* /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer
@ 2022-03-14 22:00 kernel test robot
2022-03-15 9:25 ` Masami Hiramatsu
0 siblings, 1 reply; 8+ messages in thread
From: kernel test robot @ 2022-03-14 22:00 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: kbuild-all, linux-kernel, Steven Rostedt (VMware),
Nick Desaulniers
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
head: 09688c0166e76ce2fb85e86b9d99be8b0084cdf9
commit: 7e9bf33b812471ee57a03ec7f9b544ca437cc706 ARM: kprobes: Make a frame pointer on __kretprobe_trampoline
date: 5 months ago
config: arm-randconfig-r002-20220313 (https://download.01.org/0day-ci/archive/20220315/202203150516.KTorSVVU-lkp@intel.com/config)
compiler: arm-linux-gnueabi-gcc (GCC) 11.2.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=7e9bf33b812471ee57a03ec7f9b544ca437cc706
git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
git fetch --no-tags linus master
git checkout 7e9bf33b812471ee57a03ec7f9b544ca437cc706
# save the config file to linux build tree
mkdir build_dir
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.2.0 make.cross O=build_dir ARCH=arm SHELL=/bin/bash
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
All errors (new ones prefixed by >>):
arch/arm/probes/kprobes/core.c:236:16: warning: no previous prototype for 'kprobe_handler' [-Wmissing-prototypes]
236 | void __kprobes kprobe_handler(struct pt_regs *regs)
| ^~~~~~~~~~~~~~
arch/arm/probes/kprobes/core.c: In function 'arch_init_kprobes':
arch/arm/probes/kprobes/core.c:465:12: warning: old-style function definition [-Wold-style-definition]
465 | int __init arch_init_kprobes()
| ^~~~~~~~~~~~~~~~~
/tmp/ccIWiggX.s: Assembler messages:
>> /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer
---
0-DAY CI Kernel Test Service
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer
2022-03-14 22:00 /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer kernel test robot
@ 2022-03-15 9:25 ` Masami Hiramatsu
2022-03-15 9:52 ` [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function Masami Hiramatsu
0 siblings, 1 reply; 8+ messages in thread
From: Masami Hiramatsu @ 2022-03-15 9:25 UTC (permalink / raw)
To: kernel test robot
Cc: kbuild-all, linux-kernel, Steven Rostedt (VMware),
Nick Desaulniers
On Tue, 15 Mar 2022 06:00:48 +0800
kernel test robot <lkp@intel.com> wrote:
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head: 09688c0166e76ce2fb85e86b9d99be8b0084cdf9
> commit: 7e9bf33b812471ee57a03ec7f9b544ca437cc706 ARM: kprobes: Make a frame pointer on __kretprobe_trampoline
> date: 5 months ago
> config: arm-randconfig-r002-20220313 (https://download.01.org/0day-ci/archive/20220315/202203150516.KTorSVVU-lkp@intel.com/config)
> compiler: arm-linux-gnueabi-gcc (GCC) 11.2.0
> reproduce (this is a W=1 build):
> wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
> chmod +x ~/bin/make.cross
> # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=7e9bf33b812471ee57a03ec7f9b544ca437cc706
> git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
> git fetch --no-tags linus master
> git checkout 7e9bf33b812471ee57a03ec7f9b544ca437cc706
> # save the config file to linux build tree
> mkdir build_dir
> COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.2.0 make.cross O=build_dir ARCH=arm SHELL=/bin/bash
>
> If you fix the issue, kindly add following tag as appropriate
> Reported-by: kernel test robot <lkp@intel.com>
>
> All errors (new ones prefixed by >>):
>
> arch/arm/probes/kprobes/core.c:236:16: warning: no previous prototype for 'kprobe_handler' [-Wmissing-prototypes]
> 236 | void __kprobes kprobe_handler(struct pt_regs *regs)
> | ^~~~~~~~~~~~~~
> arch/arm/probes/kprobes/core.c: In function 'arch_init_kprobes':
> arch/arm/probes/kprobes/core.c:465:12: warning: old-style function definition [-Wold-style-definition]
> 465 | int __init arch_init_kprobes()
> | ^~~~~~~~~~~~~~~~~
> /tmp/ccIWiggX.s: Assembler messages:
> >> /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer
OK, I finally reproduced it and could make a fix.
