Subject: [merged] mm-memcg-fix-race-condition-between-memcg-teardown-and-swapin.patch removed from -mm tree To: hannes@xxxxxxxxxxx,mhocko@xxxxxxx,rientjes@xxxxxxxxxx,stable@xxxxxxxxxx,mm-commits@xxxxxxxxxxxxxxx From: akpm@xxxxxxxxxxxxxxxxxxxx Date: Mon, 16 Dec 2013 12:39:42 -0800 The patch titled Subject: mm: memcg: fix race condition between memcg teardown and swapin has been removed from the -mm tree. Its filename was mm-memcg-fix-race-condition-between-memcg-teardown-and-swapin.patch This patch was dropped because it was merged into mainline or a subsystem tree ------------------------------------------------------ From: Johannes Weiner <hannes@xxxxxxxxxxx> Subject: mm: memcg: fix race condition between memcg teardown and swapin There is a race condition between a memcg being torn down and a swapin triggered from a different memcg of a page that was recorded to belong to the exiting memcg on swapout (with CONFIG_MEMCG_SWAP extension). The result is unreclaimable pages pointing to dead memcgs, which can lead to anything from endless loops in later memcg teardown (the page is charged to all hierarchical parents but is not on any LRU list) or crashes from following the dangling memcg pointer. Memcgs with tasks in them can not be torn down and usually charges don't show up in memcgs without tasks. Swapin with the CONFIG_MEMCG_SWAP extension is the notable exception because it charges the cgroup that was recorded as owner during swapout, which may be empty and in the process of being torn down when a task in another memcg triggers the swapin: teardown: swapin: lookup_swap_cgroup_id() rcu_read_lock() mem_cgroup_lookup() css_tryget() rcu_read_unlock() disable css_tryget() call_rcu() offline_css() reparent_charges() res_counter_charge() (hierarchical!) css_put() css_free() pc->mem_cgroup = dead memcg add page to dead lru Add a final reparenting step into css_free() to make sure any such raced charges are moved out of the memcg before it's finally freed. In the longer term it would be cleaner to have the css_tryget() and the res_counter charge under the same RCU lock section so that the charge reparenting is deferred until the last charge whose tryget succeeded is visible. But this will require more invasive changes that will be harder to evaluate and backport into stable, so better defer them to a separate change set. Signed-off-by: Johannes Weiner <hannes@xxxxxxxxxxx> Acked-by: Michal Hocko <mhocko@xxxxxxx> Cc: David Rientjes <rientjes@xxxxxxxxxx> Cc: <stable@xxxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- mm/memcontrol.c | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) diff -puN mm/memcontrol.c~mm-memcg-fix-race-condition-between-memcg-teardown-and-swapin mm/memcontrol.c --- a/mm/memcontrol.c~mm-memcg-fix-race-condition-between-memcg-teardown-and-swapin +++ a/mm/memcontrol.c @@ -6355,6 +6355,42 @@ static void mem_cgroup_css_offline(struc static void mem_cgroup_css_free(struct cgroup_subsys_state *css) { struct mem_cgroup *memcg = mem_cgroup_from_css(css); + /* + * XXX: css_offline() would be where we should reparent all + * memory to prepare the cgroup for destruction. However, + * memcg does not do css_tryget() and res_counter charging + * under the same RCU lock region, which means that charging + * could race with offlining. Offlining only happens to + * cgroups with no tasks in them but charges can show up + * without any tasks from the swapin path when the target + * memcg is looked up from the swapout record and not from the + * current task as it usually is. A race like this can leak + * charges and put pages with stale cgroup pointers into + * circulation: + * + * #0 #1 + * lookup_swap_cgroup_id() + * rcu_read_lock() + * mem_cgroup_lookup() + * css_tryget() + * rcu_read_unlock() + * disable css_tryget() + * call_rcu() + * offline_css() + * reparent_charges() + * res_counter_charge() + * css_put() + * css_free() + * pc->mem_cgroup = dead memcg + * add page to lru + * + * The bulk of the charges are still moved in offline_css() to + * avoid pinning a lot of pages in case a long-term reference + * like a swapout record is deferring the css_free() to long + * after offlining. But this makes sure we catch any charges + * made after offlining: + */ + mem_cgroup_reparent_charges(memcg); memcg_destroy_kmem(memcg); __mem_cgroup_free(memcg); _ Patches currently in -mm which might be from hannes@xxxxxxxxxxx are mm-page_alloc-exclude-unreclaimable-allocations-from-zone-fairness-policy.patch mm-memcg-avoid-oom-notification-when-current-needs-access-to-memory-reserves.patch proc-meminfo-provide-estimated-available-memory.patch mm-mempolicy-remove-unneeded-functions-for-uma-configs.patch memcg-fix-kmem_account_flags-check-in-memcg_can_account_kmem.patch memcg-make-memcg_update_cache_sizes-static.patch mm-memblock-debug-correct-displaying-of-upper-memory-boundary.patch mm-memblock-debug-dont-free-reserved-array-if-arch_discard_memblock.patch mm-bootmem-remove-duplicated-declaration-of-__free_pages_bootmem.patch mm-memblock-remove-unnecessary-inclusions-of-bootmemh.patch mm-memblock-drop-warn-and-use-smp_cache_bytes-as-a-default-alignment.patch mm-memblock-reorder-parameters-of-memblock_find_in_range_node.patch mm-memblock-switch-to-use-numa_no_node-instead-of-max_numnodes.patch mm-memblock-add-memblock-memory-allocation-apis.patch mm-memblock-add-memblock-memory-allocation-apis-fix.patch mm-init-use-memblock-apis-for-early-memory-allocations.patch mm-printk-use-memblock-apis-for-early-memory-allocations.patch mm-page_alloc-use-memblock-apis-for-early-memory-allocations.patch mm-power-use-memblock-apis-for-early-memory-allocations.patch lib-swiotlbc-use-memblock-apis-for-early-memory-allocations.patch lib-cpumaskc-use-memblock-apis-for-early-memory-allocations.patch mm-sparse-use-memblock-apis-for-early-memory-allocations.patch mm-hugetlb-use-memblock-apis-for-early-memory-allocations.patch mm-page_cgroup-use-memblock-apis-for-early-memory-allocations.patch mm-percpu-use-memblock-apis-for-early-memory-allocations.patch mm-memory_hotplug-use-memblock-apis-for-early-memory-allocations.patch drivers-firmware-memmapc-use-memblock-apis-for-early-memory-allocations.patch arch-arm-kernel-use-memblock-apis-for-early-memory-allocations.patch arch-arm-mm-initc-use-memblock-apis-for-early-memory-allocations.patch arch-arm-mach-omap2-omap_hwmodc-use-memblock-apis-for-early-memory-allocations.patch memcg-oom-lock-mem_cgroup_print_oom_info.patch swap-add-a-simple-detector-for-inappropriate-swapin-readahead-fix.patch linux-next.patch debugging-keep-track-of-page-owners.patch -- To unsubscribe from this list: send the line "unsubscribe mm-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html