Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1281332 > unrolled thread

[RFC] Add gup trace points support

Back to article view | Back to linux.kernel

Contents

  [RFC] Add gup trace points support Yang Shi <yang.shi@linaro.org> - 2015-12-02 00:30 +0100
    [PATCH 1/7] trace/events: Add gup trace events Yang Shi <yang.shi@linaro.org> - 2015-12-02 00:30 +0100
      Re: [PATCH 1/7] trace/events: Add gup trace events Steven Rostedt <rostedt@goodmis.org> - 2015-12-02 01:00 +0100
        Re: [PATCH 1/7] trace/events: Add gup trace events "Shi, Yang" <yang.shi@linaro.org> - 2015-12-02 01:10 +0100
          Re: [PATCH 1/7] trace/events: Add gup trace events Steven Rostedt <rostedt@goodmis.org> - 2015-12-02 01:20 +0100
            Re: [PATCH 1/7] trace/events: Add gup trace events "Shi, Yang" <yang.shi@linaro.org> - 2015-12-02 01:40 +0100
    [PATCH 4/7] mips: mm/gup: add gup trace points Yang Shi <yang.shi@linaro.org> - 2015-12-02 00:30 +0100
      Re: [PATCH 4/7] mips: mm/gup: add gup trace points Ralf Baechle <ralf@linux-mips.org> - 2015-12-02 14:30 +0100

#1281332 — [RFC] Add gup trace points support

Some background about why I think this might be useful.

When I was profiling some hugetlb related program, I got page-faults event
doubled when hugetlb is enabled. When I looked into the code, I found page-faults
come from two places, do_page_fault and gup. So, I tried to figure out which
play a role (or both) in my use case. But I can't find existing finer tracing
event for sub page-faults in current mainline kernel.

So, I added the gup trace points support to have finer tracing events for
page-faults. The below events are added:

__get_user_pages
__get_user_pages_fast
fixup_user_fault

Both __get_user_pages and fixup_user_fault call handle_mm_fault.

Just added trace points to raw version __get_user_pages since all variants
will call it finally to do real work.

Although __get_user_pages_fast doesn't call handle_mm_fault, it might be useful
to have it to distinguish between slow and fast version.


Yang Shi (7):
      trace/events: Add gup trace events
      mm/gup: add gup trace points
      x86: mm/gup: add gup trace points
      mips: mm/gup: add gup trace points
      s390: mm/gup: add gup trace points
      sh: mm/gup: add gup trace points
      sparc64: mm/gup: add gup trace points

 arch/mips/mm/gup.c         |  7 +++++++
 arch/s390/mm/gup.c         |  7 +++++++
 arch/sh/mm/gup.c           |  8 ++++++++
 arch/sparc/mm/gup.c        |  8 ++++++++
 arch/x86/mm/gup.c          |  7 +++++++
 include/trace/events/gup.h | 77 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 mm/gup.c                   |  8 ++++++++
 7 files changed, 122 insertions(+)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1281333 — [PATCH 1/7] trace/events: Add gup trace events

page-faults events record the invoke to handle_mm_fault, but the invoke
may come from do_page_fault or gup. In some use cases, the finer event count
mey be needed, so add trace events support for:

__get_user_pages
__get_user_pages_fast
fixup_user_fault

Signed-off-by: Yang Shi <yang.shi@linaro.org>
---
 include/trace/events/gup.h | 77 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 77 insertions(+)
 create mode 100644 include/trace/events/gup.h

