Skip to contentSkip to conversation

page_owner: print stacks and their counter

PATCH Superseded Oscar Salvador 1 participant 4 messages May 16, 2023

Subsystems (3)

v5 of 10 Superseded Latest is v10 →

lore
Share & tools
lore
#0 lore raw

v5 cover page_owner: print stacks and their counter

Changes v4 -> v5:
- Move page_owner_stack_threshold to page_owner file (suggested by Alexander Potapenko)
- Make the stack refcount more generic, and do not assume it will be only used for page_owner (suggested by Alexander Potapenko)

Changes v3 -> v4:
- Rebase (long time has passed)
- Use boolean instead of enum for action by Alexander Potapenko
- (I left some feedback untouched because it's been long and would like to discuss it here now instead of re-vamping and old thread)

Changes v2 -> v3:
- Replace interface in favor of seq operations (suggested by Vlastimil)
- Use debugfs interface to store/read valued (suggested by Ammar)

Hi,

page_owner is a great debug functionality tool that gets us to know about all pages that have been allocated/freed and their stacktrace. This comes very handy when e.g: debugging leaks, as with some scripting we might be able to see those stacktraces that are allocating pages but not freeing theme.

In my experience, that is one of the most useful cases, but it can get really tedious to screen through all pages aand try to reconstruct the stack <-> allocated/freed relationship. There is a lot of noise to cancel off.

This patch aims to fix that by adding a new functionality into page_owner. What this does is to create a new read-only file "page_owner_stacks", which prints only the allocating stacktraces and their counting, being that the times the stacktrace has allocated - the times it has freed.

So we have a clear overview of stacks <-> allocated/freed relationship without the need to fiddle with pages and trying to match free stacktraces with allocated stacktraces.

This is achieved by adding a new refcount_t field in the stack_record struct, incrementing that refcount_t everytime the same stacktrace allocates, and decrementing it when it frees a page. Details can be seen in the respective patches.

We also create another file called "page_owner_threshold", which let us specify a threshold, so when when reading from "page_owner_stacks", we will only see those stacktraces which counting goes beyond the threshold we specified.

One thing I am not completely happy about is to polute lib/stackdepot.c file with the stack_* functions.
We could sort that out if the stack_record struct definitions were in a header file instead of stackdepot.c.
But I am not sure about that trade-off, so suggestions are accepted.

A PoC can be found below:

# cat /sys/kernel/debug/page_owner_threshold
 0
# cat /sys/kernel/debug/page_owner_stacks > stacks_full.txt
# head -32 stacks_full.txt
 prep_new_page+0x10d/0x180
 get_page_from_freelist+0x1bd6/0x1e10
 __alloc_pages+0x194/0x360
 alloc_page_interleave+0x13/0x90
 new_slab+0x31d/0x530
 ___slab_alloc+0x5d7/0x720
 __slab_alloc.isra.85+0x4a/0x90
 kmem_cache_alloc+0x455/0x4a0
 acpi_ps_alloc_op+0x57/0x8f
 acpi_ps_create_scope_op+0x12/0x23
 acpi_ps_execute_method+0x102/0x2c1
 acpi_ns_evaluate+0x343/0x4da
 acpi_evaluate_object+0x1cb/0x392
 acpi_run_osc+0x135/0x260
 acpi_init+0x165/0x4ed
 do_one_initcall+0x3e/0x200
stack count: 2
 free_pcp_prepare+0x287/0x5c0
 free_unref_page+0x1c/0xd0
 __mmdrop+0x50/0x160
 finish_task_switch+0x249/0x2b0
 __schedule+0x2c3/0x960
 schedule+0x44/0xb0
 futex_wait_queue+0x70/0xd0
 futex_wait+0x160/0x250
 do_futex+0x11c/0x1b0
 __x64_sys_futex+0x5e/0x1d0
 do_syscall_64+0x37/0x90
 entry_SYSCALL_64_after_hwframe+0x63/0xcd

