Skip to contentSkip to conversation

lib/stackdepot: Double DEPOT_POOLS_CAP if KASAN is enabled

PATCH In review Waiman Long 3 participants 7 messages Aug 8, 2024 Library code

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 lib/stackdepot: Double DEPOT_POOLS_CAP if KASAN is enabled

As said in commit 02754e0a484a ("lib/stackdepot.c: bump stackdepot capacity from 16MB to 128MB"), KASAN uses stackdepot to memorize stacks for all kmalloc/kfree calls. So stackdepot capacity was increased 8 times to accommodate KASAN usage even thought it was claimed 4X should be enough at that time.

With commit fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records"), all stackdepot records uses a fixed size with CONFIG_STACKDEPOT_MAX_FRAMES (default=64) entries. This is merged to support evictable KASAN stack records. Commit 31639fd6cebd ("stackdepot:
use variable size records for non-evictable entries") re-enabled the use of variable size records for non-KASAN use cases, but KASAN (generic mode) still uses the large fixed size stack records.

With the default CONFIG_STACKDEPOT_MAX_FRAMES of 64, KASAN use of stackdepot space had been more than double than before. Assuming an average stack frame size of 16, a KASAN stack record is almost 4X the size of a non-KASAN one.

When a wide variety of workloads are run on a debug kernel with KASAN enabled, the following warning may sometimes be printed.

 [ 6818.650674] Stack depot reached limit capacity
 [ 6818.650730] WARNING: CPU: 1 PID: 272741 at lib/stackdepot.c:252 depot_alloc_stack+0x39e/0x3d0
   :
 [ 6818.650907] Call Trace:
 [ 6818.650909]  [<00047dd453d84b92>] depot_alloc_stack+0x3a2/0x3d0
 [ 6818.650916]  [<00047dd453d85254>] stack_depot_save_flags+0x4f4/0x5c0
 [ 6818.650920]  [<00047dd4535872c6>] kasan_save_stack+0x56/0x70
 [ 6818.650924]  [<00047dd453587328>] kasan_save_track+0x28/0x40
 [ 6818.650927]  [<00047dd45358a27a>] kasan_save_free_info+0x4a/0x70
 [ 6818.650930]  [<00047dd45358766a>] __kasan_slab_free+0x12a/0x1d0
 [ 6818.650933]  [<00047dd45350deb4>] kmem_cache_free+0x1b4/0x580
 [ 6818.650938]  [<00047dd452c520da>] __put_task_struct+0x24a/0x320
 [ 6818.650945]  [<00047dd452c6aee4>] delayed_put_task_struct+0x294/0x350
 [ 6818.650949]  [<00047dd452e9066a>] rcu_do_batch+0x6ea/0x2090
 [ 6818.650953]  [<00047dd452ea60f4>] rcu_core+0x474/0xa90
 [ 6818.650956]  [<00047dd452c780c0>] handle_softirqs+0x3c0/0xf90
 [ 6818.650960]  [<00047dd452c76fbe>] __irq_exit_rcu+0x35e/0x460
 [ 6818.650963]  [<00047dd452c79992>] irq_exit_rcu+0x22/0xb0
 [ 6818.650966]  [<00047dd454bd8128>] do_ext_irq+0xd8/0x120
 [ 6818.650972]  [<00047dd454c0ddd0>] ext_int_handler+0xb8/0xe8
 [ 6818.650979]  [<00047dd453589cf6>] kasan_check_range+0x236/0x2f0
 [ 6818.650982]  [<00047dd453378cf0>] filemap_get_pages+0x190/0xaa0
 [ 6818.650986]  [<00047dd453379940>] filemap_read+0x340/0xa70
 [ 6818.650989]  [<00047dd3d325d226>] xfs_file_buffered_read+0x2c6/0x400 [xfs]
 [ 6818.651431]  [<00047dd3d325dfe2>] xfs_file_read_iter+0x2c2/0x550 [xfs]
 [ 6818.651663]  [<00047dd45364710c>] vfs_read+0x64c/0x8c0
 [ 6818.651669]  [<00047dd453648ed8>] ksys_read+0x118/0x200
 [ 6818.651672]  [<00047dd452b6cf5a>] do_syscall+0x27a/0x380
 [ 6818.651676]  [<00047dd454bd7e74>] __do_syscall+0xf4/0x1a0
 [ 6818.651680]  [<00047dd454c0db58>] system_call+0x70/0x98

With all the recent changes in stackdepot to support new KASAN features, it is obvious that the current DEPOT_POOLS_CAP of 8192 may not be enough when KASAN is enabled. Fix this stackdepot capability issue by doubling DEPOT_POOLS_CAP if KASAN is enabled. With 4k pages, the maximum stackdepot capacity is doubled to 256 MB with KASAN enabled.

Also use the MIN() macro for defining DEPOT_MAX_POOLS to clarify the intention.