It seems that I have to avoid accessing a C function address from the inline assembler in that function.
This can be gcc-11's bug, but to avoid this issue, I decided not using inline asm in the function.
Thank you,
--
Masami Hiramatsu <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function
2022-03-15 9:25 ` Masami Hiramatsu
@ 2022-03-15 9:52 ` Masami Hiramatsu
2022-03-24 6:21 ` Masami Hiramatsu
0 siblings, 1 reply; 8+ messages in thread
From: Masami Hiramatsu @ 2022-03-15 9:52 UTC (permalink / raw)
To: Steven Rostedt
Cc: kbuild-all, linux-kernel, Masami Hiramatsu, Nick Desaulniers
Make __kretprobe_trampoline() as a pure asm function same as the x86
code does. Anyway, it is safe to define the symbol in the asm code
instead of accessing a C symbol from the inline asm.
Without this fix, building arm kernel with GCC-11 may cause below
error.
/tmp/ccIWiggX.s: Assembler messages:
>> /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer
This fixes the error reported by 0day build bot.
Reported-by: kernel test robot <lkp@intel.com>
Fixes: 7e9bf33b8124 ("ARM: kprobes: Make a frame pointer on __kretprobe_trampoline")
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
arch/arm/probes/kprobes/core.c | 57 +++++++++++++++++++++-------------------
1 file changed, 30 insertions(+), 27 deletions(-)
diff --git a/arch/arm/probes/kprobes/core.c b/arch/arm/probes/kprobes/core.c
index 4848404ba51b..51f1438456ae 100644
--- a/arch/arm/probes/kprobes/core.c
+++ b/arch/arm/probes/kprobes/core.c
@@ -373,43 +373,46 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
* should be enough for stacktrace from the return handler with or
* without pt_regs.
*/
-void __naked __kprobes __kretprobe_trampoline(void)
-{
- __asm__ __volatile__ (
+asm(
+ ".text\n"
+ ".global __kretprobe_trampoline\n"
+ ".type __kretprobe_trampoline, %function\n"
+ "__kretprobe_trampoline:\n"
#ifdef CONFIG_FRAME_POINTER
- "ldr lr, =__kretprobe_trampoline \n\t"
+ "ldr lr, =__kretprobe_trampoline \n\t"
/* __kretprobe_trampoline makes a framepointer on pt_regs. */
#ifdef CONFIG_CC_IS_CLANG
- "stmdb sp, {sp, lr, pc} \n\t"
- "sub sp, sp, #12 \n\t"
- /* In clang case, pt_regs->ip = lr. */
- "stmdb sp!, {r0 - r11, lr} \n\t"
- /* fp points regs->r11 (fp) */
- "add fp, sp, #44 \n\t"
+ "stmdb sp, {sp, lr, pc} \n\t"
+ "sub sp, sp, #12 \n\t"
+ /* In clang case, pt_regs->ip = lr. */
+ "stmdb sp!, {r0 - r11, lr} \n\t"
+ /* fp points regs->r11 (fp) */
+ "add fp, sp, #44 \n\t"
#else /* !CONFIG_CC_IS_CLANG */
- /* In gcc case, pt_regs->ip = fp. */
- "stmdb sp, {fp, sp, lr, pc} \n\t"
- "sub sp, sp, #16 \n\t"
- "stmdb sp!, {r0 - r11} \n\t"
- /* fp points regs->r15 (pc) */
- "add fp, sp, #60 \n\t"
+ /* In gcc case, pt_regs->ip = fp. */
+ "stmdb sp, {fp, sp, lr, pc} \n\t"
+ "sub sp, sp, #16 \n\t"
