diff mbox

arm64: use preempt_disable_notrace in _percpu_read/write

Message ID 1473338802-18712-1-git-send-email-zhang.chunyan@linaro.org (mailing list archive)
State New, archived
Headers show

Commit Message

Chunyan Zhang Sept. 8, 2016, 12:46 p.m. UTC
When debug preempt or preempt tracer is enabled, preempt_count_add/sub()
can be traced by function and function graph tracing, and
preempt_disable/enable() would call preempt_count_add/sub(), so in Ftrace
subsystem we should use preempt_disable/enable_notrace instead.

In the commit 345ddcc882d8 ("ftrace: Have set_ftrace_pid use the bitmap
like events do") the function this_cpu_read() was added to
trace_graph_entry(), and if this_cpu_read() calls preempt_disable(), graph
tracer will go into a recursive loop, even if the tracing_on is
disabled.

So this patch change to use preempt_enable/disable_notrace instead in
this_cpu_read().

Since Yonghui Yang helped a lot to find the root cause of this problem,
so also add his SOB.

Signed-off-by: Yonghui Yang <mark.yang@spreadtrum.com>
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
---
 arch/arm64/include/asm/percpu.h | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

Comments

Mark Rutland Sept. 8, 2016, 1:02 p.m. UTC | #1
Hi,

In future, please ensure that you include the arm64 maintainers when
sending changes to core arm64 code. I've copied Catalin and Will for you
this time.

Thanks,
Mark.

On Thu, Sep 08, 2016 at 08:46:42PM +0800, Chunyan Zhang wrote:
> When debug preempt or preempt tracer is enabled, preempt_count_add/sub()
> can be traced by function and function graph tracing, and
> preempt_disable/enable() would call preempt_count_add/sub(), so in Ftrace
> subsystem we should use preempt_disable/enable_notrace instead.
> 
> In the commit 345ddcc882d8 ("ftrace: Have set_ftrace_pid use the bitmap
> like events do") the function this_cpu_read() was added to
> trace_graph_entry(), and if this_cpu_read() calls preempt_disable(), graph
> tracer will go into a recursive loop, even if the tracing_on is
> disabled.
> 
> So this patch change to use preempt_enable/disable_notrace instead in
> this_cpu_read().
> 
> Since Yonghui Yang helped a lot to find the root cause of this problem,
> so also add his SOB.
> 
> Signed-off-by: Yonghui Yang <mark.yang@spreadtrum.com>
> Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
> ---
>  arch/arm64/include/asm/percpu.h | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/percpu.h b/arch/arm64/include/asm/percpu.h
> index 0a456be..2fee2f5 100644
> --- a/arch/arm64/include/asm/percpu.h
> +++ b/arch/arm64/include/asm/percpu.h
> @@ -199,19 +199,19 @@ static inline unsigned long __percpu_xchg(void *ptr, unsigned long val,
>  #define _percpu_read(pcp)						\
>  ({									\
>  	typeof(pcp) __retval;						\
> -	preempt_disable();						\
> +	preempt_disable_notrace();					\
>  	__retval = (typeof(pcp))__percpu_read(raw_cpu_ptr(&(pcp)), 	\
>  					      sizeof(pcp));		\
> -	preempt_enable();						\
> +	preempt_enable_notrace();					\
>  	__retval;							\
>  })
>  
>  #define _percpu_write(pcp, val)						\
>  do {									\
> -	preempt_disable();						\
> +	preempt_disable_notrace();					\
>  	__percpu_write(raw_cpu_ptr(&(pcp)), (unsigned long)(val), 	\
>  				sizeof(pcp));				\
> -	preempt_enable();						\
> +	preempt_enable_notrace();					\
>  } while(0)								\
>  
>  #define _pcp_protect(operation, pcp, val)			\
> -- 
> 2.7.4
> 
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>
Chunyan Zhang Sept. 8, 2016, 1:17 p.m. UTC | #2
Thanks Mark.

On 8 September 2016 at 21:02, Mark Rutland <mark.rutland@arm.com> wrote:
> Hi,
>
> In future, please ensure that you include the arm64 maintainers when
> sending changes to core arm64 code. I've copied Catalin and Will for you
> this time.

