Commit 6d75f366 authored by Johannes Weiner's avatar Johannes Weiner Committed by Linus Torvalds
Browse files

lib: radix-tree: check accounting of existing slot replacement users 

The bug in khugepaged fixed earlier in this series shows that radix tree
slot replacement is fragile; and it will become more so when not only
NULL<->!NULL transitions need to be caught but transitions from and to
exceptional entries as well.  We need checks.

Re-implement radix_tree_replace_slot() on top of the sanity-checked
__radix_tree_replace().  This requires existing callers to also pass the
radix tree root, but it'll warn us when somebody replaces slots with
contents that need proper accounting (transitions between NULL entries,
real entries, exceptional entries) and where a replacement through the
slot pointer would corrupt the radix tree node counts.

Link: http://lkml.kernel.org/r/20161117193021.GB23430@cmpxchg.org


Signed-off-by: default avatarJohannes Weiner <hannes@cmpxchg.org>
Suggested-by: default avatarJan Kara <jack@suse.cz>
Reviewed-by: default avatarJan Kara <jack@suse.cz>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Matthew Wilcox <mawilcox@linuxonhyperv.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent f7942430

arch/s390/mm/gmap.c

+1 −1
Original line number Diff line number Diff line
@@ -1015,7 +1015,7 @@ static inline void gmap_insert_rmap(struct gmap *sg, unsigned long vmaddr, 
	if (slot) {

		rmap->next = radix_tree_deref_slot_protected(slot,

							&sg->guest_table_lock);

		radix_tree_replace_slot(slot, rmap);

		radix_tree_replace_slot(&sg->host_to_rmap, slot, rmap);

	} else {

		rmap->next = NULL;

		radix_tree_insert(&sg->host_to_rmap, vmaddr >> PAGE_SHIFT,

drivers/sh/intc/virq.c

+1 −1
Original line number Diff line number Diff line
@@ -254,7 +254,7 @@ restart: 


		radix_tree_tag_clear(&d->tree, entry->enum_id,

				     INTC_TAG_VIRQ_NEEDS_ALLOC);

		radix_tree_replace_slot((void **)entries[i],

		radix_tree_replace_slot(&d->tree, (void **)entries[i],

					&intc_irq_xlate[irq]);

	}

fs/dax.c

+2 −2
Original line number Diff line number Diff line
@@ -342,7 +342,7 @@ static inline void *lock_slot(struct address_space *mapping, void **slot) 
		radix_tree_deref_slot_protected(slot, &mapping->tree_lock);



	entry |= RADIX_DAX_ENTRY_LOCK;

	radix_tree_replace_slot(slot, (void *)entry);

	radix_tree_replace_slot(&mapping->page_tree, slot, (void *)entry);

	return (void *)entry;

}
@@ -356,7 +356,7 @@ static inline void *unlock_slot(struct address_space *mapping, void **slot) 
		radix_tree_deref_slot_protected(slot, &mapping->tree_lock);



	entry &= ~(unsigned long)RADIX_DAX_ENTRY_LOCK;

	radix_tree_replace_slot(slot, (void *)entry);

	radix_tree_replace_slot(&mapping->page_tree, slot, (void *)entry);

	return (void *)entry;

}

include/linux/radix-tree.h

+2 −14
Original line number Diff line number Diff line
@@ -249,20 +249,6 @@ static inline int radix_tree_exception(void *arg) 
	return unlikely((unsigned long)arg & RADIX_TREE_ENTRY_MASK);

}



/**

 * radix_tree_replace_slot	- replace item in a slot

 * @pslot:	pointer to slot, returned by radix_tree_lookup_slot

 * @item:	new item to store in the slot.

 *

 * For use with radix_tree_lookup_slot().  Caller must hold tree write locked

 * across slot lookup and replacement.

