mm: introduce N_LRU_MEMORY to distinguish between normal and movable memory

From: Wu Jianguo <[email protected]>

Hi Christoph,
Now, We have node masks for both N_NORMAL_MEMORY and N_HIGH_MEMORY to
distinguish between normal and highmem on platforms such as x86. But we still don't have such a mechanism to distinguish between "normal" and "movable" memory. So this patch introduce N_LRU_MEMORY, and change the N_NORMAL_MEMORY definition.

N_NORMAL_MEMORY means non-LRU page allocs possible.
N_LRU_MEMORY means LRU page allocs possible,
node with ZONE_DMA(DMA32)/ZONE_NORMAL is marked with N_LRU_MEMORY and N_NORMAL_MEMORY, node with ZONE_MOVABLE is *only* marked with N_LRU_MEMORY,
node with ZONE_HIGHMEM is marked with N_LRU_MEMORY and N_HIGH_MEMORY. N_LRU_MEMORY also means node has any regular memory.

Thus, in SLUB allocator, we use for_each_node_state(node, N_NORMAL_MEMORY), if the node has only movable memory, then memory control structures will not need to be allocated for these nodes.
this patch will fix the bug described as follow:

When handling a memory node with only movable zone, function early_kmem_cache_node_alloc() will allocate a page from remote node but still increase object count on local node, which will trigger a BUG_ON() as below when hot-removing this memory node. Actually there's no need to create kmem_cache_node for memory node with only movable zone at all.

------------[ cut here ]------------
kernel BUG at mm/slub.c:3590!
invalid opcode: 0000 [#1] SMP
CPU 61
Modules linked in: autofs4 sunrpc cpufreq_ondemand acpi_cpufreq freq_table

mperf ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_filter ip6_tables ipv6 vfat fat dm_mirror dm_region_hash dm_log uinput iTCO_wdt iTCO_vendor_support coretemp hwmon kvm_intel kvm crc32c_intel ghash_clmulni_intel serio_raw pcspkr cdc_ether usbnet mii i2c_i801 i2c_core sg lpc_ich mfd_core shpchp ioatdma i7core_edac edac_core igb dca bnx2 ext4 mbcache jbd2 sr_mod cdrom sd_mod crc_t10dif aesni_intel cryptd aes_x86_64 aes_generic bfa scsi_transport_fc scsi_tgt pata_acpi ata_generic ata_piix megaraid_sas dm_mod [last unloaded: microcode]

Pid: 46287, comm: sh Not tainted 3.5.0-rc4-00215-g35f0828-dirty #85 IBM System x3850 X5 -[7143O3G]-/Node 1, Processor Card

RIP: 0010:[<ffffffff81160b2a>]  [<ffffffff81160b2a>]
slab_memory_callback+0x1ba/0x1c0
RSP: 0018:ffff880efdcb7c68  EFLAGS: 00010202
RAX: 0000000000000001 RBX: ffff880f7ec06100 RCX: 0000000100400001
RDX: 0000000100400002 RSI: ffff880f7ec02000 RDI: ffff880f7ec06100
RBP: ffff880efdcb7c78 R08: ffff88107b6fb098 R09: ffffffff81160a00
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000019
R13: 00000000fffffffb R14: 0000000000000000 R15: ffffffff81abe930
FS:  00007f709f342700(0000) GS:ffff880f7f3a0000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000003b5a874570 CR3: 0000000f0da20000 CR4: 00000000000007e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process sh (pid: 46287, threadinfo ffff880efdcb6000, task ffff880f0fa50000)
Stack:
 0000000000000004 ffff880efdcb7da8 ffff880efdcb7cb8 ffffffff81524af5
 0000000000000001 ffffffff81a8b620 ffffffff81a8b640 0000000000000004
 ffff880efdcb7da8 00000000ffffffff ffff880efdcb7d08 ffffffff8107a89a
Call Trace:
 [<ffffffff81524af5>] notifier_call_chain+0x55/0x80
 [<ffffffff8107a89a>] __blocking_notifier_call_chain+0x5a/0x80
 [<ffffffff8107a8d6>] blocking_notifier_call_chain+0x16/0x20
 [<ffffffff81352f0b>] memory_notify+0x1b/0x20
 [<ffffffff81507104>] offline_pages+0x624/0x700
 [<ffffffff811619de>] remove_memory+0x1e/0x20
 [<ffffffff813530cc>] memory_block_change_state+0x13c/0x2e0
 [<ffffffff81153e96>] ? alloc_pages_current+0xb6/0x120
 [<ffffffff81353332>] store_mem_state+0xc2/0xd0
 [<ffffffff8133e190>] dev_attr_store+0x20/0x30
 [<ffffffff811e2d4f>] sysfs_write_file+0xef/0x170
 [<ffffffff81173e28>] vfs_write+0xc8/0x190
 [<ffffffff81173ff1>] sys_write+0x51/0x90
 [<ffffffff81528d29>] system_call_fastpath+0x16/0x1b
Code: 8b 3d cb fd c4 00 be d0 00 00 00 e8 71 de ff ff 48 85 c0 75 9c 48 c7 c7

