* [PATCH] powerpc: kprobes: invoke handlers directly
@ 2016-11-18 11:39 Naveen N. Rao
2016-11-18 14:48 ` Masami Hiramatsu
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Naveen N. Rao @ 2016-11-18 11:39 UTC (permalink / raw)
To: Michael Ellerman, Masami Hiramatsu
Cc: Anton Blanchard, linuxppc-dev, Ananth N Mavinakayanahalli
... rather than through notify_die(), to reduce path taken for handling
kprobes. Similar to commit 6f6343f53d13 ("kprobes/x86: Call exception
handlers directly from do_int3/do_debug").
While at it, rename post_kprobe_handler() to kprobe_post_handler() for
more uniform naming.
Reported-by: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/kprobes.h | 2 ++
arch/powerpc/kernel/kprobes.c | 29 +++++++----------------------
arch/powerpc/kernel/traps.c | 20 ++++++++++++++++++++
3 files changed, 29 insertions(+), 22 deletions(-)
diff --git a/arch/powerpc/include/asm/kprobes.h b/arch/powerpc/include/asm/kprobes.h
index 2c9759bd..47c60d0 100644
--- a/arch/powerpc/include/asm/kprobes.h
+++ b/arch/powerpc/include/asm/kprobes.h
@@ -126,6 +126,8 @@ struct kprobe_ctlblk {
extern int kprobe_exceptions_notify(struct notifier_block *self,
unsigned long val, void *data);
+extern int kprobe_handler(struct pt_regs *regs);
+extern int kprobe_post_handler(struct pt_regs *regs);
extern int kprobe_fault_handler(struct pt_regs *regs, int trapnr);
#endif /* __KERNEL__ */
#endif /* _ASM_POWERPC_KPROBES_H */
diff --git a/arch/powerpc/kernel/kprobes.c b/arch/powerpc/kernel/kprobes.c
index 9479d8e..ad108b8 100644
--- a/arch/powerpc/kernel/kprobes.c
+++ b/arch/powerpc/kernel/kprobes.c
@@ -140,13 +140,16 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
regs->link = (unsigned long)kretprobe_trampoline;
}
-static int __kprobes kprobe_handler(struct pt_regs *regs)
+int __kprobes kprobe_handler(struct pt_regs *regs)
{
struct kprobe *p;
int ret = 0;
unsigned int *addr = (unsigned int *)regs->nip;
struct kprobe_ctlblk *kcb;
+ if (user_mode(regs))
+ return 0;
+
/*
* We don't want to be preempted for the entire
* duration of kprobe processing
@@ -359,12 +362,12 @@ static int __kprobes trampoline_probe_handler(struct kprobe *p,
* single-stepped a copy of the instruction. The address of this
* copy is p->ainsn.insn.
*/
-static int __kprobes post_kprobe_handler(struct pt_regs *regs)
+int __kprobes kprobe_post_handler(struct pt_regs *regs)
{
struct kprobe *cur = kprobe_running();
struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
- if (!cur)
+ if (!cur || user_mode(regs))
return 0;
/* make sure we got here for instruction we have a kprobe on */
@@ -470,25 +473,7 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
unsigned long val, void *data)
{
- struct die_args *args = (struct die_args *)data;
- int ret = NOTIFY_DONE;
-
- if (args->regs && user_mode(args->regs))
- return ret;
-
- switch (val) {
- case DIE_BPT:
- if (kprobe_handler(args->regs))
- ret = NOTIFY_STOP;
- break;
- case DIE_SSTEP:
- if (post_kprobe_handler(args->regs))
- ret = NOTIFY_STOP;
- break;
- default:
- break;
- }
- return ret;
+ return NOTIFY_DONE;
}
unsigned long arch_deref_entry_point(void *entry)
diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index 91d278c..dbcdf48 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c
@@ -824,6 +824,11 @@ void single_step_exception(struct pt_regs *regs)
clear_single_step(regs);
+#ifdef CONFIG_KPROBES
+ if (kprobe_post_handler(regs))
+ return;
+#endif
+
if (notify_die(DIE_SSTEP, "single_step", regs, 5,
5, SIGTRAP) == NOTIFY_STOP)
goto bail;
@@ -1177,6 +1182,11 @@ void program_check_exception(struct pt_regs *regs)
if (debugger_bpt(regs))
goto bail;
+#ifdef CONFIG_KPROBES
+ if (kprobe_handler(regs))
+ goto bail;
+#endif
+
/* trap exception */
if (notify_die(DIE_BPT, "breakpoint", regs, 5, 5, SIGTRAP)
== NOTIFY_STOP)
@@ -1745,6 +1755,11 @@ void DebugException(struct pt_regs *regs, unsigned long debug_status)
return;
}
+#ifdef CONFIG_KPROBES
+ if (kprobe_post_handler(regs))
+ return;
+#endif
+
if (notify_die(DIE_SSTEP, "block_step", regs, 5,
5, SIGTRAP) == NOTIFY_STOP) {
return;
@@ -1759,6 +1774,11 @@ void DebugException(struct pt_regs *regs, unsigned long debug_status)
/* Clear the instruction completion event */
mtspr(SPRN_DBSR, DBSR_IC);
+#ifdef CONFIG_KPROBES
+ if (kprobe_post_handler(regs))
+ return;
+#endif
+
if (notify_die(DIE_SSTEP, "single_step", regs, 5,
5, SIGTRAP) == NOTIFY_STOP) {
return;
--
2.10.2
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] powerpc: kprobes: invoke handlers directly
2016-11-18 11:39 [PATCH] powerpc: kprobes: invoke handlers directly Naveen N. Rao
@ 2016-11-18 14:48 ` Masami Hiramatsu
2016-11-21 4:23 ` Ananth N Mavinakayanahalli
2016-11-21 10:24 ` Michael Ellerman
2 siblings, 0 replies; 4+ messages in thread
From: Masami Hiramatsu @ 2016-11-18 14:48 UTC (permalink / raw)
To: Naveen N. Rao
Cc: Michael Ellerman, Anton Blanchard, linuxppc-dev,
Ananth N Mavinakayanahalli
On Fri, 18 Nov 2016 17:09:26 +0530
"Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> wrote:
> ... rather than through notify_die(), to reduce path taken for handling
> kprobes. Similar to commit 6f6343f53d13 ("kprobes/x86: Call exception
> handlers directly from do_int3/do_debug").
