Re: BUG: maple_tree: KCSAN: data-race in mas_topiary_replace / mtree_range_walk

From: Liam R. Howlett
Date: Fri Sep 22 2023 - 09:52:30 EST


* Mirsad Todorovac <mirsad.todorovac@xxxxxxxxxxxx> [230921 17:01]:
> Hi,
>

Hello!

> On the 6.6-rc2 vanilla torvalds tree kernel, KCSAN had reported a number of data-races:
>
> [ 6413.367326] ==================================================================
> [ 6413.367349] BUG: KCSAN: data-race in mas_topiary_replace / mtree_range_walk
>
> [ 6413.367375] write to 0xffff8883a0c5db00 of 8 bytes by task 5475 on cpu 24:
> [ 6413.367386] mas_topiary_replace (/home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:491 /home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:1701 /home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:2590)
> [ 6413.367399] mas_spanning_rebalance.isra.0 (/home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:2664 /home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:2961)
> [ 6413.367413] mas_wr_spanning_store.isra.0 (/home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:3894)
> [ 6413.367428] mas_wr_store_entry.isra.0 (/home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:4242)
> [ 6413.367442] mas_store_prealloc (/home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:256 /home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:5408)
> [ 6413.367455] vma_merge (/home/marvin/linux/kernel/torvalds2/mm/internal.h:1127 /home/marvin/linux/kernel/torvalds2/mm/mmap.c:1005)
> [ 6413.367466] mprotect_fixup (/home/marvin/linux/kernel/torvalds2/mm/mprotect.c:632)
> [ 6413.367480] do_mprotect_pkey (/home/marvin/linux/kernel/torvalds2/mm/mprotect.c:819)
> [ 6413.367494] __x64_sys_mprotect (/home/marvin/linux/kernel/torvalds2/mm/mprotect.c:837)
> [ 6413.367503] do_syscall_64 (/home/marvin/linux/kernel/torvalds2/arch/x86/entry/common.c:50 /home/marvin/linux/kernel/torvalds2/arch/x86/entry/common.c:80)
> [ 6413.367517] entry_SYSCALL_64_after_hwframe (/home/marvin/linux/kernel/torvalds2/arch/x86/entry/entry_64.S:120)
>
> [ 6413.367534] read to 0xffff8883a0c5db00 of 8 bytes by task 5558 on cpu 11:
> [ 6413.367545] mtree_range_walk (/home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:539 /home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:2789)
> [ 6413.367556] mas_walk (/home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:251 /home/marvin/linux/kernel/torvalds2/lib/maple_tree.c:4844)
> [ 6413.367567] lock_vma_under_rcu (/home/marvin/linux/kernel/torvalds2/mm/memory.c:5436)
> [ 6413.367579] do_user_addr_fault (/home/marvin/linux/kernel/torvalds2/arch/x86/mm/fault.c:1357)
> [ 6413.367593] exc_page_fault (/home/marvin/linux/kernel/torvalds2/./arch/x86/include/asm/paravirt.h:695 /home/marvin/linux/kernel/torvalds2/arch/x86/mm/fault.c:1513 /home/marvin/linux/kernel/torvalds2/arch/x86/mm/fault.c:1561)
> [ 6413.367602] asm_exc_page_fault (/home/marvin/linux/kernel/torvalds2/./arch/x86/include/asm/idtentry.h:570)
>
> [ 6413.367617] value changed: 0xffff888341d43116 -> 0xffff888340f92116
>
> [ 6413.367632] Reported by Kernel Concurrency Sanitizer on:
> [ 6413.367640] CPU: 11 PID: 5558 Comm: ThreadPoolForeg Tainted: G L 6.6.0-rc2-kcsan-00143-gb5cbe7c00aa0 #41
> [ 6413.367653] Hardware name: ASRock X670E PG Lightning/X670E PG Lightning, BIOS 1.21 04/26/2023
> [ 6413.367660] ==================================================================
>
> For your convenience, took the trouble of finding those suspicious lines of code:
>
> The write side:
>
