge Deller <deller@gmx.de>

> Cc: Ingo Molnar <mingo@redhat.com>

> Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>

> Cc: Jessica Yu <jeyu@kernel.org>

> Cc: Steven Rostedt <rostedt@goodmis.org>

> Cc: Sven Schnelle <svens@stackframe.org>

> Cc: linux-parisc@vger.kernel.org

> 


Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>


-- Steve
Steven Rostedt Nov. 4, 2019, 1:28 p.m. | #6
On Thu, 31 Oct 2019 13:00:22 +0000
Mark Rutland <mark.rutland@arm.com> wrote:

> Sure. I've folded the above into this patch, and pushed out an updated branch:

> 

>   https://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git/log/?h=arm64/ftrace-with-regs


Just to keep this change in lore, can you at a minimum reply to this
patch's thread with the new update?

Thanks!

-- Steve
Mark Rutland Nov. 4, 2019, 2 p.m. | #7
On Mon, Nov 04, 2019 at 08:28:10AM -0500, Steven Rostedt wrote:
> On Thu, 31 Oct 2019 13:00:22 +0000

> Mark Rutland <mark.rutland@arm.com> wrote:

> 

> > Sure. I've folded the above into this patch, and pushed out an updated branch:

> > 

> >   https://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git/log/?h=arm64/ftrace-with-regs

> 

> Just to keep this change in lore, can you at a minimum reply to this

> patch's thread with the new update?


The new change is below (with all else unchanged). I can send a v3 of
the series if you want the whole patch?

Thanks,
Mark.

diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index dae64600ccbf..a9c4e4721434 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -110,17 +110,17 @@
 #endif
 
 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
-#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY
-#define MCOUNT_REC()   . = ALIGN(8);                           \
-                       __start_mcount_loc = .;                 \
-                       KEEP(*(__patchable_function_entries))   \
-                       __stop_mcount_loc = .;
-#else
+/*
+ * 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
+ * FTRACE_CALLSITE_SECTION. We capture all of them here to avoid header
+ * dependencies for FTRACE_CALLSITE_SECTION's definition.
+ */
 #define MCOUNT_REC()   . = ALIGN(8);                           \
                        __start_mcount_loc = .;                 \
                        KEEP(*(__mcount_loc))                   \
+                       KEEP(*(__patchable_function_entries))   \
                        __stop_mcount_loc = .;
-#endif
 #else
 #define MCOUNT_REC()
 #endif
Mark Rutland Nov. 4, 2019, 3:51 p.m. | #8
Hi Jessica, Helge,

Are you ok with the module and parisc changes, repectively?

The kbuild test robot is happy building this for multiple architectures,
Sven has tested that this works correctly on parisc, and others have
tested other architectures.

I'd like to queue this in the arm64 tree soon if possible.

Thanks,
Mark.

On Tue, Oct 29, 2019 at 04:58:26PM +0000, Mark Rutland wrote:
> When using patchable-function-entry, the compiler will record the

> callsites into a section named "__patchable_function_entries" rather

> than "__mcount_loc". Let's abstract this difference behind a new

> FTRACE_CALLSITE_SECTION, so that architectures don't have to handle this

> explicitly (e.g. with custom module linker scripts).

> 

> As parisc currently handles this explicitly, it is fixed up accordingly,

> with its custom linker script removed. Since FTRACE_CALLSITE_SECTION is

> only defined when DYNAMIC_FTRACE is selected, the parisc module loading

> code is updated to only use the definition in that case. When

> DYNAMIC_FTRACE is not selected, modules shouldn't have this section, so

> this removes some redundant work in that case.

> 

> I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,

> and verified that the section made it into the .ko files for modules.

> 

> Signed-off-by: Mark Rutland <mark.rutland@arm.com>

> Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

> Cc: Helge Deller <deller@gmx.de>

> Cc: Ingo Molnar <mingo@redhat.com>

> Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>

> Cc: Jessica Yu <jeyu@kernel.org>

> Cc: Steven Rostedt <rostedt@goodmis.org>

> Cc: Sven Schnelle <svens@stackframe.org>

> Cc: linux-parisc@vger.kernel.org

> ---

>  arch/parisc/Makefile          |  1 -

>  arch/parisc/kernel/module.c   | 10 +++++++---

>  arch/parisc/kernel/module.lds |  7 -------

>  include/linux/ftrace.h        |  5 +++++

>  kernel/module.c               |  2 +-

>  5 files changed, 13 insertions(+), 12 deletions(-)