diff --git a/include/trace/events/gup.h b/include/trace/events/gup.h
new file mode 100644
index 0000000..37d18f9
--- /dev/null
+++ b/include/trace/events/gup.h
@@ -0,0 +1,77 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM gup
+
+#if !defined(_TRACE_GUP_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_GUP_H
+
+#include <linux/types.h>
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(gup_fixup_user_fault,
+
+	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
+			unsigned long address, unsigned int fault_flags),
+
+	TP_ARGS(tsk, mm, address, fault_flags),
+
+	TP_STRUCT__entry(
+		__array(	char,	comm,	TASK_COMM_LEN	)
+		__field(	unsigned long,	address		)
+	),
+
+	TP_fast_assign(
+		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
+		__entry->address	= address;
+	),
+
+	TP_printk("comm=%s address=%lx", __entry->comm, __entry->address)
+);
+
+TRACE_EVENT(gup_get_user_pages,
+
+	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
+			unsigned long start, unsigned long nr_pages,
+			unsigned int gup_flags, struct page **pages,
+			struct vm_area_struct **vmas, int *nonblocking),
+
+	TP_ARGS(tsk, mm, start, nr_pages, gup_flags, pages, vmas, nonblocking),
+
+	TP_STRUCT__entry(
+		__array(	char,	comm,	TASK_COMM_LEN	)
+		__field(	unsigned long,	start		)
+		__field(	unsigned long,	nr_pages	)
+	),
+
+	TP_fast_assign(
+		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
+		__entry->start		= start;
+		__entry->nr_pages	= nr_pages;
+	),
+
+	TP_printk("comm=%s start=%lx nr_pages=%lu", __entry->comm, __entry->start, __entry->nr_pages)
+);
+
+TRACE_EVENT(gup_get_user_pages_fast,
+
+	TP_PROTO(unsigned long start, int nr_pages, int write,
+			struct page **pages),
+
+	TP_ARGS(start, nr_pages, write, pages),
+
+	TP_STRUCT__entry(
+		__field(	unsigned long,	start		)
+		__field(	unsigned long,	nr_pages	)
+	),
+
+	TP_fast_assign(
+		__entry->start  	= start;
+		__entry->nr_pages	= nr_pages;
+	),
+
+	TP_printk("start=%lx nr_pages=%lu",  __entry->start, __entry->nr_pages)
+);
+
+#endif /* _TRACE_GUP_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>
-- 
2.0.2

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1281362 — Re: [PATCH 1/7] trace/events: Add gup trace events

On Tue,  1 Dec 2015 15:06:11 -0800
Yang Shi <yang.shi@linaro.org> wrote:

> page-faults events record the invoke to handle_mm_fault, but the invoke
> may come from do_page_fault or gup. In some use cases, the finer event count
> mey be needed, so add trace events support for:
> 
> __get_user_pages
> __get_user_pages_fast
> fixup_user_fault
> 
> Signed-off-by: Yang Shi <yang.shi@linaro.org>
> ---
>  include/trace/events/gup.h | 77 ++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 77 insertions(+)
>  create mode 100644 include/trace/events/gup.h
> 
> diff --git a/include/trace/events/gup.h b/include/trace/events/gup.h
> new file mode 100644
> index 0000000..37d18f9
> --- /dev/null
> +++ b/include/trace/events/gup.h
> @@ -0,0 +1,77 @@
> +#undef TRACE_SYSTEM
> +#define TRACE_SYSTEM gup
> +
> +#if !defined(_TRACE_GUP_H) || defined(TRACE_HEADER_MULTI_READ)
> +#define _TRACE_GUP_H
> +
> +#include <linux/types.h>
> +#include <linux/tracepoint.h>
> +
> +TRACE_EVENT(gup_fixup_user_fault,
> +
> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
> +			unsigned long address, unsigned int fault_flags),
> +
> +	TP_ARGS(tsk, mm, address, fault_flags),
> +
> +	TP_STRUCT__entry(
> +		__array(	char,	comm,	TASK_COMM_LEN	)

Why save the comm? The tracing infrastructure should keep track of that.

> +		__field(	unsigned long,	address		)
> +	),
> +
> +	TP_fast_assign(
> +		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
> +		__entry->address	= address;
> +	),
> +
> +	TP_printk("comm=%s address=%lx", __entry->comm, __entry->address)
> +);
> +
> +TRACE_EVENT(gup_get_user_pages,
> +
> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
> +			unsigned long start, unsigned long nr_pages,
> +			unsigned int gup_flags, struct page **pages,
> +			struct vm_area_struct **vmas, int *nonblocking),
> +
> +	TP_ARGS(tsk, mm, start, nr_pages, gup_flags, pages, vmas, nonblocking),

Why so many arguments? Most are not used.