Fixes fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records")
---
 lib/stackdepot.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

[v3] Include <linux/minmax.h"> to avoid potential undefined macro problem.

lib/stackdepot.c +5 -3
index 5ed34cc963fc..1f1673b357b5 100644
@@ -20,6 +20,7 @@
 #include <linux/kernel.h>
 #include <linux/kmsan.h>
 #include <linux/list.h>
+#include <linux/minmax.h>
 #include <linux/mm.h>
 #include <linux/mutex.h>
 #include <linux/poison.h>
@@ -36,11 +37,12 @@
 #include <linux/memblock.h>
 #include <linux/kasan-enabled.h>
 
-#define DEPOT_POOLS_CAP 8192
+/* KASAN is a big user of stackdepot, double the cap if KASAN is enabled */
+#define DEPOT_POOLS_CAP (8192 * (IS_ENABLED(CONFIG_KASAN) ? 2 : 1))
+
 /* The pool_index is offset by 1 so the first record does not have a 0 handle. */
 #define DEPOT_MAX_POOLS \
-	(((1LL << (DEPOT_POOL_INDEX_BITS)) - 1 < DEPOT_POOLS_CAP) ? \
-	 (1LL << (DEPOT_POOL_INDEX_BITS)) - 1 : DEPOT_POOLS_CAP)
+	MIN((1LL << (DEPOT_POOL_INDEX_BITS)) - 1, DEPOT_POOLS_CAP)
 
 static bool stack_depot_disabled;
 static bool __stack_depot_early_init_requested __initdata = IS_ENABLED(CONFIG_STACKDEPOT_ALWAYS_INIT);
Signature · 1 line
2.43.5
#1 lore raw
Quoted 7 lines · Waiman Long

As said in commit 02754e0a484a ("lib/stackdepot.c: bump stackdepot capacity from 16MB to 128MB"), KASAN uses stackdepot to memorize stacks for all kmalloc/kfree calls. So stackdepot capacity was increased 8 times to accommodate KASAN usage even thought it was claimed 4X should be enough at that time.

Quoted 7 lines

With commit fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records"), all stackdepot records uses a fixed size with CONFIG_STACKDEPOT_MAX_FRAMES (default=64) entries. This is merged to support evictable KASAN stack records. Commit 31639fd6cebd ("stackdepot:
use variable size records for non-evictable entries") re-enabled the use of variable size records for non-KASAN use cases, but KASAN (generic mode) still uses the large fixed size stack records.

No, since commit 711d349174fd ("kasan: revert eviction of stack traces in generic mode") Generic KASAN doesn't use fixed-sized slots.

Quoted 4 lines

With the default CONFIG_STACKDEPOT_MAX_FRAMES of 64, KASAN use of stackdepot space had been more than double than before. Assuming an average stack frame size of 16, a KASAN stack record is almost 4X the size of a non-KASAN one.

So this is not correct.

Quoted 34 lines

When a wide variety of workloads are run on a debug kernel with KASAN enabled, the following warning may sometimes be printed.

 [ 6818.650674] Stack depot reached limit capacity
 [ 6818.650730] WARNING: CPU: 1 PID: 272741 at lib/stackdepot.c:252 depot_alloc_stack+0x39e/0x3d0
   :
 [ 6818.650907] Call Trace:
 [ 6818.650909]  [<00047dd453d84b92>] depot_alloc_stack+0x3a2/0x3d0
 [ 6818.650916]  [<00047dd453d85254>] stack_depot_save_flags+0x4f4/0x5c0
 [ 6818.650920]  [<00047dd4535872c6>] kasan_save_stack+0x56/0x70
 [ 6818.650924]  [<00047dd453587328>] kasan_save_track+0x28/0x40
 [ 6818.650927]  [<00047dd45358a27a>] kasan_save_free_info+0x4a/0x70
 [ 6818.650930]  [<00047dd45358766a>] __kasan_slab_free+0x12a/0x1d0
 [ 6818.650933]  [<00047dd45350deb4>] kmem_cache_free+0x1b4/0x580
 [ 6818.650938]  [<00047dd452c520da>] __put_task_struct+0x24a/0x320
 [ 6818.650945]  [<00047dd452c6aee4>] delayed_put_task_struct+0x294/0x350
 [ 6818.650949]  [<00047dd452e9066a>] rcu_do_batch+0x6ea/0x2090
 [ 6818.650953]  [<00047dd452ea60f4>] rcu_core+0x474/0xa90
 [ 6818.650956]  [<00047dd452c780c0>] handle_softirqs+0x3c0/0xf90
 [ 6818.650960]  [<00047dd452c76fbe>] __irq_exit_rcu+0x35e/0x460
 [ 6818.650963]  [<00047dd452c79992>] irq_exit_rcu+0x22/0xb0
 [ 6818.650966]  [<00047dd454bd8128>] do_ext_irq+0xd8/0x120
 [ 6818.650972]  [<00047dd454c0ddd0>] ext_int_handler+0xb8/0xe8
 [ 6818.650979]  [<00047dd453589cf6>] kasan_check_range+0x236/0x2f0
 [ 6818.650982]  [<00047dd453378cf0>] filemap_get_pages+0x190/0xaa0
 [ 6818.650986]  [<00047dd453379940>] filemap_read+0x340/0xa70
 [ 6818.650989]  [<00047dd3d325d226>] xfs_file_buffered_read+0x2c6/0x400 [xfs]
 [ 6818.651431]  [<00047dd3d325dfe2>] xfs_file_read_iter+0x2c2/0x550 [xfs]
 [ 6818.651663]  [<00047dd45364710c>] vfs_read+0x64c/0x8c0
 [ 6818.651669]  [<00047dd453648ed8>] ksys_read+0x118/0x200
 [ 6818.651672]  [<00047dd452b6cf5a>] do_syscall+0x27a/0x380
 [ 6818.651676]  [<00047dd454bd7e74>] __do_syscall+0xf4/0x1a0
 [ 6818.651680]  [<00047dd454c0db58>] system_call+0x70/0x98
Quoted 5 lines

With all the recent changes in stackdepot to support new KASAN features, it is obvious that the current DEPOT_POOLS_CAP of 8192 may not be enough when KASAN is enabled. Fix this stackdepot capability issue by doubling DEPOT_POOLS_CAP if KASAN is enabled. With 4k pages, the maximum stackdepot capacity is doubled to 256 MB with KASAN enabled.

