Skip to contentSkip to conversation

fork: annotate a data race in vm_area_dup()

PATCH Applied Qian Cai 3 participants 3 messages Feb 19, 2020 – Feb 20, 2020

Subsystems (4)
lore
Share & tools
lore

Qian Cai

author
#0 lore raw

PATCH v2 1/1 fork: annotate a data race in vm_area_dup()

struct vm_area_struct could be accessed concurrently as noticed by KCSAN,

write to 0xffff9cf8bba08ad8 of 8 bytes by task 14263 on cpu 35:

  vma_interval_tree_insert+0x101/0x150:
  rb_insert_augmented_cached at include/linux/rbtree_augmented.h:58
  (inlined by) vma_interval_tree_insert at mm/interval_tree.c:23
  __vma_link_file+0x6e/0xe0
  __vma_link_file at mm/mmap.c:629
  vma_link+0xa2/0x120
  mmap_region+0x753/0xb90
  do_mmap+0x45c/0x710
  vm_mmap_pgoff+0xc0/0x130
  ksys_mmap_pgoff+0x1d1/0x300
  __x64_sys_mmap+0x33/0x40
  do_syscall_64+0x91/0xc44
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

read to 0xffff9cf8bba08a80 of 200 bytes by task 14262 on cpu 122:

  vm_area_dup+0x6a/0xe0
  vm_area_dup at kernel/fork.c:362
  __split_vma+0x72/0x2a0
  __split_vma at mm/mmap.c:2661
  split_vma+0x5a/0x80
  mprotect_fixup+0x368/0x3f0
  do_mprotect_pkey+0x263/0x420
  __x64_sys_mprotect+0x51/0x70
  do_syscall_64+0x91/0xc44
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

vm_area_dup() blindly copies all fields of original VMA to the new one. This includes coping vm_area_struct::shared.rb which is normally protected by i_mmap_lock. But this is fine because the read value will be overwritten on the following __vma_link_file() under proper protection. Thus, mark it as an intentional data race and insert a few assertions for the fields that should not be modified concurrently.

---

v2: Insert a few assertions thanks to Marco.
Update the commit log thanks to Kirill.

 kernel/fork.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
kernel/fork.c +7 -1
index cb2ae49e497e..67a5d691ffa8 100644
@@ -359,7 +359,13 @@ struct vm_area_struct *vm_area_dup(struct vm_area_struct *orig)
 	struct vm_area_struct *new = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
 
 	if (new) {
-		*new = *orig;
+		ASSERT_EXCLUSIVE_WRITER(orig->vm_flags);
+		ASSERT_EXCLUSIVE_WRITER(orig->vm_file);
+		/*
+		 * orig->shared.rb may be modified concurrently, but the clone
+		 * will be reinitialized.
+		 */
+		*new = data_race(*orig);
 		INIT_LIST_HEAD(&new->anon_vma_chain);
 		new->vm_next = new->vm_prev = NULL;
 	}
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 30 lines · Qian Cai

struct vm_area_struct could be accessed concurrently as noticed by KCSAN,

write to 0xffff9cf8bba08ad8 of 8 bytes by task 14263 on cpu 35:

  vma_interval_tree_insert+0x101/0x150:
  rb_insert_augmented_cached at include/linux/rbtree_augmented.h:58
  (inlined by) vma_interval_tree_insert at mm/interval_tree.c:23
  __vma_link_file+0x6e/0xe0
  __vma_link_file at mm/mmap.c:629
  vma_link+0xa2/0x120
  mmap_region+0x753/0xb90
  do_mmap+0x45c/0x710
  vm_mmap_pgoff+0xc0/0x130
  ksys_mmap_pgoff+0x1d1/0x300
  __x64_sys_mmap+0x33/0x40
  do_syscall_64+0x91/0xc44
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

read to 0xffff9cf8bba08a80 of 200 bytes by task 14262 on cpu 122:

  vm_area_dup+0x6a/0xe0
  vm_area_dup at kernel/fork.c:362
  __split_vma+0x72/0x2a0
  __split_vma at mm/mmap.c:2661
  split_vma+0x5a/0x80
  mprotect_fixup+0x368/0x3f0
  do_mprotect_pkey+0x263/0x420
  __x64_sys_mprotect+0x51/0x70
  do_syscall_64+0x91/0xc44
  entry_SYSCALL_64_after_hwframe+0x49/0xbe
Quoted 8 lines

vm_area_dup() blindly copies all fields of original VMA to the new one. This includes coping vm_area_struct::shared.rb which is normally protected by i_mmap_lock. But this is fine because the read value will be overwritten on the following __vma_link_file() under proper protection. Thus, mark it as an intentional data race and insert a few assertions for the fields that should not be modified concurrently.