 */

static inline void radix_tree_replace_slot(void **pslot, void *item)

{

	BUG_ON(radix_tree_is_internal_node(item));

	rcu_assign_pointer(*pslot, item);

}



int __radix_tree_create(struct radix_tree_root *root, unsigned long index,

			unsigned order, struct radix_tree_node **nodep,

			void ***slotp);
@@ -280,6 +266,8 @@ void **radix_tree_lookup_slot(struct radix_tree_root *, unsigned long); 
void __radix_tree_replace(struct radix_tree_root *root,

			  struct radix_tree_node *node,

			  void **slot, void *item);

void radix_tree_replace_slot(struct radix_tree_root *root,

			     void **slot, void *item);

bool __radix_tree_delete_node(struct radix_tree_root *root,

			      struct radix_tree_node *node);

void *radix_tree_delete_item(struct radix_tree_root *, unsigned long, void *);

lib/radix-tree.c

+49 −14
Original line number Diff line number Diff line
@@ -753,19 +753,10 @@ void *radix_tree_lookup(struct radix_tree_root *root, unsigned long index) 
}

EXPORT_SYMBOL(radix_tree_lookup);



/**

 * __radix_tree_replace		- replace item in a slot

 * @root:	radix tree root

 * @node:	pointer to tree node

 * @slot:	pointer to slot in @node

 * @item:	new item to store in the slot.

 *

 * For use with __radix_tree_lookup().  Caller must hold tree write locked

 * across slot lookup and replacement.

 */

void __radix_tree_replace(struct radix_tree_root *root,

static void replace_slot(struct radix_tree_root *root,

			 struct radix_tree_node *node,

			  void **slot, void *item)

			 void **slot, void *item,

			 bool warn_typeswitch)

{

	void *old = rcu_dereference_raw(*slot);

	int exceptional;
@@ -776,7 +767,7 @@ void __radix_tree_replace(struct radix_tree_root *root, 
	exceptional = !!radix_tree_exceptional_entry(item) -

		      !!radix_tree_exceptional_entry(old);



	WARN_ON_ONCE(exceptional && !node && slot != (void **)&root->rnode);

	WARN_ON_ONCE(warn_typeswitch && exceptional);



	if (node)

		node->exceptional += exceptional;
@@ -784,6 +775,50 @@ void __radix_tree_replace(struct radix_tree_root *root, 
	rcu_assign_pointer(*slot, item);

}



/**

 * __radix_tree_replace		- replace item in a slot

 * @root:	radix tree root

 * @node:	pointer to tree node

 * @slot:	pointer to slot in @node

 * @item:	new item to store in the slot.

 *

 * For use with __radix_tree_lookup().  Caller must hold tree write locked

 * across slot lookup and replacement.

 */

void __radix_tree_replace(struct radix_tree_root *root,

			  struct radix_tree_node *node,

			  void **slot, void *item)

{

	/*

	 * This function supports replacing exceptional entries, but

	 * that needs accounting against the node unless the slot is

	 * root->rnode.

	 */

	replace_slot(root, node, slot, item,

		     !node && slot != (void **)&root->rnode);

}



/**

 * radix_tree_replace_slot	- replace item in a slot

 * @root:	radix tree root

 * @slot:	pointer to slot

 * @item:	new item to store in the slot.

 *

 * For use with radix_tree_lookup_slot(), radix_tree_gang_lookup_slot(),

 * radix_tree_gang_lookup_tag_slot().  Caller must hold tree write locked

 * across slot lookup and replacement.

 *

 * NOTE: This cannot be used to switch between non-entries (empty slots),

 * regular entries, and exceptional entries, as that requires accounting

 * inside the radix tree node. When switching from one type of entry to

 * another, use __radix_tree_lookup() and __radix_tree_replace().

 */

void radix_tree_replace_slot(struct radix_tree_root *root,

			     void **slot, void *item)

{

	replace_slot(root, NULL, slot, item, true);

}



/**

 *	radix_tree_tag_set - set a tag on a radix tree node

 *	@root:		radix tree root

CRA Git | Maintained and supported by SUSTech CRA and CCSE