It is possible that the stack depot runs out of space due to a truly large number of unique stack traces, but I would first make sure that is indeed the case. The one thing to check would be to dump all the stack traces from the stack depot when it overflows, and check whether they make sense. There have been cases in the past, when e.g. the task context part of a stack trace from an interrupt didn't get stripped properly, and thus almost each stack trace from an interrupt was considered unique by the stack depot. Perhaps, something similar started happening again.

Quoted 43 lines

Also use the MIN() macro for defining DEPOT_MAX_POOLS to clarify the intention.

Fixes fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records")
---
 lib/stackdepot.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

[v3] Include <linux/minmax.h"> to avoid potential undefined macro problem.

lib/stackdepot.c +5 -4
index 5ed34cc963fc..1f1673b357b5 100644
@@ -20,6 +20,7 @@
 #include <linux/kernel.h>
 #include <linux/kmsan.h>
 #include <linux/list.h>
+#include <linux/minmax.h>
 #include <linux/mm.h>
 #include <linux/mutex.h>
 #include <linux/poison.h>
@@ -36,11 +37,12 @@
 #include <linux/memblock.h>
 #include <linux/kasan-enabled.h>
 
-#define DEPOT_POOLS_CAP 8192
+/* KASAN is a big user of stackdepot, double the cap if KASAN is enabled */
+#define DEPOT_POOLS_CAP (8192 * (IS_ENABLED(CONFIG_KASAN) ? 2 : 1))
+
 /* The pool_index is offset by 1 so the first record does not have a 0 handle. */
 #define DEPOT_MAX_POOLS \
-       (((1LL << (DEPOT_POOL_INDEX_BITS)) - 1 < DEPOT_POOLS_CAP) ? \
-        (1LL << (DEPOT_POOL_INDEX_BITS)) - 1 : DEPOT_POOLS_CAP)
+       MIN((1LL << (DEPOT_POOL_INDEX_BITS)) - 1, DEPOT_POOLS_CAP)
 
 static bool stack_depot_disabled;
 static bool __stack_depot_early_init_requested __initdata = IS_ENABLED(CONFIG_STACKDEPOT_ALWAYS_INIT);
--

2.43.5

#2 lore raw
Quoted 14 lines · Andrey Konovalov
Quoted 6 lines · Waiman Long

As said in commit 02754e0a484a ("lib/stackdepot.c: bump stackdepot capacity from 16MB to 128MB"), KASAN uses stackdepot to memorize stacks for all kmalloc/kfree calls. So stackdepot capacity was increased 8 times to accommodate KASAN usage even thought it was claimed 4X should be enough at that time.

Quoted 7 lines

With commit fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records"), all stackdepot records uses a fixed size with CONFIG_STACKDEPOT_MAX_FRAMES (default=64) entries. This is merged to support evictable KASAN stack records. Commit 31639fd6cebd ("stackdepot:
use variable size records for non-evictable entries") re-enabled the use of variable size records for non-KASAN use cases, but KASAN (generic mode) still uses the large fixed size stack records.

Quoted 8 lines

No, since commit 711d349174fd ("kasan: revert eviction of stack traces in generic mode") Generic KASAN doesn't use fixed-sized slots.

Quoted 4 lines

With the default CONFIG_STACKDEPOT_MAX_FRAMES of 64, KASAN use of stackdepot space had been more than double than before. Assuming an average stack frame size of 16, a KASAN stack record is almost 4X the size of a non-KASAN one.

So this is not correct.

Yes, you are right. I missed the KASAN commit 711d349174fd in my analysis. I will update the commit log in the next version.

Quoted 39 lines
Quoted 34 lines

When a wide variety of workloads are run on a debug kernel with KASAN enabled, the following warning may sometimes be printed.