-- Steve

> +
> +	TP_STRUCT__entry(
> +		__array(	char,	comm,	TASK_COMM_LEN	)
> +		__field(	unsigned long,	start		)
> +		__field(	unsigned long,	nr_pages	)
> +	),
> +
> +	TP_fast_assign(
> +		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
> +		__entry->start		= start;
> +		__entry->nr_pages	= nr_pages;
> +	),
> +
> +	TP_printk("comm=%s start=%lx nr_pages=%lu", __entry->comm, __entry->start, __entry->nr_pages)
> +);
> +
> +TRACE_EVENT(gup_get_user_pages_fast,
> +
> +	TP_PROTO(unsigned long start, int nr_pages, int write,
> +			struct page **pages),
> +
> +	TP_ARGS(start, nr_pages, write, pages),
> +
> +	TP_STRUCT__entry(
> +		__field(	unsigned long,	start		)
> +		__field(	unsigned long,	nr_pages	)
> +	),
> +
> +	TP_fast_assign(
> +		__entry->start  	= start;
> +		__entry->nr_pages	= nr_pages;
> +	),
> +
> +	TP_printk("start=%lx nr_pages=%lu",  __entry->start, __entry->nr_pages)
> +);
> +
> +#endif /* _TRACE_GUP_H */
> +
> +/* This part must be outside protection */
> +#include <trace/define_trace.h>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1281369 — Re: [PATCH 1/7] trace/events: Add gup trace events

On 12/1/2015 3:56 PM, Steven Rostedt wrote:
> On Tue,  1 Dec 2015 15:06:11 -0800
> Yang Shi <yang.shi@linaro.org> wrote:
>
>> page-faults events record the invoke to handle_mm_fault, but the invoke
>> may come from do_page_fault or gup. In some use cases, the finer event count
>> mey be needed, so add trace events support for:
>>
>> __get_user_pages
>> __get_user_pages_fast
>> fixup_user_fault
>>
>> Signed-off-by: Yang Shi <yang.shi@linaro.org>
>> ---
>>   include/trace/events/gup.h | 77 ++++++++++++++++++++++++++++++++++++++++++++++
>>   1 file changed, 77 insertions(+)
>>   create mode 100644 include/trace/events/gup.h
>>
>> diff --git a/include/trace/events/gup.h b/include/trace/events/gup.h
>> new file mode 100644
>> index 0000000..37d18f9
>> --- /dev/null
>> +++ b/include/trace/events/gup.h
>> @@ -0,0 +1,77 @@
>> +#undef TRACE_SYSTEM
>> +#define TRACE_SYSTEM gup
>> +
>> +#if !defined(_TRACE_GUP_H) || defined(TRACE_HEADER_MULTI_READ)
>> +#define _TRACE_GUP_H
>> +
>> +#include <linux/types.h>
>> +#include <linux/tracepoint.h>
>> +
>> +TRACE_EVENT(gup_fixup_user_fault,
>> +
>> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
>> +			unsigned long address, unsigned int fault_flags),
>> +
>> +	TP_ARGS(tsk, mm, address, fault_flags),
>> +
>> +	TP_STRUCT__entry(
>> +		__array(	char,	comm,	TASK_COMM_LEN	)
>
> Why save the comm? The tracing infrastructure should keep track of that.

The code is referred to kmem.h which has comm copied. If it is 
unnecessary, it definitely could be removed.

>
>> +		__field(	unsigned long,	address		)
>> +	),
>> +
>> +	TP_fast_assign(
>> +		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
>> +		__entry->address	= address;
>> +	),
>> +
>> +	TP_printk("comm=%s address=%lx", __entry->comm, __entry->address)
>> +);
>> +
>> +TRACE_EVENT(gup_get_user_pages,
>> +
>> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
>> +			unsigned long start, unsigned long nr_pages,
>> +			unsigned int gup_flags, struct page **pages,
>> +			struct vm_area_struct **vmas, int *nonblocking),
>> +
>> +	TP_ARGS(tsk, mm, start, nr_pages, gup_flags, pages, vmas, nonblocking),
>
> Why so many arguments? Most are not used.

