Merge branch 'akpm' (patches from Andrew)

Merge fixes from Andrew Morton: "20 fixes" * emailed patches from Andrew Morton <akpm@linux-foundation.org>: Documentation/sysctl/vm.txt: update numa_zonelist_order description lib/stackdepot.c: allow the stack trace hash to be zero rapidio: fix potential NULL pointer dereference mm/memory-failure: fix race with compound page split/merge ocfs2/dlm: return zero if deref_done message is successfully handled Ananth has moved kcov: don't profile branches in kcov kcov: don't trace the code coverage code mm: wake kcompactd before kswapd's short sleep .mailmap: add Frank Rowand mm/hwpoison: fix wrong num_poisoned_pages accounting mm: call swap_slot_free_notify() with page lock held mm: vmscan: reclaim highmem zone if buffer_heads is over limit numa: fix /proc/<pid>/numa_maps for THP mm/huge_memory: replace VM_NO_THP VM_BUG_ON with actual VMA check mailmap: fix Krzysztof Kozlowski's misspelled name thp: keep huge zero page pinned until tlb flush mm: exclude HugeTLB pages from THP page_mapped() logic kexec: export OFFSET(page.compound_head) to find out compound tail page kexec: update VMCOREINFO for compound_order/dtor

Linus Torvalds committed 10 years ago

Commit 1d003af2 parents 92c19ea 7c88a2918 files · +151 −47Browse files

Unified · Split

18 files

.mailmap +4 −0
Documentation/sysctl/vm.txt +10 −9
MAINTAINERS +1 −1
drivers/rapidio/devices/rio_mport_cdev.c +2 −2
fs/ocfs2/dlm/dlmmaster.c +2 −0
fs/proc/task_mmu.c +30 −3
include/linux/huge_mm.h +5 −0
include/linux/mm.h +4 −0
kernel/kcov.c +2 −1
kernel/kexec_core.c +5 −2
lib/stackdepot.c +0 −4
mm/huge_memory.c +5 −7
mm/memory-failure.c +9 −1
mm/memory.c +40 −0
mm/migrate.c +7 −1
mm/page_io.c +5 −1
mm/swap.c +5 −0
mm/vmscan.c +15 −15

.mailmap+4−0

 Felix Moeller <felix@derklecks.de>
 Filipe Lautert <filipe@icewall.org>
 Franck Bui-Huu <vagabon.xyz@gmail.com>
+Frank Rowand <frowand.list@gmail.com> <frowand@mvista.com>
+Frank Rowand <frowand.list@gmail.com> <frank.rowand@am.sony.com>
+Frank Rowand <frowand.list@gmail.com> <frank.rowand@sonymobile.com>
 Frank Zago <fzago@systemfabricworks.com>
 Greg Kroah-Hartman <greg@echidna.(none)>
 Greg Kroah-Hartman <gregkh@suse.de>
 Kenneth W Chen <kenneth.w.chen@intel.com>
 Konstantin Khlebnikov <koct9i@gmail.com> <k.khlebnikov@samsung.com>
 Koushik <raghavendra.koushik@neterion.com>
+Krzysztof Kozlowski <krzk@kernel.org> <k.kozlowski.k@gmail.com>
 Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
 Leonid I Ananiev <leonid.i.ananiev@intel.com>
 Linas Vepstas <linas@austin.ibm.com>

Documentation/sysctl/vm.txt+10−9

@@ -581,15 +581,16 @@ Specify "[Nn]ode" for node order · 580 unmodified lines
581	"Zone Order" orders the zonelists by zone type, then by node within each	581	"Zone Order" orders the zonelists by zone type, then by node within each
582	zone. Specify "[Zz]one" for zone order.	582	zone. Specify "[Zz]one" for zone order.
583		583
584	Specify "[Dd]efault" to request automatic configuration. Autoconfiguration	584	Specify "[Dd]efault" to request automatic configuration.
585	will select "node" order in following case.	585
586	(1) if the DMA zone does not exist or	586	On 32-bit, the Normal zone needs to be preserved for allocations accessible
587	(2) if the DMA zone comprises greater than 50% of the available memory or	587	by the kernel, so "zone" order will be selected.
588	(3) if any node's DMA zone comprises greater than 70% of its local memory and	588
589	the amount of local memory is big enough.	589	On 64-bit, devices that require DMA32/DMA are relatively rare, so "node"
590		590	order will be selected.
591	Otherwise, "zone" order will be selected. Default order is recommended unless	591
592	this is causing problems for your system/application.	592	Default order is recommended unless this is causing problems for your
		593	system/application.
593		594
594	==============================================================	595	==============================================================
595		596

