--- /dev/null
+From mgorman@suse.de Tue Jan 7 09:51:39 2014
+From: Mel Gorman <mgorman@suse.de>
+Date: Tue, 7 Jan 2014 14:00:42 +0000
+Subject: mm: numa: avoid unnecessary work on the failure path
+To: gregkh@linuxfoundation.org
+Cc: athorlton@sgi.com, riel@redhat.com, chegu_vinod@hp.com, Mel Gorman <mgorman@suse.de>, stable@vger.kernel.org
+Message-ID: <1389103248-17617-8-git-send-email-mgorman@suse.de>
+
+From: Mel Gorman <mgorman@suse.de>
+
+commit eb4489f69f224356193364dc2762aa009738ca7f upstream.
+
+If a PMD changes during a THP migration then migration aborts but the
+failure path is doing more work than is necessary.
+
+Signed-off-by: Mel Gorman <mgorman@suse.de>
+Reviewed-by: Rik van Riel <riel@redhat.com>
+Cc: Alex Thorlton <athorlton@sgi.com>
+Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
+Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ mm/migrate.c | 4 +++-
+ 1 file changed, 3 insertions(+), 1 deletion(-)
+
+--- a/mm/migrate.c
++++ b/mm/migrate.c
+@@ -1726,7 +1726,8 @@ fail_putback:
+ putback_lru_page(page);
+ mod_zone_page_state(page_zone(page),
+ NR_ISOLATED_ANON + page_lru, -HPAGE_PMD_NR);
+- goto out_fail;
++
++ goto out_unlock;
+ }
+
+ /*
+@@ -1800,6 +1801,7 @@ out_dropref:
+ }
+ spin_unlock(&mm->page_table_lock);
+
++out_unlock:
+ unlock_page(page);
+ put_page(page);
+ return 0;
--- /dev/null
+From mgorman@suse.de Tue Jan 7 09:49:21 2014
+From: Mel Gorman <mgorman@suse.de>
+Date: Tue, 7 Jan 2014 14:00:39 +0000
+Subject: mm: numa: do not clear PMD during PTE update scan
+To: gregkh@linuxfoundation.org
+Cc: athorlton@sgi.com, riel@redhat.com, chegu_vinod@hp.com, Mel Gorman <mgorman@suse.de>, stable@vger.kernel.org
+Message-ID: <1389103248-17617-5-git-send-email-mgorman@suse.de>
+
+From: Mel Gorman <mgorman@suse.de>
+
+commit 5a6dac3ec5f583cc8ee7bc53b5500a207c4ca433 upstream.
+
+If the PMD is flushed then a parallel fault in handle_mm_fault() will
+enter the pmd_none and do_huge_pmd_anonymous_page() path where it'll
+attempt to insert a huge zero page. This is wasteful so the patch
+avoids clearing the PMD when setting pmd_numa.
+
+Signed-off-by: Mel Gorman <mgorman@suse.de>
+Reviewed-by: Rik van Riel <riel@redhat.com>
+Cc: Alex Thorlton <athorlton@sgi.com>
+Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
+Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ mm/huge_memory.c | 6 ++++--
+ 1 file changed, 4 insertions(+), 2 deletions(-)
+
+--- a/mm/huge_memory.c
++++ b/mm/huge_memory.c
+@@ -1474,20 +1474,22 @@ int change_huge_pmd(struct vm_area_struc
+
+ if (__pmd_trans_huge_lock(pmd, vma) == 1) {
+ pmd_t entry;
+- entry = pmdp_get_and_clear(mm, addr, pmd);
+ if (!prot_numa) {
++ entry = pmdp_get_and_clear(mm, addr, pmd);
+ entry = pmd_modify(entry, newprot);
+ BUG_ON(pmd_write(entry));
++ set_pmd_at(mm, addr, pmd, entry);
+ } else {
+ struct page *page = pmd_page(*pmd);
++ entry = *pmd;
+
+ /* only check non-shared pages */
+ if (page_mapcount(page) == 1 &&
+ !pmd_numa(*pmd)) {
+ entry = pmd_mknuma(entry);
++ set_pmd_at(mm, addr, pmd, entry);
+ }
+ }
+- set_pmd_at(mm, addr, pmd, entry);
+ spin_unlock(&vma->vm_mm->page_table_lock);
+ ret = 1;
+ }
--- /dev/null
+From mgorman@suse.de Tue Jan 7 09:50:13 2014
+From: Mel Gorman <mgorman@suse.de>
+Date: Tue, 7 Jan 2014 14:00:40 +0000
+Subject: mm: numa: do not clear PTE for pte_numa update
+To: gregkh@linuxfoundation.org
+Cc: athorlton@sgi.com, riel@redhat.com, chegu_vinod@hp.com, Mel Gorman <mgorman@suse.de>, stable@vger.kernel.org
+Message-ID: <1389103248-17617-6-git-send-email-mgorman@suse.de>
+
+From: Mel Gorman <mgorman@suse.de>
+
+commit 0c5f83c23ca703d32f930393825487257a5cde6d upstream.
