#include <linux/suspend.h>
#include <linux/slab.h>
#include <linux/swapops.h>
+#include <linux/hugetlb.h>
#include "internal.h"
int sysctl_memory_failure_early_kill __read_mostly = 0;
* signal.
*/
static int kill_proc_ao(struct task_struct *t, unsigned long addr, int trapno,
- unsigned long pfn)
+ unsigned long pfn, struct page *page)
{
struct siginfo si;
int ret;
#ifdef __ARCH_SI_TRAPNO
si.si_trapno = trapno;
#endif
- si.si_addr_lsb = PAGE_SHIFT;
+ si.si_addr_lsb = compound_order(compound_head(page)) + PAGE_SHIFT;
/*
* Don't use force here, it's convenient if the signal
* can be temporarily blocked.
int nr;
do {
nr = shrink_slab(1000, GFP_KERNEL, 1000);
- if (page_count(p) == 0)
+ if (page_count(p) == 1)
break;
} while (nr > 10);
}
* wrong earlier.
*/
static void kill_procs_ao(struct list_head *to_kill, int doit, int trapno,
- int fail, unsigned long pfn)
+ int fail, struct page *page, unsigned long pfn)
{
struct to_kill *tk, *next;
* process anyways.
*/
else if (kill_proc_ao(tk->tsk, tk->addr, trapno,
- pfn) < 0)
+ pfn, page) < 0)
printk(KERN_ERR
"MCE %#lx: Cannot send advisory machine check signal to %s:%d\n",
pfn, tk->tsk->comm, tk->tsk->pid);
/*
* Huge pages. Needs work.
* Issues:
- * No rmap support so we cannot find the original mapper. In theory could walk
- * all MMs and look for the mappings, but that would be non atomic and racy.
- * Need rmap for hugepages for this. Alternatively we could employ a heuristic,
- * like just walking the current process and hoping it has it mapped (that
- * should be usually true for the common "shared database cache" case)
- * Should handle free huge pages and dequeue them too, but this needs to
- * handle huge page accounting correctly.
+ * - Error on hugepage is contained in hugepage unit (not in raw page unit.)
+ * To narrow down kill region to one page, we need to break up pmd.
*/
static int me_huge_page(struct page *p, unsigned long pfn)
{
- return FAILED;
+ int res = 0;
+ struct page *hpage = compound_head(p);
+ /*
+ * We can safely recover from error on free or reserved (i.e.
+ * not in-use) hugepage by dequeuing it from freelist.
+ * To check whether a hugepage is in-use or not, we can't use
+ * page->lru because it can be used in other hugepage operations,
+ * such as __unmap_hugepage_range() and gather_surplus_pages().
+ * So instead we use page_mapping() and PageAnon().
+ * We assume that this function is called with page lock held,
+ * so there is no race between isolation and mapping/unmapping.
+ */
+ if (!(page_mapping(hpage) || PageAnon(hpage))) {
+ res = dequeue_hwpoisoned_huge_page(hpage);
+ if (!res)
+ return RECOVERED;
+ }
+ return DELAYED;
}
/*
int ret;
int i;
int kill = 1;
+ struct page *hpage = compound_head(p);
if (PageReserved(p) || PageSlab(p))
return SWAP_SUCCESS;
* This check implies we don't kill processes if their pages
* are in the swap cache early. Those are always late kills.
*/
- if (!page_mapped(p))
+ if (!page_mapped(hpage))
return SWAP_SUCCESS;
- if (PageCompound(p) || PageKsm(p))
+ if (PageKsm(p))
return SWAP_FAIL;
if (PageSwapCache(p)) {
* XXX: the dirty test could be racy: set_page_dirty() may not always
* be called inside page lock (it's recommended but not enforced).
*/
- mapping = page_mapping(p);
- if (!PageDirty(p) && mapping && mapping_cap_writeback_dirty(mapping)) {
- if (page_mkclean(p)) {
- SetPageDirty(p);
+ mapping = page_mapping(hpage);
+ if (!PageDirty(hpage) && mapping &&
+ mapping_cap_writeback_dirty(mapping)) {
+ if (page_mkclean(hpage)) {
+ SetPageDirty(hpage);
} else {
kill = 0;
ttu |= TTU_IGNORE_HWPOISON;
* there's nothing that can be done.
*/
if (kill)
- collect_procs(p, &tokill);
+ collect_procs(hpage, &tokill);
/*
* try_to_unmap can fail temporarily due to races.
* Try a few times (RED-PEN better strategy?)
*/
for (i = 0; i < N_UNMAP_TRIES; i++) {
- ret = try_to_unmap(p, ttu);
+ ret = try_to_unmap(hpage, ttu);
if (ret == SWAP_SUCCESS)
break;
pr_debug("MCE %#lx: try_to_unmap retry needed %d\n", pfn, ret);
if (ret != SWAP_SUCCESS)
printk(KERN_ERR "MCE %#lx: failed to unmap page (mapcount=%d)\n",
- pfn, page_mapcount(p));
+ pfn, page_mapcount(hpage));
/*
* Now that the dirty bit has been propagated to the
* use a more force-full uncatchable kill to prevent
* any accesses to the poisoned memory.
