The patch titled
Subject: mm/rmap: try_to_unmap_one() do not call mmu_notifier under ptl v2
has been added to the -mm tree. Its filename is
mm-rmap-try_to_unmap_one-do-not-call-mmu_notifier-under-ptl-v2.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/mm-rmap-try_to_unmap_one-do-not-call-mmu_notifier-under-ptl-v2.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/mm-rmap-try_to_unmap_one-do-not-call-mmu_notifier-under-ptl-v2.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Jérôme Glisse <jglisse@xxxxxxxxxx>
Subject: mm/rmap: try_to_unmap_one() do not call mmu_notifier under ptl v2
MMU notifiers can sleep, but in try_to_unmap_one() we call
mmu_notifier_invalidate_page() under page table lock.
Let's instead use mmu_notifier_invalidate_range() outside
page_vma_mapped_walk() loop.
Link: http://lkml.kernel.org/r/20170809204333.27485-1-jglisse@xxxxxxxxxx
Fixes: c7ab0d2fdc84 ("mm: convert try_to_unmap_one() to use page_vma_mapped_walk()")
Signed-off-by: Jérôme Glisse <jglisse@xxxxxxxxxx>
Cc: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/rmap.c | 31 +++++++++++++++++--------------
1 file changed, 17 insertions(+), 14 deletions(-)
diff -puN mm/rmap.c~mm-rmap-try_to_unmap_one-do-not-call-mmu_notifier-under-ptl-v2 mm/rmap.c
--- a/mm/rmap.c~mm-rmap-try_to_unmap_one-do-not-call-mmu_notifier-under-ptl-v2
+++ a/mm/rmap.c
@@ -1328,7 +1328,7 @@ static bool try_to_unmap_one(struct page
};
pte_t pteval;
struct page *subpage;
- bool ret = true;
+ bool ret = true, invalidation_needed = false;
enum ttu_flags flags = (enum ttu_flags)arg;
/* munlock has nothing to gain from examining un-locked vmas */
@@ -1368,11 +1368,9 @@ static bool try_to_unmap_one(struct page
VM_BUG_ON_PAGE(!pvmw.pte, page);
subpage = page - page_to_pfn(page) + pte_pfn(*pvmw.pte);
- address = pvmw.address;
-
if (!(flags & TTU_IGNORE_ACCESS)) {
- if (ptep_clear_flush_young_notify(vma, address,
+ if (ptep_clear_flush_young_notify(vma, pvmw.address,
pvmw.pte)) {
ret = false;
page_vma_mapped_walk_done(&pvmw);
@@ -1381,7 +1379,7 @@ static bool try_to_unmap_one(struct page
}
/* Nuke the page table entry. */
- flush_cache_page(vma, address, pte_pfn(*pvmw.pte));
+ flush_cache_page(vma, pvmw.address, pte_pfn(*pvmw.pte));
if (should_defer_flush(mm, flags)) {
/*
* We clear the PTE but do not flush so potentially
@@ -1391,11 +1389,12 @@ static bool try_to_unmap_one(struct page
* transition on a cached TLB entry is written through
* and traps if the PTE is unmapped.
*/
- pteval = ptep_get_and_clear(mm, address, pvmw.pte);
+ pteval = ptep_get_and_clear(mm, pvmw.address,
+ pvmw.pte);
set_tlb_ubc_flush_pending(mm, pte_dirty(pteval));
} else {
- pteval = ptep_clear_flush(vma, address, pvmw.pte);
+ pteval = ptep_clear_flush(vma, pvmw.address, pvmw.pte);
}
/* Move the dirty bit to the page. Now the pte is gone. */
@@ -1410,12 +1409,12 @@ static bool try_to_unmap_one(struct page
if (PageHuge(page)) {
int nr = 1 << compound_order(page);
hugetlb_count_sub(nr, mm);
- set_huge_swap_pte_at(mm, address,
+ set_huge_swap_pte_at(mm, pvmw.address,
pvmw.pte, pteval,
vma_mmu_pagesize(vma));
} else {
dec_mm_counter(mm, mm_counter(page));
- set_pte_at(mm, address, pvmw.pte, pteval);
+ set_pte_at(mm, pvmw.address, pvmw.pte, pteval);
}
} else if (pte_unused(pteval)) {
@@ -1439,7 +1438,7 @@ static bool try_to_unmap_one(struct page
swp_pte = swp_entry_to_pte(entry);
if (pte_soft_dirty(pteval))
swp_pte = pte_swp_mksoft_dirty(swp_pte);
- set_pte_at(mm, address, pvmw.pte, swp_pte);
+ set_pte_at(mm, pvmw.address, pvmw.pte, swp_pte);
} else if (PageAnon(page)) {
swp_entry_t entry = { .val = page_private(subpage) };
pte_t swp_pte;
@@ -1465,7 +1464,7 @@ static bool try_to_unmap_one(struct page
* If the page was redirtied, it cannot be
* discarded. Remap the page to page table.
*/
- set_pte_at(mm, address, pvmw.pte, pteval);
+ set_pte_at(mm, pvmw.address, pvmw.pte, pteval);
SetPageSwapBacked(page);
ret = false;
page_vma_mapped_walk_done(&pvmw);
@@ -1473,7 +1472,7 @@ static bool try_to_unmap_one(struct page
}
if (swap_duplicate(entry) < 0) {
- set_pte_at(mm, address, pvmw.pte, pteval);
+ set_pte_at(mm, pvmw.address, pvmw.pte, pteval);
ret = false;
page_vma_mapped_walk_done(&pvmw);
break;
@@ -1489,14 +1488,18 @@ static bool try_to_unmap_one(struct page
swp_pte = swp_entry_to_pte(entry);
if (pte_soft_dirty(pteval))
swp_pte = pte_swp_mksoft_dirty(swp_pte);
- set_pte_at(mm, address, pvmw.pte, swp_pte);
+ set_pte_at(mm, pvmw.address, pvmw.pte, swp_pte);
} else
dec_mm_counter(mm, mm_counter_file(page));
discard:
page_remove_rmap(subpage, PageHuge(page));
put_page(page);
- mmu_notifier_invalidate_page(mm, address);
+ invalidation_needed = true;
}
+
+ if (invalidation_needed)
+ mmu_notifier_invalidate_range(mm, address,
+ address + (1UL << compound_order(page)));
return ret;
}
_
Patches currently in -mm which might be from jglisse@xxxxxxxxxx are
mm-rmap-try_to_unmap_one-do-not-call-mmu_notifier-under-ptl-v2.patch
lib-interval_tree-fast-overlap-detection-fix.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
