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


Groups > linux.kernel > #1214651 > unrolled thread

[PATCH 2/3] mm, compaction: export tracepoints zone names to userspace

Started byVlastimil Babka <vbabka@suse.cz>
First post2015-08-27 17:30 +0200
Last post2015-08-31 17:40 +0200
Articles 5 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace Vlastimil Babka <vbabka@suse.cz> - 2015-08-27 17:30 +0200
    Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to  userspace Steven Rostedt <rostedt@goodmis.org> - 2015-08-31 17:00 +0200
      Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to  userspace Vlastimil Babka <vbabka@suse.cz> - 2015-08-31 17:10 +0200
        Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to  userspace Steven Rostedt <rostedt@goodmis.org> - 2015-08-31 17:20 +0200
    Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to  userspace Steven Rostedt <rostedt@goodmis.org> - 2015-08-31 17:40 +0200

#1214651 — [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace

FromVlastimil Babka <vbabka@suse.cz>
Date2015-08-27 17:30 +0200
Subject[PATCH 2/3] mm, compaction: export tracepoints zone names to userspace
Message-ID<q27hM-8oR-27@gated-at.bofh.it>
Some compaction tracepoints use zone->name to print which zone is being
compacted. This works for in-kernel printing, but not userspace trace printing
of raw captured trace such as via trace-cmd report.

This patch uses zone_idx() instead of zone->name as the raw value, and when
printing, converts the zone_type to string using the appropriate EM() macros
and some ugly tricks to overcome the problem that half the values depend on
CONFIG_ options and one does not simply use #ifdef inside of #define.

trace-cmd output before:
transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
zone=ffffffff81815d7a order=9 ret=partial

after:
transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
zone=Normal   order=9 ret=partial

Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: David Rientjes <rientjes@google.com>
---
 include/trace/events/compaction.h | 38 ++++++++++++++++++++++++++++++++------
 1 file changed, 32 insertions(+), 6 deletions(-)

diff --git a/include/trace/events/compaction.h b/include/trace/events/compaction.h
index 1275a55..8daa8fa 100644
--- a/include/trace/events/compaction.h
+++ b/include/trace/events/compaction.h
@@ -18,6 +18,31 @@
 	EM( COMPACT_NO_SUITABLE_PAGE,	"no_suitable_page")	\
 	EMe(COMPACT_NOT_SUITABLE_ZONE,	"not_suitable_zone")
 
+#ifdef CONFIG_ZONE_DMA
+#define IFDEF_ZONE_DMA(X) X
+#else
+#define IFDEF_ZONE_DMA(X)
+#endif
+
+#ifdef CONFIG_ZONE_DMA32
+#define IFDEF_ZONE_DMA32(X) X
+#else
+#define IFDEF_ZONE_DMA32(X)
+#endif
+
+#ifdef CONFIG_ZONE_HIGHMEM_
+#define IFDEF_ZONE_HIGHMEM(X) X
+#else
+#define IFDEF_ZONE_HIGHMEM(X)
+#endif
+
+#define ZONE_TYPE						\
+	IFDEF_ZONE_DMA(		EM (ZONE_DMA,	 "DMA"))	\
+	IFDEF_ZONE_DMA32(	EM (ZONE_DMA32,	 "DMA32"))	\
+				EM (ZONE_NORMAL, "Normal")	\
+	IFDEF_ZONE_HIGHMEM(	EM (ZONE_HIGHMEM,"HighMem"))	\
+				EMe(ZONE_MOVABLE,"Movable")
+
 /*
  * First define the enums in the above macros to be exported to userspace
  * via TRACE_DEFINE_ENUM().
@@ -28,6 +53,7 @@
 #define EMe(a, b)	TRACE_DEFINE_ENUM(a);
 
 COMPACTION_STATUS
+ZONE_TYPE
 
 /*
  * Now redefine the EM() and EMe() macros to map the enums to the strings
@@ -230,21 +256,21 @@ DECLARE_EVENT_CLASS(mm_compaction_suitable_template,
 
 	TP_STRUCT__entry(
 		__field(int, nid)
-		__field(char *, name)
+		__field(enum zone_type, idx)
 		__field(int, order)
 		__field(int, ret)
 	),
 