> lib/maple_tree.c:491
> --------------------
> 456 /*
> 457 * mas_set_parent() - Set the parent node and encode the slot
> 458 * @enode: The encoded maple node.
> 459 * @parent: The encoded maple node that is the parent of @enode.
> 460 * @slot: The slot that @enode resides in @parent.
> 461 *
> 462 * Slot number is encoded in the enode->parent bit 3-6 or 2-6, depending on the
> 463 * parent type.
> 464 */
> 465 static inline
> 466 void mas_set_parent(struct ma_state *mas, struct maple_enode *enode,
> 467 const struct maple_enode *parent, unsigned char slot)
> 468 {
> 469 unsigned long val = (unsigned long)parent;
> 470 unsigned long shift;
> 471 unsigned long type;
> 472 enum maple_type p_type = mte_node_type(parent);
> 473
> 474 MAS_BUG_ON(mas, p_type == maple_dense);
> 475 MAS_BUG_ON(mas, p_type == maple_leaf_64);
> 476
> 477 switch (p_type) {
> 478 case maple_range_64:
> 479 case maple_arange_64:
> 480 shift = MAPLE_PARENT_SLOT_SHIFT;
> 481 type = MAPLE_PARENT_RANGE64;
> 482 break;
> 483 default:
> 484 case maple_dense:
> 485 case maple_leaf_64:
> 486 shift = type = 0;
> 487 break;
> 488 }
> 489
> 490 val &= ~MAPLE_NODE_MASK; /* Clear all node metadata in parent */
> → 491 val |= (slot << shift) | type;
> 492 mte_to_node(enode)->parent = ma_parent_ptr(val);
> 493 }
>
> lib/maple_tree.c:1701
> ---------------------
> 1682 /*
> 1683 * mas_adopt_children() - Set the parent pointer of all nodes in @parent to
> 1684 * @parent with the slot encoded.
> 1685 * @mas - the maple state (for the tree)
> 1686 * @parent - the maple encoded node containing the children.
> 1687 */
> 1688 static inline void mas_adopt_children(struct ma_state *mas,
> 1689 struct maple_enode *parent)
> 1690 {
> 1691 enum maple_type type = mte_node_type(parent);
> 1692 struct maple_node *node = mte_to_node(parent);
> 1693 void __rcu **slots = ma_slots(node, type);
> 1694 unsigned long *pivots = ma_pivots(node, type);
> 1695 struct maple_enode *child;
> 1696 unsigned char offset;
> 1697
> 1698 offset = ma_data_end(node, type, pivots, mas->max);
> 1699 do {
> 1700 child = mas_slot_locked(mas, slots, offset);
> → 1701 mas_set_parent(mas, child, parent, offset);
> 1702 } while (offset--);
> 1703 }
>
> 2562 static inline void mas_topiary_replace(struct ma_state *mas,
> 2563 struct maple_enode *old_enode)
> 2564 {
> 2565 struct ma_state tmp[3], tmp_next[3];
> 2566 MA_TOPIARY(subtrees, mas->tree);
> 2567 bool in_rcu;
> 2568 int i, n;
> 2569
> 2570 /* Place data in tree & then mark node as old */
> 2571 mas_put_in_tree(mas, old_enode);
> 2572
> 2573 /* Update the parent pointers in the tree */
> 2574 tmp[0] = *mas;
> 2575 tmp[0].offset = 0;
> 2576 tmp[1].node = MAS_NONE;
> 2577 tmp[2].node = MAS_NONE;
> 2578 while (!mte_is_leaf(tmp[0].node)) {
> 2579 n = 0;
> 2580 for (i = 0; i < 3; i++) {
> 2581 if (mas_is_none(&tmp[i]))
> 2582 continue;
> 2583
> 2584 while (n < 3) {
> 2585 if (!mas_find_child(&tmp[i], &tmp_next[n]))
> 2586 break;
> 2587 n++;
> 2588 }
> 2589
> → 2590 mas_adopt_children(&tmp[i], tmp[i].node);
> 2591 }
> 2592
> 2593 if (MAS_WARN_ON(mas, n == 0))
> 2594 break;
> 2595
> 2596 while (n < 3)
> 2597 tmp_next[n++].node = MAS_NONE;
> 2598
> 2599 for (i = 0; i < 3; i++)
> 2600 tmp[i] = tmp_next[i];
> 2601 }
> 2602
> 2603 /* Collect the old nodes that need to be discarded */
> 2604 if (mte_is_leaf(old_enode))
> 2605 return mas_free(mas, old_enode);
> 2606
> 2607 tmp[0] = *mas;
> 2608 tmp[0].offset = 0;
> 2609 tmp[0].node = old_enode;
> 2610 tmp[1].node = MAS_NONE;
> 2611 tmp[2].node = MAS_NONE;