c0 7f a5 81 e8 c0 89 f1 ff b8 0d 80 00 00 e9 69 fe ff ff <0f> 0b eb fe 66 90 55 48 89 e5 41 57 41 56 41 55 41 54 53 48 83
RIP [<ffffffff81160b2a>] slab_memory_callback+0x1ba/0x1c0
RSP <ffff880efdcb7c68>

---[ end trace 749e9e9a67c78c12 ]---

---
 arch/alpha/mm/numa.c     |    2 +-
 arch/m32r/mm/discontig.c |    2 +-
 arch/m68k/mm/motorola.c  |    2 +-
 arch/parisc/mm/init.c    |    2 +-
 arch/tile/kernel/setup.c |    2 +-
 arch/x86/mm/init_64.c    |    2 +-
 drivers/base/node.c      |    4 +++-
 include/linux/nodemask.h |   12 ++++++++++--
 mm/page_alloc.c          |   14 ++++++++++++--
 9 files changed, 31 insertions(+), 11 deletions(-)

index 3973ae3..8402b29 100644
@@ -313,7 +313,7 @@ void __init paging_init(void)
 			zones_size[ZONE_DMA] = dma_local_pfn;
 			zones_size[ZONE_NORMAL] = (end_pfn - start_pfn) - dma_local_pfn;
 		}
-		node_set_state(nid, N_NORMAL_MEMORY);
+		node_set_state(nid, N_LRU_MEMORY);
 		free_area_init_node(nid, zones_size, start_pfn, NULL);
 	}
 

index 2c468e8..4d76e19 100644
@@ -149,7 +149,7 @@ unsigned long __init zone_sizes_init(void)
 		zholes_size[ZONE_DMA] = mp->holes;
 		holes += zholes_size[ZONE_DMA];
 
-		node_set_state(nid, N_NORMAL_MEMORY);
+		node_set_state(nid, N_LRU_MEMORY);
 		free_area_init_node(nid, zones_size, start_pfn, zholes_size);
 	}
 

index 0dafa69..31a0b00 100644
@@ -300,7 +300,7 @@ void __init paging_init(void)
 		free_area_init_node(i, zones_size,
 				    m68k_memory[i].addr >> PAGE_SHIFT, NULL);
 		if (node_present_pages(i))
-			node_set_state(i, N_NORMAL_MEMORY);
+			node_set_state(i, N_LRU_MEMORY);
 	}
 }
 

index 3ac462d..ad286fd 100644
@@ -277,7 +277,7 @@ static void __init setup_bootmem(void)
 	memset(pfnnid_map, 0xff, sizeof(pfnnid_map));
 
 	for (i = 0; i < npmem_ranges; i++) {
-		node_set_state(i, N_NORMAL_MEMORY);
+		node_set_state(i, N_LRU_MEMORY);
 		node_set_online(i);
 	}
 #endif

index 6a649a4..464672d 100644
@@ -748,7 +748,7 @@ static void __init zone_sizes_init(void)
 
 		/* Track the type of memory on each node */
 		if (zones_size[ZONE_NORMAL] || zones_size[ZONE_DMA])
-			node_set_state(i, N_NORMAL_MEMORY);
+			node_set_state(i, N_LRU_MEMORY);
 #ifdef CONFIG_HIGHMEM
 		if (end != start)
 			node_set_state(i, N_HIGH_MEMORY);

index 2b6b4a3..43bf392 100644
@@ -625,7 +625,7 @@ void __init paging_init(void)
 	 *	 numa support is not compiled in, and later node_set_state
 	 *	 will not set it back.
 	 */
-	node_clear_state(0, N_NORMAL_MEMORY);
+	node_clear_state(0, N_LRU_MEMORY);
 
 	zone_sizes_init();
 }

index af1a177..4a631f9 100644
@@ -617,6 +617,7 @@ static struct node_attr node_state_attr[] = {
 	_NODE_ATTR(possible, N_POSSIBLE),
 	_NODE_ATTR(online, N_ONLINE),
 	_NODE_ATTR(has_normal_memory, N_NORMAL_MEMORY),
+	_NODE_ATTR(has_lru_memory, N_LRU_MEMORY),
 	_NODE_ATTR(has_cpu, N_CPU),
 #ifdef CONFIG_HIGHMEM
 	_NODE_ATTR(has_high_memory, N_HIGH_MEMORY),