 	TP_fast_assign(
 		__entry->nid = zone_to_nid(zone);
-		__entry->name = (char *)zone->name;
+		__entry->idx = zone_idx(zone);
 		__entry->order = order;
 		__entry->ret = ret;
 	),
 
 	TP_printk("node=%d zone=%-8s order=%d ret=%s",
 		__entry->nid,
-		__entry->name,
+		__print_symbolic(__entry->idx, ZONE_TYPE),
 		__entry->order,
 		__print_symbolic(__entry->ret, COMPACTION_STATUS))
 );
@@ -276,7 +302,7 @@ DECLARE_EVENT_CLASS(mm_compaction_defer_template,
 
 	TP_STRUCT__entry(
 		__field(int, nid)
-		__field(char *, name)
+		__field(enum zone_type, idx)
 		__field(int, order)
 		__field(unsigned int, considered)
 		__field(unsigned int, defer_shift)
@@ -285,7 +311,7 @@ DECLARE_EVENT_CLASS(mm_compaction_defer_template,
 
 	TP_fast_assign(
 		__entry->nid = zone_to_nid(zone);
-		__entry->name = (char *)zone->name;
+		__entry->idx = zone_idx(zone);
 		__entry->order = order;
 		__entry->considered = zone->compact_considered;
 		__entry->defer_shift = zone->compact_defer_shift;
@@ -294,7 +320,7 @@ DECLARE_EVENT_CLASS(mm_compaction_defer_template,
 
 	TP_printk("node=%d zone=%-8s order=%d order_failed=%d consider=%u limit=%lu",
 		__entry->nid,
-		__entry->name,
+		__print_symbolic(__entry->idx, ZONE_TYPE),
 		__entry->order,
 		__entry->order_failed,
 		__entry->considered,
-- 
2.5.0

--
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]


#1216243 — Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace

FromSteven Rostedt <rostedt@goodmis.org>
Date2015-08-31 17:00 +0200
SubjectRe: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace
Message-ID<q3yIX-2rV-41@gated-at.bofh.it>
In reply to#1214651
On Thu, 27 Aug 2015 17:24:03 +0200
Vlastimil Babka <vbabka@suse.cz> wrote:

> Some compaction tracepoints use zone->name to print which zone is being
> compacted. This works for in-kernel printing, but not userspace trace printing
> of raw captured trace such as via trace-cmd report.
> 
> This patch uses zone_idx() instead of zone->name as the raw value, and when
> printing, converts the zone_type to string using the appropriate EM() macros
> and some ugly tricks to overcome the problem that half the values depend on
> CONFIG_ options and one does not simply use #ifdef inside of #define.
> 
> trace-cmd output before:
> transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
> zone=ffffffff81815d7a order=9 ret=partial
> 
> after:
> transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
> zone=Normal   order=9 ret=partial
> 
> Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> Cc: Steven Rostedt <rostedt@goodmis.org>
> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: Mel Gorman <mgorman@suse.de>
> Cc: David Rientjes <rientjes@google.com>
> ---
>  include/trace/events/compaction.h | 38 ++++++++++++++++++++++++++++++++------
>  1 file changed, 32 insertions(+), 6 deletions(-)
> 
> diff --git a/include/trace/events/compaction.h b/include/trace/events/compaction.h
> index 1275a55..8daa8fa 100644
> --- a/include/trace/events/compaction.h
> +++ b/include/trace/events/compaction.h
> @@ -18,6 +18,31 @@
>  	EM( COMPACT_NO_SUITABLE_PAGE,	"no_suitable_page")	\
>  	EMe(COMPACT_NOT_SUITABLE_ZONE,	"not_suitable_zone")
>  
> +#ifdef CONFIG_ZONE_DMA
> +#define IFDEF_ZONE_DMA(X) X
> +#else
> +#define IFDEF_ZONE_DMA(X)
> +#endif
> +
> +#ifdef CONFIG_ZONE_DMA32
> +#define IFDEF_ZONE_DMA32(X) X
> +#else
> +#define IFDEF_ZONE_DMA32(X)
> +#endif
> +
> +#ifdef CONFIG_ZONE_HIGHMEM_
> +#define IFDEF_ZONE_HIGHMEM(X) X
> +#else
> +#define IFDEF_ZONE_HIGHMEM(X)
> +#endif
> +
> +#define ZONE_TYPE						\
> +	IFDEF_ZONE_DMA(		EM (ZONE_DMA,	 "DMA"))	\
> +	IFDEF_ZONE_DMA32(	EM (ZONE_DMA32,	 "DMA32"))	\
> +				EM (ZONE_NORMAL, "Normal")	\
> +	IFDEF_ZONE_HIGHMEM(	EM (ZONE_HIGHMEM,"HighMem"))	\
> +				EMe(ZONE_MOVABLE,"Movable")
> +