stack count: 1

# echo 10000 > /sys/kernel/debug/page_owner_threshold
# cat /sys/kernel/debug/page_owner_stacks > stacks_10000.txt
# cat stacks_10000.txt 
 prep_new_page+0x10d/0x180
 get_page_from_freelist+0x1bd6/0x1e10
 __alloc_pages+0x194/0x360
 folio_alloc+0x17/0x40
 page_cache_ra_unbounded+0x96/0x170
 filemap_get_pages+0x23d/0x5e0
 filemap_read+0xbf/0x3a0
 __kernel_read+0x136/0x2f0
 kernel_read_file+0x197/0x2d0
 kernel_read_file_from_fd+0x54/0x90
 __do_sys_finit_module+0x89/0x120
 do_syscall_64+0x37/0x90
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
stack count: 36195
 prep_new_page+0x10d/0x180
 get_page_from_freelist+0x1bd6/0x1e10
 __alloc_pages+0x194/0x360
 folio_alloc+0x17/0x40
 page_cache_ra_unbounded+0x96/0x170
 filemap_get_pages+0x23d/0x5e0
 filemap_read+0xbf/0x3a0
 new_sync_read+0x106/0x180
 vfs_read+0x16f/0x190
 ksys_read+0xa5/0xe0
 do_syscall_64+0x37/0x90
 entry_SYSCALL_64_after_hwframe+0x63/0xcd

stack count: 44484

 prep_new_page+0x10d/0x180
 get_page_from_freelist+0x1bd6/0x1e10
 __alloc_pages+0x194/0x360
 folio_alloc+0x17/0x40
 page_cache_ra_unbounded+0x96/0x170
 filemap_get_pages+0xdd/0x5e0
 filemap_read+0xbf/0x3a0
 new_sync_read+0x106/0x180
 vfs_read+0x16f/0x190
 ksys_read+0xa5/0xe0
 do_syscall_64+0x37/0x90
 entry_SYSCALL_64_after_hwframe+0x63/0xcd

stack count: 17874

Oscar Salvador (3):
lib/stackdepot: Add a refcount field in stack_record
mm, page_owner: Add page_owner_stacks file to print out only stacks
and their counte
mm,page_owner: Filter out stacks by a threshold counter

 include/linux/stackdepot.h |   8 +++
 lib/stackdepot.c           | 128 ++++++++++++++++++++++++++++++++++---
 mm/page_owner.c            |  54 ++++++++++++++++
 3 files changed, 180 insertions(+), 10 deletions(-)
Signature · 1 line
2.35.3
#1 lore raw

PATCH v5 1/3 lib/stackdepot: Add a refcount field in stack_record

We want to filter out page_owner output and print only those stacks that have been repeated beyond a certain threshold. This gives us the chance to get rid of a lot of noise.
In order to do that, we need to keep track of how many repeated stacks (for allocation) do we have, so we add a new refcount_t field in the stack_record struct.

Note that this might increase the size of the struct for some architectures.
E.g: x86_64 is not affected due to alignment, but x86 32bits might.

The alternative would be to have some kind of struct like this:

struct track_stacks {
	struct stack_record *stack;
	struct track_stacks *next;
	refcount_t stack_count;

But ithat would imply to perform more allocations and glue everything together, which would make the code more complex, so I think that going with a new field in the struct stack_record is good enough.

Note that on __set_page_owner_handle(), page_owner->handle is set, and on __reset_page_owner(), page_owner->free_handle is set.

We are interested in page_owner->handle, so when __set_page_owner() gets called, we derive the stack_record struct from page_owner->handle, and we increment its refcount_t field; and when __reset_page_owner() gets called, we derive its stack_record from page_owner->handle() and we decrement its refcount_t field.