Queued for safekeeping on -rcu. I had to adjust a bit to get it to apply on -rcu, please see below. In my experience, git should have no trouble figuring it out. ;-)

Thanx, Paul

------------------------------------------------------------------------

commit 1228aca56f2a25b67876d8a819437b620a6e1cee

Author: Qian Cai <[email protected]>
Date:   Wed Feb 19 11:00:54 2020 -0800
    fork: Annotate a data race in vm_area_dup()
    
    struct vm_area_struct could be accessed concurrently as noticed by
    KCSAN,
    
     write to 0xffff9cf8bba08ad8 of 8 bytes by task 14263 on cpu 35:
      vma_interval_tree_insert+0x101/0x150:
      rb_insert_augmented_cached at include/linux/rbtree_augmented.h:58
      (inlined by) vma_interval_tree_insert at mm/interval_tree.c:23
      __vma_link_file+0x6e/0xe0
      __vma_link_file at mm/mmap.c:629
      vma_link+0xa2/0x120
      mmap_region+0x753/0xb90
      do_mmap+0x45c/0x710
      vm_mmap_pgoff+0xc0/0x130
      ksys_mmap_pgoff+0x1d1/0x300
      __x64_sys_mmap+0x33/0x40
      do_syscall_64+0x91/0xc44
      entry_SYSCALL_64_after_hwframe+0x49/0xbe
    
     read to 0xffff9cf8bba08a80 of 200 bytes by task 14262 on cpu 122:
      vm_area_dup+0x6a/0xe0
      vm_area_dup at kernel/fork.c:362
      __split_vma+0x72/0x2a0
      __split_vma at mm/mmap.c:2661
      split_vma+0x5a/0x80
      mprotect_fixup+0x368/0x3f0
      do_mprotect_pkey+0x263/0x420
      __x64_sys_mprotect+0x51/0x70
      do_syscall_64+0x91/0xc44
      entry_SYSCALL_64_after_hwframe+0x49/0xbe
    
    vm_area_dup() blindly copies all fields of original VMA to the new one.
    This includes coping vm_area_struct::shared.rb which is normally
    protected by i_mmap_lock. But this is fine because the read value will
    be overwritten on the following __vma_link_file() under proper
    protection. Thus, mark it as an intentional data race and insert a few
    assertions for the fields that should not be modified concurrently.
    
    Signed-off-by: Qian Cai <[email protected]>
    Signed-off-by: Paul E. McKenney <[email protected]>
kernel/fork.c +7 -1
index 60a1295..e592e6f 100644
@@ -359,7 +359,13 @@ struct vm_area_struct *vm_area_dup(struct vm_area_struct *orig)
 	struct vm_area_struct *new = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
 
 	if (new) {
-		*new = *orig;
+		ASSERT_EXCLUSIVE_WRITER(orig->vm_flags);
+		ASSERT_EXCLUSIVE_WRITER(orig->vm_file);
+		/*
+		 * orig->shared.rb may be modified concurrently, but the clone
+		 * will be reinitialized.
+		 */
+		*new = data_race(*orig);
 		INIT_LIST_HEAD(&new->anon_vma_chain);
 	}
 	return new;
Acked-by on v2 1/1
#2 lore raw
In reply to Qian Cai
Quoted 30 lines · Qian Cai

struct vm_area_struct could be accessed concurrently as noticed by KCSAN,

write to 0xffff9cf8bba08ad8 of 8 bytes by task 14263 on cpu 35:

  vma_interval_tree_insert+0x101/0x150:
  rb_insert_augmented_cached at include/linux/rbtree_augmented.h:58
  (inlined by) vma_interval_tree_insert at mm/interval_tree.c:23
  __vma_link_file+0x6e/0xe0
  __vma_link_file at mm/mmap.c:629
  vma_link+0xa2/0x120
  mmap_region+0x753/0xb90
  do_mmap+0x45c/0x710
  vm_mmap_pgoff+0xc0/0x130
  ksys_mmap_pgoff+0x1d1/0x300
  __x64_sys_mmap+0x33/0x40
  do_syscall_64+0x91/0xc44
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

read to 0xffff9cf8bba08a80 of 200 bytes by task 14262 on cpu 122:

  vm_area_dup+0x6a/0xe0
  vm_area_dup at kernel/fork.c:362
  __split_vma+0x72/0x2a0
  __split_vma at mm/mmap.c:2661
  split_vma+0x5a/0x80
  mprotect_fixup+0x368/0x3f0
  do_mprotect_pkey+0x263/0x420
  __x64_sys_mprotect+0x51/0x70
  do_syscall_64+0x91/0xc44
  entry_SYSCALL_64_after_hwframe+0x49/0xbe
Quoted 8 lines

vm_area_dup() blindly copies all fields of original VMA to the new one. This includes coping vm_area_struct::shared.rb which is normally protected by i_mmap_lock. But this is fine because the read value will be overwritten on the following __vma_link_file() under proper protection. Thus, mark it as an intentional data race and insert a few assertions for the fields that should not be modified concurrently.

Signature · 1 line
 Kirill A. Shutemov