>  delete mode 100644 arch/parisc/kernel/module.lds

> 

> diff --git a/arch/parisc/Makefile b/arch/parisc/Makefile

> index 36b834f1c933..dca8f2de8cf5 100644

> --- a/arch/parisc/Makefile

> +++ b/arch/parisc/Makefile

> @@ -60,7 +60,6 @@ KBUILD_CFLAGS += -DCC_USING_PATCHABLE_FUNCTION_ENTRY=1 \

>  		 -DFTRACE_PATCHABLE_FUNCTION_SIZE=$(NOP_COUNT)

>  

>  CC_FLAGS_FTRACE := -fpatchable-function-entry=$(NOP_COUNT),$(shell echo $$(($(NOP_COUNT)-1)))

> -KBUILD_LDS_MODULE += $(srctree)/arch/parisc/kernel/module.lds

>  endif

>  

>  OBJCOPY_FLAGS =-O binary -R .note -R .comment -S

> diff --git a/arch/parisc/kernel/module.c b/arch/parisc/kernel/module.c

> index ac5f34993b53..1c50093e2ebe 100644

> --- a/arch/parisc/kernel/module.c

> +++ b/arch/parisc/kernel/module.c

> @@ -43,6 +43,7 @@

>  #include <linux/elf.h>

>  #include <linux/vmalloc.h>

>  #include <linux/fs.h>

> +#include <linux/ftrace.h>

>  #include <linux/string.h>

>  #include <linux/kernel.h>

>  #include <linux/bug.h>

> @@ -862,7 +863,7 @@ int module_finalize(const Elf_Ehdr *hdr,

>  	const char *strtab = NULL;

>  	const Elf_Shdr *s;

>  	char *secstrings;

> -	int err, symindex = -1;

> +	int symindex = -1;

>  	Elf_Sym *newptr, *oldptr;

>  	Elf_Shdr *symhdr = NULL;

>  #ifdef DEBUG

> @@ -946,11 +947,13 @@ int module_finalize(const Elf_Ehdr *hdr,

>  			/* patch .altinstructions */

>  			apply_alternatives(aseg, aseg + s->sh_size, me->name);

>  

> +#ifdef CONFIG_DYNAMIC_FTRACE

>  		/* For 32 bit kernels we're compiling modules with

>  		 * -ffunction-sections so we must relocate the addresses in the

> -		 *__mcount_loc section.

> +		 *  ftrace callsite section.

>  		 */

> -		if (symindex != -1 && !strcmp(secname, "__mcount_loc")) {

> +		if (symindex != -1 && !strcmp(secname, FTRACE_CALLSITE_SECTION)) {

> +			int err;

>  			if (s->sh_type == SHT_REL)

>  				err = apply_relocate((Elf_Shdr *)sechdrs,

>  							strtab, symindex,

> @@ -962,6 +965,7 @@ int module_finalize(const Elf_Ehdr *hdr,

>  			if (err)

>  				return err;

>  		}

> +#endif

>  	}

>  	return 0;

>  }

> diff --git a/arch/parisc/kernel/module.lds b/arch/parisc/kernel/module.lds

> deleted file mode 100644

> index 1a9a92aca5c8..000000000000

> --- a/arch/parisc/kernel/module.lds

> +++ /dev/null

> @@ -1,7 +0,0 @@

> -/* SPDX-License-Identifier: GPL-2.0 */

> -

> -SECTIONS {

> -	__mcount_loc : {

> -		*(__patchable_function_entries)

> -	}

> -}

> diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h

> index 9867d90d635e..9141f2263286 100644

> --- a/include/linux/ftrace.h

> +++ b/include/linux/ftrace.h

> @@ -738,6 +738,11 @@ static inline unsigned long get_lock_parent_ip(void)

>  

>  #ifdef CONFIG_FTRACE_MCOUNT_RECORD

>  extern void ftrace_init(void);

> +#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY

> +#define FTRACE_CALLSITE_SECTION	"__patchable_function_entries"

> +#else

> +#define FTRACE_CALLSITE_SECTION	"__mcount_loc"

> +#endif

>  #else

>  static inline void ftrace_init(void) { }

>  #endif

> diff --git a/kernel/module.c b/kernel/module.c

> index ff2d7359a418..acf7962936c4 100644

> --- a/kernel/module.c

> +++ b/kernel/module.c

> @@ -3222,7 +3222,7 @@ static int find_module_sections(struct module *mod, struct load_info *info)

>  #endif

>  #ifdef CONFIG_FTRACE_MCOUNT_RECORD

>  	/* sechdrs[0].sh_size is always zero */

> -	mod->ftrace_callsites = section_objs(info, "__mcount_loc",

> +	mod->ftrace_callsites = section_objs(info, FTRACE_CALLSITE_SECTION,

>  					     sizeof(*mod->ftrace_callsites),

>  					     &mod->num_ftrace_callsites);

>  #endif

> -- 

> 2.11.0

>
Helge Deller Nov. 4, 2019, 8:58 p.m. | #9
Hi Mark,

On 04.11.19 16:51, Mark Rutland wrote:
> Hi Jessica, Helge,

>

> Are you ok with the module and parisc changes, repectively?