+
+The TLB must be flushed if the PTE is updated but change_pte_range is
+clearing the PTE while marking PTEs pte_numa without necessarily
+flushing the TLB if it reinserts the same entry. Without the flush,
+it's conceivable that two processors have different TLBs for the same
+virtual address and at the very least it would generate spurious faults.
+
+This patch only unmaps the pages in change_pte_range for a full
+protection change.
+
+[riel@redhat.com: write pte_numa pte back to the page tables]
+Signed-off-by: Mel Gorman <mgorman@suse.de>
+Signed-off-by: Rik van Riel <riel@redhat.com>
+Reviewed-by: Rik van Riel <riel@redhat.com>
+Cc: Alex Thorlton <athorlton@sgi.com>
+Cc: Chegu Vinod <chegu_vinod@hp.com>
+Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
+Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ mm/mprotect.c | 9 +++++++--
+ 1 file changed, 7 insertions(+), 2 deletions(-)
+
+--- a/mm/mprotect.c
++++ b/mm/mprotect.c
+@@ -54,13 +54,14 @@ static unsigned long change_pte_range(st
+ pte_t ptent;
+ bool updated = false;
+
+- ptent = ptep_modify_prot_start(mm, addr, pte);
+ if (!prot_numa) {
++ ptent = ptep_modify_prot_start(mm, addr, pte);
+ ptent = pte_modify(ptent, newprot);
+ updated = true;
+ } else {
+ struct page *page;
+
++ ptent = *pte;
+ page = vm_normal_page(vma, addr, oldpte);
+ if (page) {
+ int this_nid = page_to_nid(page);
+@@ -73,6 +74,7 @@ static unsigned long change_pte_range(st
+ if (!pte_numa(oldpte) &&
+ page_mapcount(page) == 1) {
+ ptent = pte_mknuma(ptent);
++ set_pte_at(mm, addr, pte, ptent);
+ updated = true;
+ }
+ }
+@@ -89,7 +91,10 @@ static unsigned long change_pte_range(st
+
+ if (updated)
+ pages++;
+- ptep_modify_prot_commit(mm, addr, pte, ptent);
++
++ /* Only !prot_numa always clears the pte */
++ if (!prot_numa)
++ ptep_modify_prot_commit(mm, addr, pte, ptent);
+ } else if (IS_ENABLED(CONFIG_MIGRATION) && !pte_file(oldpte)) {
+ swp_entry_t entry = pte_to_swp_entry(oldpte);
+
--- /dev/null
+From mgorman@suse.de Tue Jan 7 09:51:09 2014
+From: Mel Gorman <mgorman@suse.de>
+Date: Tue, 7 Jan 2014 14:00:41 +0000
+Subject: mm: numa: ensure anon_vma is locked to prevent parallel THP splits
+To: gregkh@linuxfoundation.org
+Cc: athorlton@sgi.com, riel@redhat.com, chegu_vinod@hp.com, Mel Gorman <mgorman@suse.de>, stable@vger.kernel.org
+Message-ID: <1389103248-17617-7-git-send-email-mgorman@suse.de>
+
+From: Mel Gorman <mgorman@suse.de>
+
+commit c3a489cac38d43ea6dc4ac240473b44b46deecf7 upstream.
+
+The anon_vma lock prevents parallel THP splits and any associated
+complexity that arises when handling splits during THP migration. This
+patch checks if the lock was successfully acquired and bails from THP
+migration if it failed for any reason.
+
+Signed-off-by: Mel Gorman <mgorman@suse.de>
+Reviewed-by: Rik van Riel <riel@redhat.com>
+Cc: Alex Thorlton <athorlton@sgi.com>
+Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
+Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ mm/huge_memory.c | 7 +++++++
+ 1 file changed, 7 insertions(+)
+
+--- a/mm/huge_memory.c
++++ b/mm/huge_memory.c
+@@ -1342,6 +1342,13 @@ int do_huge_pmd_numa_page(struct mm_stru
+ goto out_unlock;
+ }
+
++ /* Bail if we fail to protect against THP splits for any reason */
++ if (unlikely(!anon_vma)) {
++ put_page(page);
++ page_nid = -1;
++ goto clear_pmdnuma;
++ }
++
+ /*
+ * Migrate the THP to the requested node, returns with page unlocked
+ * and pmd_numa cleared.
mm-numa-serialise-parallel-get_user_page-against-thp-migration.patch
mm-numa-call-mmu-notifiers-on-thp-migration.patch
mm-clear-pmd_numa-before-invalidating.patch
+mm-numa-do-not-clear-pmd-during-pte-update-scan.patch
+mm-numa-do-not-clear-pte-for-pte_numa-update.patch
+mm-numa-ensure-anon_vma-is-locked-to-prevent-parallel-thp-splits.patch
+mm-numa-avoid-unnecessary-work-on-the-failure-path.patch
mm-mempolicy-correct-putback-method-for-isolate-pages-if-failed.patch
mm-compaction-respect-ignore_skip_hint-in-update_pageblock_skip.patch
mm-memory-failure.c-recheck-pagehuge-after-hugetlb-page-migrate-successfully.patch
projects / thirdparty / kernel / stable-queue.git / commitdiff
