mm: pass nid to reserve_bootmem_region()

Commit Message

Yajun Deng June 15, 2023, 11:03 a.m. UTC

early_pfn_to_nid() is called frequently in init_reserved_page(),
it is used to convert PFN to node id. These PFN may belong to the same
memory region, they have the same node id. It is not necessary to call
early_pfn_to_nid() for each PFN.

Pass nid to reserve_bootmem_region() and drop the call to
early_pfn_to_nid() in init_reserved_page().

The following data was tested on my machine, it has 190GB RAM,
and enable CONFIG_DEFERRED_STRUCT_PAGE_INIT.

before the patch:
memmap_init_reserved_pages()  67ms

after the patch:
memmap_init_reserved_pages()  20ms

Signed-off-by: Yajun Deng <[e-mail]>
---
 include/linux/mm.h |  3 ++-
 mm/memblock.c      |  9 ++++++---
 mm/mm_init.c       | 31 +++++++++++++++++++------------
 3 files changed, 27 insertions(+), 16 deletions(-)

Comments

kernel test robot June 15, 2023, 5:42 p.m. UTC | #1

Hi Yajun,

kernel test robot noticed the following build errors:

[auto build test ERROR on akpm-mm/mm-everything]

url:    https://github.com/intel-lab-lkp/linux/commits/Yajun-Deng/mm-pass-nid-to-reserve_bootmem_region/20230615-190631
base:   https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
patch link:    https://lore.kernel.org/r/20230615110355.1694461-1-yajun.deng%40linux.dev
patch subject: [PATCH] mm: pass nid to reserve_bootmem_region()
config: arm-randconfig-r023-20230614 (https://download.01.org/0day-ci/archive/20230616/[e-mail]/config)
compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
reproduce (this is a W=1 build):
        mkdir -p ~/bin
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # install arm cross compiling tool for clang build
        # apt-get install binutils-arm-linux-gnueabi
        git remote add akpm-mm https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git
        git fetch akpm-mm mm-everything
        git checkout akpm-mm/mm-everything
        b4 shazam https://lore.kernel.org/r/[e-mail]
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang ~/bin/make.cross W=1 O=build_dir ARCH=arm olddefconfig
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang ~/bin/make.cross W=1 O=build_dir ARCH=arm SHELL=/bin/bash

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[e-mail]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[e-mail]/

All errors (new ones prefixed by >>):

>> mm/memblock.c:2098:18: error: no member named 'nid' in 'struct memblock_region'
                           nid = region->nid;
                                 ~~~~~~  ^
   1 error generated.


vim +2098 mm/memblock.c

  2080	
  2081	static void __init memmap_init_reserved_pages(void)
  2082	{
  2083		struct memblock_region *region;
  2084		phys_addr_t start, end;
  2085		int nid;
  2086		u64 i;
  2087	
  2088		/* initialize struct pages for the reserved regions */
  2089		__for_each_mem_range(i, &memblock.reserved, NULL, NUMA_NO_NODE,
  2090				     MEMBLOCK_NONE, &start, &end, &nid)
  2091			reserve_bootmem_region(start, end, nid);
  2092	
  2093		/* and also treat struct pages for the NOMAP regions as PageReserved */
  2094		for_each_mem_region(region) {
  2095			if (memblock_is_nomap(region)) {
  2096				start = region->base;
  2097				end = start + region->size;
> 2098				nid = region->nid;
  2099				reserve_bootmem_region(start, end, nid);
  2100			}
  2101		}
  2102	}
  2103

Patch

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 17317b1673b0..39e72ca6bf22 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2964,7 +2964,8 @@  extern unsigned long free_reserved_area(void *start, void *end,
 
 extern void adjust_managed_page_count(struct page *page, long count);
 
-extern void reserve_bootmem_region(phys_addr_t start, phys_addr_t end);
+extern void reserve_bootmem_region(phys_addr_t start,
+				   phys_addr_t end, int nid);
 
 /* Free the reserved page into the buddy system, so it gets managed. */
 static inline void free_reserved_page(struct page *page)
diff --git a/mm/memblock.c b/mm/memblock.c
index ff0da1858778..9e9b9a3473f9 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c
@@ -2091,18 +2091,21 @@  static void __init memmap_init_reserved_pages(void)
 {
 	struct memblock_region *region;
 	phys_addr_t start, end;
+	int nid;
 	u64 i;
 