My understanding to TP_ARGS may be not right. Doesn't it require all the 
args defined by the function? If not, it could definitely be shrunk. 
Just need keep the args used by TP_printk?

Thanks,
Yang

>
> -- Steve
>
>> +
>> +	TP_STRUCT__entry(
>> +		__array(	char,	comm,	TASK_COMM_LEN	)
>> +		__field(	unsigned long,	start		)
>> +		__field(	unsigned long,	nr_pages	)
>> +	),
>> +
>> +	TP_fast_assign(
>> +		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
>> +		__entry->start		= start;
>> +		__entry->nr_pages	= nr_pages;
>> +	),
>> +
>> +	TP_printk("comm=%s start=%lx nr_pages=%lu", __entry->comm, __entry->start, __entry->nr_pages)
>> +);
>> +
>> +TRACE_EVENT(gup_get_user_pages_fast,
>> +
>> +	TP_PROTO(unsigned long start, int nr_pages, int write,
>> +			struct page **pages),
>> +
>> +	TP_ARGS(start, nr_pages, write, pages),
>> +
>> +	TP_STRUCT__entry(
>> +		__field(	unsigned long,	start		)
>> +		__field(	unsigned long,	nr_pages	)
>> +	),
>> +
>> +	TP_fast_assign(
>> +		__entry->start  	= start;
>> +		__entry->nr_pages	= nr_pages;
>> +	),
>> +
>> +	TP_printk("start=%lx nr_pages=%lu",  __entry->start, __entry->nr_pages)
>> +);
>> +
>> +#endif /* _TRACE_GUP_H */
>> +
>> +/* This part must be outside protection */
>> +#include <trace/define_trace.h>
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1281371 — Re: [PATCH 1/7] trace/events: Add gup trace events

On Tue, 01 Dec 2015 16:07:44 -0800
"Shi, Yang" <yang.shi@linaro.org> wrote:

> On 12/1/2015 3:56 PM, Steven Rostedt wrote:
> > On Tue,  1 Dec 2015 15:06:11 -0800
> > Yang Shi <yang.shi@linaro.org> wrote:
> >  
> >> page-faults events record the invoke to handle_mm_fault, but the invoke
> >> may come from do_page_fault or gup. In some use cases, the finer event count
> >> mey be needed, so add trace events support for:
> >>
> >> __get_user_pages
> >> __get_user_pages_fast
> >> fixup_user_fault
> >>
> >> Signed-off-by: Yang Shi <yang.shi@linaro.org>
> >> ---
> >>   include/trace/events/gup.h | 77 ++++++++++++++++++++++++++++++++++++++++++++++
> >>   1 file changed, 77 insertions(+)
> >>   create mode 100644 include/trace/events/gup.h
> >>
> >> diff --git a/include/trace/events/gup.h b/include/trace/events/gup.h
> >> new file mode 100644
> >> index 0000000..37d18f9
> >> --- /dev/null
> >> +++ b/include/trace/events/gup.h
> >> @@ -0,0 +1,77 @@
> >> +#undef TRACE_SYSTEM
> >> +#define TRACE_SYSTEM gup
> >> +
> >> +#if !defined(_TRACE_GUP_H) || defined(TRACE_HEADER_MULTI_READ)
> >> +#define _TRACE_GUP_H
> >> +
> >> +#include <linux/types.h>
> >> +#include <linux/tracepoint.h>
> >> +
> >> +TRACE_EVENT(gup_fixup_user_fault,
> >> +
> >> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
> >> +			unsigned long address, unsigned int fault_flags),
> >> +
> >> +	TP_ARGS(tsk, mm, address, fault_flags),
> >> +
> >> +	TP_STRUCT__entry(
> >> +		__array(	char,	comm,	TASK_COMM_LEN	)  
> >
> > Why save the comm? The tracing infrastructure should keep track of that.  
> 
> The code is referred to kmem.h which has comm copied. If it is 
> unnecessary, it definitely could be removed.