+ "stmdb sp!, {r0 - r11} \n\t"
+ /* fp points regs->r15 (pc) */
+ "add fp, sp, #60 \n\t"
#endif /* CONFIG_CC_IS_CLANG */
#else /* !CONFIG_FRAME_POINTER */
- "sub sp, sp, #16 \n\t"
- "stmdb sp!, {r0 - r11} \n\t"
+ "sub sp, sp, #16 \n\t"
+ "stmdb sp!, {r0 - r11} \n\t"
#endif /* CONFIG_FRAME_POINTER */
- "mov r0, sp \n\t"
- "bl trampoline_handler \n\t"
- "mov lr, r0 \n\t"
- "ldmia sp!, {r0 - r11} \n\t"
- "add sp, sp, #16 \n\t"
+ "mov r0, sp \n\t"
+ "bl trampoline_handler \n\t"
+ "mov lr, r0 \n\t"
+ "ldmia sp!, {r0 - r11} \n\t"
+ "add sp, sp, #16 \n\t"
#ifdef CONFIG_THUMB2_KERNEL
- "bx lr \n\t"
+ "bx lr \n\t"
#else
- "mov pc, lr \n\t"
+ "mov pc, lr \n\t"
#endif
- : : : "memory");
-}
+ ".size __kretprobe_trampoline, .-__kretprobe_trampoline\n"
+);
+NOKPROBE_SYMBOL(__kretprobe_trampoline);
/* Called from __kretprobe_trampoline */
static __used __kprobes void *trampoline_handler(struct pt_regs *regs)
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function
2022-03-15 9:52 ` [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function Masami Hiramatsu
@ 2022-03-24 6:21 ` Masami Hiramatsu
2022-03-24 13:23 ` Masami Hiramatsu
0 siblings, 1 reply; 8+ messages in thread
From: Masami Hiramatsu @ 2022-03-24 6:21 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: Steven Rostedt, kbuild-all, linux-kernel, Nick Desaulniers,
Ard Biesheuvel
Hi,
I found a better solution for this issue from Ard :-)
https://lore.kernel.org/all/20220203082204.1176734-6-ardb@kernel.org/T/#u
I should use mov_l instead of ldr for loading the symbol address.
Thank you,
On Tue, 15 Mar 2022 18:52:16 +0900
Masami Hiramatsu <mhiramat@kernel.org> wrote:
> Make __kretprobe_trampoline() as a pure asm function same as the x86
> code does. Anyway, it is safe to define the symbol in the asm code
> instead of accessing a C symbol from the inline asm.
>
> Without this fix, building arm kernel with GCC-11 may cause below
> error.
>
> /tmp/ccIWiggX.s: Assembler messages:
> >> /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer
>
> This fixes the error reported by 0day build bot.
>
> Reported-by: kernel test robot <lkp@intel.com>
> Fixes: 7e9bf33b8124 ("ARM: kprobes: Make a frame pointer on __kretprobe_trampoline")
> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> ---
> arch/arm/probes/kprobes/core.c | 57 +++++++++++++++++++++-------------------
> 1 file changed, 30 insertions(+), 27 deletions(-)
>
> diff --git a/arch/arm/probes/kprobes/core.c b/arch/arm/probes/kprobes/core.c
> index 4848404ba51b..51f1438456ae 100644
> --- a/arch/arm/probes/kprobes/core.c
> +++ b/arch/arm/probes/kprobes/core.c
> @@ -373,43 +373,46 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
> * should be enough for stacktrace from the return handler with or
> * without pt_regs.
