In register_mem_sect_under_node() the system_state's value is checked to detect whether the call is made during boot time or during an hot-plug operation. Unfortunately, that check against SYSTEM_BOOTING is wrong because regular memory is registered at SYSTEM_SCHEDULING state. In addition, memory hot-plug operation can be triggered at this system state by the ACPI [1]. So checking against the system state is not enough. The consequence is that on system with interleaved node's ranges like this: Early memory node ranges node 1: [mem 0x0000000000000000-0x000000011fffffff] node 2: [mem 0x0000000120000000-0x000000014fffffff] node 1: [mem 0x0000000150000000-0x00000001ffffffff] node 0: [mem 0x0000000200000000-0x000000048fffffff] node 2: [mem 0x0000000490000000-0x00000007ffffffff] This can be seen on PowerPC LPAR after multiple memory hot-plug and hot-unplug operations are done. At the next reboot the node's memory ranges can be interleaved and since the call to link_mem_sections() is made in topology_init() while the system is in the SYSTEM_SCHEDULING state, the node's id is not checked, and the sections registered to multiple nodes: $ ls -l /sys/devices/system/memory/memory21/node* total 0 lrwxrwxrwx 1 root root 0 Aug 24 05:27 node1 -> ../../node/node1 lrwxrwxrwx 1 root root 0 Aug 24 05:27 node2 -> ../../node/node2 In that case, the system is able to boot but if later one of theses memory blocks is hot-unplugged and then hot-plugged, the sysfs inconsistency is detected and this is triggering a BUG_ON(): kernel BUG at /Users/laurent/src/linux-ppc/mm/memory_hotplug.c:1084! Oops: Exception in kernel mode, sig: 5 [#1] LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries Modules linked in: rpadlpar_io rpaphp pseries_rng rng_core vmx_crypto gf128mul binfmt_misc ip_tables x_tables xfs libcrc32c crc32c_vpmsum autofs4 CPU: 8 PID: 10256 Comm: drmgr Not tainted 5.9.0-rc1+ #25 Call Trace: add_memory_resource+0x23c/0x340 (unreliable) __add_memory+0x5c/0xf0 dlpar_add_lmb+0x1b4/0x500 dlpar_memory+0x1f8/0xb80 handle_dlpar_errorlog+0xc0/0x190 dlpar_store+0x198/0x4a0 kobj_attr_store+0x30/0x50 sysfs_kf_write+0x64/0x90 kernfs_fop_write+0x1b0/0x290 vfs_write+0xe8/0x290 ksys_write+0xdc/0x130 system_call_exception+0x160/0x270 system_call_common+0xf0/0x27c This patch addresses the root cause by not relying on the system_state value to detect whether the call is due to a hot-plug operation. An extra parameter is added to link_mem_sections() detailing whether the operation is due to a hot-plug operation. [1] According to Oscar Salvador, using this qemu command line, ACPI memory hotplug operations are raised at SYSTEM_SCHEDULING state: $QEMU -enable-kvm -machine pc -smp 4,sockets=4,cores=1,threads=1 -cpu host -monitor pty \ -m size=$MEM,slots=255,maxmem=4294967296k \ -numa node,nodeid=0,cpus=0-3,mem=512 -numa node,nodeid=1,mem=512 \ -object memory-backend-ram,id=memdimm0,size=134217728 -device pc-dimm,node=0,memdev=memdimm0,id=dimm0,slot=0 \ -object memory-backend-ram,id=memdimm1,size=134217728 -device pc-dimm,node=0,memdev=memdimm1,id=dimm1,slot=1 \ -object memory-backend-ram,id=memdimm2,size=134217728 -device pc-dimm,node=0,memdev=memdimm2,id=dimm2,slot=2 \ -object memory-backend-ram,id=memdimm3,size=134217728 -device pc-dimm,node=0,memdev=memdimm3,id=dimm3,slot=3 \ -object memory-backend-ram,id=memdimm4,size=134217728 -device pc-dimm,node=1,memdev=memdimm4,id=dimm4,slot=4 \ -object memory-backend-ram,id=memdimm5,size=134217728 -device pc-dimm,node=1,memdev=memdimm5,id=dimm5,slot=5 \ -object memory-backend-ram,id=memdimm6,size=134217728 -device pc-dimm,node=1,memdev=memdimm6,id=dimm6,slot=6 \