  [ 6818.650674] Stack depot reached limit capacity
  [ 6818.650730] WARNING: CPU: 1 PID: 272741 at lib/stackdepot.c:252 depot_alloc_stack+0x39e/0x3d0
    :
  [ 6818.650907] Call Trace:
  [ 6818.650909]  [<00047dd453d84b92>] depot_alloc_stack+0x3a2/0x3d0
  [ 6818.650916]  [<00047dd453d85254>] stack_depot_save_flags+0x4f4/0x5c0
  [ 6818.650920]  [<00047dd4535872c6>] kasan_save_stack+0x56/0x70
  [ 6818.650924]  [<00047dd453587328>] kasan_save_track+0x28/0x40
  [ 6818.650927]  [<00047dd45358a27a>] kasan_save_free_info+0x4a/0x70
  [ 6818.650930]  [<00047dd45358766a>] __kasan_slab_free+0x12a/0x1d0
  [ 6818.650933]  [<00047dd45350deb4>] kmem_cache_free+0x1b4/0x580
  [ 6818.650938]  [<00047dd452c520da>] __put_task_struct+0x24a/0x320
  [ 6818.650945]  [<00047dd452c6aee4>] delayed_put_task_struct+0x294/0x350
  [ 6818.650949]  [<00047dd452e9066a>] rcu_do_batch+0x6ea/0x2090
  [ 6818.650953]  [<00047dd452ea60f4>] rcu_core+0x474/0xa90
  [ 6818.650956]  [<00047dd452c780c0>] handle_softirqs+0x3c0/0xf90
  [ 6818.650960]  [<00047dd452c76fbe>] __irq_exit_rcu+0x35e/0x460
  [ 6818.650963]  [<00047dd452c79992>] irq_exit_rcu+0x22/0xb0
  [ 6818.650966]  [<00047dd454bd8128>] do_ext_irq+0xd8/0x120
  [ 6818.650972]  [<00047dd454c0ddd0>] ext_int_handler+0xb8/0xe8
  [ 6818.650979]  [<00047dd453589cf6>] kasan_check_range+0x236/0x2f0
  [ 6818.650982]  [<00047dd453378cf0>] filemap_get_pages+0x190/0xaa0
  [ 6818.650986]  [<00047dd453379940>] filemap_read+0x340/0xa70
  [ 6818.650989]  [<00047dd3d325d226>] xfs_file_buffered_read+0x2c6/0x400 [xfs]
  [ 6818.651431]  [<00047dd3d325dfe2>] xfs_file_read_iter+0x2c2/0x550 [xfs]
  [ 6818.651663]  [<00047dd45364710c>] vfs_read+0x64c/0x8c0
  [ 6818.651669]  [<00047dd453648ed8>] ksys_read+0x118/0x200
  [ 6818.651672]  [<00047dd452b6cf5a>] do_syscall+0x27a/0x380
  [ 6818.651676]  [<00047dd454bd7e74>] __do_syscall+0xf4/0x1a0
  [ 6818.651680]  [<00047dd454c0db58>] system_call+0x70/0x98
Quoted 5 lines

With all the recent changes in stackdepot to support new KASAN features, it is obvious that the current DEPOT_POOLS_CAP of 8192 may not be enough when KASAN is enabled. Fix this stackdepot capability issue by doubling DEPOT_POOLS_CAP if KASAN is enabled. With 4k pages, the maximum stackdepot capacity is doubled to 256 MB with KASAN enabled.

Quoted 9 lines

It is possible that the stack depot runs out of space due to a truly large number of unique stack traces, but I would first make sure that is indeed the case. The one thing to check would be to dump all the stack traces from the stack depot when it overflows, and check whether they make sense. There have been cases in the past, when e.g. the task context part of a stack trace from an interrupt didn't get stripped properly, and thus almost each stack trace from an interrupt was considered unique by the stack depot. Perhaps, something similar started happening again.

It could be.

Anyway, I got a bug report from our QE team and they have seen it once so far. So it is not an easily reproducible problem. I need to do further investigation first. Thanks for your quick response.

Cheers,
Longman

#3 lore raw
Quoted 76 lines · Waiman Long
Quoted 14 lines · Andrey Konovalov
Quoted 6 lines · Waiman Long

As said in commit 02754e0a484a ("lib/stackdepot.c: bump stackdepot capacity from 16MB to 128MB"), KASAN uses stackdepot to memorize stacks for all kmalloc/kfree calls. So stackdepot capacity was increased 8 times to accommodate KASAN usage even thought it was claimed 4X should be enough at that time.

Quoted 7 lines

With commit fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records"), all stackdepot records uses a fixed size with CONFIG_STACKDEPOT_MAX_FRAMES (default=64) entries. This is merged to support evictable KASAN stack records. Commit 31639fd6cebd ("stackdepot:
use variable size records for non-evictable entries") re-enabled the use of variable size records for non-KASAN use cases, but KASAN (generic mode) still uses the large fixed size stack records.

Quoted 8 lines

No, since commit 711d349174fd ("kasan: revert eviction of stack traces in generic mode") Generic KASAN doesn't use fixed-sized slots.

Quoted 4 lines

With the default CONFIG_STACKDEPOT_MAX_FRAMES of 64, KASAN use of stackdepot space had been more than double than before. Assuming an average stack frame size of 16, a KASAN stack record is almost 4X the size of a non-KASAN one.

So this is not correct.

Yes, you are right. I missed the KASAN commit 711d349174fd in my analysis. I will update the commit log in the next version.

Quoted 39 lines
Quoted 34 lines

When a wide variety of workloads are run on a debug kernel with KASAN enabled, the following warning may sometimes be printed.