>
> While at it, rename post_kprobe_handler() to kprobe_post_handler() for
> more uniform naming.
Looks good to me.
Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org>
Thanks!
>
> Reported-by: Masami Hiramatsu <mhiramat@kernel.org>
> Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
> ---
> arch/powerpc/include/asm/kprobes.h | 2 ++
> arch/powerpc/kernel/kprobes.c | 29 +++++++----------------------
> arch/powerpc/kernel/traps.c | 20 ++++++++++++++++++++
> 3 files changed, 29 insertions(+), 22 deletions(-)
>
> diff --git a/arch/powerpc/include/asm/kprobes.h b/arch/powerpc/include/asm/kprobes.h
> index 2c9759bd..47c60d0 100644
> --- a/arch/powerpc/include/asm/kprobes.h
> +++ b/arch/powerpc/include/asm/kprobes.h
> @@ -126,6 +126,8 @@ struct kprobe_ctlblk {
>
> extern int kprobe_exceptions_notify(struct notifier_block *self,
> unsigned long val, void *data);
> +extern int kprobe_handler(struct pt_regs *regs);
> +extern int kprobe_post_handler(struct pt_regs *regs);
> extern int kprobe_fault_handler(struct pt_regs *regs, int trapnr);
> #endif /* __KERNEL__ */
> #endif /* _ASM_POWERPC_KPROBES_H */
> diff --git a/arch/powerpc/kernel/kprobes.c b/arch/powerpc/kernel/kprobes.c
> index 9479d8e..ad108b8 100644
> --- a/arch/powerpc/kernel/kprobes.c
> +++ b/arch/powerpc/kernel/kprobes.c
> @@ -140,13 +140,16 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
> regs->link = (unsigned long)kretprobe_trampoline;
> }
>
> -static int __kprobes kprobe_handler(struct pt_regs *regs)
> +int __kprobes kprobe_handler(struct pt_regs *regs)
> {
> struct kprobe *p;
> int ret = 0;
> unsigned int *addr = (unsigned int *)regs->nip;
> struct kprobe_ctlblk *kcb;
>
> + if (user_mode(regs))
> + return 0;
> +
> /*
> * We don't want to be preempted for the entire
> * duration of kprobe processing
> @@ -359,12 +362,12 @@ static int __kprobes trampoline_probe_handler(struct kprobe *p,
> * single-stepped a copy of the instruction. The address of this
> * copy is p->ainsn.insn.