Unified · Split
| @@ -761,14 +761,36 @@ static int __ref get_nid_for_pfn(unsigned long pfn) · 760 unmodified lines | ||
| 761 | 761 | return pfn_to_nid(pfn); |
| 762 | 762 | } |
| 763 | 763 | |
| 764 | +static int do_register_memory_block_under_node(int nid, | |
| 765 | + struct memory_block *mem_blk) | |
| 766 | +{ | |
| 767 | + int ret; | |
| 768 | + | |
| 769 | + /* | |
| 770 | + * If this memory block spans multiple nodes, we only indicate | |
| 771 | + * the last processed node. | |
| 772 | + */ | |
| 773 | + mem_blk->nid = nid; | |
| 774 | + | |
| 775 | + ret = sysfs_create_link_nowarn(&node_devices[nid]->dev.kobj, | |
| 776 | + &mem_blk->dev.kobj, | |
| 777 | + kobject_name(&mem_blk->dev.kobj)); | |
| 778 | + if (ret) | |
| 779 | + return ret; | |
| 780 | + | |
| 781 | + return sysfs_create_link_nowarn(&mem_blk->dev.kobj, | |
| 782 | + &node_devices[nid]->dev.kobj, | |
| 783 | + kobject_name(&node_devices[nid]->dev.kobj)); | |
| 784 | +} | |
| 785 | + | |
| 764 | 786 | /* register memory section under specified node if it spans that node */ |
| 765 | −static int register_mem_sect_under_node(struct memory_block *mem_blk, | |
| 766 | − void *arg) | |
| 787 | +static int register_mem_block_under_node_early(struct memory_block *mem_blk, | |
| 788 | + void *arg) | |
| 767 | 789 | { |
| 768 | 790 | unsigned long memory_block_pfns = memory_block_size_bytes() / PAGE_SIZE; |
| 769 | 791 | unsigned long start_pfn = section_nr_to_pfn(mem_blk->start_section_nr); |
| 770 | 792 | unsigned long end_pfn = start_pfn + memory_block_pfns - 1; |
| 771 | − int ret, nid = *(int *)arg; | |
| 793 | + int nid = *(int *)arg; | |
| 772 | 794 | unsigned long pfn; |
| 773 | 795 | |
| 774 | 796 | for (pfn = start_pfn; pfn <= end_pfn; pfn++) { |
| @@ -785,38 +807,33 @@ static int register_mem_sect_under_node(struct memory_block *mem_blk, · 10 unmodified lines | ||
| 785 | 807 | } |
| 786 | 808 | |
| 787 | 809 | /* |
| 788 | − * We need to check if page belongs to nid only for the boot | |
| 789 | − * case, during hotplug we know that all pages in the memory | |
| 790 | − * block belong to the same node. | |
| 791 | − */ | |
| 792 | − if (system_state == SYSTEM_BOOTING) { | |
| 793 | − page_nid = get_nid_for_pfn(pfn); | |
| 794 | − if (page_nid < 0) | |
| 795 | − continue; | |
| 796 | − if (page_nid != nid) | |
| 797 | − continue; | |
| 798 | − } | |
| 799 | − | |
| 800 | − /* | |
| 801 | − * If this memory block spans multiple nodes, we only indicate | |
| 802 | − * the last processed node. | |
| 810 | + * We need to check if page belongs to nid only at the boot | |
| 811 | + * case because node's ranges can be interleaved. | |
| 803 | 812 | */ |
| 804 | − mem_blk->nid = nid; | |
| 805 | − | |
| 806 | − ret = sysfs_create_link_nowarn(&node_devices[nid]->dev.kobj, | |
| 807 | − &mem_blk->dev.kobj, | |
| 808 | − kobject_name(&mem_blk->dev.kobj)); | |
| 809 | − if (ret) | |
| 810 | − return ret; | |
| 813 | + page_nid = get_nid_for_pfn(pfn); | |
| 814 | + if (page_nid < 0) | |
| 815 | + continue; | |
| 816 | + if (page_nid != nid) | |
| 817 | + continue; | |
| 811 | 818 | |
| 812 | − return sysfs_create_link_nowarn(&mem_blk->dev.kobj, | |
| 813 | − &node_devices[nid]->dev.kobj, | |
| 814 | − kobject_name(&node_devices[nid]->dev.kobj)); | |
| 819 | + return do_register_memory_block_under_node(nid, mem_blk); | |
| 815 | 820 | } |