Sometimes comm isn't that reliable. But really, the only tracepoint
that should record it is sched_switch, and sched_wakeup. With those
two, the rest of the trace points should be fine.

> 
> >  
> >> +		__field(	unsigned long,	address		)
> >> +	),
> >> +
> >> +	TP_fast_assign(
> >> +		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
> >> +		__entry->address	= address;
> >> +	),
> >> +
> >> +	TP_printk("comm=%s address=%lx", __entry->comm, __entry->address)
> >> +);
> >> +
> >> +TRACE_EVENT(gup_get_user_pages,
> >> +
> >> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
> >> +			unsigned long start, unsigned long nr_pages,
> >> +			unsigned int gup_flags, struct page **pages,
> >> +			struct vm_area_struct **vmas, int *nonblocking),
> >> +
> >> +	TP_ARGS(tsk, mm, start, nr_pages, gup_flags, pages, vmas, nonblocking),  
> >
> > Why so many arguments? Most are not used.  
> 
> My understanding to TP_ARGS may be not right. Doesn't it require all the 
> args defined by the function? If not, it could definitely be shrunk. 
> Just need keep the args used by TP_printk?

It only needs what is used by TP_fast_assign().

-- Steve
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1281374 — Re: [PATCH 1/7] trace/events: Add gup trace events

On 12/1/2015 4:18 PM, Steven Rostedt wrote:
> On Tue, 01 Dec 2015 16:07:44 -0800
> "Shi, Yang" <yang.shi@linaro.org> wrote:
>
>> On 12/1/2015 3:56 PM, Steven Rostedt wrote:
>>> On Tue,  1 Dec 2015 15:06:11 -0800
>>> Yang Shi <yang.shi@linaro.org> wrote:
>>>
>>>> page-faults events record the invoke to handle_mm_fault, but the invoke
>>>> may come from do_page_fault or gup. In some use cases, the finer event count
>>>> mey be needed, so add trace events support for:
>>>>
>>>> __get_user_pages
>>>> __get_user_pages_fast
>>>> fixup_user_fault
>>>>
>>>> Signed-off-by: Yang Shi <yang.shi@linaro.org>
>>>> ---
>>>>    include/trace/events/gup.h | 77 ++++++++++++++++++++++++++++++++++++++++++++++
>>>>    1 file changed, 77 insertions(+)
>>>>    create mode 100644 include/trace/events/gup.h
>>>>
>>>> diff --git a/include/trace/events/gup.h b/include/trace/events/gup.h
>>>> new file mode 100644
>>>> index 0000000..37d18f9
>>>> --- /dev/null
>>>> +++ b/include/trace/events/gup.h
>>>> @@ -0,0 +1,77 @@
>>>> +#undef TRACE_SYSTEM
>>>> +#define TRACE_SYSTEM gup
>>>> +
>>>> +#if !defined(_TRACE_GUP_H) || defined(TRACE_HEADER_MULTI_READ)
>>>> +#define _TRACE_GUP_H
>>>> +
>>>> +#include <linux/types.h>
>>>> +#include <linux/tracepoint.h>
>>>> +
>>>> +TRACE_EVENT(gup_fixup_user_fault,
>>>> +
>>>> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
>>>> +			unsigned long address, unsigned int fault_flags),
>>>> +
>>>> +	TP_ARGS(tsk, mm, address, fault_flags),
>>>> +
>>>> +	TP_STRUCT__entry(
>>>> +		__array(	char,	comm,	TASK_COMM_LEN	)
>>>
>>> Why save the comm? The tracing infrastructure should keep track of that.
>>
>> The code is referred to kmem.h which has comm copied. If it is
>> unnecessary, it definitely could be removed.
>
> Sometimes comm isn't that reliable. But really, the only tracepoint
> that should record it is sched_switch, and sched_wakeup. With those
> two, the rest of the trace points should be fine.
>
>>
>>>
>>>> +		__field(	unsigned long,	address		)
>>>> +	),
>>>> +
>>>> +	TP_fast_assign(
>>>> +		memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
>>>> +		__entry->address	= address;
>>>> +	),
>>>> +
>>>> +	TP_printk("comm=%s address=%lx", __entry->comm, __entry->address)
>>>> +);
>>>> +
>>>> +TRACE_EVENT(gup_get_user_pages,
>>>> +
>>>> +	TP_PROTO(struct task_struct *tsk, struct mm_struct *mm,
>>>> +			unsigned long start, unsigned long nr_pages,
>>>> +			unsigned int gup_flags, struct page **pages,
>>>> +			struct vm_area_struct **vmas, int *nonblocking),
>>>> +
>>>> +	TP_ARGS(tsk, mm, start, nr_pages, gup_flags, pages, vmas, nonblocking),
>>>
>>> Why so many arguments? Most are not used.
>>
>> My understanding to TP_ARGS may be not right. Doesn't it require all the
>> args defined by the function? If not, it could definitely be shrunk.
>> Just need keep the args used by TP_printk?
>
> It only needs what is used by TP_fast_assign().

