linux_dsm_epyc7002/include/linux/balloon_compaction.h
Michael S. Tsirkin c7cdff0e86 virtio_balloon: fix deadlock on OOM
fill_balloon doing memory allocations under balloon_lock
can cause a deadlock when leak_balloon is called from
virtballoon_oom_notify and tries to take same lock.

To fix, split page allocation and enqueue and do allocations outside the lock.

Here's a detailed analysis of the deadlock by Tetsuo Handa:

In leak_balloon(), mutex_lock(&vb->balloon_lock) is called in order to
serialize against fill_balloon(). But in fill_balloon(),
alloc_page(GFP_HIGHUSER[_MOVABLE] | __GFP_NOMEMALLOC | __GFP_NORETRY) is
called with vb->balloon_lock mutex held. Since GFP_HIGHUSER[_MOVABLE]
implies __GFP_DIRECT_RECLAIM | __GFP_IO | __GFP_FS, despite __GFP_NORETRY
is specified, this allocation attempt might indirectly depend on somebody
else's __GFP_DIRECT_RECLAIM memory allocation. And such indirect
__GFP_DIRECT_RECLAIM memory allocation might call leak_balloon() via
virtballoon_oom_notify() via blocking_notifier_call_chain() callback via
out_of_memory() when it reached __alloc_pages_may_oom() and held oom_lock
mutex. Since vb->balloon_lock mutex is already held by fill_balloon(), it
will cause OOM lockup.

  Thread1                                       Thread2
    fill_balloon()
      takes a balloon_lock
      balloon_page_enqueue()
        alloc_page(GFP_HIGHUSER_MOVABLE)
          direct reclaim (__GFP_FS context)       takes a fs lock
            waits for that fs lock                  alloc_page(GFP_NOFS)
                                                      __alloc_pages_may_oom()
                                                        takes the oom_lock
                                                        out_of_memory()
                                                          blocking_notifier_call_chain()
                                                            leak_balloon()
                                                              tries to take that balloon_lock and deadlocks

Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Wei Wang <wei.w.wang@intel.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2017-11-14 23:57:38 +02:00