Sure, please add my:
Acked-by: Helge Deller <deller@gmx.de>


Helge

> The kbuild test robot is happy building this for multiple architectures,

> Sven has tested that this works correctly on parisc, and others have

> tested other architectures.

>

> I'd like to queue this in the arm64 tree soon if possible.

>

> Thanks,

> Mark.

>

> On Tue, Oct 29, 2019 at 04:58:26PM +0000, Mark Rutland wrote:

>> When using patchable-function-entry, the compiler will record the

>> callsites into a section named "__patchable_function_entries" rather

>> than "__mcount_loc". Let's abstract this difference behind a new

>> FTRACE_CALLSITE_SECTION, so that architectures don't have to handle this

>> explicitly (e.g. with custom module linker scripts).

>>

>> As parisc currently handles this explicitly, it is fixed up accordingly,

>> with its custom linker script removed. Since FTRACE_CALLSITE_SECTION is

>> only defined when DYNAMIC_FTRACE is selected, the parisc module loading

>> code is updated to only use the definition in that case. When

>> DYNAMIC_FTRACE is not selected, modules shouldn't have this section, so

>> this removes some redundant work in that case.

>>

>> I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,

>> and verified that the section made it into the .ko files for modules.

>>

>> Signed-off-by: Mark Rutland <mark.rutland@arm.com>

>> Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

>> Cc: Helge Deller <deller@gmx.de>

>> Cc: Ingo Molnar <mingo@redhat.com>

>> Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>

>> Cc: Jessica Yu <jeyu@kernel.org>

>> Cc: Steven Rostedt <rostedt@goodmis.org>

>> Cc: Sven Schnelle <svens@stackframe.org>

>> Cc: linux-parisc@vger.kernel.org

>> ---

>>  arch/parisc/Makefile          |  1 -

>>  arch/parisc/kernel/module.c   | 10 +++++++---

>>  arch/parisc/kernel/module.lds |  7 -------

>>  include/linux/ftrace.h        |  5 +++++

>>  kernel/module.c               |  2 +-

>>  5 files changed, 13 insertions(+), 12 deletions(-)

>>  delete mode 100644 arch/parisc/kernel/module.lds

>>

>> diff --git a/arch/parisc/Makefile b/arch/parisc/Makefile

>> index 36b834f1c933..dca8f2de8cf5 100644

>> --- a/arch/parisc/Makefile

>> +++ b/arch/parisc/Makefile

>> @@ -60,7 +60,6 @@ KBUILD_CFLAGS += -DCC_USING_PATCHABLE_FUNCTION_ENTRY=1 \

>>  		 -DFTRACE_PATCHABLE_FUNCTION_SIZE=$(NOP_COUNT)

>>

>>  CC_FLAGS_FTRACE := -fpatchable-function-entry=$(NOP_COUNT),$(shell echo $$(($(NOP_COUNT)-1)))

>> -KBUILD_LDS_MODULE += $(srctree)/arch/parisc/kernel/module.lds

>>  endif

>>

>>  OBJCOPY_FLAGS =-O binary -R .note -R .comment -S

>> diff --git a/arch/parisc/kernel/module.c b/arch/parisc/kernel/module.c

>> index ac5f34993b53..1c50093e2ebe 100644

>> --- a/arch/parisc/kernel/module.c

>> +++ b/arch/parisc/kernel/module.c

>> @@ -43,6 +43,7 @@

>>  #include <linux/elf.h>

>>  #include <linux/vmalloc.h>

>>  #include <linux/fs.h>

>> +#include <linux/ftrace.h>

>>  #include <linux/string.h>

>>  #include <linux/kernel.h>

>>  #include <linux/bug.h>

>> @@ -862,7 +863,7 @@ int module_finalize(const Elf_Ehdr *hdr,

>>  	const char *strtab = NULL;

>>  	const Elf_Shdr *s;

>>  	char *secstrings;

>> -	int err, symindex = -1;

>> +	int symindex = -1;

>>  	Elf_Sym *newptr, *oldptr;

>>  	Elf_Shdr *symhdr = NULL;

>>  #ifdef DEBUG

>> @@ -946,11 +947,13 @@ int module_finalize(const Elf_Ehdr *hdr,

>>  			/* patch .altinstructions */

>>  			apply_alternatives(aseg, aseg + s->sh_size, me->name);

>>

>> +#ifdef CONFIG_DYNAMIC_FTRACE

>>  		/* For 32 bit kernels we're compiling modules with

>>  		 * -ffunction-sections so we must relocate the addresses in the

>> -		 *__mcount_loc section.