> */
> -void __naked __kprobes __kretprobe_trampoline(void)
> -{
> - __asm__ __volatile__ (
> +asm(
> + ".text\n"
> + ".global __kretprobe_trampoline\n"
> + ".type __kretprobe_trampoline, %function\n"
> + "__kretprobe_trampoline:\n"
> #ifdef CONFIG_FRAME_POINTER
> - "ldr lr, =__kretprobe_trampoline \n\t"
> + "ldr lr, =__kretprobe_trampoline \n\t"
> /* __kretprobe_trampoline makes a framepointer on pt_regs. */
> #ifdef CONFIG_CC_IS_CLANG
> - "stmdb sp, {sp, lr, pc} \n\t"
> - "sub sp, sp, #12 \n\t"
> - /* In clang case, pt_regs->ip = lr. */
> - "stmdb sp!, {r0 - r11, lr} \n\t"
> - /* fp points regs->r11 (fp) */
> - "add fp, sp, #44 \n\t"
> + "stmdb sp, {sp, lr, pc} \n\t"
> + "sub sp, sp, #12 \n\t"
> + /* In clang case, pt_regs->ip = lr. */
> + "stmdb sp!, {r0 - r11, lr} \n\t"
> + /* fp points regs->r11 (fp) */
> + "add fp, sp, #44 \n\t"
> #else /* !CONFIG_CC_IS_CLANG */
> - /* In gcc case, pt_regs->ip = fp. */
> - "stmdb sp, {fp, sp, lr, pc} \n\t"
> - "sub sp, sp, #16 \n\t"
> - "stmdb sp!, {r0 - r11} \n\t"
> - /* fp points regs->r15 (pc) */
> - "add fp, sp, #60 \n\t"
> + /* In gcc case, pt_regs->ip = fp. */
> + "stmdb sp, {fp, sp, lr, pc} \n\t"
> + "sub sp, sp, #16 \n\t"
> + "stmdb sp!, {r0 - r11} \n\t"
> + /* fp points regs->r15 (pc) */
> + "add fp, sp, #60 \n\t"
> #endif /* CONFIG_CC_IS_CLANG */
> #else /* !CONFIG_FRAME_POINTER */
> - "sub sp, sp, #16 \n\t"
> - "stmdb sp!, {r0 - r11} \n\t"
> + "sub sp, sp, #16 \n\t"
> + "stmdb sp!, {r0 - r11} \n\t"
> #endif /* CONFIG_FRAME_POINTER */
> - "mov r0, sp \n\t"
> - "bl trampoline_handler \n\t"
> - "mov lr, r0 \n\t"
> - "ldmia sp!, {r0 - r11} \n\t"
> - "add sp, sp, #16 \n\t"
> + "mov r0, sp \n\t"
> + "bl trampoline_handler \n\t"
> + "mov lr, r0 \n\t"
> + "ldmia sp!, {r0 - r11} \n\t"
> + "add sp, sp, #16 \n\t"
> #ifdef CONFIG_THUMB2_KERNEL
> - "bx lr \n\t"
> + "bx lr \n\t"
> #else
> - "mov pc, lr \n\t"
> + "mov pc, lr \n\t"
> #endif
> - : : : "memory");
> -}
> + ".size __kretprobe_trampoline, .-__kretprobe_trampoline\n"
> +);
> +NOKPROBE_SYMBOL(__kretprobe_trampoline);
>
> /* Called from __kretprobe_trampoline */
> static __used __kprobes void *trampoline_handler(struct pt_regs *regs)
>
--
Masami Hiramatsu <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function
2022-03-24 6:21 ` Masami Hiramatsu
@ 2022-03-24 13:23 ` Masami Hiramatsu
2022-03-24 13:34 ` Ard Biesheuvel
0 siblings, 1 reply; 8+ messages in thread
From: Masami Hiramatsu @ 2022-03-24 13:23 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: Steven Rostedt, kbuild-all, linux-kernel, Nick Desaulniers, ardb
On Thu, 24 Mar 2022 15:21:08 +0900
Masami Hiramatsu <mhiramat@kernel.org> wrote:
> Hi,
>
> I found a better solution for this issue from Ard :-)
>
> https://lore.kernel.org/all/20220203082204.1176734-6-ardb@kernel.org/T/#u
>
> I should use mov_l instead of ldr for loading the symbol address.
Hm, these macros are only for the pure assembly file (.S), so we have to
split this in a asm file to use that.
Thanks,
>
> Thank you,
>
> On Tue, 15 Mar 2022 18:52:16 +0900
> Masami Hiramatsu <mhiramat@kernel.org> wrote:
>
> > Make __kretprobe_trampoline() as a pure asm function same as the x86
> > code does. Anyway, it is safe to define the symbol in the asm code
> > instead of accessing a C symbol from the inline asm.
> >
> > Without this fix, building arm kernel with GCC-11 may cause below
> > error.
> >
> > /tmp/ccIWiggX.s: Assembler messages:
> > >> /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer
> >
> > This fixes the error reported by 0day build bot.