  [ 6818.650674] Stack depot reached limit capacity
  [ 6818.650730] WARNING: CPU: 1 PID: 272741 at lib/stackdepot.c:252 depot_alloc_stack+0x39e/0x3d0
    :
  [ 6818.650907] Call Trace:
  [ 6818.650909]  [<00047dd453d84b92>] depot_alloc_stack+0x3a2/0x3d0
  [ 6818.650916]  [<00047dd453d85254>] stack_depot_save_flags+0x4f4/0x5c0
  [ 6818.650920]  [<00047dd4535872c6>] kasan_save_stack+0x56/0x70
  [ 6818.650924]  [<00047dd453587328>] kasan_save_track+0x28/0x40
  [ 6818.650927]  [<00047dd45358a27a>] kasan_save_free_info+0x4a/0x70
  [ 6818.650930]  [<00047dd45358766a>] __kasan_slab_free+0x12a/0x1d0
  [ 6818.650933]  [<00047dd45350deb4>] kmem_cache_free+0x1b4/0x580
  [ 6818.650938]  [<00047dd452c520da>] __put_task_struct+0x24a/0x320
  [ 6818.650945]  [<00047dd452c6aee4>] delayed_put_task_struct+0x294/0x350
  [ 6818.650949]  [<00047dd452e9066a>] rcu_do_batch+0x6ea/0x2090
  [ 6818.650953]  [<00047dd452ea60f4>] rcu_core+0x474/0xa90
  [ 6818.650956]  [<00047dd452c780c0>] handle_softirqs+0x3c0/0xf90
  [ 6818.650960]  [<00047dd452c76fbe>] __irq_exit_rcu+0x35e/0x460
  [ 6818.650963]  [<00047dd452c79992>] irq_exit_rcu+0x22/0xb0
  [ 6818.650966]  [<00047dd454bd8128>] do_ext_irq+0xd8/0x120
  [ 6818.650972]  [<00047dd454c0ddd0>] ext_int_handler+0xb8/0xe8
  [ 6818.650979]  [<00047dd453589cf6>] kasan_check_range+0x236/0x2f0
  [ 6818.650982]  [<00047dd453378cf0>] filemap_get_pages+0x190/0xaa0
  [ 6818.650986]  [<00047dd453379940>] filemap_read+0x340/0xa70
  [ 6818.650989]  [<00047dd3d325d226>] xfs_file_buffered_read+0x2c6/0x400 [xfs]
  [ 6818.651431]  [<00047dd3d325dfe2>] xfs_file_read_iter+0x2c2/0x550 [xfs]
  [ 6818.651663]  [<00047dd45364710c>] vfs_read+0x64c/0x8c0
  [ 6818.651669]  [<00047dd453648ed8>] ksys_read+0x118/0x200
  [ 6818.651672]  [<00047dd452b6cf5a>] do_syscall+0x27a/0x380
  [ 6818.651676]  [<00047dd454bd7e74>] __do_syscall+0xf4/0x1a0
  [ 6818.651680]  [<00047dd454c0db58>] system_call+0x70/0x98
Quoted 5 lines

With all the recent changes in stackdepot to support new KASAN features, it is obvious that the current DEPOT_POOLS_CAP of 8192 may not be enough when KASAN is enabled. Fix this stackdepot capability issue by doubling DEPOT_POOLS_CAP if KASAN is enabled. With 4k pages, the maximum stackdepot capacity is doubled to 256 MB with KASAN enabled.

Quoted 9 lines

It is possible that the stack depot runs out of space due to a truly large number of unique stack traces, but I would first make sure that is indeed the case. The one thing to check would be to dump all the stack traces from the stack depot when it overflows, and check whether they make sense. There have been cases in the past, when e.g. the task context part of a stack trace from an interrupt didn't get stripped properly, and thus almost each stack trace from an interrupt was considered unique by the stack depot. Perhaps, something similar started happening again.

Quoted 5 lines

It could be.

Anyway, I got a bug report from our QE team and they have seen it once so far. So it is not an easily reproducible problem. I need to do further investigation first. Thanks for your quick response.

If you have access to the environment where the issue is happening, you can just make the stack depot keep dumping all new stack traces being saved and monitor them for some time. If there's a problem with interrupt contexts, you should be able to see it even without waiting for the stack depot to overflow.

Thanks!

#4 lore raw
Quoted 77 lines · Andrey Konovalov
Quoted 76 lines · Waiman Long
Quoted 14 lines · Andrey Konovalov
Quoted 6 lines · Waiman Long

As said in commit 02754e0a484a ("lib/stackdepot.c: bump stackdepot capacity from 16MB to 128MB"), KASAN uses stackdepot to memorize stacks for all kmalloc/kfree calls. So stackdepot capacity was increased 8 times to accommodate KASAN usage even thought it was claimed 4X should be enough at that time.

Quoted 7 lines

With commit fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records"), all stackdepot records uses a fixed size with CONFIG_STACKDEPOT_MAX_FRAMES (default=64) entries. This is merged to support evictable KASAN stack records. Commit 31639fd6cebd ("stackdepot:
use variable size records for non-evictable entries") re-enabled the use of variable size records for non-KASAN use cases, but KASAN (generic mode) still uses the large fixed size stack records.

Quoted 8 lines

No, since commit 711d349174fd ("kasan: revert eviction of stack traces in generic mode") Generic KASAN doesn't use fixed-sized slots.

Quoted 4 lines

With the default CONFIG_STACKDEPOT_MAX_FRAMES of 64, KASAN use of stackdepot space had been more than double than before. Assuming an average stack frame size of 16, a KASAN stack record is almost 4X the size of a non-KASAN one.

So this is not correct.

Yes, you are right. I missed the KASAN commit 711d349174fd in my analysis. I will update the commit log in the next version.

Quoted 39 lines
Quoted 34 lines

When a wide variety of workloads are run on a debug kernel with KASAN enabled, the following warning may sometimes be printed.