> */
> -static int __kprobes post_kprobe_handler(struct pt_regs *regs)
> +int __kprobes kprobe_post_handler(struct pt_regs *regs)
> {
> struct kprobe *cur = kprobe_running();
> struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
>
> - if (!cur)
> + if (!cur || user_mode(regs))
> return 0;
>
> /* make sure we got here for instruction we have a kprobe on */
> @@ -470,25 +473,7 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
> int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
> unsigned long val, void *data)
> {
> - struct die_args *args = (struct die_args *)data;
> - int ret = NOTIFY_DONE;
> -
> - if (args->regs && user_mode(args->regs))
> - return ret;
> -
> - switch (val) {
> - case DIE_BPT:
> - if (kprobe_handler(args->regs))
> - ret = NOTIFY_STOP;
> - break;
> - case DIE_SSTEP:
> - if (post_kprobe_handler(args->regs))
> - ret = NOTIFY_STOP;
> - break;
> - default:
> - break;
> - }
> - return ret;
> + return NOTIFY_DONE;
> }
>
> unsigned long arch_deref_entry_point(void *entry)
> diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
> index 91d278c..dbcdf48 100644
> --- a/arch/powerpc/kernel/traps.c
> +++ b/arch/powerpc/kernel/traps.c
> @@ -824,6 +824,11 @@ void single_step_exception(struct pt_regs *regs)
>
> clear_single_step(regs);
>
> +#ifdef CONFIG_KPROBES
> + if (kprobe_post_handler(regs))
> + return;
> +#endif
> +
> if (notify_die(DIE_SSTEP, "single_step", regs, 5,
> 5, SIGTRAP) == NOTIFY_STOP)
> goto bail;
> @@ -1177,6 +1182,11 @@ void program_check_exception(struct pt_regs *regs)
> if (debugger_bpt(regs))
> goto bail;
>
> +#ifdef CONFIG_KPROBES
> + if (kprobe_handler(regs))
> + goto bail;
> +#endif
> +
> /* trap exception */
> if (notify_die(DIE_BPT, "breakpoint", regs, 5, 5, SIGTRAP)
> == NOTIFY_STOP)
> @@ -1745,6 +1755,11 @@ void DebugException(struct pt_regs *regs, unsigned long debug_status)
> return;
> }
>
> +#ifdef CONFIG_KPROBES
> + if (kprobe_post_handler(regs))
> + return;
> +#endif
> +
> if (notify_die(DIE_SSTEP, "block_step", regs, 5,
> 5, SIGTRAP) == NOTIFY_STOP) {
> return;
> @@ -1759,6 +1774,11 @@ void DebugException(struct pt_regs *regs, unsigned long debug_status)
> /* Clear the instruction completion event */
> mtspr(SPRN_DBSR, DBSR_IC);
>
> +#ifdef CONFIG_KPROBES
> + if (kprobe_post_handler(regs))
> + return;
> +#endif
> +
> if (notify_die(DIE_SSTEP, "single_step", regs, 5,
> 5, SIGTRAP) == NOTIFY_STOP) {
> return;
> --
> 2.10.2
>
--
Masami Hiramatsu <mhiramat@kernel.org>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] powerpc: kprobes: invoke handlers directly
2016-11-18 11:39 [PATCH] powerpc: kprobes: invoke handlers directly Naveen N. Rao
2016-11-18 14:48 ` Masami Hiramatsu
@ 2016-11-21 4:23 ` Ananth N Mavinakayanahalli
2016-11-21 10:24 ` Michael Ellerman
2 siblings, 0 replies; 4+ messages in thread
From: Ananth N Mavinakayanahalli @ 2016-11-21 4:23 UTC (permalink / raw)
To: Naveen N. Rao
Cc: Michael Ellerman, Masami Hiramatsu, Anton Blanchard, linuxppc-dev
On Fri, Nov 18, 2016 at 05:09:26PM +0530, Naveen N. Rao wrote:
> ... rather than through notify_die(), to reduce path taken for handling
> kprobes. Similar to commit 6f6343f53d13 ("kprobes/x86: Call exception
> handlers directly from do_int3/do_debug").
>
> While at it, rename post_kprobe_handler() to kprobe_post_handler() for
> more uniform naming.
>
> Reported-by: Masami Hiramatsu <mhiramat@kernel.org>
> Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Acked-by: Ananth N Mavinakayanahalli <ananth@linux.vnet.ibm.com>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] powerpc: kprobes: invoke handlers directly
2016-11-18 11:39 [PATCH] powerpc: kprobes: invoke handlers directly Naveen N. Rao
2016-11-18 14:48 ` Masami Hiramatsu
2016-11-21 4:23 ` Ananth N Mavinakayanahalli
@ 2016-11-21 10:24 ` Michael Ellerman
2 siblings, 0 replies; 4+ messages in thread
From: Michael Ellerman @ 2016-11-21 10:24 UTC (permalink / raw)
To: Naveen N. Rao, Masami Hiramatsu
Cc: Anton Blanchard, linuxppc-dev, Ananth N Mavinakayanahalli
"Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> writes:
> +#ifdef CONFIG_KPROBES
> + if (kprobe_post_handler(regs))
> + return;
> +#endif
...
> +#ifdef CONFIG_KPROBES
> + if (kprobe_handler(regs))
> + goto bail;
> +#endif
...
> +#ifdef CONFIG_KPROBES
> + if (kprobe_post_handler(regs))
> + return;
> +#endif
...
> +#ifdef CONFIG_KPROBES
> + if (kprobe_post_handler(regs))
> + return;
> +#endif
> +
Sorry too many ifdefs.
Can you please add empty static inline versions for CONFIG_KPROBES=n
that always return false. The compiler will be smart enough to elide
them entirely.
cheers
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2016-11-21 10:24 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-11-18 11:39 [PATCH] powerpc: kprobes: invoke handlers directly Naveen N. Rao
2016-11-18 14:48 ` Masami Hiramatsu
2016-11-21 4:23 ` Ananth N Mavinakayanahalli
2016-11-21 10:24 ` Michael Ellerman
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.