Hmm, have you tried to compile this with CONFIG_ZONE_HIGHMEM disabled,
and CONFIG_ZONE_DMA and/or CONFIG_ZONE_DMA32 enabled?

The EMe() macro must come last, as it doesn't have the ending comma and
the __print_symbolic() can fail to compile due to it.

-- Steve


>  /*
>   * First define the enums in the above macros to be exported to userspace
>   * via TRACE_DEFINE_ENUM().
> @@ -28,6 +53,7 @@
--
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]


#1216247 — Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace

FromVlastimil Babka <vbabka@suse.cz>
Date2015-08-31 17:10 +0200
SubjectRe: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace
Message-ID<q3ySC-2SC-13@gated-at.bofh.it>
In reply to#1216243
On 08/31/2015 04:58 PM, Steven Rostedt wrote:
> On Thu, 27 Aug 2015 17:24:03 +0200
> Vlastimil Babka <vbabka@suse.cz> wrote:
>
>> Some compaction tracepoints use zone->name to print which zone is being
>> compacted. This works for in-kernel printing, but not userspace trace printing
>> of raw captured trace such as via trace-cmd report.
>>
>> This patch uses zone_idx() instead of zone->name as the raw value, and when
>> printing, converts the zone_type to string using the appropriate EM() macros
>> and some ugly tricks to overcome the problem that half the values depend on
>> CONFIG_ options and one does not simply use #ifdef inside of #define.
>>
>> trace-cmd output before:
>> transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
>> zone=ffffffff81815d7a order=9 ret=partial
>>
>> after:
>> transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
>> zone=Normal   order=9 ret=partial
>>
>> Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
>> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
>> Cc: Steven Rostedt <rostedt@goodmis.org>
>> Cc: Ingo Molnar <mingo@redhat.com>
>> Cc: Mel Gorman <mgorman@suse.de>
>> Cc: David Rientjes <rientjes@google.com>
>> ---
>>   include/trace/events/compaction.h | 38 ++++++++++++++++++++++++++++++++------
>>   1 file changed, 32 insertions(+), 6 deletions(-)
>>
>> diff --git a/include/trace/events/compaction.h b/include/trace/events/compaction.h
>> index 1275a55..8daa8fa 100644
>> --- a/include/trace/events/compaction.h
>> +++ b/include/trace/events/compaction.h
>> @@ -18,6 +18,31 @@
>>   	EM( COMPACT_NO_SUITABLE_PAGE,	"no_suitable_page")	\
>>   	EMe(COMPACT_NOT_SUITABLE_ZONE,	"not_suitable_zone")
>>
>> +#ifdef CONFIG_ZONE_DMA
>> +#define IFDEF_ZONE_DMA(X) X
>> +#else
>> +#define IFDEF_ZONE_DMA(X)
>> +#endif
>> +
>> +#ifdef CONFIG_ZONE_DMA32
>> +#define IFDEF_ZONE_DMA32(X) X
>> +#else
>> +#define IFDEF_ZONE_DMA32(X)
>> +#endif
>> +
>> +#ifdef CONFIG_ZONE_HIGHMEM_
>> +#define IFDEF_ZONE_HIGHMEM(X) X
>> +#else
>> +#define IFDEF_ZONE_HIGHMEM(X)
>> +#endif
>> +
>> +#define ZONE_TYPE						\
>> +	IFDEF_ZONE_DMA(		EM (ZONE_DMA,	 "DMA"))	\
>> +	IFDEF_ZONE_DMA32(	EM (ZONE_DMA32,	 "DMA32"))	\
>> +				EM (ZONE_NORMAL, "Normal")	\
>> +	IFDEF_ZONE_HIGHMEM(	EM (ZONE_HIGHMEM,"HighMem"))	\
>> +				EMe(ZONE_MOVABLE,"Movable")
>> +
>
> Hmm, have you tried to compile this with CONFIG_ZONE_HIGHMEM disabled,
> and CONFIG_ZONE_DMA and/or CONFIG_ZONE_DMA32 enabled?