   [ 6818.650674] Stack depot reached limit capacity
   [ 6818.650730] WARNING: CPU: 1 PID: 272741 at lib/stackdepot.c:252 depot_alloc_stack+0x39e/0x3d0
     :
   [ 6818.650907] Call Trace:
   [ 6818.650909]  [<00047dd453d84b92>] depot_alloc_stack+0x3a2/0x3d0
   [ 6818.650916]  [<00047dd453d85254>] stack_depot_save_flags+0x4f4/0x5c0
   [ 6818.650920]  [<00047dd4535872c6>] kasan_save_stack+0x56/0x70
   [ 6818.650924]  [<00047dd453587328>] kasan_save_track+0x28/0x40
   [ 6818.650927]  [<00047dd45358a27a>] kasan_save_free_info+0x4a/0x70
   [ 6818.650930]  [<00047dd45358766a>] __kasan_slab_free+0x12a/0x1d0
   [ 6818.650933]  [<00047dd45350deb4>] kmem_cache_free+0x1b4/0x580
   [ 6818.650938]  [<00047dd452c520da>] __put_task_struct+0x24a/0x320
   [ 6818.650945]  [<00047dd452c6aee4>] delayed_put_task_struct+0x294/0x350
   [ 6818.650949]  [<00047dd452e9066a>] rcu_do_batch+0x6ea/0x2090
   [ 6818.650953]  [<00047dd452ea60f4>] rcu_core+0x474/0xa90
   [ 6818.650956]  [<00047dd452c780c0>] handle_softirqs+0x3c0/0xf90
   [ 6818.650960]  [<00047dd452c76fbe>] __irq_exit_rcu+0x35e/0x460
   [ 6818.650963]  [<00047dd452c79992>] irq_exit_rcu+0x22/0xb0
   [ 6818.650966]  [<00047dd454bd8128>] do_ext_irq+0xd8/0x120
   [ 6818.650972]  [<00047dd454c0ddd0>] ext_int_handler+0xb8/0xe8
   [ 6818.650979]  [<00047dd453589cf6>] kasan_check_range+0x236/0x2f0
   [ 6818.650982]  [<00047dd453378cf0>] filemap_get_pages+0x190/0xaa0
   [ 6818.650986]  [<00047dd453379940>] filemap_read+0x340/0xa70
   [ 6818.650989]  [<00047dd3d325d226>] xfs_file_buffered_read+0x2c6/0x400 [xfs]
   [ 6818.651431]  [<00047dd3d325dfe2>] xfs_file_read_iter+0x2c2/0x550 [xfs]
   [ 6818.651663]  [<00047dd45364710c>] vfs_read+0x64c/0x8c0
   [ 6818.651669]  [<00047dd453648ed8>] ksys_read+0x118/0x200
   [ 6818.651672]  [<00047dd452b6cf5a>] do_syscall+0x27a/0x380
   [ 6818.651676]  [<00047dd454bd7e74>] __do_syscall+0xf4/0x1a0
   [ 6818.651680]  [<00047dd454c0db58>] system_call+0x70/0x98
Quoted 5 lines

With all the recent changes in stackdepot to support new KASAN features, it is obvious that the current DEPOT_POOLS_CAP of 8192 may not be enough when KASAN is enabled. Fix this stackdepot capability issue by doubling DEPOT_POOLS_CAP if KASAN is enabled. With 4k pages, the maximum stackdepot capacity is doubled to 256 MB with KASAN enabled.

Quoted 9 lines

It is possible that the stack depot runs out of space due to a truly large number of unique stack traces, but I would first make sure that is indeed the case. The one thing to check would be to dump all the stack traces from the stack depot when it overflows, and check whether they make sense. There have been cases in the past, when e.g. the task context part of a stack trace from an interrupt didn't get stripped properly, and thus almost each stack trace from an interrupt was considered unique by the stack depot. Perhaps, something similar started happening again.

It could be.

Quoted 8 lines
Quoted 3 lines

Anyway, I got a bug report from our QE team and they have seen it once so far. So it is not an easily reproducible problem. I need to do further investigation first. Thanks for your quick response.

If you have access to the environment where the issue is happening, you can just make the stack depot keep dumping all new stack traces being saved and monitor them for some time. If there's a problem with interrupt contexts, you should be able to see it even without waiting for the stack depot to overflow.

I see. Thanks for the suggestion.

Cheers,
Longman

#5 lore raw
In reply to Andrey Konovalov
Quoted 39 lines · Andrey Konovalov
Quoted 34 lines

When a wide variety of workloads are run on a debug kernel with KASAN enabled, the following warning may sometimes be printed.