> >
> > Reported-by: kernel test robot <lkp@intel.com>
> > Fixes: 7e9bf33b8124 ("ARM: kprobes: Make a frame pointer on __kretprobe_trampoline")
> > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> > ---
> > arch/arm/probes/kprobes/core.c | 57 +++++++++++++++++++++-------------------
> > 1 file changed, 30 insertions(+), 27 deletions(-)
> >
> > diff --git a/arch/arm/probes/kprobes/core.c b/arch/arm/probes/kprobes/core.c
> > index 4848404ba51b..51f1438456ae 100644
> > --- a/arch/arm/probes/kprobes/core.c
> > +++ b/arch/arm/probes/kprobes/core.c
> > @@ -373,43 +373,46 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
> > * should be enough for stacktrace from the return handler with or
> > * without pt_regs.
> > */
> > -void __naked __kprobes __kretprobe_trampoline(void)
> > -{
> > - __asm__ __volatile__ (
> > +asm(
> > + ".text\n"
> > + ".global __kretprobe_trampoline\n"
> > + ".type __kretprobe_trampoline, %function\n"
> > + "__kretprobe_trampoline:\n"
> > #ifdef CONFIG_FRAME_POINTER
> > - "ldr lr, =__kretprobe_trampoline \n\t"
> > + "ldr lr, =__kretprobe_trampoline \n\t"
> > /* __kretprobe_trampoline makes a framepointer on pt_regs. */
> > #ifdef CONFIG_CC_IS_CLANG
> > - "stmdb sp, {sp, lr, pc} \n\t"
> > - "sub sp, sp, #12 \n\t"
> > - /* In clang case, pt_regs->ip = lr. */
> > - "stmdb sp!, {r0 - r11, lr} \n\t"
> > - /* fp points regs->r11 (fp) */
> > - "add fp, sp, #44 \n\t"
> > + "stmdb sp, {sp, lr, pc} \n\t"
> > + "sub sp, sp, #12 \n\t"
> > + /* In clang case, pt_regs->ip = lr. */
> > + "stmdb sp!, {r0 - r11, lr} \n\t"
> > + /* fp points regs->r11 (fp) */
> > + "add fp, sp, #44 \n\t"
> > #else /* !CONFIG_CC_IS_CLANG */
> > - /* In gcc case, pt_regs->ip = fp. */
> > - "stmdb sp, {fp, sp, lr, pc} \n\t"
> > - "sub sp, sp, #16 \n\t"
> > - "stmdb sp!, {r0 - r11} \n\t"
> > - /* fp points regs->r15 (pc) */
> > - "add fp, sp, #60 \n\t"
> > + /* In gcc case, pt_regs->ip = fp. */
> > + "stmdb sp, {fp, sp, lr, pc} \n\t"
> > + "sub sp, sp, #16 \n\t"
> > + "stmdb sp!, {r0 - r11} \n\t"
> > + /* fp points regs->r15 (pc) */
> > + "add fp, sp, #60 \n\t"
> > #endif /* CONFIG_CC_IS_CLANG */
> > #else /* !CONFIG_FRAME_POINTER */
> > - "sub sp, sp, #16 \n\t"
> > - "stmdb sp!, {r0 - r11} \n\t"
> > + "sub sp, sp, #16 \n\t"
> > + "stmdb sp!, {r0 - r11} \n\t"
> > #endif /* CONFIG_FRAME_POINTER */
> > - "mov r0, sp \n\t"
> > - "bl trampoline_handler \n\t"
> > - "mov lr, r0 \n\t"
> > - "ldmia sp!, {r0 - r11} \n\t"
> > - "add sp, sp, #16 \n\t"
> > + "mov r0, sp \n\t"
> > + "bl trampoline_handler \n\t"
> > + "mov lr, r0 \n\t"
> > + "ldmia sp!, {r0 - r11} \n\t"
> > + "add sp, sp, #16 \n\t"
> > #ifdef CONFIG_THUMB2_KERNEL
> > - "bx lr \n\t"
> > + "bx lr \n\t"
> > #else
> > - "mov pc, lr \n\t"
> > + "mov pc, lr \n\t"
> > #endif
> > - : : : "memory");
> > -}
> > + ".size __kretprobe_trampoline, .-__kretprobe_trampoline\n"
> > +);
> > +NOKPROBE_SYMBOL(__kretprobe_trampoline);
> >
> > /* Called from __kretprobe_trampoline */