*/
- kill_procs_ao(&tokill, !!PageDirty(p), trapno,
- ret != SWAP_SUCCESS, pfn);
+ kill_procs_ao(&tokill, !!PageDirty(hpage), trapno,
+ ret != SWAP_SUCCESS, p, pfn);
return ret;
}
+static void set_page_hwpoison_huge_page(struct page *hpage)
+{
+ int i;
+ int nr_pages = 1 << compound_order(hpage);
+ for (i = 0; i < nr_pages; i++)
+ SetPageHWPoison(hpage + i);
+}
+
+static void clear_page_hwpoison_huge_page(struct page *hpage)
+{
+ int i;
+ int nr_pages = 1 << compound_order(hpage);
+ for (i = 0; i < nr_pages; i++)
+ ClearPageHWPoison(hpage + i);
+}
+
int __memory_failure(unsigned long pfn, int trapno, int flags)
{
struct page_state *ps;
struct page *p;
+ struct page *hpage;
int res;
+ unsigned int nr_pages;
if (!sysctl_memory_failure_recovery)
panic("Memory failure from trap %d on page %lx", trapno, pfn);
}
p = pfn_to_page(pfn);
+ hpage = compound_head(p);
if (TestSetPageHWPoison(p)) {
printk(KERN_ERR "MCE %#lx: already hardware poisoned\n", pfn);
return 0;
}
- atomic_long_add(1, &mce_bad_pages);
+ nr_pages = 1 << compound_order(hpage);
+ atomic_long_add(nr_pages, &mce_bad_pages);
/*
* We need/can do nothing about count=0 pages.
* 1) it's a free page, and therefore in safe hand:
* prep_new_page() will be the gate keeper.
- * 2) it's part of a non-compound high order page.
+ * 2) it's a free hugepage, which is also safe:
+ * an affected hugepage will be dequeued from hugepage freelist,
+ * so there's no concern about reusing it ever after.
+ * 3) it's part of a non-compound high order page.
* Implies some kernel user: cannot stop them from
* R/W the page; let's pray that the page has been
* used and will be freed some time later.
* that may make page_freeze_refs()/page_unfreeze_refs() mismatch.
*/
if (!(flags & MF_COUNT_INCREASED) &&
- !get_page_unless_zero(compound_head(p))) {
+ !get_page_unless_zero(hpage)) {
if (is_free_buddy_page(p)) {
action_result(pfn, "free buddy", DELAYED);
return 0;
+ } else if (PageHuge(hpage)) {
+ /*
+ * Check "just unpoisoned", "filter hit", and
+ * "race with other subpage."
+ */
+ lock_page_nosync(hpage);
+ if (!PageHWPoison(hpage)
+ || (hwpoison_filter(p) && TestClearPageHWPoison(p))
+ || (p != hpage && TestSetPageHWPoison(hpage))) {
+ atomic_long_sub(nr_pages, &mce_bad_pages);
+ return 0;
+ }
+ set_page_hwpoison_huge_page(hpage);
+ res = dequeue_hwpoisoned_huge_page(hpage);
+ action_result(pfn, "free huge",
+ res ? IGNORED : DELAYED);
+ unlock_page(hpage);
+ return res;
} else {
action_result(pfn, "high order kernel", IGNORED);
return -EBUSY;
* The check (unnecessarily) ignores LRU pages being isolated and
* walked by the page reclaim code, however that's not a big loss.
*/
- if (!PageLRU(p))
+ if (!PageLRU(p) && !PageHuge(p))
shake_page(p, 0);
- if (!PageLRU(p)) {
+ if (!PageLRU(p) && !PageHuge(p)) {
/*
* shake_page could have turned it free.
*/
* It's very difficult to mess with pages currently under IO
* and in many cases impossible, so we just avoid it here.
*/
- lock_page_nosync(p);
+ lock_page_nosync(hpage);
/*
* unpoison always clear PG_hwpoison inside page lock
}
if (hwpoison_filter(p)) {
if (TestClearPageHWPoison(p))
- atomic_long_dec(&mce_bad_pages);
- unlock_page(p);
- put_page(p);
+ atomic_long_sub(nr_pages, &mce_bad_pages);
+ unlock_page(hpage);
+ put_page(hpage);
+ return 0;
+ }
+
+ /*
+ * For error on the tail page, we should set PG_hwpoison
+ * on the head page to show that the hugepage is hwpoisoned
+ */
+ if (PageTail(p) && TestSetPageHWPoison(hpage)) {
+ action_result(pfn, "hugepage already hardware poisoned",
+ IGNORED);
+ unlock_page(hpage);
+ put_page(hpage);
return 0;
}
+ /*
+ * Set PG_hwpoison on all pages in an error hugepage,
+ * because containment is done in hugepage unit for now.
