- 16 Jan, 2016 40 commits
-
-
Kirill A. Shutemov authored
We're going to allow mapping of individual 4k pages of THP compound. It means we need to track mapcount on per small page basis. Straight-forward approach is to use ->_mapcount in all subpages to track how many time this subpage is mapped with PMDs or PTEs combined. But this is rather expensive: mapping or unmapping of a THP page with PMD would require HPAGE_PMD_NR atomic operations instead of single we have now. The idea is to store separately how many times the page was mapped as whole -- compound_mapcount. This frees up ->_mapcount in subpages to track PTE mapcount. We use the same approach as with compound page destructor and compound order to store compound_mapcount: use space in first tail page, ->mapping this time. Any time we map/unmap whole compound page (THP or hugetlb) -- we increment/decrement compound_mapcount. When we map part of compound page with PTE we operate on ->_mapcount of the subpage. page_mapcount() counts both: PTE and PMD mappings of the page. Basically, we have mapcount for a subpage spread over two counters. It makes tricky to detect when last mapcount for a page goes away. We introduced PageDoubleMap() for this. When we split THP PMD for the first time and there's other PMD mapping left we offset up ->_mapcount in all subpages by one and set PG_double_map on the compound page. These additional references go away with last compound_mapcount. This approach provides a way to detect when last mapcount goes away on per small page basis without introducing new overhead for most common cases. [akpm@linux-foundation.org: fix typo in comment] [mhocko@suse.com: ignore partial THP when moving task] Signed-off-by:
Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by:
Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by:
Jerome Marchand <jmarchan@redhat.com> Cc: Sasha Levin <sasha.levin@oracle.com> Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Cc: Jerome Marchand <jmarchan@redhat.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by:
Michal Hocko <mhocko@suse.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. Signed-off-by:
Sasha Levin <sasha.levin@oracle.com> Tested-by:
Vlastimil Babka <vbabka@suse.cz> Acked-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. pmdp_splitting_flush() is not needed too: on splitting PMD we will do pmdp_clear_flush() + set_pte_at(). pmdp_clear_flush() will do IPI as needed for fast_gup. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. pmdp_splitting_flush() is not needed too: on splitting PMD we will do pmdp_clear_flush() + set_pte_at(). pmdp_clear_flush() will do IPI as needed for fast_gup. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. pmdp_splitting_flush() is not needed too: on splitting PMD we will do pmdp_clear_flush() + set_pte_at(). pmdp_clear_flush() will do IPI as needed for fast_gup. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. pmdp_splitting_flush() is not needed too: on splitting PMD we will do pmdp_clear_flush() + set_pte_at(). pmdp_clear_flush() will do IPI as needed for fast_gup. [arnd@arndb.de: fix unterminated ifdef in header file] Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we don't need to mark PMDs splitting. Let's drop code to handle this. pmdp_splitting_flush() is not needed too: on splitting PMD we will do pmdp_clear_flush() + set_pte_at(). pmdp_clear_flush() will do IPI as needed for fast_gup. Signed-off-by:
-
Kirill A. Shutemov authored
We are going to use migration entries to stabilize page counts. It means we don't need compound_lock() for that. Signed-off-by:
-
Kirill A. Shutemov authored
We don't need special code to stabilize THP. If you've got reference to any subpage of THP it will not be split under you. New split_huge_page() also accepts tail pages: no need in special code to get reference to head page. Signed-off-by:
-
Kirill A. Shutemov authored
With new THP refcounting, we don't need tricks to stabilize huge page. If we've got reference to tail page, it can't split under us. This patch effectively reverts a5b338f2 ("thp: update futex compound knowledge"). Signed-off-by:
Artem Savkov <artem.savkov@gmail.com> Signed-off-by:
-
Kirill A. Shutemov authored
Tail page refcounting is utterly complicated and painful to support. It uses ->_mapcount on tail pages to store how many times this page is pinned. get_page() bumps ->_mapcount on tail page in addition to ->_count on head. This information is required by split_huge_page() to be able to distribute pins from head of compound page to tails during the split. We will need ->_mapcount to account PTE mappings of subpages of the compound page. We eliminate need in current meaning of ->_mapcount in tail pages by forbidding split entirely if the page is pinned. The only user of tail page refcounting is THP which is marked BROKEN for now. Let's drop all this mess. It makes get_page() and put_page() much simpler. Signed-off-by:
-
Kirill A. Shutemov authored
We will re-introduce new version with new refcounting later in patchset. Signed-off-by:
-
Kirill A. Shutemov authored
Up to this point we tried to keep patchset bisectable, but next patches are going to change how core of THP refcounting work. It would be beneficial to split the change into several patches and make it more reviewable. Unfortunately, I don't see how we can achieve that while keeping THP working. Let's hide THP under CONFIG_BROKEN for now and bring it back when new refcounting get established. Signed-off-by:
-
Kirill A. Shutemov authored
The patch replaces THP_SPLIT with tree events: THP_SPLIT_PAGE, THP_SPLIT_PAGE_FAILED and THP_SPLIT_PMD. It reflects the fact that we are going to be able split PMD without the compound page and that split_huge_page() can fail. Signed-off-by:
Christoph Lameter <cl@linux.com> Tested-by:
-
Kirill A. Shutemov authored
We are going to decouple splitting THP PMD from splitting underlying compound page. This patch renames split_huge_page_pmd*() functions to split_huge_pmd*() to reflect the fact that it doesn't imply page splitting, only PMD. Signed-off-by:
-
Kirill A. Shutemov authored
Prepare khugepaged to see compound pages mapped with pte. For now we won't collapse the pmd table with such pte. khugepaged is subject for future rework wrt new refcounting. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting THP can belong to several VMAs. This makes tricky to track THP pages, when they partially mlocked. It can lead to leaking mlocked pages to non-VM_LOCKED vmas and other problems. With this patch we will split all pages on mlock and avoid fault-in/collapse new THP in VM_LOCKED vmas. I've tried alternative approach: do not mark THP pages mlocked and keep them on normal LRUs. This way vmscan could try to split huge pages on memory pressure and free up subpages which doesn't belong to VM_LOCKED vmas. But this is user-visible change: we screw up Mlocked accouting reported in meminfo, so I had to leave this approach aside. We can bring something better later, but this should be good enough for now. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we are going to see THP tail pages mapped with PTE. Generic fast GUP rely on page_cache_get_speculative() to obtain reference on page. page_cache_get_speculative() always fails on tail pages, because ->_count on tail pages is always zero. Let's handle tail pages in gup_pte_range(). New split_huge_page() will rely on migration entries to freeze page's counts. Recheck PTE value after page_cache_get_speculative() on head page should be enough to serialize against split. Signed-off-by:
-
Kirill A. Shutemov authored
We need to prepare kernel to allow transhuge pages to be mapped with ptes too. We need to handle FOLL_SPLIT in follow_page_pte(). Also we use split_huge_page() directly instead of split_huge_page_pmd(). split_huge_page_pmd() will gone. Signed-off-by:
-
Kirill A. Shutemov authored
With new refcounting we will be able map the same compound page with PTEs and PMDs. It requires adjustment to conditions when we can reuse the page on write-protection fault. For PTE fault we can't reuse the page if it's part of huge page. For PMD we can only reuse the page if nobody else maps the huge page or it's part. We can do it by checking page_mapcount() on each sub-page, but it's expensive. The cheaper way is to check page_count() to be equal 1: every mapcount takes page reference, so this way we can guarantee, that the PMD is the only mapping. This approach can give false negative if somebody pinned the page, but that doesn't affect correctness. Signed-off-by:
-
Kirill A. Shutemov authored
As with rmap, with new refcounting we cannot rely on PageTransHuge() to check if we need to charge size of huge page form the cgroup. We need to get information from caller to know whether it was mapped with PMD or PTE. We do uncharge when last reference on the page gone. At that point if we see PageTransHuge() it means we need to unchange whole huge page. The tricky part is partial unmap -- when we try to unmap part of huge page. We don't do a special handing of this situation, meaning we don't uncharge the part of huge page unless last user is gone or split_huge_page() is triggered. In case of cgroup memory pressure happens the partial unmapped page will be split through shrinker. This should be good enough. Signed-off-by:
-
Kirill A. Shutemov authored
We're going to allow mapping of individual 4k pages of THP compound page. It means we cannot rely on PageTransHuge() check to decide if map/unmap small page or THP. The patch adds new argument to rmap functions to indicate whether we want to operate on whole compound page or only the small page. [n-horiguchi@ah.jp.nec.com: fix mapcount mismatch in hugepage migration] Signed-off-by:
Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by:
-
Kirill A. Shutemov authored
The goal of this patchset is to make refcounting on THP pages cheaper with simpler semantics and allow the same THP compound page to be mapped with PMD and PTEs. This is required to get reasonable THP-pagecache implementation. With the new refcounting design it's much easier to protect against split_huge_page(): simple reference on a page will make you the deal. It makes gup_fast() implementation simpler and doesn't require special-case in futex code to handle tail THP pages. It should improve THP utilization over the system since splitting THP in one process doesn't necessary lead to splitting the page in all other processes have the page mapped. The patchset drastically lower complexity of get_page()/put_page() codepaths. I encourage people look on this code before-and-after to justify time budget on reviewing this patchset. This patch (of 37): With new refcounting all subpages of the compound page are not necessary have the same mapcount. We need to take into account mapcount of every sub-page. Signed-off-by:
-
Kirill A. Shutemov authored
Nobody uses them. Signed-off-by:
-
Kirill A. Shutemov authored
We don't define meaning of page->mapping for tail pages. Currently it's always NULL, which can be inconsistent with head page and potentially lead to problems. Let's poison the pointer to catch all illigal uses. page_rmapping(), page_mapping() and page_anon_vma() are changed to look on head page. The only illegal use I've caught so far is __GPF_COMP pages from sound subsystem, mapped with PTEs. do_shared_fault() is changed to use page_rmapping() instead of direct access to fault_page->mapping. Signed-off-by:
Jérôme Glisse <jglisse@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Christoph Lameter <cl@linux.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Jerome Marchand <jmarchan@redhat.com> Signed-off-by:
-
Kirill A. Shutemov authored
PageAnon() and PageKsm() look at lower bits of page->mapping to check if the page is Anon or KSM. page->mapping can be overloaded in tail pages. Let's always look at head page to avoid false-positives. Signed-off-by:
-
Kirill A. Shutemov authored
We use PG_uptodate on head pages on transparent huge page. Let's use PF_NO_TAIL. Signed-off-by:
-
Kirill A. Shutemov authored
So far, only IA64 uses PG_uncached and only on non-compound pages. Signed-off-by:
-
Kirill A. Shutemov authored
Transparent huge pages can be mlocked -- whole compund page at once. Something went wrong if we're trying to mlock() tail page. Let's use PF_NO_TAIL. Signed-off-by:
-
Kirill A. Shutemov authored
Swap cannot handle compound pages so far. Transparent huge pages are split on the way to swap. Signed-off-by:
-
Kirill A. Shutemov authored
PG_swapbacked is used for transparent huge pages. For head pages only. Let's use PF_NO_TAIL policy. Signed-off-by:
-
Kirill A. Shutemov authored
As far as I can see there's no users of PG_reserved on compound pages. Let's use PF_NO_COMPOUND here. Signed-off-by:
-
Kirill A. Shutemov authored
PG_pinned and PG_savepinned are about page table's pages which are never compound. I'm not so sure about PG_foreign, but it seems we shouldn't see compound pages there too. Let's use PF_NO_COMPOUND for all of them. Signed-off-by:
-
Kirill A. Shutemov authored
SL*B uses compound pages and marks head pages with PG_slab. __SetPageSlab() and __ClearPageSlab() are never called for tail pages. The same situation with PG_slob_free in SLOB allocator. PF_NO_TAIL is appropriate for these flags. Signed-off-by:
-
Kirill A. Shutemov authored
Only head pages are ever on LRU. Let's use PF_HEAD policy to avoid any confusion for all LRU-related flags. Signed-off-by:
-
Kirill A. Shutemov authored
It seems we don't have compound page on FS/IO path currently. Use PF_NO_COMPOUND to catch if we have. The odd exception is PG_dirty: sound uses compound pages and maps them with PTEs. PF_NO_COMPOUND triggers VM_BUG_ON() in set_page_dirty() on handling shared fault. Let's use PF_HEAD for PG_dirty. Signed-off-by:
-
Kirill A. Shutemov authored
lock_page() must operate on the whole compound page. It doesn't make much sense to lock part of compound page. Change code to use head page's PG_locked, if tail page is passed. This patch also gets rid of custom helper functions -- __set_page_locked() and __clear_page_locked(). They are replaced with helpers generated by __SETPAGEFLAG/__CLEARPAGEFLAG. Tail pages to these helper would trigger VM_BUG_ON(). SLUB uses PG_locked as a bit spin locked. IIUC, tail pages should never appear there. VM_BUG_ON() is added to make sure that this assumption is correct. [akpm@linux-foundation.org: fix fs/cifs/file.c] Signed-off-by:
-