> > static __used __kprobes void *trampoline_handler(struct pt_regs *regs)
> >
>
>
> --
> Masami Hiramatsu <mhiramat@kernel.org>
--
Masami Hiramatsu <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function
2022-03-24 13:23 ` Masami Hiramatsu
@ 2022-03-24 13:34 ` Ard Biesheuvel
2022-03-24 13:35 ` Ard Biesheuvel
2022-03-24 15:18 ` Masami Hiramatsu
0 siblings, 2 replies; 8+ messages in thread
From: Ard Biesheuvel @ 2022-03-24 13:34 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: Steven Rostedt, kbuild-all, Linux Kernel Mailing List,
Nick Desaulniers
--- a/arch/arm/probes/kprobes/core.c
+++ b/arch/arm/probes/kprobes/core.c
@@ -377,7 +377,7 @@ void __naked __kprobes __kretprobe_trampoline(void)
{
__asm__ __volatile__ (
#ifdef CONFIG_FRAME_POINTER
- "ldr lr, =__kretprobe_trampoline \n\t"
+ "ldr lr, .L__kretprobe_trampoline \n\t"
/* __kretprobe_trampoline makes a framepointer on pt_regs. */
#ifdef CONFIG_CC_IS_CLANG
"stmdb sp, {sp, lr, pc} \n\t"
@@ -407,6 +407,11 @@ void __naked __kprobes __kretprobe_trampoline(void)
"bx lr \n\t"
#else
"mov pc, lr \n\t"
+#endif
+#ifdef CONFIG_FRAME_POINTER
+ ".align 2 \n\t"
+ ".L__kretprobe_trampoline: \n\t"
+ ".long __kretprobe_trampoline \n\t"
#endif
: : : "memory");
}
On Thu, 24 Mar 2022 at 14:23, Masami Hiramatsu <mhiramat@kernel.org> wrote:
>
> On Thu, 24 Mar 2022 15:21:08 +0900
> Masami Hiramatsu <mhiramat@kernel.org> wrote:
>
> > Hi,
> >
> > I found a better solution for this issue from Ard :-)
> >
> > https://lore.kernel.org/all/20220203082204.1176734-6-ardb@kernel.org/T/#u
> >
> > I should use mov_l instead of ldr for loading the symbol address.
>
> Hm, these macros are only for the pure assembly file (.S), so we have to
> split this in a asm file to use that.
>
What about the below?
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function
2022-03-24 13:34 ` Ard Biesheuvel
@ 2022-03-24 13:35 ` Ard Biesheuvel
2022-03-24 15:18 ` Masami Hiramatsu
1 sibling, 0 replies; 8+ messages in thread
From: Ard Biesheuvel @ 2022-03-24 13:35 UTC (permalink / raw)
To: Masami Hiramatsu
Cc: Steven Rostedt, kbuild-all, Linux Kernel Mailing List,
Nick Desaulniers
On Thu, 24 Mar 2022 at 14:34, Ard Biesheuvel <ardb@kernel.org> wrote:
>
> --- a/arch/arm/probes/kprobes/core.c
> +++ b/arch/arm/probes/kprobes/core.c
> @@ -377,7 +377,7 @@ void __naked __kprobes __kretprobe_trampoline(void)
> {
> __asm__ __volatile__ (
> #ifdef CONFIG_FRAME_POINTER
> - "ldr lr, =__kretprobe_trampoline \n\t"
> + "ldr lr, .L__kretprobe_trampoline \n\t"
> /* __kretprobe_trampoline makes a framepointer on pt_regs. */
> #ifdef CONFIG_CC_IS_CLANG
> "stmdb sp, {sp, lr, pc} \n\t"
> @@ -407,6 +407,11 @@ void __naked __kprobes __kretprobe_trampoline(void)
> "bx lr \n\t"
> #else
> "mov pc, lr \n\t"
> +#endif
> +#ifdef CONFIG_FRAME_POINTER
> + ".align 2 \n\t"
> + ".L__kretprobe_trampoline: \n\t"
> + ".long __kretprobe_trampoline \n\t"
> #endif
> : : : "memory");
> }
>
>
Pardon the copy/paste error.