---
 include/linux/stackdepot.h |  2 ++
 lib/stackdepot.c           | 53 +++++++++++++++++++++++++++++++-------
 mm/page_owner.c            |  6 +++++
 3 files changed, 51 insertions(+), 10 deletions(-)
include/linux/stackdepot.h +2 -0
index e58306783d8e..6ba4fcdb0c5f 100644
@@ -94,6 +94,8 @@ static inline int stack_depot_early_init(void)	{ return 0; }
 depot_stack_handle_t __stack_depot_save(unsigned long *entries,
 					unsigned int nr_entries,
 					gfp_t gfp_flags, bool can_alloc);
+void stack_depot_inc_count(depot_stack_handle_t handle);
+void stack_depot_dec_count(depot_stack_handle_t handle);
 
 /**
  * stack_depot_save - Save a stack trace to stack depot
lib/stackdepot.c +43 -10

90 lines not on this page · raw message

mm/page_owner.c +6 -0
index 31169b3e7f06..2d5d07013e4e 100644
@@ -139,6 +139,7 @@ void __reset_page_owner(struct page *page, unsigned short order)
 	int i;
 	struct page_ext *page_ext;
 	depot_stack_handle_t handle;
+	depot_stack_handle_t alloc_handle;
 	struct page_owner *page_owner;
 	u64 free_ts_nsec = local_clock();
 
@@ -146,6 +147,9 @@ void __reset_page_owner(struct page *page, unsigned short order)
 	if (unlikely(!page_ext))
 		return;
 
+	page_owner = get_page_owner(page_ext);
+	alloc_handle = page_owner->handle;
+
 	handle = save_stack(GFP_NOWAIT | __GFP_NOWARN);
 	for (i = 0; i < (1 << order); i++) {
 		__clear_bit(PAGE_EXT_OWNER_ALLOCATED, &page_ext->flags);
@@ -155,6 +159,7 @@ void __reset_page_owner(struct page *page, unsigned short order)
 		page_ext = page_ext_next(page_ext);
 	}
 	page_ext_put(page_ext);
+	stack_depot_dec_count(alloc_handle);
 }
 
 static inline void __set_page_owner_handle(struct page_ext *page_ext,
@@ -196,6 +201,7 @@ noinline void __set_page_owner(struct page *page, unsigned short order,
 		return;
 	__set_page_owner_handle(page_ext, handle, order, gfp_mask);
 	page_ext_put(page_ext);
+	stack_depot_inc_count(handle);
 }
 
 void __set_page_owner_migrate_reason(struct page *page, int reason)
Signature · 1 line
2.35.3
#2 lore raw

PATCH v5 2/3 mm, page_owner: Add page_owner_stacks file to print out only stacks and their counte

In reply to Oscar Salvador

We might be only interested in knowing about stacks <-> count relationship, so instead of having to fiddle with page_owner output and screen through pfns, let us add a new file called 'page_owner_stacks' that does just that.
By cating such file, we will get all the stacktraces followed by its counter, so we can have a more global view.

---
 include/linux/stackdepot.h |  6 ++++
 lib/stackdepot.c           | 72 ++++++++++++++++++++++++++++++++++++++
 mm/page_owner.c            | 27 ++++++++++++++
 3 files changed, 105 insertions(+)
include/linux/stackdepot.h +6 -0
index 6ba4fcdb0c5f..7e9d0e9ec66b 100644
@@ -112,6 +112,12 @@ void stack_depot_dec_count(depot_stack_handle_t handle);
 depot_stack_handle_t stack_depot_save(unsigned long *entries,
 				      unsigned int nr_entries, gfp_t gfp_flags);
 
+#ifdef CONFIG_PAGE_OWNER
+void *stack_start(struct seq_file *m, loff_t *ppos);
+void *stack_next(struct seq_file *m, void *v, loff_t *ppos);
+int stack_print(struct seq_file *m, void *v);
+#endif
+
 /**
  * stack_depot_fetch - Fetch a stack trace from stack depot
  *
lib/stackdepot.c +72 -0

84 lines not on this page · raw message

mm/page_owner.c +27 -0
index 2d5d07013e4e..2d97f6b34ea6 100644
@@ -719,6 +719,30 @@ static const struct file_operations proc_page_owner_operations = {
 	.llseek		= lseek_page_owner,
 };
 
+static void stack_stop(struct seq_file *m, void *v)
+{
+}
+
+static const struct seq_operations page_owner_stack_op = {
+	.start  = stack_start,
+	.next   = stack_next,
+	.stop   = stack_stop,
+	.show   = stack_print
+};
+
+static int page_owner_stack_open(struct inode *inode, struct file *file)
+{
+	return seq_open_private(file, &page_owner_stack_op,
+				sizeof(unsigned long));
+}
+
+const struct file_operations page_owner_stack_operations = {
+	.open           = page_owner_stack_open,
+	.read           = seq_read,
+	.llseek         = seq_lseek,
+	.release        = seq_release,
+};
+
 static int __init pageowner_init(void)
 {
 	if (!static_branch_unlikely(&page_owner_inited)) {
@@ -729,6 +753,9 @@ static int __init pageowner_init(void)
 	debugfs_create_file("page_owner", 0400, NULL, NULL,
 			    &proc_page_owner_operations);
 
+	debugfs_create_file("page_owner_stacks", 0400, NULL, NULL,
+			     &page_owner_stack_operations);
+
 	return 0;
 }
 late_initcall(pageowner_init)
Signature · 1 line
2.35.3
#3 lore raw

PATCH v5 3/3 mm,page_owner: Filter out stacks by a threshold counter

In reply to Oscar Salvador

We want to be able to filter out the output on a threshold basis, in this way we can get rid of a lot of noise and focus only on those stacks which have an allegedly high counter.

We can control the threshold value by a new file called 'page_owner_threshold', which is 0 by default.

---
 lib/stackdepot.c |  5 ++++-
 mm/page_owner.c  | 21 +++++++++++++++++++++
 2 files changed, 25 insertions(+), 1 deletion(-)
lib/stackdepot.c +4 -1
index c4af2e946500..7053221ce1d8 100644
@@ -488,6 +488,9 @@ static struct stack_record *stack_depot_getstack(depot_stack_handle_t handle)
 }
 
 #ifdef CONFIG_PAGE_OWNER
+
+extern unsigned long page_owner_stack_threshold;
+
 void *stack_start(struct seq_file *m, loff_t *ppos)
 {
 	unsigned long *table = m->private;
@@ -543,7 +546,7 @@ int stack_print(struct seq_file *m, void *v)
 
 	if (!stack->size || stack->size < 0 ||
 	    stack->size > PAGE_SIZE || stack->handle.valid != 1 ||
-	    refcount_read(&stack->count) < 1)
+	    refcount_read(&stack->count) < page_owner_stack_threshold)
 		return 0;
 
 	buf = kzalloc(PAGE_SIZE, GFP_KERNEL);
mm/page_owner.c +21 -0
index 2d97f6b34ea6..28c519fc9372 100644
@@ -743,6 +743,23 @@ const struct file_operations page_owner_stack_operations = {
 	.release        = seq_release,
 };
 
+unsigned long page_owner_stack_threshold;
+
+int page_owner_threshold_get(void *data, u64 *val)
+{
+	*val = page_owner_stack_threshold;
+	return 0;
+}
+
+int page_owner_threshold_set(void *data, u64 val)
+{
+	page_owner_stack_threshold = val;
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(proc_page_owner_threshold, &page_owner_threshold_get,
+			&page_owner_threshold_set, "%llu");
+
 static int __init pageowner_init(void)
 {
 	if (!static_branch_unlikely(&page_owner_inited)) {
@@ -755,6 +772,10 @@ static int __init pageowner_init(void)
 
 	debugfs_create_file("page_owner_stacks", 0400, NULL, NULL,
 			     &page_owner_stack_operations);
+	debugfs_create_file("page_owner_threshold", 0600, NULL, NULL,
+			    &proc_page_owner_threshold);
+
+	page_owner_stack_threshold = 0;
 
 	return 0;
 }
Signature · 1 line
2.35.3