MAINTAINERS+1−1

@@ -6400,7 +6400,7 @@ F: mm/kmemleak.c · 6399 unmodified lines
6400	F: mm/kmemleak-test.c	6400	F: mm/kmemleak-test.c
6401		6401
6402	KPROBES	6402	KPROBES
6403	M: Ananth N Mavinakayanahalli <ananth@in.ibm.com>	6403	M: Ananth N Mavinakayanahalli <ananth@linux.vnet.ibm.com>
6404	M: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>	6404	M: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
6405	M: "David S. Miller" <davem@davemloft.net>	6405	M: "David S. Miller" <davem@davemloft.net>
6406	M: Masami Hiramatsu <mhiramat@kernel.org>	6406	M: Masami Hiramatsu <mhiramat@kernel.org>

drivers/rapidio/devices/rio_mport_cdev.c+2−2

@@ -2669,9 +2669,9 @@ static int __init mport_init(void) · 2668 unmodified lines
2669		2669
2670	/* Create device class needed by udev */	2670	/* Create device class needed by udev */
2671	dev_class = class_create(THIS_MODULE, DRV_NAME);	2671	dev_class = class_create(THIS_MODULE, DRV_NAME);
2672	if (!dev_class) {	2672	if (IS_ERR(dev_class)) {
2673	rmcd_error("Unable to create " DRV_NAME " class");	2673	rmcd_error("Unable to create " DRV_NAME " class");
2674	return -EINVAL;	2674	return PTR_ERR(dev_class);
2675	}	2675	}
2676		2676
2677	ret = alloc_chrdev_region(&dev_number, 0, RIO_MAX_MPORTS, DRV_NAME);	2677	ret = alloc_chrdev_region(&dev_number, 0, RIO_MAX_MPORTS, DRV_NAME);

fs/ocfs2/dlm/dlmmaster.c+2−0

@@ -2455,6 +2455,8 @@ int dlm_deref_lockres_done_handler(struct o2net_msg msg, u32 len, void data, · 2454 unmodified lines
2455		2455
2456	spin_unlock(&dlm->spinlock);	2456	spin_unlock(&dlm->spinlock);
2457		2457
		2458	ret = 0;
		2459
2458	done:	2460	done:
2459	dlm_put(dlm);	2461	dlm_put(dlm);
2460	return ret;	2462	return ret;

fs/proc/task_mmu.c+30−3

@@ -1518,6 +1518,32 @@ static struct page can_gather_numa_stats(pte_t pte, struct vm_area_struct vma, · 1517 unmodified lines
1518	return page;	1518	return page;
1519	}	1519	}
1520		1520
		1521	#ifdef CONFIG_TRANSPARENT_HUGEPAGE
		1522	static struct page *can_gather_numa_stats_pmd(pmd_t pmd,
		1523	struct vm_area_struct *vma,
		1524	unsigned long addr)
		1525	{
		1526	struct page *page;
		1527	int nid;
		1528
		1529	if (!pmd_present(pmd))
		1530	return NULL;
		1531
		1532	page = vm_normal_page_pmd(vma, addr, pmd);
		1533	if (!page)
		1534	return NULL;
		1535
		1536	if (PageReserved(page))
		1537	return NULL;
		1538
		1539	nid = page_to_nid(page);
		1540	if (!node_isset(nid, node_states[N_MEMORY]))
		1541	return NULL;
		1542
		1543	return page;
		1544	}
		1545	#endif
		1546
1521	static int gather_pte_stats(pmd_t *pmd, unsigned long addr,	1547	static int gather_pte_stats(pmd_t *pmd, unsigned long addr,
1522	unsigned long end, struct mm_walk *walk)	1548	unsigned long end, struct mm_walk *walk)
1523	{	1549	{
@@ -1527,14 +1553,14 @@ static int gather_pte_stats(pmd_t *pmd, unsigned long addr, · 3 unmodified lines
1527	pte_t *orig_pte;	1553	pte_t *orig_pte;
1528	pte_t *pte;	1554	pte_t *pte;
1529		1555
		1556	#ifdef CONFIG_TRANSPARENT_HUGEPAGE
1530	ptl = pmd_trans_huge_lock(pmd, vma);	1557	ptl = pmd_trans_huge_lock(pmd, vma);
1531	if (ptl) {	1558	if (ptl) {
1532	pte_t huge_pte = (pte_t )pmd;
1533	struct page *page;	1559	struct page *page;
1534		1560
1535	page = can_gather_numa_stats(huge_pte, vma, addr);	1561	page = can_gather_numa_stats_pmd(*pmd, vma, addr);
1536	if (page)	1562	if (page)
1537	gather_stats(page, md, pte_dirty(huge_pte),	1563	gather_stats(page, md, pmd_dirty(*pmd),
1538	HPAGE_PMD_SIZE/PAGE_SIZE);	1564	HPAGE_PMD_SIZE/PAGE_SIZE);
1539	spin_unlock(ptl);	1565	spin_unlock(ptl);
1540	return 0;	1566	return 0;
@@ -1542,6 +1568,7 @@ static int gather_pte_stats(pmd_t *pmd, unsigned long addr, · 1 unmodified line
1542		1568
1543	if (pmd_trans_unstable(pmd))	1569	if (pmd_trans_unstable(pmd))
1544	return 0;	1570	return 0;
		1571	#endif
1545	orig_pte = pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);	1572	orig_pte = pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
1546	do {	1573	do {
1547	struct page page = can_gather_numa_stats(pte, vma, addr);	1574	struct page page = can_gather_numa_stats(pte, vma, addr);