| 816 | 821 | /* mem section does not span the specified node */ |
| 817 | 822 | return 0; |
| 818 | 823 | } |
| 819 | 824 | |
| 825 | +/* | |
| 826 | + * During hotplug we know that all pages in the memory block belong to the same | |
| 827 | + * node. | |
| 828 | + */ | |
| 829 | +static int register_mem_block_under_node_hotplug(struct memory_block *mem_blk, | |
| 830 | + void *arg) | |
| 831 | +{ | |
| 832 | + int nid = *(int *)arg; | |
| 833 | + | |
| 834 | + return do_register_memory_block_under_node(nid, mem_blk); | |
| 835 | +} | |
| 836 | + | |
| 820 | 837 | /* |
| 821 | 838 | * Unregister a memory block device under the node it spans. Memory blocks |
| 822 | 839 | * with multiple nodes cannot be offlined and therefore also never be removed. |
| @@ -832,11 +849,19 @@ void unregister_memory_block_under_nodes(struct memory_block *mem_blk) · 9 unmodified lines | ||
| 832 | 849 | kobject_name(&node_devices[mem_blk->nid]->dev.kobj)); |
| 833 | 850 | } |
| 834 | 851 | |
| 835 | −int link_mem_sections(int nid, unsigned long start_pfn, unsigned long end_pfn) | |
| 852 | +int link_mem_sections(int nid, unsigned long start_pfn, unsigned long end_pfn, | |
| 853 | + enum meminit_context context) | |
| 836 | 854 | { |
| 855 | + walk_memory_blocks_func_t func; | |
| 856 | + | |
| 857 | + if (context == MEMINIT_HOTPLUG) | |
| 858 | + func = register_mem_block_under_node_hotplug; | |
| 859 | + else | |
| 860 | + func = register_mem_block_under_node_early; | |
| 861 | + | |
| 837 | 862 | return walk_memory_blocks(PFN_PHYS(start_pfn), |
| 838 | 863 | PFN_PHYS(end_pfn - start_pfn), (void *)&nid, |
| 839 | − register_mem_sect_under_node); | |
| 864 | + func); | |
| 840 | 865 | } |
| 841 | 866 | |
| 842 | 867 | #ifdef CONFIG_HUGETLBFS |
| @@ -99,11 +99,13 @@ extern struct node *node_devices[]; · 98 unmodified lines | ||
| 99 | 99 | typedef void (*node_registration_func_t)(struct node *); |
| 100 | 100 | |
| 101 | 101 | #if defined(CONFIG_MEMORY_HOTPLUG_SPARSE) && defined(CONFIG_NUMA) |
| 102 | −extern int link_mem_sections(int nid, unsigned long start_pfn, | |
| 103 | − unsigned long end_pfn); | |
| 102 | +int link_mem_sections(int nid, unsigned long start_pfn, | |
| 103 | + unsigned long end_pfn, | |
| 104 | + enum meminit_context context); | |
| 104 | 105 | #else |
| 105 | 106 | static inline int link_mem_sections(int nid, unsigned long start_pfn, |
| 106 | − unsigned long end_pfn) | |
| 107 | + unsigned long end_pfn, | |
| 108 | + enum meminit_context context) | |
| 107 | 109 | { |
| 108 | 110 | return 0; |
| 109 | 111 | } |
| @@ -128,7 +130,8 @@ static inline int register_one_node(int nid) · 18 unmodified lines | ||
| 128 | 130 | if (error) |
| 129 | 131 | return error; |
| 130 | 132 | /* link memory sections under this node */ |
| 131 | − error = link_mem_sections(nid, start_pfn, end_pfn); | |
| 133 | + error = link_mem_sections(nid, start_pfn, end_pfn, | |
| 134 | + MEMINIT_EARLY); | |
| 132 | 135 | } |
| 133 | 136 | |
| 134 | 137 | return error; |
| @@ -1080,7 +1080,8 @@ int __ref add_memory_resource(int nid, struct resource *res) · 1079 unmodified lines | ||
| 1080 | 1080 | } |
| 1081 | 1081 | |
| 1082 | 1082 | /* link memory sections under this node.*/ |
| 1083 | − ret = link_mem_sections(nid, PFN_DOWN(start), PFN_UP(start + size - 1)); | |
| 1083 | + ret = link_mem_sections(nid, PFN_DOWN(start), PFN_UP(start + size - 1), | |
| 1084 | + MEMINIT_HOTPLUG); | |
| 1084 | 1085 | BUG_ON(ret); |
| 1085 | 1086 | |
| 1086 | 1087 | /* create new memmap entry */ |