Yep, that's standard x86_64 situation (highmem disabled, dma+dma32 enabled).

> The EMe() macro must come last, as it doesn't have the ending comma and
> the __print_symbolic() can fail to compile due to it.

Thanks to ZONE_MOVABLE being unconditional, EMe(ZONE_MOVABLE...) is 
always last. Otherwise the macros would get even more ugly...

> -- Steve
>
>
>>   /*
>>    * First define the enums in the above macros to be exported to userspace
>>    * via TRACE_DEFINE_ENUM().
>> @@ -28,6 +53,7 @@

--
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]


#1216257 — Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace

FromSteven Rostedt <rostedt@goodmis.org>
Date2015-08-31 17:20 +0200
SubjectRe: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace
Message-ID<q3z2i-33U-21@gated-at.bofh.it>
In reply to#1216247
On Mon, 31 Aug 2015 17:02:38 +0200
Vlastimil Babka <vbabka@suse.cz> wrote:


> >> +#define ZONE_TYPE						\
> >> +	IFDEF_ZONE_DMA(		EM (ZONE_DMA,	 "DMA"))	\
> >> +	IFDEF_ZONE_DMA32(	EM (ZONE_DMA32,	 "DMA32"))	\
> >> +				EM (ZONE_NORMAL, "Normal")	\
> >> +	IFDEF_ZONE_HIGHMEM(	EM (ZONE_HIGHMEM,"HighMem"))	\
> >> +				EMe(ZONE_MOVABLE,"Movable")
> >> +
> >
> > Hmm, have you tried to compile this with CONFIG_ZONE_HIGHMEM disabled,
> > and CONFIG_ZONE_DMA and/or CONFIG_ZONE_DMA32 enabled?
> 
> Yep, that's standard x86_64 situation (highmem disabled, dma+dma32 enabled).
> 
> > The EMe() macro must come last, as it doesn't have the ending comma and
> > the __print_symbolic() can fail to compile due to it.
> 
> Thanks to ZONE_MOVABLE being unconditional, EMe(ZONE_MOVABLE...) is 
> always last. Otherwise the macros would get even more ugly...

Ah! My mistake was to see where the end parenthesis of
IFDEF_ZONE_HIGHMEM() laid. It looked to me that it encompassed
ZONE_MOVABLE. But obviously it does not.

-- 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]


#1216271 — Re: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace

FromSteven Rostedt <rostedt@goodmis.org>
Date2015-08-31 17:40 +0200
SubjectRe: [PATCH 2/3] mm, compaction: export tracepoints zone names to userspace
Message-ID<q3zlF-3rM-37@gated-at.bofh.it>
In reply to#1214651
On Thu, 27 Aug 2015 17:24:03 +0200
Vlastimil Babka <vbabka@suse.cz> wrote:

> Some compaction tracepoints use zone->name to print which zone is being
> compacted. This works for in-kernel printing, but not userspace trace printing
> of raw captured trace such as via trace-cmd report.
> 
> This patch uses zone_idx() instead of zone->name as the raw value, and when
> printing, converts the zone_type to string using the appropriate EM() macros
> and some ugly tricks to overcome the problem that half the values depend on
> CONFIG_ options and one does not simply use #ifdef inside of #define.
> 
> trace-cmd output before:
> transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
> zone=ffffffff81815d7a order=9 ret=partial
> 
> after:
> transhuge-stres-4235  [000]   453.149280: mm_compaction_finished: node=0
> zone=Normal   order=9 ret=partial
> 
> Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> Cc: Steven Rostedt <rostedt@goodmis.org>

Reviewed-by: Steven Rostedt <rostedt@goodmis.org>

-- Steve

> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: Mel Gorman <mgorman@suse.de>
> Cc: David Rientjes <rientjes@google.com>
> ---
>  include/trace/events/compaction.h | 38 ++++++++++++++++++++++++++++++++------
>  1 file changed, 32 insertions(+), 6 deletions(-)
> 
--
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