Sorry about this.

Chunyan

>
> Thanks,
> Mark.
>
> On Thu, Sep 08, 2016 at 08:46:42PM +0800, Chunyan Zhang wrote:
>> When debug preempt or preempt tracer is enabled, preempt_count_add/sub()
>> can be traced by function and function graph tracing, and
>> preempt_disable/enable() would call preempt_count_add/sub(), so in Ftrace
>> subsystem we should use preempt_disable/enable_notrace instead.
>>
>> In the commit 345ddcc882d8 ("ftrace: Have set_ftrace_pid use the bitmap
>> like events do") the function this_cpu_read() was added to
>> trace_graph_entry(), and if this_cpu_read() calls preempt_disable(), graph
>> tracer will go into a recursive loop, even if the tracing_on is
>> disabled.
>>
>> So this patch change to use preempt_enable/disable_notrace instead in
>> this_cpu_read().
>>
>> Since Yonghui Yang helped a lot to find the root cause of this problem,
>> so also add his SOB.
>>
>> Signed-off-by: Yonghui Yang <mark.yang@spreadtrum.com>
>> Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
>> ---
>>  arch/arm64/include/asm/percpu.h | 8 ++++----
>>  1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/arch/arm64/include/asm/percpu.h b/arch/arm64/include/asm/percpu.h
>> index 0a456be..2fee2f5 100644
>> --- a/arch/arm64/include/asm/percpu.h
>> +++ b/arch/arm64/include/asm/percpu.h
>> @@ -199,19 +199,19 @@ static inline unsigned long __percpu_xchg(void *ptr, unsigned long val,
>>  #define _percpu_read(pcp)                                            \
>>  ({                                                                   \
>>       typeof(pcp) __retval;                                           \
>> -     preempt_disable();                                              \
>> +     preempt_disable_notrace();                                      \
>>       __retval = (typeof(pcp))__percpu_read(raw_cpu_ptr(&(pcp)),      \
>>                                             sizeof(pcp));             \
>> -     preempt_enable();                                               \
>> +     preempt_enable_notrace();                                       \
>>       __retval;                                                       \
>>  })
>>
>>  #define _percpu_write(pcp, val)                                              \
>>  do {                                                                 \
>> -     preempt_disable();                                              \
>> +     preempt_disable_notrace();                                      \
>>       __percpu_write(raw_cpu_ptr(&(pcp)), (unsigned long)(val),       \
>>                               sizeof(pcp));                           \
>> -     preempt_enable();                                               \
>> +     preempt_enable_notrace();                                       \
>>  } while(0)                                                           \
>>
>>  #define _pcp_protect(operation, pcp, val)                    \
>> --
>> 2.7.4
>>
>>
>> _______________________________________________
>> linux-arm-kernel mailing list
>> linux-arm-kernel@lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>>
Will Deacon Sept. 9, 2016, 10:07 a.m. UTC | #3
On Thu, Sep 08, 2016 at 08:46:42PM +0800, Chunyan Zhang wrote:
> When debug preempt or preempt tracer is enabled, preempt_count_add/sub()
> can be traced by function and function graph tracing, and
> preempt_disable/enable() would call preempt_count_add/sub(), so in Ftrace
> subsystem we should use preempt_disable/enable_notrace instead.
> 
> In the commit 345ddcc882d8 ("ftrace: Have set_ftrace_pid use the bitmap
> like events do") the function this_cpu_read() was added to
> trace_graph_entry(), and if this_cpu_read() calls preempt_disable(), graph
> tracer will go into a recursive loop, even if the tracing_on is
> disabled.
> 
> So this patch change to use preempt_enable/disable_notrace instead in
> this_cpu_read().
> 
> Since Yonghui Yang helped a lot to find the root cause of this problem,
> so also add his SOB.
> 
> Signed-off-by: Yonghui Yang <mark.yang@spreadtrum.com>
> Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
> ---
>  arch/arm64/include/asm/percpu.h | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)

Looks good to me:

Acked-by: Will Deacon <will.deacon@arm.com>

However, don't you need to make a similar change to asm-generic/percpu.h
for other architectures (e.g. arch/arm/)?