  [ 6818.650674] Stack depot reached limit capacity
  [ 6818.650730] WARNING: CPU: 1 PID: 272741 at lib/stackdepot.c:252 depot_alloc_stack+0x39e/0x3d0
    :
  [ 6818.650907] Call Trace:
  [ 6818.650909]  [<00047dd453d84b92>] depot_alloc_stack+0x3a2/0x3d0
  [ 6818.650916]  [<00047dd453d85254>] stack_depot_save_flags+0x4f4/0x5c0
  [ 6818.650920]  [<00047dd4535872c6>] kasan_save_stack+0x56/0x70
  [ 6818.650924]  [<00047dd453587328>] kasan_save_track+0x28/0x40
  [ 6818.650927]  [<00047dd45358a27a>] kasan_save_free_info+0x4a/0x70
  [ 6818.650930]  [<00047dd45358766a>] __kasan_slab_free+0x12a/0x1d0
  [ 6818.650933]  [<00047dd45350deb4>] kmem_cache_free+0x1b4/0x580
  [ 6818.650938]  [<00047dd452c520da>] __put_task_struct+0x24a/0x320
  [ 6818.650945]  [<00047dd452c6aee4>] delayed_put_task_struct+0x294/0x350
  [ 6818.650949]  [<00047dd452e9066a>] rcu_do_batch+0x6ea/0x2090
  [ 6818.650953]  [<00047dd452ea60f4>] rcu_core+0x474/0xa90
  [ 6818.650956]  [<00047dd452c780c0>] handle_softirqs+0x3c0/0xf90
  [ 6818.650960]  [<00047dd452c76fbe>] __irq_exit_rcu+0x35e/0x460
  [ 6818.650963]  [<00047dd452c79992>] irq_exit_rcu+0x22/0xb0
  [ 6818.650966]  [<00047dd454bd8128>] do_ext_irq+0xd8/0x120
  [ 6818.650972]  [<00047dd454c0ddd0>] ext_int_handler+0xb8/0xe8
  [ 6818.650979]  [<00047dd453589cf6>] kasan_check_range+0x236/0x2f0
  [ 6818.650982]  [<00047dd453378cf0>] filemap_get_pages+0x190/0xaa0
  [ 6818.650986]  [<00047dd453379940>] filemap_read+0x340/0xa70
  [ 6818.650989]  [<00047dd3d325d226>] xfs_file_buffered_read+0x2c6/0x400 [xfs]
  [ 6818.651431]  [<00047dd3d325dfe2>] xfs_file_read_iter+0x2c2/0x550 [xfs]
  [ 6818.651663]  [<00047dd45364710c>] vfs_read+0x64c/0x8c0
  [ 6818.651669]  [<00047dd453648ed8>] ksys_read+0x118/0x200
  [ 6818.651672]  [<00047dd452b6cf5a>] do_syscall+0x27a/0x380
  [ 6818.651676]  [<00047dd454bd7e74>] __do_syscall+0xf4/0x1a0
  [ 6818.651680]  [<00047dd454c0db58>] system_call+0x70/0x98
Quoted 5 lines

With all the recent changes in stackdepot to support new KASAN features, it is obvious that the current DEPOT_POOLS_CAP of 8192 may not be enough when KASAN is enabled. Fix this stackdepot capability issue by doubling DEPOT_POOLS_CAP if KASAN is enabled. With 4k pages, the maximum stackdepot capacity is doubled to 256 MB with KASAN enabled.

Quoted 9 lines

It is possible that the stack depot runs out of space due to a truly large number of unique stack traces, but I would first make sure that is indeed the case. The one thing to check would be to dump all the stack traces from the stack depot when it overflows, and check whether they make sense. There have been cases in the past, when e.g. the task context part of a stack trace from an interrupt didn't get stripped properly, and thus almost each stack trace from an interrupt was considered unique by the stack depot. Perhaps, something similar started happening again.

You are right. Indeed this problem is caused by mixing the task and interrupt stack traces and treating them as unique traces causing a lot more partially duplicated stack traces to be stored and thus overflowing the current stackdepot capacity.

This problem happens in a s390 machine and it looks like __irqentry_text_start and __irqentry_text_end aren't properly set up causing filter_irq_stacks() to fail in identifying the proper interrupt boundary.

It may be a defect in the s390 architectural code or something is wrong in the way the s390 kernel is being built. I am not proficient in the s390 arch. So I will ask our s390 experts to take a further look at this.

Thanks a lot for the helping me to figure this out. I am going to drop this patch.

Cheers,
Longman

#6 lore raw
In reply to Waiman Long
Quoted 60 lines · Waiman Long

As said in commit 02754e0a484a ("lib/stackdepot.c: bump stackdepot capacity from 16MB to 128MB"), KASAN uses stackdepot to memorize stacks for all kmalloc/kfree calls. So stackdepot capacity was increased 8 times to accommodate KASAN usage even thought it was claimed 4X should be enough at that time.

With commit fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records"), all stackdepot records uses a fixed size with CONFIG_STACKDEPOT_MAX_FRAMES (default=64) entries. This is merged to support evictable KASAN stack records. Commit 31639fd6cebd ("stackdepot:
use variable size records for non-evictable entries") re-enabled the use of variable size records for non-KASAN use cases, but KASAN (generic mode) still uses the large fixed size stack records.

With the default CONFIG_STACKDEPOT_MAX_FRAMES of 64, KASAN use of stackdepot space had been more than double than before. Assuming an average stack frame size of 16, a KASAN stack record is almost 4X the size of a non-KASAN one.

When a wide variety of workloads are run on a debug kernel with KASAN enabled, the following warning may sometimes be printed.

