mm: convert totalram_pages and totalhigh_pages variables to atomic

totalram_pages and totalhigh_pages are made static inline function.

Main motivation was that managed_page_count_lock handling was complicating
things.  It was discussed in length here,
https://lore.kernel.org/patchwork/patch/995739/#1181785 So it seemes
better to remove the lock and convert variables to atomic, with preventing
poteintial store-to-read tearing as a bonus.

Link: http://lkml.kernel.org/r/1542090790-21750-4-git-send-email-arunks@codeaurora.org
Signed-off-by: Arun KS <arunks@codeaurora.org>
Suggested-by: Michal Hocko <mhocko@suse.com>
Suggested-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

arch/csky/mm/init.c

@@ -71,7 +71,7 @@ void free_initrd_mem(unsigned long start, unsigned long end)
 		ClearPageReserved(virt_to_page(start));
 		init_page_count(virt_to_page(start));
 		free_page(start);
-		totalram_pages++;
+		totalram_pages_inc();
 	}
 }
 #endif
@@ -88,7 +88,7 @@ void free_initmem(void)
 		ClearPageReserved(virt_to_page(addr));
 		init_page_count(virt_to_page(addr));
 		free_page(addr);
-		totalram_pages++;
+		totalram_pages_inc();
 		addr += PAGE_SIZE;
 	}
 

arch/powerpc/platforms/pseries/cmm.c

@@ -208,7 +208,7 @@ static long cmm_alloc_pages(long nr)
 
 		pa->page[pa->index++] = addr;
 		loaned_pages++;
-		totalram_pages--;
+		totalram_pages_dec();
 		spin_unlock(&cmm_lock);
 		nr--;
 	}
@@ -247,7 +247,7 @@ static long cmm_free_pages(long nr)
 		free_page(addr);
 		loaned_pages--;
 		nr--;
-		totalram_pages++;
+		totalram_pages_inc();
 	}
 	spin_unlock(&cmm_lock);
 	cmm_dbg("End request with %ld pages unfulfilled\n", nr);
@@ -291,7 +291,7 @@ static void cmm_get_mpp(void)
 	int rc;
 	struct hvcall_mpp_data mpp_data;
 	signed long active_pages_target, page_loan_request, target;
-	signed long total_pages = totalram_pages + loaned_pages;
+	signed long total_pages = totalram_pages() + loaned_pages;
 	signed long min_mem_pages = (min_mem_mb * 1024 * 1024) / PAGE_SIZE;
 
 	rc = h_get_mpp(&mpp_data);
@@ -322,7 +322,7 @@ static void cmm_get_mpp(void)
 
 	cmm_dbg("delta = %ld, loaned = %lu, target = %lu, oom = %lu, totalram = %lu\n",
 		page_loan_request, loaned_pages, loaned_pages_target,
-		oom_freed_pages, totalram_pages);
+		oom_freed_pages, totalram_pages());
 }
 
 static struct notifier_block cmm_oom_nb = {
@@ -581,7 +581,7 @@ static int cmm_mem_going_offline(void *arg)
 			free_page(pa_curr->page[idx]);
 			freed++;
 			loaned_pages--;
-			totalram_pages++;
+			totalram_pages_inc();
 			pa_curr->page[idx] = pa_last->page[--pa_last->index];
 			if (pa_last->index == 0) {
 				if (pa_curr == pa_last)

arch/s390/mm/init.c

@@ -59,7 +59,7 @@ static void __init setup_zero_pages(void)
 	order = 7;
 
 	/* Limit number of empty zero pages for small memory sizes */
-	while (order > 2 && (totalram_pages >> 10) < (1UL << order))
+	while (order > 2 && (totalram_pages() >> 10) < (1UL << order))
 		order--;
 
 	empty_zero_page = __get_free_pages(GFP_KERNEL | __GFP_ZERO, order);

arch/um/kernel/mem.c

@@ -51,7 +51,7 @@ void __init mem_init(void)
 
 	/* this will put all low memory onto the freelists */
 	memblock_free_all();
-	max_low_pfn = totalram_pages;
+	max_low_pfn = totalram_pages();
 	max_pfn = max_low_pfn;
 	mem_init_print_info(NULL);
 	kmalloc_ok = 1;

arch/x86/kernel/cpu/microcode/core.c

@@ -434,7 +434,7 @@ static ssize_t microcode_write(struct file *file, const char __user *buf,
 			       size_t len, loff_t *ppos)
 {
 	ssize_t ret = -EINVAL;
-	unsigned long nr_pages = totalram_pages;
+	unsigned long nr_pages = totalram_pages();
 
 	if ((len >> PAGE_SHIFT) > nr_pages) {
 		pr_err("too much data (max %ld pages)\n", nr_pages);

drivers/char/agp/backend.c

@@ -115,9 +115,9 @@ static int agp_find_max(void)
 	long memory, index, result;
 
 #if PAGE_SHIFT < 20
-	memory = totalram_pages >> (20 - PAGE_SHIFT);
+	memory = totalram_pages() >> (20 - PAGE_SHIFT);
 #else
-	memory = totalram_pages << (PAGE_SHIFT - 20);
+	memory = totalram_pages() << (PAGE_SHIFT - 20);
 #endif
 	index = 1;
 

drivers/gpu/drm/i915/i915_gem.c

@@ -2539,7 +2539,7 @@ static int i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	 * If there's no chance of allocating enough pages for the whole
 	 * object, bail early.
 	 */
-	if (page_count > totalram_pages)
+	if (page_count > totalram_pages())
 		return -ENOMEM;
 
 	st = kmalloc(sizeof(*st), GFP_KERNEL);

drivers/gpu/drm/i915/selftests/i915_gem_gtt.c

@@ -170,7 +170,7 @@ static int igt_ppgtt_alloc(void *arg)
 	 * This should ensure that we do not run into the oomkiller during
 	 * the test and take down the machine wilfully.
 	 */
-	limit = totalram_pages << PAGE_SHIFT;
+	limit = totalram_pages() << PAGE_SHIFT;
 	limit = min(ppgtt->vm.total, limit);
 
 	/* Check we can allocate the entire range */
@@ -1244,7 +1244,7 @@ static int exercise_mock(struct drm_i915_private *i915,
 				     u64 hole_start, u64 hole_end,
 				     unsigned long end_time))
 {
-	const u64 limit = totalram_pages << PAGE_SHIFT;
+	const u64 limit = totalram_pages() << PAGE_SHIFT;
 	struct i915_gem_context *ctx;
 	struct i915_hw_ppgtt *ppgtt;
 	IGT_TIMEOUT(end_time);

drivers/hv/hv_balloon.c

@@ -1090,7 +1090,7 @@ static void process_info(struct hv_dynmem_device *dm, struct dm_info_msg *msg)
 static unsigned long compute_balloon_floor(void)
 {
 	unsigned long min_pages;
-	unsigned long nr_pages = totalram_pages;
+	unsigned long nr_pages = totalram_pages();
 #define MB2PAGES(mb) ((mb) << (20 - PAGE_SHIFT))
 	/* Simple continuous piecewiese linear function:
 	 *  max MiB -> min MiB  gradient

drivers/md/dm-bufio.c

@@ -1887,7 +1887,7 @@ static int __init dm_bufio_init(void)
 	dm_bufio_allocated_vmalloc = 0;
 	dm_bufio_current_allocated = 0;
 
-	mem = (__u64)mult_frac(totalram_pages - totalhigh_pages,
+	mem = (__u64)mult_frac(totalram_pages() - totalhigh_pages(),
 			       DM_BUFIO_MEMORY_PERCENT, 100) << PAGE_SHIFT;
 
 	if (mem > ULONG_MAX)

drivers/md/dm-crypt.c

@@ -2158,7 +2158,7 @@ static int crypt_wipe_key(struct crypt_config *cc)
 
 static void crypt_calculate_pages_per_client(void)
 {
-	unsigned long pages = (totalram_pages - totalhigh_pages) * DM_CRYPT_MEMORY_PERCENT / 100;
+	unsigned long pages = (totalram_pages() - totalhigh_pages()) * DM_CRYPT_MEMORY_PERCENT / 100;
 
 	if (!dm_crypt_clients_n)
 		return;

drivers/md/dm-integrity.c

@@ -2843,7 +2843,7 @@ static int create_journal(struct dm_integrity_c *ic, char **error)
 	journal_pages = roundup((__u64)ic->journal_sections * ic->journal_section_sectors,
 				PAGE_SIZE >> SECTOR_SHIFT) >> (PAGE_SHIFT - SECTOR_SHIFT);
 	journal_desc_size = journal_pages * sizeof(struct page_list);
-	if (journal_pages >= totalram_pages - totalhigh_pages || journal_desc_size > ULONG_MAX) {
+	if (journal_pages >= totalram_pages() - totalhigh_pages() || journal_desc_size > ULONG_MAX) {
 		*error = "Journal doesn't fit into memory";
 		r = -ENOMEM;
 		goto bad;

drivers/md/dm-stats.c

@@ -85,7 +85,7 @@ static bool __check_shared_memory(size_t alloc_size)
 	a = shared_memory_amount + alloc_size;
 	if (a < shared_memory_amount)
 		return false;
-	if (a >> PAGE_SHIFT > totalram_pages / DM_STATS_MEMORY_FACTOR)
+	if (a >> PAGE_SHIFT > totalram_pages() / DM_STATS_MEMORY_FACTOR)
 		return false;
 #ifdef CONFIG_MMU
 	if (a > (VMALLOC_END - VMALLOC_START) / DM_STATS_VMALLOC_FACTOR)

drivers/media/platform/mtk-vpu/mtk_vpu.c

@@ -855,7 +855,7 @@ static int mtk_vpu_probe(struct platform_device *pdev)
 	/* Set PTCM to 96K and DTCM to 32K */
 	vpu_cfg_writel(vpu, 0x2, VPU_TCM_CFG);
 
-	vpu->enable_4GB = !!(totalram_pages > (SZ_2G >> PAGE_SHIFT));
+	vpu->enable_4GB = !!(totalram_pages() > (SZ_2G >> PAGE_SHIFT));
 	dev_info(dev, "4GB mode %u\n", vpu->enable_4GB);
 
 	if (vpu->enable_4GB) {

drivers/misc/vmw_balloon.c

@@ -570,7 +570,7 @@ static int vmballoon_send_get_target(struct vmballoon *b)
 	unsigned long status;
 	unsigned long limit;
 
-	limit = totalram_pages;
+	limit = totalram_pages();
 
 	/* Ensure limit fits in 32-bits */
 	if (limit != (u32)limit)

drivers/parisc/ccio-dma.c

@@ -1251,7 +1251,7 @@ ccio_ioc_init(struct ioc *ioc)
 	** Hot-Plug/Removal of PCI cards. (aka PCI OLARD).
 	*/
 
-	iova_space_size = (u32) (totalram_pages / count_parisc_driver(&ccio_driver));
+	iova_space_size = (u32) (totalram_pages() / count_parisc_driver(&ccio_driver));
 
 	/* limit IOVA space size to 1MB-1GB */
 
@@ -1290,7 +1290,7 @@ ccio_ioc_init(struct ioc *ioc)
 
 	DBG_INIT("%s() hpa 0x%p mem %luMB IOV %dMB (%d bits)\n",
 			__func__, ioc->ioc_regs,
-			(unsigned long) totalram_pages >> (20 - PAGE_SHIFT),
+			(unsigned long) totalram_pages() >> (20 - PAGE_SHIFT),
 			iova_space_size>>20,
 			iov_order + PAGE_SHIFT);
 

drivers/parisc/sba_iommu.c

@@ -1414,7 +1414,7 @@ sba_ioc_init(struct parisc_device *sba, struct ioc *ioc, int ioc_num)
 	** for DMA hints - ergo only 30 bits max.
 	*/
 
-	iova_space_size = (u32) (totalram_pages/global_ioc_cnt);
+	iova_space_size = (u32) (totalram_pages()/global_ioc_cnt);
 
 	/* limit IOVA space size to 1MB-1GB */
 	if (iova_space_size < (1 << (20 - PAGE_SHIFT))) {
@@ -1439,7 +1439,7 @@ sba_ioc_init(struct parisc_device *sba, struct ioc *ioc, int ioc_num)
 	DBG_INIT("%s() hpa 0x%lx mem %ldMB IOV %dMB (%d bits)\n",
 			__func__,
 			ioc->ioc_hpa,
-			(unsigned long) totalram_pages >> (20 - PAGE_SHIFT),
+			(unsigned long) totalram_pages() >> (20 - PAGE_SHIFT),
 			iova_space_size>>20,
 			iov_order + PAGE_SHIFT);
 

drivers/staging/android/ion/ion_system_heap.c

@@ -110,7 +110,7 @@ static int ion_system_heap_allocate(struct ion_heap *heap,
 	unsigned long size_remaining = PAGE_ALIGN(size);
 	unsigned int max_order = orders[0];
 
-	if (size / PAGE_SIZE > totalram_pages / 2)
+	if (size / PAGE_SIZE > totalram_pages() / 2)
 		return -ENOMEM;
 
 	INIT_LIST_HEAD(&pages);

drivers/xen/xen-selfballoon.c

@@ -189,7 +189,7 @@ static void selfballoon_process(struct work_struct *work)
 	bool reset_timer = false;
 
 	if (xen_selfballooning_enabled) {
-		cur_pages = totalram_pages;
+		cur_pages = totalram_pages();
 		tgt_pages = cur_pages; /* default is no change */
 		goal_pages = vm_memory_committed() +
 				totalreserve_pages +
@@ -227,7 +227,7 @@ static void selfballoon_process(struct work_struct *work)
 		if (tgt_pages < floor_pages)
 			tgt_pages = floor_pages;
 		balloon_set_new_target(tgt_pages +
-			balloon_stats.current_pages - totalram_pages);
+			balloon_stats.current_pages - totalram_pages());
 		reset_timer = true;
 	}
 #ifdef CONFIG_FRONTSWAP
@@ -569,7 +569,7 @@ int xen_selfballoon_init(bool use_selfballooning, bool use_frontswap_selfshrink)
 	 * much more reliably and response faster in some cases.
 	 */
 	if (!selfballoon_reserved_mb) {
-		reserve_pages = totalram_pages / 10;
+		reserve_pages = totalram_pages() / 10;
 		selfballoon_reserved_mb = PAGES2MB(reserve_pages);
 	}
 	schedule_delayed_work(&selfballoon_worker, selfballoon_interval * HZ);

fs/ceph/super.h

@@ -810,7 +810,7 @@ static inline int default_congestion_kb(void)
 	 * This allows larger machines to have larger/more transfers.
 	 * Limit the default to 256M
 	 */
-	congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10);
+	congestion_kb = (16*int_sqrt(totalram_pages())) << (PAGE_SHIFT-10);
 	if (congestion_kb > 256*1024)
 		congestion_kb = 256*1024;
 

fs/file_table.c

@@ -380,7 +380,7 @@ void __init files_init(void)
 void __init files_maxfiles_init(void)
 {
 	unsigned long n;
-	unsigned long nr_pages = totalram_pages;
+	unsigned long nr_pages = totalram_pages();
 	unsigned long memreserve = (nr_pages - nr_free_pages()) * 3/2;
 
 	memreserve = min(memreserve, nr_pages - 1);

fs/fuse/inode.c

@@ -824,7 +824,7 @@ static const struct super_operations fuse_super_operations = {
 static void sanitize_global_limit(unsigned *limit)
 {
 	if (*limit == 0)
-		*limit = ((totalram_pages << PAGE_SHIFT) >> 13) /
+		*limit = ((totalram_pages() << PAGE_SHIFT) >> 13) /
 			 sizeof(struct fuse_req);
 
 	if (*limit >= 1 << 16)

fs/nfs/write.c

@@ -2121,7 +2121,7 @@ int __init nfs_init_writepagecache(void)
 	 * This allows larger machines to have larger/more transfers.
 	 * Limit the default to 256M
 	 */
-	nfs_congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10);
+	nfs_congestion_kb = (16*int_sqrt(totalram_pages())) << (PAGE_SHIFT-10);
 	if (nfs_congestion_kb > 256*1024)
 		nfs_congestion_kb = 256*1024;
 

fs/nfsd/nfscache.c

@@ -99,7 +99,7 @@ static unsigned int
 nfsd_cache_size_limit(void)
 {
 	unsigned int limit;
-	unsigned long low_pages = totalram_pages - totalhigh_pages;
+	unsigned long low_pages = totalram_pages() - totalhigh_pages();
 
 	limit = (16 * int_sqrt(low_pages)) << (PAGE_SHIFT-10);
 	return min_t(unsigned int, limit, 256*1024);

fs/ntfs/malloc.h

@@ -47,7 +47,7 @@ static inline void *__ntfs_malloc(unsigned long size, gfp_t gfp_mask)
 		return kmalloc(PAGE_SIZE, gfp_mask & ~__GFP_HIGHMEM);
 		/* return (void *)__get_free_page(gfp_mask); */
 	}
-	if (likely((size >> PAGE_SHIFT) < totalram_pages))
+	if (likely((size >> PAGE_SHIFT) < totalram_pages()))
 		return __vmalloc(size, gfp_mask, PAGE_KERNEL);
 	return NULL;
 }

fs/proc/base.c

@@ -530,7 +530,7 @@ static const struct file_operations proc_lstats_operations = {
 static int proc_oom_score(struct seq_file *m, struct pid_namespace *ns,
 			  struct pid *pid, struct task_struct *task)
 {
-	unsigned long totalpages = totalram_pages + total_swap_pages;
+	unsigned long totalpages = totalram_pages() + total_swap_pages;
 	unsigned long points = 0;
 
 	points = oom_badness(task, NULL, NULL, totalpages) *

include/linux/highmem.h

@@ -36,7 +36,31 @@ static inline void invalidate_kernel_vmap_range(void *vaddr, int size)
 
 /* declarations for linux/mm/highmem.c */
 unsigned int nr_free_highpages(void);
-extern unsigned long totalhigh_pages;
+extern atomic_long_t _totalhigh_pages;
+static inline unsigned long totalhigh_pages(void)
+{
+	return (unsigned long)atomic_long_read(&_totalhigh_pages);
+}
+
+static inline void totalhigh_pages_inc(void)
+{
+       atomic_long_inc(&_totalhigh_pages);
+}
+
+static inline void totalhigh_pages_dec(void)
+{
+       atomic_long_dec(&_totalhigh_pages);
+}
+
+static inline void totalhigh_pages_add(long count)
+{
+       atomic_long_add(count, &_totalhigh_pages);
+}
+
+static inline void totalhigh_pages_set(long val)
+{
+       atomic_long_set(&_totalhigh_pages, val);
+}
 
 void kmap_flush_unused(void);
 
@@ -51,7 +75,7 @@ static inline struct page *kmap_to_page(void *addr)
 	return virt_to_page(addr);
 }
 
-#define totalhigh_pages 0UL
+static inline unsigned long totalhigh_pages(void) { return 0UL; }
 
 #ifndef ARCH_HAS_KMAP
 static inline void *kmap(struct page *page)