thp: fix splitting of hwpoisoned hugepages

[mv-sheeva.git] / mm / huge_memory.c

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index f4f6041176a47e56549007ee69f364f7ea9f5058..e187454d82f666a70bab08762ca92cd60b8f848c 100644 (file)
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -16,6 +16,7 @@
 #include <linux/kthread.h>
 #include <linux/khugepaged.h>
 #include <linux/freezer.h>
+#include <linux/mman.h>
 #include <asm/tlb.h>
 #include <asm/pgalloc.h>
 #include "internal.h"
@@ -1202,6 +1203,8 @@ static void __split_huge_page_refcount(struct page *page)
                BUG_ON(!PageDirty(page_tail));
                BUG_ON(!PageSwapBacked(page_tail));
 
+               mem_cgroup_split_huge_fixup(page, page_tail);
+
                lru_add_page_tail(zone, page, page_tail);
        }
 
@@ -1388,18 +1391,49 @@ out:
        return ret;
 }
 
-int hugepage_madvise(unsigned long *vm_flags)
+int hugepage_madvise(struct vm_area_struct *vma,
+                    unsigned long *vm_flags, int advice)
 {
-       /*
-        * Be somewhat over-protective like KSM for now!
-        */
-       if (*vm_flags & (VM_HUGEPAGE | VM_SHARED  | VM_MAYSHARE   |
-                        VM_PFNMAP   | VM_IO      | VM_DONTEXPAND |
-                        VM_RESERVED | VM_HUGETLB | VM_INSERTPAGE |
-                        VM_MIXEDMAP | VM_SAO))
-               return -EINVAL;
-
-       *vm_flags |= VM_HUGEPAGE;
+       switch (advice) {
+       case MADV_HUGEPAGE:
+               /*
+                * Be somewhat over-protective like KSM for now!
+                */
+               if (*vm_flags & (VM_HUGEPAGE |
+                                VM_SHARED   | VM_MAYSHARE   |
+                                VM_PFNMAP   | VM_IO      | VM_DONTEXPAND |
+                                VM_RESERVED | VM_HUGETLB | VM_INSERTPAGE |
+                                VM_MIXEDMAP | VM_SAO))
+                       return -EINVAL;
+               *vm_flags &= ~VM_NOHUGEPAGE;
+               *vm_flags |= VM_HUGEPAGE;
+               /*
+                * If the vma become good for khugepaged to scan,
+                * register it here without waiting a page fault that
+                * may not happen any time soon.
+                */
+               if (unlikely(khugepaged_enter_vma_merge(vma)))
+                       return -ENOMEM;
+               break;
+       case MADV_NOHUGEPAGE:
+               /*
+                * Be somewhat over-protective like KSM for now!
+                */
+               if (*vm_flags & (VM_NOHUGEPAGE |
+                                VM_SHARED   | VM_MAYSHARE   |
+                                VM_PFNMAP   | VM_IO      | VM_DONTEXPAND |
+                                VM_RESERVED | VM_HUGETLB | VM_INSERTPAGE |
+                                VM_MIXEDMAP | VM_SAO))
+                       return -EINVAL;
+               *vm_flags &= ~VM_HUGEPAGE;
+               *vm_flags |= VM_NOHUGEPAGE;
+               /*
+                * Setting VM_NOHUGEPAGE will prevent khugepaged from scanning
+                * this vma even if we leave the mm registered in khugepaged if
+                * it got registered before VM_NOHUGEPAGE was set.
+                */
+               break;
+       }
 
        return 0;
 }
@@ -1765,7 +1799,8 @@ static void collapse_huge_page(struct mm_struct *mm,
        if (address < hstart || address + HPAGE_PMD_SIZE > hend)
                goto out;
 
-       if (!(vma->vm_flags & VM_HUGEPAGE) && !khugepaged_always())
+       if ((!(vma->vm_flags & VM_HUGEPAGE) && !khugepaged_always()) ||
+           (vma->vm_flags & VM_NOHUGEPAGE))
                goto out;
 
        /* VM_PFNMAP vmas may have vm_ops null but vm_file set */
@@ -1804,9 +1839,9 @@ static void collapse_huge_page(struct mm_struct *mm,
        spin_lock(ptl);
        isolated = __collapse_huge_page_isolate(vma, address, pte);
        spin_unlock(ptl);
-       pte_unmap(pte);
 
        if (unlikely(!isolated)) {
+               pte_unmap(pte);
                spin_lock(&mm->page_table_lock);
                BUG_ON(!pmd_none(*pmd));
                set_pmd_at(mm, address, pmd, _pmd);
@@ -1823,6 +1858,7 @@ static void collapse_huge_page(struct mm_struct *mm,
        anon_vma_unlock(vma->anon_vma);
 
        __collapse_huge_page_copy(pte, new_page, vma, address, ptl);
+       pte_unmap(pte);
        __SetPageUptodate(new_page);
        pgtable = pmd_pgtable(_pmd);
        VM_BUG_ON(page_count(pgtable) != 1);
@@ -1988,8 +2024,9 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages,
                        break;
                }
 
-               if (!(vma->vm_flags & VM_HUGEPAGE) &&
-                   !khugepaged_always()) {
+               if ((!(vma->vm_flags & VM_HUGEPAGE) &&
+                    !khugepaged_always()) ||
+                   (vma->vm_flags & VM_NOHUGEPAGE)) {
                        progress++;
                        continue;
                }