 	/* initialize struct pages for the reserved regions */
-	for_each_reserved_mem_range(i, &start, &end)
-		reserve_bootmem_region(start, end);
+	__for_each_mem_range(i, &memblock.reserved, NULL, NUMA_NO_NODE,
+			     MEMBLOCK_NONE, &start, &end, &nid)
+		reserve_bootmem_region(start, end, nid);
 
 	/* and also treat struct pages for the NOMAP regions as PageReserved */
 	for_each_mem_region(region) {
 		if (memblock_is_nomap(region)) {
 			start = region->base;
 			end = start + region->size;
-			reserve_bootmem_region(start, end);
+			nid = region->nid;
+			reserve_bootmem_region(start, end, nid);
 		}
 	}
 }
diff --git a/mm/mm_init.c b/mm/mm_init.c
index d393631599a7..1499efbebc6f 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -646,10 +646,8 @@  static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
 }
 
 /* Returns true if the struct page for the pfn is initialised */
-static inline bool __meminit early_page_initialised(unsigned long pfn)
+static inline bool __meminit early_page_initialised(unsigned long pfn, int nid)
 {
-	int nid = early_pfn_to_nid(pfn);
-
 	if (node_online(nid) && pfn >= NODE_DATA(nid)->first_deferred_pfn)
 		return false;
 
@@ -695,15 +693,14 @@  defer_init(int nid, unsigned long pfn, unsigned long end_pfn)
 	return false;
 }
 
-static void __meminit init_reserved_page(unsigned long pfn)
+static void __meminit init_reserved_page(unsigned long pfn, int nid)
 {
 	pg_data_t *pgdat;
-	int nid, zid;
+	int zid;
 
-	if (early_page_initialised(pfn))
+	if (early_page_initialised(pfn, nid))
 		return;
 
-	nid = early_pfn_to_nid(pfn);
 	pgdat = NODE_DATA(nid);
 
 	for (zid = 0; zid < MAX_NR_ZONES; zid++) {
@@ -717,7 +714,7 @@  static void __meminit init_reserved_page(unsigned long pfn)
 #else
 static inline void pgdat_set_deferred_range(pg_data_t *pgdat) {}
 
-static inline bool early_page_initialised(unsigned long pfn)
+static inline bool early_page_initialised(unsigned long pfn, int nid)
 {
 	return true;
 }
@@ -727,7 +724,7 @@  static inline bool defer_init(int nid, unsigned long pfn, unsigned long end_pfn)
 	return false;
 }
 
-static inline void init_reserved_page(unsigned long pfn)
+static inline void init_reserved_page(unsigned long pfn, int nid)
 {
 }
 #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
@@ -738,16 +735,20 @@  static inline void init_reserved_page(unsigned long pfn)
  * marks the pages PageReserved. The remaining valid pages are later
  * sent to the buddy page allocator.
  */
-void __meminit reserve_bootmem_region(phys_addr_t start, phys_addr_t end)
+void __meminit reserve_bootmem_region(phys_addr_t start,
+				      phys_addr_t end, int nid)
 {
 	unsigned long start_pfn = PFN_DOWN(start);
 	unsigned long end_pfn = PFN_UP(end);
 
+	if (nid == MAX_NUMNODES)
+		nid = first_online_node;
+
 	for (; start_pfn < end_pfn; start_pfn++) {
 		if (pfn_valid(start_pfn)) {
 			struct page *page = pfn_to_page(start_pfn);
 
-			init_reserved_page(start_pfn);
+			init_reserved_page(start_pfn, nid);
 
 			/* Avoid false-positive PageTail() */
 			INIT_LIST_HEAD(&page->lru);
@@ -2579,7 +2580,13 @@  void __init set_dma_reserve(unsigned long new_dma_reserve)
 void __init memblock_free_pages(struct page *page, unsigned long pfn,
 							unsigned int order)
 {
-	if (!early_page_initialised(pfn))
+	int nid = 0;
+
+#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
+	nid = early_pfn_to_nid(pfn);
+#endif
+
+	if (!early_page_initialised(pfn, nid))
 		return;
 	if (!kmsan_memblock_free_pages(page, order)) {
 		/* KMSAN will take care of these pages. */