include/linux/huge_mm.h+5−0

 }
 struct page *get_huge_zero_page(void);
+void put_huge_zero_page(void);
 #else /* CONFIG_TRANSPARENT_HUGEPAGE */
 #define HPAGE_PMD_SHIFT ({ BUILD_BUG(); 0; })
 	return false;
 }
+static inline void put_huge_zero_page(void)
+{
+	BUILD_BUG();
+}
 static inline struct page *follow_devmap_pmd(struct vm_area_struct *vma,
 		unsigned long addr, pmd_t *pmd, int flags)

include/linux/mm.h+4−0

 	page = compound_head(page);
 	if (atomic_read(compound_mapcount_ptr(page)) >= 0)
 		return true;
+	if (PageHuge(page))
+		return false;
 	for (i = 0; i < hpage_nr_pages(page); i++) {
 		if (atomic_read(&page[i]._mapcount) >= 0)
 			return true;
 struct page *vm_normal_page(struct vm_area_struct *vma, unsigned long addr,
 		pte_t pte);
+struct page *vm_normal_page_pmd(struct vm_area_struct *vma, unsigned long addr,
+				pmd_t pmd);
 int zap_vma_ptes(struct vm_area_struct *vma, unsigned long address,
 		unsigned long size);

kernel/kcov.c+2−1

 #define pr_fmt(fmt) "kcov: " fmt
+#define DISABLE_BRANCH_PROFILING
 #include <linux/compiler.h>
 #include <linux/types.h>
 #include <linux/file.h>
  * Entry point from instrumented code.
  * This is called once per basic-block/edge.
  */
-void __sanitizer_cov_trace_pc(void)
+void notrace __sanitizer_cov_trace_pc(void)
 {
 	struct task_struct *t;
 	enum kcov_mode mode;

kernel/kexec_core.c+5−2

 	VMCOREINFO_OFFSET(page, lru);
 	VMCOREINFO_OFFSET(page, _mapcount);
 	VMCOREINFO_OFFSET(page, private);
+	VMCOREINFO_OFFSET(page, compound_dtor);
+	VMCOREINFO_OFFSET(page, compound_order);
+	VMCOREINFO_OFFSET(page, compound_head);
 	VMCOREINFO_OFFSET(pglist_data, node_zones);
 	VMCOREINFO_OFFSET(pglist_data, nr_zones);
 #ifdef CONFIG_FLAT_NODE_MEM_MAP
 #ifdef CONFIG_X86
 	VMCOREINFO_NUMBER(KERNEL_IMAGE_SIZE);
 #endif
-#ifdef CONFIG_HUGETLBFS
+#ifdef CONFIG_HUGETLB_PAGE
-	VMCOREINFO_SYMBOL(free_huge_page);
+	VMCOREINFO_NUMBER(HUGETLB_PAGE_DTOR);
 #endif
 	arch_crash_save_vmcoreinfo();