> What about the below?
was meant to refer to the diff above.
> On Thu, 24 Mar 2022 at 14:23, Masami Hiramatsu <mhiramat@kernel.org> wrote:
> >
> > On Thu, 24 Mar 2022 15:21:08 +0900
> > Masami Hiramatsu <mhiramat@kernel.org> wrote:
> >
> > > Hi,
> > >
> > > I found a better solution for this issue from Ard :-)
> > >
> > > https://lore.kernel.org/all/20220203082204.1176734-6-ardb@kernel.org/T/#u
> > >
> > > I should use mov_l instead of ldr for loading the symbol address.
> >
> > Hm, these macros are only for the pure assembly file (.S), so we have to
> > split this in a asm file to use that.
> >
>
>
> What about the below?
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function
2022-03-24 13:34 ` Ard Biesheuvel
2022-03-24 13:35 ` Ard Biesheuvel
@ 2022-03-24 15:18 ` Masami Hiramatsu
1 sibling, 0 replies; 8+ messages in thread
From: Masami Hiramatsu @ 2022-03-24 15:18 UTC (permalink / raw)
To: Ard Biesheuvel
Cc: Steven Rostedt, kbuild-all, Linux Kernel Mailing List,
Nick Desaulniers
Hi Ard,
On Thu, 24 Mar 2022 14:34:46 +0100
Ard Biesheuvel <ardb@kernel.org> wrote:
> --- a/arch/arm/probes/kprobes/core.c
> +++ b/arch/arm/probes/kprobes/core.c
> @@ -377,7 +377,7 @@ void __naked __kprobes __kretprobe_trampoline(void)
> {
> __asm__ __volatile__ (
> #ifdef CONFIG_FRAME_POINTER
> - "ldr lr, =__kretprobe_trampoline \n\t"
> + "ldr lr, .L__kretprobe_trampoline \n\t"
> /* __kretprobe_trampoline makes a framepointer on pt_regs. */
> #ifdef CONFIG_CC_IS_CLANG
> "stmdb sp, {sp, lr, pc} \n\t"
> @@ -407,6 +407,11 @@ void __naked __kprobes __kretprobe_trampoline(void)
> "bx lr \n\t"
> #else
> "mov pc, lr \n\t"
> +#endif
> +#ifdef CONFIG_FRAME_POINTER
> + ".align 2 \n\t"
> + ".L__kretprobe_trampoline: \n\t"
> + ".long __kretprobe_trampoline \n\t"
> #endif
> : : : "memory");
> }
Yes, I confirmed this works too :-)
Can you send the patch with my Tested-by ?
Thank you!
>
>
> On Thu, 24 Mar 2022 at 14:23, Masami Hiramatsu <mhiramat@kernel.org> wrote:
> >
> > On Thu, 24 Mar 2022 15:21:08 +0900
> > Masami Hiramatsu <mhiramat@kernel.org> wrote:
> >
> > > Hi,
> > >
> > > I found a better solution for this issue from Ard :-)
> > >
> > > https://lore.kernel.org/all/20220203082204.1176734-6-ardb@kernel.org/T/#u
> > >
> > > I should use mov_l instead of ldr for loading the symbol address.
> >
> > Hm, these macros are only for the pure assembly file (.S), so we have to
> > split this in a asm file to use that.
> >
>
>
> What about the below?
--
Masami Hiramatsu <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2022-03-24 15:18 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-03-14 22:00 /tmp/ccIWiggX.s:22: Error: invalid literal constant: pool needs to be closer kernel test robot
2022-03-15 9:25 ` Masami Hiramatsu
2022-03-15 9:52 ` [PATCH] ARM: kprobes: Make __kretprobe_trampoline as a pure asm function Masami Hiramatsu
2022-03-24 6:21 ` Masami Hiramatsu
2022-03-24 13:23 ` Masami Hiramatsu
2022-03-24 13:34 ` Ard Biesheuvel
2022-03-24 13:35 ` Ard Biesheuvel
2022-03-24 15:18 ` Masami Hiramatsu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox