Re: slab-out-of-bounds due to "KVM: Dynamically size memslot array based on number of used slots"

From: Sean Christopherson
Date: Fri Mar 20 2020 - 00:34:09 EST

Next message: Zenghui Yu: "Re: [PATCH v5 23/23] KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs"
Previous message: Vinay Simha B N: "Re: graph connection to node is not bidirectional kernel-5.6.0-rc6"
In reply to: Qian Cai: "slab-out-of-bounds due to "KVM: Dynamically size memslot array based on number of used slots""
Next in thread: Qian Cai: "Re: slab-out-of-bounds due to "KVM: Dynamically size memslot array based on number of used slots""
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Thu, Mar 19, 2020 at 11:59:23PM -0400, Qian Cai wrote:
> Reverted the linux-next commit 36947254e5f98 (âKVM: Dynamically size memslot array based on number of used slotsâ)
> fixed illegal slab object redzone accesses.
>
> [6727.939776][ T1818] BUG: KASAN: slab-out-of-bounds in gfn_to_hva+0xc1/0x2b0 [kvm]
> search_memslots at include/linux/kvm_host.h:1035

Drat. I'm guessing lru_slot is out of range after a memslot is deleted.
This should fix the issue, though it may not be the most proper fix, e.g.
it might be better to reset lru_slot when deleting a memslot. I'll try and
reproduce tomorrow, unless you can confirm this does the trick.

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index e9a7d3476a12..b8f5d3b2fe5c 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1021,6 +1021,12 @@ search_memslots(struct kvm_memslots *slots, gfn_t gfn)
int slot = atomic_read(&slots->lru_slot);
struct kvm_memory_slot *memslots = slots->memslots;

+ if (!slots->used_slots)
+ return NULL;
+
+ if (slot >= slots->used_slots)
+ slot = 0;
+
if (gfn >= memslots[slot].base_gfn &&
gfn < memslots[slot].base_gfn + memslots[slot].npages)
return &memslots[slot];

> (inlined by) __gfn_to_memslot at include/linux/kvm_host.h:1060
> (inlined by) gfn_to_memslot at virt/kvm/kvm_main.c:1594
> (inlined by) gfn_to_hva at virt/kvm/kvm_main.c:1674
> [ 6727.975376][ T1818] Read of size 8 at addr ffff8885351c4c10 by task qemu-system-x86/1818
> [ 6728.012769][ T1818]
> [ 6728.022690][ T1818] CPU: 11 PID: 1818 Comm: qemu-system-x86 Not tainted 5.6.0-rc6-next-20200319+ #5
> [ 6728.069063][ T1818] Hardware name: HP Synergy 480 Gen9/Synergy 480 Gen9 Compute Module, BIOS I37 10/21/2019
> [ 6728.117009][ T1818] Call Trace:
> [ 6728.131504][ T1818] dump_stack+0xa1/0xea
> [ 6728.149757][ T1818] print_address_description.constprop.0+0x33/0x270
> [ 6728.179315][ T1818] __kasan_report.cold+0x78/0xd0
> [ 6728.201386][ T1818] ? kvm_mmu_notifier_invalidate_range_start+0x50/0x150 [kvm]
> [ 6728.235146][ T1818] ? gfn_to_hva+0xc1/0x2b0 [kvm]
> [ 6728.256942][ T1818] kasan_report+0x41/0x60
> [ 6728.276172][ T1818] ? gfn_to_hva+0xc1/0x2b0 [kvm]
> [ 6728.298089][ T1818] __asan_load8+0x86/0xa0
> [ 6728.317366][ T1818] gfn_to_hva+0xc1/0x2b0 [kvm]
> [ 6728.338687][ T1818] kvm_arch_mmu_notifier_invalidate_range+0x21/0x50 [kvm]
> [ 6728.370595][ T1818] kvm_mmu_notifier_invalidate_range_start+0xfd/0x150 [kvm]
> [ 6728.402904][ T1818] mn_hlist_invalidate_range_start+0x109/0x2b0
> [ 6728.430526][ T1818] ? kvm_flush_remote_tlbs+0x170/0x170 [kvm]
> [ 6728.457312][ T1818] __mmu_notifier_invalidate_range_start+0x74/0x80
> [ 6728.486687][ T1818] wp_page_copy+0xd9f/0x12e0
> [ 6728.506872][ T1818] ? __lock_release+0x14d/0x2b0
> [ 6728.528484][ T1818] ? follow_pfn+0x160/0x160
> [ 6728.548555][ T1818] ? __kasan_check_read+0x11/0x20
> [ 6728.573773][ T1818] ? check_chain_key+0x1d3/0x2c0
> [ 6728.599226][ T1818] ? __kasan_check_read+0x11/0x20
> [ 6728.622648][ T1818] ? do_raw_spin_unlock+0x98/0xf0
> [ 6728.644878][ T1818] do_wp_page+0x37d/0xac0
> [ 6728.664110][ T1818] ? wp_page_shared+0x4e0/0x4e0
> [ 6728.685702][ T1818] handle_pte_fault+0x3b0/0x3e0
> [ 6728.707315][ T1818] __handle_mm_fault+0x5e7/0xaf0
> [ 6728.729495][ T1818] ? apply_to_existing_page_range+0x20/0x20
> [ 6728.755780][ T1818] handle_mm_fault+0x10b/0x320
> [ 6728.776840][ T1818] do_user_addr_fault+0x22c/0x5a0
> [ 6728.799089][ T1818] do_page_fault+0x9d/0x421
> [ 6728.818910][ T1818] page_fault+0x34/0x40
> [ 6728.837424][ T1818] RIP: 0033:0x55cc277d6cc4
> [ 6728.856867][ T1818] Code: 00 00 00 53 48 89 f3 48 83 ec 08 e8 a6 37 e1 ff 48 8b 15 27 a8 70 00 48 89 68 10 48 89 58 18 48 c7 00 00 00 00 00 48 89 50 08 <48> 89 02 48 89 05 0a a8 70 00 48 83 c4 08 5b 5d c3 66 66 2e 0f 1f
> [ 6728.946881][ T1818] RSP: 002b:00007ffe79363de0 EFLAGS: 00010206
> [ 6728.973883][ T1818] RAX: 000055cc29963830 RBX: 0000000000000000 RCX: 0000000000000030
> [ 6729.009698][ T1818] RDX: 000055cc27ee14d0 RSI: 0000000000000020 RDI: 000055cc29963830
> [ 6729.045698][ T1818] RBP: 000055cc277534b0 R08: 000055cc29963830 R09: 00007fc044602a40
> [ 6729.084027][ T1818] R10: 000055cc29963880 R11: 0000000000000000 R12: 0000000000000001
> [ 6729.125556][ T1818] R13: 000055cc29965180 R14: 0000000000000001 R15: 0000000000000001
> [ 6729.160970][ T1818]
> [ 6729.170932][ T1818] Allocated by task 1818:
> [ 6729.190130][ T1818] save_stack+0x23/0x50
> [ 6729.208517][ T1818] __kasan_kmalloc.constprop.0+0xcf/0xe0
> [ 6729.233816][ T1818] kasan_kmalloc+0x9/0x10
> [ 6729.252852][ T1818] __kmalloc_node+0x1c3/0x410
> [ 6729.273517][ T1818] kvmalloc_node+0x7b/0x90
> [ 6729.293002][ T1818] kvm_create_vm+0x264/0x810 [kvm]
> kvmalloc at include/linux/mm.h:759
> (inlined by) kvzalloc at include/linux/mm.h:767
> (inlined by) kvm_alloc_memslots at virt/kvm/kvm_main.c:562
> (inlined by) kvm_create_vm at virt/kvm/kvm_main.c:697
> [ 6729.315774][ T1818] kvm_dev_ioctl_create_vm+0x14/0xf0 [kvm]
> [ 6729.341840][ T1818] kvm_dev_ioctl+0x6b/0x90 [kvm]
> [ 6729.363821][ T1818] ksys_ioctl+0xa3/0xc0
> [ 6729.382166][ T1818] __x64_sys_ioctl+0x43/0x4c
> [ 6729.403174][ T1818] do_syscall_64+0x69/0xf4
> [ 6729.423242][ T1818] entry_SYSCALL_64_after_hwframe+0x49/0xb3
> [ 6729.449985][ T1818]
> [ 6729.459896][ T1818] Freed by task 0:
> [ 6729.476312][ T1818] (stack is not available)