Thanks, will fix them in V2.

Yang

>
> -- Steve
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1281334 — [PATCH 4/7] mips: mm/gup: add gup trace points

Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Signed-off-by: Yang Shi <yang.shi@linaro.org>
---
 arch/mips/mm/gup.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/mips/mm/gup.c b/arch/mips/mm/gup.c
index 349995d..3c5b8c8 100644
--- a/arch/mips/mm/gup.c
+++ b/arch/mips/mm/gup.c
@@ -12,6 +12,9 @@
 #include <linux/swap.h>
 #include <linux/hugetlb.h>
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/gup.h>
+
 #include <asm/cpu-features.h>
 #include <asm/pgtable.h>
 
@@ -211,6 +214,8 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 					(void __user *)start, len)))
 		return 0;
 
+	trace_gup_get_user_pages_fast(start, nr_pages, write, pages);
+
 	/*
 	 * XXX: batch / limit 'nr', to avoid large irq off latency
 	 * needs some instrumenting to determine the common sizes used by
@@ -277,6 +282,8 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	if (end < start || cpu_has_dc_aliases)
 		goto slow_irqon;
 
+	trace_gup_get_user_pages_fast(start, nr_pages, write, pages);
+
 	/* XXX: batch / limit 'nr' */
 	local_irq_disable();
 	pgdp = pgd_offset(mm, addr);
-- 
2.0.2

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1281780 — Re: [PATCH 4/7] mips: mm/gup: add gup trace points

On Tue, Dec 01, 2015 at 03:06:14PM -0800, Yang Shi wrote:

>  arch/mips/mm/gup.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/arch/mips/mm/gup.c b/arch/mips/mm/gup.c
> index 349995d..3c5b8c8 100644
> --- a/arch/mips/mm/gup.c
> +++ b/arch/mips/mm/gup.c
> @@ -12,6 +12,9 @@
>  #include <linux/swap.h>
>  #include <linux/hugetlb.h>
>  
> +#define CREATE_TRACE_POINTS
> +#include <trace/events/gup.h>
> +
>  #include <asm/cpu-features.h>
>  #include <asm/pgtable.h>
>  
> @@ -211,6 +214,8 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
>  					(void __user *)start, len)))
>  		return 0;
>  
> +	trace_gup_get_user_pages_fast(start, nr_pages, write, pages);
> +
>  	/*
>  	 * XXX: batch / limit 'nr', to avoid large irq off latency
>  	 * needs some instrumenting to determine the common sizes used by
> @@ -277,6 +282,8 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
>  	if (end < start || cpu_has_dc_aliases)
>  		goto slow_irqon;
>  
> +	trace_gup_get_user_pages_fast(start, nr_pages, write, pages);
> +
>  	/* XXX: batch / limit 'nr' */
>  	local_irq_disable();
>  	pgdp = pgd_offset(mm, addr);

Acked-by: Ralf Baechle <ralf@linux-mips.org>

Please feel free to merge this upstream with the remainder of the
series once it's been acked.

  Ralf
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web