lib/stackdepot.c+0−4

@@ -210,10 +210,6 @@ depot_stack_handle_t depot_save_stack(struct stack_trace *trace, · 209 unmodified lines
210	goto fast_exit;	210	goto fast_exit;
211		211
212	hash = hash_stack(trace->entries, trace->nr_entries);	212	hash = hash_stack(trace->entries, trace->nr_entries);
213	/* Bad luck, we won't store this stack. */
214	if (hash == 0)
215	goto exit;
216
217	bucket = &stack_table[hash & STACK_HASH_MASK];	213	bucket = &stack_table[hash & STACK_HASH_MASK];
218		214
219	/*	215	/*

mm/huge_memory.c+5−7

 	return READ_ONCE(huge_zero_page);
 }
-static void put_huge_zero_page(void)
+void put_huge_zero_page(void)
 {
 	/*
 	 * Counter should never go to zero here. Only shrinker can put
 	if (vma_is_dax(vma)) {
 		spin_unlock(ptl);
 		if (is_huge_zero_pmd(orig_pmd))
-			put_huge_zero_page();
+			tlb_remove_page(tlb, pmd_page(orig_pmd));
 	} else if (is_huge_zero_pmd(orig_pmd)) {
 		pte_free(tlb->mm, pgtable_trans_huge_withdraw(tlb->mm, pmd));
 		atomic_long_dec(&tlb->mm->nr_ptes);
 		spin_unlock(ptl);
-		put_huge_zero_page();
+		tlb_remove_page(tlb, pmd_page(orig_pmd));
 	} else {
 		struct page *page = pmd_page(orig_pmd);
 		page_remove_rmap(page, true);
 		 * page fault if needed.
 		 */
 		return 0;
-	if (vma->vm_ops)
+	if (vma->vm_ops || (vm_flags & VM_NO_THP))
 		/* khugepaged not yet working on file or special mappings */
 		return 0;
-	VM_BUG_ON_VMA(vm_flags & VM_NO_THP, vma);
 	hstart = (vma->vm_start + ~HPAGE_PMD_MASK) & HPAGE_PMD_MASK;
 	hend = vma->vm_end & HPAGE_PMD_MASK;
 	if (hstart < hend)
 		return false;
 	if (is_vma_temporary_stack(vma))
 		return false;
-	VM_BUG_ON_VMA(vma->vm_flags & VM_NO_THP, vma);
+	return !(vma->vm_flags & VM_NO_THP);
-	return true;
 }
 static void collapse_huge_page(struct mm_struct *mm,

mm/memory-failure.c+9−1

@@ -888,7 +888,15 @@ int get_hwpoison_page(struct page *page) · 887 unmodified lines
888	}	888	}
889	}	889	}
890		890
891	return get_page_unless_zero(head);	891	if (get_page_unless_zero(head)) {
		892	if (head == compound_head(page))
		893	return 1;
		894
		895	pr_info("MCE: %#lx cannot catch tail\n", page_to_pfn(page));
		896	put_page(head);
		897	}
		898
		899	return 0;
892	}	900	}
893	EXPORT_SYMBOL_GPL(get_hwpoison_page);	901	EXPORT_SYMBOL_GPL(get_hwpoison_page);
894		902