> 2612 in_rcu = mt_in_rcu(mas->tree);
> 2613 do {
> 2614 n = 0;
> 2615 for (i = 0; i < 3; i++) {
> 2616 if (mas_is_none(&tmp[i]))
> 2617 continue;
>
> The read side:
>
> 527 /*
> 528 * ma_dead_node() - check if the @enode is dead.
> 529 * @enode: The encoded maple node
> 530 *
> 531 * Return: true if dead, false otherwise.
> 532 */
> 533 static inline bool ma_dead_node(const struct maple_node *node)
> 534 {
> 535 struct maple_node *parent;
> 536
> 537 /* Do not reorder reads from the node prior to the parent check */
> 538 smp_rmb();
> → 539 parent = (void *)((unsigned long) node->parent & ~MAPLE_NODE_MASK);
> 540 return (parent == node);
> 541 }
>
> 2767 static inline void *mtree_range_walk(struct ma_state *mas)
> 2768 {
> 2769 unsigned long *pivots;
> 2770 unsigned char offset;
> 2771 struct maple_node *node;
> 2772 struct maple_enode *next, *last;
> 2773 enum maple_type type;
> 2774 void __rcu **slots;
> 2775 unsigned char end;
> 2776 unsigned long max, min;
> 2777 unsigned long prev_max, prev_min;
> 2778
> 2779 next = mas->node;
> 2780 min = mas->min;
> 2781 max = mas->max;
> 2782 do {
> 2783 offset = 0;
> 2784 last = next;
> 2785 node = mte_to_node(next);
> 2786 type = mte_node_type(next);
> 2787 pivots = ma_pivots(node, type);
> 2788 end = ma_data_end(node, type, pivots, max);
> → 2789 if (unlikely(ma_dead_node(node)))
> 2790 goto dead_node;
> 2791
> 2792 if (pivots[offset] >= mas->index) {
> 2793 prev_max = max;
> 2794 prev_min = min;
> 2795 max = pivots[offset];
> 2796 goto next;
> 2797 }
> 2798
> 2799 do {
> 2800 offset++;
> 2801 } while ((offset < end) && (pivots[offset] < mas->index));
>
> As it is evident, ma_dead_node() expands to:
>
> 527 /*
> 528 * ma_dead_node() - check if the @enode is dead.
> 529 * @enode: The encoded maple node
> 530 *
> 531 * Return: true if dead, false otherwise.
> 532 */
> 533 static inline bool ma_dead_node(const struct maple_node *node)
> 534 {
> 535 struct maple_node *parent;
> 536
> 537 /* Do not reorder reads from the node prior to the parent check */
> 538 smp_rmb();
> → 539 parent = (void *)((unsigned long) node->parent & ~MAPLE_NODE_MASK);
> 540 return (parent == node);
> 541 }
>
> as above, but the smb_rmb() protection is here, so the KCSAN warning should be double-checked for
> validity.
>
> But I do not really understand maple trees to their depth, I am only reporting the symptomatic
> outlook of the reported data-race.

This is the most complex part of the maple tree, and unique to the
ability to store a range over multiple existing entries. I recently
rewrote this particular section to avoid a potential live-lock issue.

I am confident that the parent pointer will not be set to the node
pointer as checked by ma_dead_node(). But in an abundance of caution
and to resolve this potential false-positive, I am looking at this in
more detail.

This race is to see if the walk down the tree into unchanged data will
be seen before it is placed in the new subtree with its data also
unchanged. Since we know the parent can never be the node, I feel this
is safe - but I'm not willing to live with the complaints from kasan.

>
> This is all-in-all a very interested subject, and I wish there was a way to just slurp all those
> interesting kernel intrinsics into the brain, but it just ain't that easy. Forgive me for being
> open ...

I appreciate that and your detailed analysis with code pointers of where
this happens. Is this easy to recreate? If so, how? Can you attach
your kernel config?

Thanks,
Liam