> [ 6729.495726][ T1818]
> [ 6729.505710][ T1818] The buggy address belongs to the object at ffff8885351c4800
> [ 6729.505710][ T1818] which belongs to the cache kmalloc-2k(976:session-c1.scope) of size 2048
> [ 6729.579319][ T1818] The buggy address is located 1040 bytes inside of
> [ 6729.579319][ T1818] 2048-byte region [ffff8885351c4800, ffff8885351c5000)
> [ 6729.647086][ T1818] The buggy address belongs to the page:
> [ 6729.671752][ T1818] page:ffffea0014d47100 refcount:1 mapcount:0 mapping:00000000b234685e index:0x0 head:ffffea0014d47100 order:1 compound_mapcount:0
> [ 6729.734134][ T1818] flags: 0x5fffe000010200(slab|head)
> [ 6729.757725][ T1818] raw: 005fffe000010200 ffffea0019590308 ffffea0013df5508 ffff888f05490ac0
> [ 6729.796528][ T1818] raw: 0000000000000000 0000000000010001 00000001ffffffff 0000000000000000
> [ 6729.835200][ T1818] page dumped because: kasan: bad access detected
> [ 6729.864026][ T1818] page_owner tracks the page as allocated
> [ 6729.889565][ T1818] page last allocated via order 1, migratetype Unmovable, gfp_mask 0x52cc0(GFP_KERNEL|__GFP_NOWARN|__GFP_NORETRY|__GFP_COMP)
> [ 6729.948959][ T1818] prep_new_page+0x23e/0x270
> [ 6729.969540][ T1818] get_page_from_freelist+0x4f7/0x580
> [ 6729.993824][ T1818] __alloc_pages_nodemask+0x2de/0x780
> [ 6730.017658][ T1818] alloc_pages_current+0x9e/0x170
> [ 6730.039922][ T1818] alloc_slab_page+0x321/0x580
> [ 6730.061106][ T1818] allocate_slab+0x7f/0x3c0
> [ 6730.081229][ T1818] new_slab+0x4a/0x70
> [ 6730.100073][ T1818] ___slab_alloc+0x294/0x380
> [ 6730.122658][ T1818] __kmalloc_node+0x227/0x410
> [ 6730.146452][ T1818] kvmalloc_node+0x7b/0x90
> [ 6730.166644][ T1818] kvm_create_vm+0x264/0x810 [kvm]
> [ 6730.189396][ T1818] kvm_dev_ioctl_create_vm+0x14/0xf0 [kvm]
> [ 6730.215647][ T1818] kvm_dev_ioctl+0x6b/0x90 [kvm]
> [ 6730.237526][ T1818] ksys_ioctl+0xa3/0xc0
> [ 6730.255755][ T1818] __x64_sys_ioctl+0x43/0x4c
> [ 6730.276150][ T1818] do_syscall_64+0x69/0xf4
> [ 6730.295871][ T1818] page last free stack trace:
> [ 6730.316511][ T1818] __free_pages_ok+0x678/0x760
> [ 6730.337440][ T1818] __free_pages+0xa2/0xe0
> [ 6730.356259][ T1818] free_pages+0x53/0x70
> [ 6730.374572][ T1818] stack_depot_save+0x9f/0x213
> [ 6730.395761][ T1818] save_stack+0x42/0x50
> [ 6730.414042][ T1818] __kasan_kmalloc.constprop.0+0xcf/0xe0
> [ 6730.439088][ T1818] kasan_slab_alloc+0xe/0x10
> [ 6730.459349][ T1818] kmem_cache_alloc+0x136/0x470
> [ 6730.481158][ T1818] mem_pool_alloc+0x38/0x170
> [ 6730.501470][ T1818] create_object.isra.0+0x27/0x430
> [ 6730.524151][ T1818] kmemleak_alloc_percpu+0xbb/0x150
> [ 6730.547342][ T1818] pcpu_alloc+0x537/0xad0
> [ 6730.566651][ T1818] __alloc_percpu+0x15/0x20
> [ 6730.588291][ T1818] init_srcu_struct_fields+0x208/0x220
> [ 6730.612543][ T1818] __init_srcu_struct+0x5c/0x70
> [ 6730.636590][ T1818] kvm_create_vm+0x20d/0x810 [kvm]
> [ 6730.662421][ T1818]
> [ 6730.672838][ T1818] Memory state around the buggy address:
> [ 6730.697743][ T1818] ffff8885351c4b00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> [ 6730.734279][ T1818] ffff8885351c4b80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> [ 6730.770594][ T1818] >ffff8885351c4c00: 00 00 fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> [ 6730.806845][ T1818] ^
> [ 6730.827097][ T1818] ffff8885351c4c80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> [ 6730.863311][ T1818] ffff8885351c4d00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> [ 6730.899350][ T1818] ==================================================================
> [ 6730.935728][ T1818] Disabling lock debugging due to kernel taint

Next message: Zenghui Yu: "Re: [PATCH v5 23/23] KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs"
Previous message: Vinay Simha B N: "Re: graph connection to node is not bidirectional kernel-5.6.0-rc6"
In reply to: Qian Cai: "slab-out-of-bounds due to "KVM: Dynamically size memslot array based on number of used slots""
Next in thread: Qian Cai: "Re: slab-out-of-bounds due to "KVM: Dynamically size memslot array based on number of used slots""
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]