@@ -628,8 +629,9 @@ static struct attribute *node_state_attrs[] = {
 	&node_state_attr[1].attr.attr,
 	&node_state_attr[2].attr.attr,
 	&node_state_attr[3].attr.attr,
-#ifdef CONFIG_HIGHMEM
 	&node_state_attr[4].attr.attr,
+#ifdef CONFIG_HIGHMEM
+	&node_state_attr[5].attr.attr,
 #endif
 	NULL
 };

index 7afc363..92cee40 100644
@@ -370,15 +370,23 @@ static inline void __nodes_fold(nodemask_t *dstp, const nodemask_t *origp,
 
 /*
  * Bitmasks that are kept for all the nodes.
+ * N_NORMAL_MEMORY means non-LRU page allocs possible.
+ * N_LRU_MEMORY means LRU page allocs possible,
+ * node with ZONE_DMA/ZONE_DMA32/ZONE_NORMAL is marked with
+ * N_LRU_MEMORY and N_NORMAL_MEMORY,
+ * node with ZONE_MOVABLE is *only* marked with N_LRU_MEMORY,
+ * node with ZONE_HIGHMEM is marked with N_LRU_MEMORY and N_HIGH_MEMORY.
+ * N_LRU_MEMORY also means node has any regular memory.
  */
 enum node_states {
 	N_POSSIBLE,		/* The node could become online at some point */
 	N_ONLINE,		/* The node is online */
-	N_NORMAL_MEMORY,	/* The node has regular memory */
+	N_NORMAL_MEMORY,	/* The node has normal memory */
+	N_LRU_MEMORY,		/* The node has regular memory */
 #ifdef CONFIG_HIGHMEM
 	N_HIGH_MEMORY,		/* The node has regular or high memory */
 #else
-	N_HIGH_MEMORY = N_NORMAL_MEMORY,
+	N_HIGH_MEMORY = N_LRU_MEMORY,
 #endif
 	N_CPU,		/* The node has one or more cpus */
 	NR_NODE_STATES

index 009ac28..43261fc 100644
@@ -87,6 +87,7 @@ nodemask_t node_states[NR_NODE_STATES] __read_mostly = {
 	[N_ONLINE] = { { [0] = 1UL } },
 #ifndef CONFIG_NUMA
 	[N_NORMAL_MEMORY] = { { [0] = 1UL } },
+	[N_LRU_MEMORY] = { { [0] = 1UL } },
 #ifdef CONFIG_HIGHMEM
 	[N_HIGH_MEMORY] = { { [0] = 1UL } },
 #endif

@@ -4796,16 +4797,25 @@ out:
 /* Any regular memory on that node ? */
 static void __init check_for_regular_memory(pg_data_t *pgdat)
 {
-#ifdef CONFIG_HIGHMEM
+	struct zone *zone;
 	enum zone_type zone_type;
 
+	/*
+	 * node with ZONE_DMA/ZONE_DMA32/ZONE_NORMAL
+	 * is mark with N_NORMAL_MEMORY, means non-LUR page allocs possible.
+	 */
 	for (zone_type = 0; zone_type <= ZONE_NORMAL; zone_type++) {
-		struct zone *zone = &pgdat->node_zones[zone_type];
+		zone = &pgdat->node_zones[zone_type];
 		if (zone->present_pages) {
 			node_set_state(zone_to_nid(zone), N_NORMAL_MEMORY);
 			break;
 		}
 	}
+
+#ifdef CONFIG_HIGHMEM
+	zone = &pgdat->node_zones[ZONE_MOVABLE];
+	if (zone->present_pages)
+		node_set_state(zone_to_nid(zone), N_LRU_MEMORY);
 #endif
 }
 

1.7.6.1




.




--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Hmmm... It may be better to say

N_NORMAL_MEMORY Allocations are allowed for pages that will not be
managed via a LRU and that cannot be moved by the page migration logic.

N_LRU_MEMORY Allocations are possible for pages that are managed via LRUs

N_HIGH_MEMORY Allocations are allowed for pages that are only temporarliy mapped into kernel address space.

Any node that has the ability to allocate memory at all has at least N_LRU_MEMORY set.