230 lines
7.2 KiB
C

/* SPDX-License-Identifier: GPL-2.0 */
/*
* include/linux/balloon_compaction.h
*
* Common interface definitions for making balloon pages movable by compaction.
*
* Despite being perfectly possible to perform ballooned pages migration, they
* make a special corner case to compaction scans because balloon pages are not
* enlisted at any LRU list like the other pages we do compact / migrate.
*
* As the page isolation scanning step a compaction thread does is a lockless
* procedure (from a page standpoint), it might bring some racy situations while
* performing balloon page compaction. In order to sort out these racy scenarios
* and safely perform balloon's page compaction and migration we must, always,
* ensure following these three simple rules:
*
* i. when updating a balloon's page ->mapping element, strictly do it under
* the following lock order, independently of the far superior
* locking scheme (lru_lock, balloon_lock):
* +-page_lock(page);
* +--spin_lock_irq(&b_dev_info->pages_lock);
* ... page->mapping updates here ...
*
* ii. before isolating or dequeueing a balloon page from the balloon device
* pages list, the page reference counter must be raised by one and the
* extra refcount must be dropped when the page is enqueued back into
* the balloon device page list, thus a balloon page keeps its reference
* counter raised only while it is under our special handling;
*
* iii. after the lockless scan step have selected a potential balloon page for
* isolation, re-test the PageBalloon mark and the PagePrivate flag
* under the proper page lock, to ensure isolating a valid balloon page
* (not yet isolated, nor under release procedure)
*
* iv. isolation or dequeueing procedure must clear PagePrivate flag under
* page lock together with removing page from balloon device page list.
*
* The functions provided by this interface are placed to help on coping with
* the aforementioned balloon page corner case, as well as to ensure the simple
* set of exposed rules are satisfied while we are dealing with balloon pages
* compaction / migration.
*
* Copyright (C) 2012, Red Hat, Inc. Rafael Aquini <aquini@redhat.com>
*/
#ifndef _LINUX_BALLOON_COMPACTION_H
#define _LINUX_BALLOON_COMPACTION_H
#include <linux/pagemap.h>
#include <linux/page-flags.h>
#include <linux/migrate.h>
#include <linux/gfp.h>
#include <linux/err.h>
#include <linux/fs.h>
#include <linux/list.h>
/*
* Balloon device information descriptor.
* This struct is used to allow the common balloon compaction interface
* procedures to find the proper balloon device holding memory pages they'll
* have to cope for page compaction / migration, as well as it serves the
* balloon driver as a page book-keeper for its registered balloon devices.
*/
struct balloon_dev_info {
unsigned long isolated_pages; /* # of isolated pages for migration */
spinlock_t pages_lock; /* Protection to pages list */
struct list_head pages; /* Pages enqueued & handled to Host */
int (*migratepage)(struct balloon_dev_info *, struct page *newpage,
struct page *page, enum migrate_mode mode);
struct inode *inode;
};
extern struct page *balloon_page_alloc(void);
extern void balloon_page_enqueue(struct balloon_dev_info *b_dev_info,
struct page *page);
extern struct page *balloon_page_dequeue(struct balloon_dev_info *b_dev_info);
static inline void balloon_devinfo_init(struct balloon_dev_info *balloon)
{
balloon->isolated_pages = 0;
spin_lock_init(&balloon->pages_lock);
INIT_LIST_HEAD(&balloon->pages);
balloon->migratepage = NULL;
balloon->inode = NULL;
}
#ifdef CONFIG_BALLOON_COMPACTION
extern const struct address_space_operations balloon_aops;
extern bool balloon_page_isolate(struct page *page,
isolate_mode_t mode);
extern void balloon_page_putback(struct page *page);
extern int balloon_page_migrate(struct address_space *mapping,
struct page *newpage,
struct page *page, enum migrate_mode mode);
/*
* balloon_page_insert - insert a page into the balloon's page list and make
* the page->private assignment accordingly.
* @balloon : pointer to balloon device
* @page : page to be assigned as a 'balloon page'
*
* Caller must ensure the page is locked and the spin_lock protecting balloon
* pages list is held before inserting a page into the balloon device.
*/
static inline void balloon_page_insert(struct balloon_dev_info *balloon,
struct page *page)
{
__SetPageBalloon(page);
__SetPageMovable(page, balloon->inode->i_mapping);
set_page_private(page, (unsigned long)balloon);
list_add(&page->lru, &balloon->pages);
}
/*
* balloon_page_delete - delete a page from balloon's page list and clear
* the page->private assignement accordingly.
* @page : page to be released from balloon's page list
*
* Caller must ensure the page is locked and the spin_lock protecting balloon
* pages list is held before deleting a page from the balloon device.
*/
static inline void balloon_page_delete(struct page *page)
{
__ClearPageBalloon(page);
__ClearPageMovable(page);
set_page_private(page, 0);
/*
* No touch page.lru field once @page has been isolated
* because VM is using the field.
*/
if (!PageIsolated(page))
list_del(&page->lru);
}
/*
* balloon_page_device - get the b_dev_info descriptor for the balloon device
* that enqueues the given page.
*/
static inline struct balloon_dev_info *balloon_page_device(struct page *page)
{
return (struct balloon_dev_info *)page_private(page);
}
static inline gfp_t balloon_mapping_gfp_mask(void)
{
return GFP_HIGHUSER_MOVABLE;
}
#else /* !CONFIG_BALLOON_COMPACTION */
static inline void balloon_page_insert(struct balloon_dev_info *balloon,
struct page *page)
{
__SetPageBalloon(page);
list_add(&page->lru, &balloon->pages);
}
static inline void balloon_page_delete(struct page *page)
{
__ClearPageBalloon(page);
list_del(&page->lru);
}
static inline bool __is_movable_balloon_page(struct page *page)
{
return false;
}
static inline bool balloon_page_movable(struct page *page)
{
return false;
}
static inline bool isolated_balloon_page(struct page *page)
{
return false;
}
static inline bool balloon_page_isolate(struct page *page)
{
return false;
}
static inline void balloon_page_putback(struct page *page)
{
return;
}
static inline int balloon_page_migrate(struct page *newpage,
struct page *page, enum migrate_mode mode)
{
return 0;
}
static inline gfp_t balloon_mapping_gfp_mask(void)
{
return GFP_HIGHUSER;
}
#endif /* CONFIG_BALLOON_COMPACTION */
/*
* balloon_page_push - insert a page into a page list.
* @head : pointer to list
* @page : page to be added
*
* Caller must ensure the page is private and protect the list.
*/
static inline void balloon_page_push(struct list_head *pages, struct page *page)
{
list_add(&page->lru, pages);
}
/*
* balloon_page_pop - remove a page from a page list.
* @head : pointer to list
* @page : page to be added
*
* Caller must ensure the page is private and protect the list.
*/
static inline struct page *balloon_page_pop(struct list_head *pages)
{
struct page *page = list_first_entry_or_null(pages, struct page, lru);
if (!page)
return NULL;
list_del(&page->lru);
return page;
}
#endif /* _LINUX_BALLOON_COMPACTION_H */