>> +		 *  ftrace callsite section.

>>  		 */

>> -		if (symindex != -1 && !strcmp(secname, "__mcount_loc")) {

>> +		if (symindex != -1 && !strcmp(secname, FTRACE_CALLSITE_SECTION)) {

>> +			int err;

>>  			if (s->sh_type == SHT_REL)

>>  				err = apply_relocate((Elf_Shdr *)sechdrs,

>>  							strtab, symindex,

>> @@ -962,6 +965,7 @@ int module_finalize(const Elf_Ehdr *hdr,

>>  			if (err)

>>  				return err;

>>  		}

>> +#endif

>>  	}

>>  	return 0;

>>  }

>> diff --git a/arch/parisc/kernel/module.lds b/arch/parisc/kernel/module.lds

>> deleted file mode 100644

>> index 1a9a92aca5c8..000000000000

>> --- a/arch/parisc/kernel/module.lds

>> +++ /dev/null

>> @@ -1,7 +0,0 @@

>> -/* SPDX-License-Identifier: GPL-2.0 */

>> -

>> -SECTIONS {

>> -	__mcount_loc : {

>> -		*(__patchable_function_entries)

>> -	}

>> -}

>> diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h

>> index 9867d90d635e..9141f2263286 100644

>> --- a/include/linux/ftrace.h

>> +++ b/include/linux/ftrace.h

>> @@ -738,6 +738,11 @@ static inline unsigned long get_lock_parent_ip(void)

>>

>>  #ifdef CONFIG_FTRACE_MCOUNT_RECORD

>>  extern void ftrace_init(void);

>> +#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY

>> +#define FTRACE_CALLSITE_SECTION	"__patchable_function_entries"

>> +#else

>> +#define FTRACE_CALLSITE_SECTION	"__mcount_loc"

>> +#endif

>>  #else

>>  static inline void ftrace_init(void) { }

>>  #endif

>> diff --git a/kernel/module.c b/kernel/module.c

>> index ff2d7359a418..acf7962936c4 100644

>> --- a/kernel/module.c

>> +++ b/kernel/module.c

>> @@ -3222,7 +3222,7 @@ static int find_module_sections(struct module *mod, struct load_info *info)

>>  #endif

>>  #ifdef CONFIG_FTRACE_MCOUNT_RECORD

>>  	/* sechdrs[0].sh_size is always zero */

>> -	mod->ftrace_callsites = section_objs(info, "__mcount_loc",

>> +	mod->ftrace_callsites = section_objs(info, FTRACE_CALLSITE_SECTION,

>>  					     sizeof(*mod->ftrace_callsites),

>>  					     &mod->num_ftrace_callsites);

>>  #endif

>> --

>> 2.11.0

>>
Miroslav Benes Nov. 5, 2019, 8:59 a.m. | #10
On Tue, 29 Oct 2019, Mark Rutland wrote:

> When using patchable-function-entry, the compiler will record the

> callsites into a section named "__patchable_function_entries" rather

> than "__mcount_loc". Let's abstract this difference behind a new

> FTRACE_CALLSITE_SECTION, so that architectures don't have to handle this

> explicitly (e.g. with custom module linker scripts).

> 

> As parisc currently handles this explicitly, it is fixed up accordingly,

> with its custom linker script removed. Since FTRACE_CALLSITE_SECTION is

> only defined when DYNAMIC_FTRACE is selected, the parisc module loading

> code is updated to only use the definition in that case. When

> DYNAMIC_FTRACE is not selected, modules shouldn't have this section, so

> this removes some redundant work in that case.

> 

> I built parisc generic-{32,64}bit_defconfig with DYNAMIC_FTRACE enabled,

> and verified that the section made it into the .ko files for modules.