mm/memory.c+40−0

@@ -789,6 +789,46 @@ out: · 788 unmodified lines
789	return pfn_to_page(pfn);	789	return pfn_to_page(pfn);
790	}	790	}
791		791
		792	#ifdef CONFIG_TRANSPARENT_HUGEPAGE
		793	struct page vm_normal_page_pmd(struct vm_area_struct vma, unsigned long addr,
		794	pmd_t pmd)
		795	{
		796	unsigned long pfn = pmd_pfn(pmd);
		797
		798	/*
		799	* There is no pmd_special() but there may be special pmds, e.g.
		800	* in a direct-access (dax) mapping, so let's just replicate the
		801	* !HAVE_PTE_SPECIAL case from vm_normal_page() here.
		802	*/
		803	if (unlikely(vma->vm_flags & (VM_PFNMAP\|VM_MIXEDMAP))) {
		804	if (vma->vm_flags & VM_MIXEDMAP) {
		805	if (!pfn_valid(pfn))
		806	return NULL;
		807	goto out;
		808	} else {
		809	unsigned long off;
		810	off = (addr - vma->vm_start) >> PAGE_SHIFT;
		811	if (pfn == vma->vm_pgoff + off)
		812	return NULL;
		813	if (!is_cow_mapping(vma->vm_flags))
		814	return NULL;
		815	}
		816	}
		817
		818	if (is_zero_pfn(pfn))
		819	return NULL;
		820	if (unlikely(pfn > highest_memmap_pfn))
		821	return NULL;
		822
		823	/*
		824	* NOTE! We still have PageReserved() pages in the page tables.
		825	* eg. VDSO mappings can cause them to exist.
		826	*/
		827	out:
		828	return pfn_to_page(pfn);
		829	}
		830	#endif
		831
792	/*	832	/*
793	* copy one vm_area from one task to the other. Assumes the page tables	833	* copy one vm_area from one task to the other. Assumes the page tables
794	* already present in the new task to be cleared in the whole range	834	* already present in the new task to be cleared in the whole range

mm/migrate.c+7−1

 		dec_zone_page_state(page, NR_ISOLATED_ANON +
 				page_is_file_cache(page));
 		/* Soft-offlined page shouldn't go through lru cache list */
-		if (reason == MR_MEMORY_FAILURE) {
+		if (reason == MR_MEMORY_FAILURE && rc == MIGRATEPAGE_SUCCESS) {
+			/*
+			 * With this release, we free successfully migrated
+			 * page and set PG_HWPoison on just freed page
+			 * intentionally. Although it's rather weird, it's how
+			 * HWPoison flag works at the moment.
+			 */
 			put_page(page);
 			if (!test_set_page_hwpoison(page))
 				num_poisoned_pages_inc();

mm/page_io.c+5−1

 	ret = bdev_read_page(sis->bdev, swap_page_sector(page), page);
 	if (!ret) {
-		swap_slot_free_notify(page);
+		if (trylock_page(page)) {
+			swap_slot_free_notify(page);
+			unlock_page(page);
+		}
 		count_vm_event(PSWPIN);
 		return 0;
 	}

mm/swap.c+5−0

 			zone = NULL;
 		}
+		if (is_huge_zero_page(page)) {
+			put_huge_zero_page();
+			continue;
+		}
 		page = compound_head(page);
 		if (!put_page_testzero(page))
 			continue;

mm/vmscan.c+15−15

 		sc->gfp_mask |= __GFP_HIGHMEM;
 	for_each_zone_zonelist_nodemask(zone, z, zonelist,
-					requested_highidx, sc->nodemask) {
+					gfp_zone(sc->gfp_mask), sc->nodemask) {
 		enum zone_type classzone_idx;
 		if (!populated_zone(zone))
 	/* Try to sleep for a short interval */
 	if (prepare_kswapd_sleep(pgdat, order, remaining,
 						balanced_classzone_idx)) {
+		/*
+		 * Compaction records what page blocks it recently failed to
+		 * isolate pages from and skips them in the future scanning.
+		 * When kswapd is going to sleep, it is reasonable to assume
+		 * that pages and compaction may succeed so reset the cache.
+		 */
+		reset_isolation_suitable(pgdat);
+		/*
+		 * We have freed the memory, now we should compact it to make
+		 * allocation of the requested order possible.
+		 */
+		wakeup_kcompactd(pgdat, order, classzone_idx);
 		remaining = schedule_timeout(HZ/10);
 		finish_wait(&pgdat->kswapd_wait, &wait);
 		prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
 		 */
 		set_pgdat_percpu_threshold(pgdat, calculate_normal_threshold);
-		/*
-		 * Compaction records what page blocks it recently failed to
-		 * isolate pages from and skips them in the future scanning.
-		 * When kswapd is going to sleep, it is reasonable to assume
-		 * that pages and compaction may succeed so reset the cache.
-		 */
-		reset_isolation_suitable(pgdat);
-		/*
-		 * We have freed the memory, now we should compact it to make
-		 * allocation of the requested order possible.
-		 */
-		wakeup_kcompactd(pgdat, order, classzone_idx);
 		if (!kthread_should_stop())
 			schedule();

Merge branch 'akpm' (patches from Andrew)

Keyboard shortcuts

General

Issues and pull requests

Issues, pull requests and Inbox

Repository code