Will
Catalin Marinas Sept. 9, 2016, 11:35 a.m. UTC | #4
On Thu, Sep 08, 2016 at 08:46:42PM +0800, Chunyan Zhang wrote:
> When debug preempt or preempt tracer is enabled, preempt_count_add/sub()
> can be traced by function and function graph tracing, and
> preempt_disable/enable() would call preempt_count_add/sub(), so in Ftrace
> subsystem we should use preempt_disable/enable_notrace instead.
> 
> In the commit 345ddcc882d8 ("ftrace: Have set_ftrace_pid use the bitmap
> like events do") the function this_cpu_read() was added to
> trace_graph_entry(), and if this_cpu_read() calls preempt_disable(), graph
> tracer will go into a recursive loop, even if the tracing_on is
> disabled.
> 
> So this patch change to use preempt_enable/disable_notrace instead in
> this_cpu_read().
> 
> Since Yonghui Yang helped a lot to find the root cause of this problem,
> so also add his SOB.
> 
> Signed-off-by: Yonghui Yang <mark.yang@spreadtrum.com>
> Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>

Queued for 4.8-rc6. Thanks.
Chunyan Zhang Sept. 9, 2016, 11:43 a.m. UTC | #5
On 9 September 2016 at 18:07, Will Deacon <will.deacon@arm.com> wrote:
> On Thu, Sep 08, 2016 at 08:46:42PM +0800, Chunyan Zhang wrote:
>> When debug preempt or preempt tracer is enabled, preempt_count_add/sub()
>> can be traced by function and function graph tracing, and
>> preempt_disable/enable() would call preempt_count_add/sub(), so in Ftrace
>> subsystem we should use preempt_disable/enable_notrace instead.
>>
>> In the commit 345ddcc882d8 ("ftrace: Have set_ftrace_pid use the bitmap
>> like events do") the function this_cpu_read() was added to
>> trace_graph_entry(), and if this_cpu_read() calls preempt_disable(), graph
>> tracer will go into a recursive loop, even if the tracing_on is
>> disabled.
>>
>> So this patch change to use preempt_enable/disable_notrace instead in
>> this_cpu_read().
>>
>> Since Yonghui Yang helped a lot to find the root cause of this problem,
>> so also add his SOB.
>>
>> Signed-off-by: Yonghui Yang <mark.yang@spreadtrum.com>
>> Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
>> ---
>>  arch/arm64/include/asm/percpu.h | 8 ++++----
>>  1 file changed, 4 insertions(+), 4 deletions(-)
>
> Looks good to me:
>
> Acked-by: Will Deacon <will.deacon@arm.com>
>
> However, don't you need to make a similar change to asm-generic/percpu.h
> for other architectures (e.g. arch/arm/)?

Yes, I will send out another patch to fix that.

Thanks,
Chunyan

>
> Will
diff mbox

Patch

diff --git a/arch/arm64/include/asm/percpu.h b/arch/arm64/include/asm/percpu.h
index 0a456be..2fee2f5 100644
--- a/arch/arm64/include/asm/percpu.h
+++ b/arch/arm64/include/asm/percpu.h
@@ -199,19 +199,19 @@  static inline unsigned long __percpu_xchg(void *ptr, unsigned long val,
 #define _percpu_read(pcp)						\
 ({									\
 	typeof(pcp) __retval;						\
-	preempt_disable();						\
+	preempt_disable_notrace();					\
 	__retval = (typeof(pcp))__percpu_read(raw_cpu_ptr(&(pcp)), 	\
 					      sizeof(pcp));		\
-	preempt_enable();						\
+	preempt_enable_notrace();					\
 	__retval;							\
 })
 
 #define _percpu_write(pcp, val)						\
 do {									\
-	preempt_disable();						\
+	preempt_disable_notrace();					\
 	__percpu_write(raw_cpu_ptr(&(pcp)), (unsigned long)(val), 	\
 				sizeof(pcp));				\
-	preempt_enable();						\
+	preempt_enable_notrace();					\
 } while(0)								\
 
 #define _pcp_protect(operation, pcp, val)			\