> 

> Signed-off-by: Mark Rutland <mark.rutland@arm.com>

> Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

> Cc: Helge Deller <deller@gmx.de>

> Cc: Ingo Molnar <mingo@redhat.com>

> Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>

> Cc: Jessica Yu <jeyu@kernel.org>

> Cc: Steven Rostedt <rostedt@goodmis.org>

> Cc: Sven Schnelle <svens@stackframe.org>

> Cc: linux-parisc@vger.kernel.org


For the updated patch

Reviewed-by: Miroslav Benes <mbenes@suse.cz>


M

Patch

diff --git a/arch/parisc/Makefile b/arch/parisc/Makefile
index 36b834f1c933..dca8f2de8cf5 100644
--- a/arch/parisc/Makefile
+++ b/arch/parisc/Makefile
@@ -60,7 +60,6 @@  KBUILD_CFLAGS += -DCC_USING_PATCHABLE_FUNCTION_ENTRY=1 \
 		 -DFTRACE_PATCHABLE_FUNCTION_SIZE=$(NOP_COUNT)
 
 CC_FLAGS_FTRACE := -fpatchable-function-entry=$(NOP_COUNT),$(shell echo $$(($(NOP_COUNT)-1)))
-KBUILD_LDS_MODULE += $(srctree)/arch/parisc/kernel/module.lds
 endif
 
 OBJCOPY_FLAGS =-O binary -R .note -R .comment -S
diff --git a/arch/parisc/kernel/module.c b/arch/parisc/kernel/module.c
index ac5f34993b53..1c50093e2ebe 100644
--- a/arch/parisc/kernel/module.c
+++ b/arch/parisc/kernel/module.c
@@ -43,6 +43,7 @@ 
 #include <linux/elf.h>
 #include <linux/vmalloc.h>
 #include <linux/fs.h>
+#include <linux/ftrace.h>
 #include <linux/string.h>
 #include <linux/kernel.h>
 #include <linux/bug.h>
@@ -862,7 +863,7 @@  int module_finalize(const Elf_Ehdr *hdr,
 	const char *strtab = NULL;
 	const Elf_Shdr *s;
 	char *secstrings;
-	int err, symindex = -1;
+	int symindex = -1;
 	Elf_Sym *newptr, *oldptr;
 	Elf_Shdr *symhdr = NULL;
 #ifdef DEBUG
@@ -946,11 +947,13 @@  int module_finalize(const Elf_Ehdr *hdr,
 			/* patch .altinstructions */
 			apply_alternatives(aseg, aseg + s->sh_size, me->name);
 
+#ifdef CONFIG_DYNAMIC_FTRACE
 		/* For 32 bit kernels we're compiling modules with
 		 * -ffunction-sections so we must relocate the addresses in the
-		 *__mcount_loc section.
+		 *  ftrace callsite section.
 		 */
-		if (symindex != -1 && !strcmp(secname, "__mcount_loc")) {
+		if (symindex != -1 && !strcmp(secname, FTRACE_CALLSITE_SECTION)) {
+			int err;
 			if (s->sh_type == SHT_REL)
 				err = apply_relocate((Elf_Shdr *)sechdrs,
 							strtab, symindex,
@@ -962,6 +965,7 @@  int module_finalize(const Elf_Ehdr *hdr,
 			if (err)
 				return err;
 		}
+#endif
 	}
 	return 0;
 }
diff --git a/arch/parisc/kernel/module.lds b/arch/parisc/kernel/module.lds
deleted file mode 100644
index 1a9a92aca5c8..000000000000
--- a/arch/parisc/kernel/module.lds
+++ /dev/null
@@ -1,7 +0,0 @@ 
-/* SPDX-License-Identifier: GPL-2.0 */
-
-SECTIONS {
-	__mcount_loc : {
-		*(__patchable_function_entries)
-	}
-}
diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h
index 9867d90d635e..9141f2263286 100644
--- a/include/linux/ftrace.h
+++ b/include/linux/ftrace.h
@@ -738,6 +738,11 @@  static inline unsigned long get_lock_parent_ip(void)
 
 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
 extern void ftrace_init(void);
+#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY
+#define FTRACE_CALLSITE_SECTION	"__patchable_function_entries"
+#else
+#define FTRACE_CALLSITE_SECTION	"__mcount_loc"
+#endif
 #else
 static inline void ftrace_init(void) { }
 #endif
diff --git a/kernel/module.c b/kernel/module.c
index ff2d7359a418..acf7962936c4 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -3222,7 +3222,7 @@  static int find_module_sections(struct module *mod, struct load_info *info)
 #endif
 #ifdef CONFIG_FTRACE_MCOUNT_RECORD
 	/* sechdrs[0].sh_size is always zero */
-	mod->ftrace_callsites = section_objs(info, "__mcount_loc",
+	mod->ftrace_callsites = section_objs(info, FTRACE_CALLSITE_SECTION,
 					     sizeof(*mod->ftrace_callsites),
 					     &mod->num_ftrace_callsites);
 #endif