mm: add show num_poisoned_pages when oom

PATCH Superseded Xishi Qiu 4 participants 5 messages Dec 9, 2013

Subsystems (2)

Memory management Library code

lore

Share & tools

lore

Xishi Qiu

author

2013-12-09 09:53 UTC #0 lore raw

PATCH 1/1 mm: add show num_poisoned_pages when oom

Show num_poisoned_pages when oom, it is helpful to find the reason.

Signed-off-by Xishi Qiu <[email protected]>

---
 lib/show_mem.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

lib/show_mem.c +3 -0

index 5847a49..1cbdcd8 100644
@@ -46,4 +46,7 @@ void show_mem(unsigned int filter)
 	printk("%lu pages in pagetable cache\n",
 		quicklist_total_size());
 #endif
+#ifdef CONFIG_MEMORY_FAILURE
+	printk("%lu pages poisoned\n", atomic_long_read(&num_poisoned_pages));
+#endif
 }

Signature · 7 lines

1.7.1

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Michal Hocko

Acked-by on v1 1/1

2013-12-09 15:50 UTC #1 lore raw

Quoted 3 lines · Xishi Qiu

Show num_poisoned_pages when oom, it is helpful to find the reason.
Signed-off-by Xishi Qiu <[email protected]>

I would not expect the number of these pages would be too high to matter in real life but having the information cannot be harmful in any way.

Acked-by Michal Hocko <[email protected]>

Quoted 19 lines

---
 lib/show_mem.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

lib/show_mem.c +3 -0

index 5847a49..1cbdcd8 100644
@@ -46,4 +46,7 @@ void show_mem(unsigned int filter)
 	printk("%lu pages in pagetable cache\n",
 		quicklist_total_size());
 #endif
+#ifdef CONFIG_MEMORY_FAILURE
+	printk("%lu pages poisoned\n", atomic_long_read(&num_poisoned_pages));
+#endif
 }

Signature · 1 line

1.7.1

Signature · 8 lines

Michal Hocko
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

David Rientjes

Acked-by on v1 1/1

2013-12-09 19:50 UTC #2 lore raw

In reply to Xishi Qiu

Quoted 3 lines · Xishi Qiu

Show num_poisoned_pages when oom, it is helpful to find the reason.
Signed-off-by Xishi Qiu <[email protected]>

Acked-by David Rientjes <[email protected]>

Although this patch says it's for oom conditions, it will be emitted anytime show_mem() is called which sounds good as well.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Naoya Horiguchi

2013-12-09 21:49 UTC #3 lore raw

In reply to Xishi Qiu

Quoted 12 lines · Xishi Qiu

Show num_poisoned_pages when oom, it is helpful to find the reason.
Signed-off-by Xishi Qiu <[email protected]>
---
 lib/show_mem.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)
lib/show_mem.c +0 -0
index 5847a49..1cbdcd8 100644

lib/show_mem.c void show_mem(unsigned int filter) · line 52

Quoted 7 lines

lib/show_mem.c +3 -0

@@ -46,4 +46,7 @@ void show_mem(unsigned int filter)
 	printk("%lu pages in pagetable cache\n",
 		quicklist_total_size());
 #endif
+#ifdef CONFIG_MEMORY_FAILURE
+	printk("%lu pages poisoned\n", atomic_long_read(&num_poisoned_pages));
+#endif
 }

I think that just "poisoned" could be confusing because this word seems to be used also in other context (like slab and list_debug handling.) "hwpoisoned" or "hardware corrupted" (which is the same label in /proc/meminfo) looks better to me.

Thanks,
Naoya Horiguchi

David Rientjes

2013-12-09 22:05 UTC #4 lore raw

Quoted 16 lines · Naoya Horiguchi

Quoted 16 lines
Show num_poisoned_pages when oom, it is helpful to find the reason.
Signed-off-by Xishi Qiu <[email protected]>
---
 lib/show_mem.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)
lib/show_mem.c +1 -0
index 5847a49..1cbdcd8 100644
@@ -46,4 +46,7 @@ void show_mem(unsigned int filter)
 	printk("%lu pages in pagetable cache\n",
 		quicklist_total_size());
 #endif
+#ifdef CONFIG_MEMORY_FAILURE

Quoted 9 lines

lib/show_mem.c void show_mem(unsigned int filter) · line 52
Quoted 3 lines
lib/show_mem.c +2 -0
+	printk("%lu pages poisoned\n", atomic_long_read(&num_poisoned_pages));
+#endif
 }
I think that just "poisoned" could be confusing because this word seems to be used also in other context (like slab and list_debug handling.) "hwpoisoned" or "hardware corrupted" (which is the same label in /proc/meminfo) looks better to me.

Ah, good point.