+ * Since we have done TestSetPageHWPoison() for the head page with
+ * page lock held, we can safely set PG_hwpoison bits on tail pages.
+ */
+ if (PageHuge(p))
+ set_page_hwpoison_huge_page(hpage);
wait_on_page_writeback(p);
}
}
out:
- unlock_page(p);
+ unlock_page(hpage);
return res;
}
EXPORT_SYMBOL_GPL(__memory_failure);
struct page *page;
struct page *p;
int freeit = 0;
+ unsigned int nr_pages;
if (!pfn_valid(pfn))
return -ENXIO;
return 0;
}
+ nr_pages = 1 << compound_order(page);
+
if (!get_page_unless_zero(page)) {
+ /*
+ * Since HWPoisoned hugepage should have non-zero refcount,
+ * race between memory failure and unpoison seems to happen.
+ * In such case unpoison fails and memory failure runs
+ * to the end.
+ */
+ if (PageHuge(page)) {
+ pr_debug("MCE: Memory failure is now running on free hugepage %#lx\n", pfn);
+ return 0;
+ }
if (TestClearPageHWPoison(p))
- atomic_long_dec(&mce_bad_pages);
+ atomic_long_sub(nr_pages, &mce_bad_pages);
pr_debug("MCE: Software-unpoisoned free page %#lx\n", pfn);
return 0;
}
* the PG_hwpoison page will be caught and isolated on the entrance to
* the free buddy page pool.
*/
- if (TestClearPageHWPoison(p)) {
+ if (TestClearPageHWPoison(page)) {
pr_debug("MCE: Software-unpoisoned page %#lx\n", pfn);
- atomic_long_dec(&mce_bad_pages);
+ atomic_long_sub(nr_pages, &mce_bad_pages);
freeit = 1;
}
+ if (PageHuge(p))
+ clear_page_hwpoison_huge_page(page);
unlock_page(page);
put_page(page);
static struct page *new_page(struct page *p, unsigned long private, int **x)
{
int nid = page_to_nid(p);
- return alloc_pages_exact_node(nid, GFP_HIGHUSER_MOVABLE, 0);
+ if (PageHuge(p))
+ return alloc_huge_page_node(page_hstate(compound_head(p)),
+ nid);
+ else
+ return alloc_pages_exact_node(nid, GFP_HIGHUSER_MOVABLE, 0);
}
/*
* was free.
*/
set_migratetype_isolate(p);
+ /*
+ * When the target page is a free hugepage, just remove it
+ * from free hugepage list.
+ */
if (!get_page_unless_zero(compound_head(p))) {
- if (is_free_buddy_page(p)) {
+ if (PageHuge(p)) {
+ pr_debug("get_any_page: %#lx free huge page\n", pfn);
+ ret = dequeue_hwpoisoned_huge_page(compound_head(p));
+ } else if (is_free_buddy_page(p)) {
pr_debug("get_any_page: %#lx free buddy page\n", pfn);
/* Set hwpoison bit while page is still isolated */
SetPageHWPoison(p);
return ret;
}
+static int soft_offline_huge_page(struct page *page, int flags)
+{
+ int ret;
+ unsigned long pfn = page_to_pfn(page);
+ struct page *hpage = compound_head(page);
+ LIST_HEAD(pagelist);
+
+ ret = get_any_page(page, pfn, flags);
+ if (ret < 0)
+ return ret;
+ if (ret == 0)
+ goto done;
+
+ if (PageHWPoison(hpage)) {
+ put_page(hpage);
+ pr_debug("soft offline: %#lx hugepage already poisoned\n", pfn);
+ return -EBUSY;
+ }
+
+ /* Keep page count to indicate a given hugepage is isolated. */
+
+ list_add(&hpage->lru, &pagelist);
+ ret = migrate_huge_pages(&pagelist, new_page, MPOL_MF_MOVE_ALL, 0);
+ if (ret) {
+ pr_debug("soft offline: %#lx: migration failed %d, type %lx\n",
+ pfn, ret, page->flags);
+ if (ret > 0)
+ ret = -EIO;
+ return ret;
+ }
+done:
+ if (!PageHWPoison(hpage))
+ atomic_long_add(1 << compound_order(hpage), &mce_bad_pages);
+ set_page_hwpoison_huge_page(hpage);
+ dequeue_hwpoisoned_huge_page(hpage);
+ /* keep elevated page count for bad page */
+ return ret;
+}
+
/**
* soft_offline_page - Soft offline a page.
* @page: page to offline
int ret;
unsigned long pfn = page_to_pfn(page);
+ if (PageHuge(page))
+ return soft_offline_huge_page(page, flags);
+
ret = get_any_page(page, pfn, flags);
if (ret < 0)
return ret;