 [ 6818.650674] Stack depot reached limit capacity
 [ 6818.650730] WARNING: CPU: 1 PID: 272741 at lib/stackdepot.c:252 depot_alloc_stack+0x39e/0x3d0
   :
 [ 6818.650907] Call Trace:
 [ 6818.650909]  [<00047dd453d84b92>] depot_alloc_stack+0x3a2/0x3d0
 [ 6818.650916]  [<00047dd453d85254>] stack_depot_save_flags+0x4f4/0x5c0
 [ 6818.650920]  [<00047dd4535872c6>] kasan_save_stack+0x56/0x70
 [ 6818.650924]  [<00047dd453587328>] kasan_save_track+0x28/0x40
 [ 6818.650927]  [<00047dd45358a27a>] kasan_save_free_info+0x4a/0x70
 [ 6818.650930]  [<00047dd45358766a>] __kasan_slab_free+0x12a/0x1d0
 [ 6818.650933]  [<00047dd45350deb4>] kmem_cache_free+0x1b4/0x580
 [ 6818.650938]  [<00047dd452c520da>] __put_task_struct+0x24a/0x320
 [ 6818.650945]  [<00047dd452c6aee4>] delayed_put_task_struct+0x294/0x350
 [ 6818.650949]  [<00047dd452e9066a>] rcu_do_batch+0x6ea/0x2090
 [ 6818.650953]  [<00047dd452ea60f4>] rcu_core+0x474/0xa90
 [ 6818.650956]  [<00047dd452c780c0>] handle_softirqs+0x3c0/0xf90
 [ 6818.650960]  [<00047dd452c76fbe>] __irq_exit_rcu+0x35e/0x460
 [ 6818.650963]  [<00047dd452c79992>] irq_exit_rcu+0x22/0xb0
 [ 6818.650966]  [<00047dd454bd8128>] do_ext_irq+0xd8/0x120
 [ 6818.650972]  [<00047dd454c0ddd0>] ext_int_handler+0xb8/0xe8
 [ 6818.650979]  [<00047dd453589cf6>] kasan_check_range+0x236/0x2f0
 [ 6818.650982]  [<00047dd453378cf0>] filemap_get_pages+0x190/0xaa0
 [ 6818.650986]  [<00047dd453379940>] filemap_read+0x340/0xa70
 [ 6818.650989]  [<00047dd3d325d226>] xfs_file_buffered_read+0x2c6/0x400 [xfs]
 [ 6818.651431]  [<00047dd3d325dfe2>] xfs_file_read_iter+0x2c2/0x550 [xfs]
 [ 6818.651663]  [<00047dd45364710c>] vfs_read+0x64c/0x8c0
 [ 6818.651669]  [<00047dd453648ed8>] ksys_read+0x118/0x200
 [ 6818.651672]  [<00047dd452b6cf5a>] do_syscall+0x27a/0x380
 [ 6818.651676]  [<00047dd454bd7e74>] __do_syscall+0xf4/0x1a0
 [ 6818.651680]  [<00047dd454c0db58>] system_call+0x70/0x98

With all the recent changes in stackdepot to support new KASAN features, it is obvious that the current DEPOT_POOLS_CAP of 8192 may not be enough when KASAN is enabled. Fix this stackdepot capability issue by doubling DEPOT_POOLS_CAP if KASAN is enabled. With 4k pages, the maximum stackdepot capacity is doubled to 256 MB with KASAN enabled.

Quoted 5 lines

Also use the MIN() macro for defining DEPOT_MAX_POOLS to clarify the intention.

Fixes fc60e0caa94d ("lib/stackdepot: use fixed-sized slots for stack records")

Hi Waiman,

I am getting this issue on ARM64:

lib/stackdepot.c:44:9: error: implicit declaration of function ‘MIN’ [-Werror=implicit-function-declaration]

   44 |         MIN((1LL << (DEPOT_POOL_INDEX_BITS)) - 1, DEPOT_POOLS_CAP)
      |         ^~~
lib/stackdepot.c:66:26: note: in expansion of macro ‘DEPOT_MAX_POOLS’
   66 | static void *stack_pools[DEPOT_MAX_POOLS];
      |                          ^~~~~~~~~~~~~~~
lib/stackdepot.c:66:14: error: variably modified ‘stack_pools’ at file scope
   66 | static void *stack_pools[DEPOT_MAX_POOLS];
      |              ^~~~~~~~~~~

I've temporarily added this to get the build to pass, though I'm not sure if it's the correct fix,

lib/stackdepot.c +2 -0
index beeb70b57710..f210e3202927 100644
@@ -36,6 +36,8 @@
 #include <linux/memblock.h>
 #include <linux/kasan-enabled.h>
 
+#define MIN(a,b) ((a)<(b) ? (a):(b))
+
 /* KASAN is a big user of stackdepot, double the cap if KASAN is enabled */
 #define DEPOT_POOLS_CAP (8192 * (IS_ENABLED(CONFIG_KASAN) ? 2 : 1))
 
Quoted 35 lines
---
 lib/stackdepot.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

[v3] Include <linux/minmax.h"> to avoid potential undefined macro problem.

lib/stackdepot.c +5 -4
index 5ed34cc963fc..1f1673b357b5 100644
@@ -20,6 +20,7 @@
 #include <linux/kernel.h>
 #include <linux/kmsan.h>
 #include <linux/list.h>
+#include <linux/minmax.h>
 #include <linux/mm.h>
 #include <linux/mutex.h>
 #include <linux/poison.h>
@@ -36,11 +37,12 @@
 #include <linux/memblock.h>
 #include <linux/kasan-enabled.h>
 
-#define DEPOT_POOLS_CAP 8192
+/* KASAN is a big user of stackdepot, double the cap if KASAN is enabled */
+#define DEPOT_POOLS_CAP (8192 * (IS_ENABLED(CONFIG_KASAN) ? 2 : 1))
+
 /* The pool_index is offset by 1 so the first record does not have a 0 handle. */
 #define DEPOT_MAX_POOLS \
-       (((1LL << (DEPOT_POOL_INDEX_BITS)) - 1 < DEPOT_POOLS_CAP) ? \
-        (1LL << (DEPOT_POOL_INDEX_BITS)) - 1 : DEPOT_POOLS_CAP)
+       MIN((1LL << (DEPOT_POOL_INDEX_BITS)) - 1, DEPOT_POOLS_CAP)
 
 static bool stack_depot_disabled;
 static bool __stack_depot_early_init_requested __initdata = IS_ENABLED(CONFIG_STACKDEPOT_ALWAYS_INIT);
--
Quoted 3 lines

2.43.5

Thanks
Barry