From 293f293637b55db4f9f522a5a72514e98a541076 Mon Sep 17 00:00:00 2001 From: Suzuki K Poulose Date: Thu, 8 Sep 2016 16:25:49 +0100 Subject: [PATCH] kvm-arm: Unmap shadow pagetables properly On arm/arm64, we depend on the kvm_unmap_hva* callbacks (via mmu_notifiers::invalidate_*) to unmap the stage2 pagetables when the userspace buffer gets unmapped. However, when the Hypervisor process exits without explicit unmap of the guest buffers, the only notifier we get is kvm_arch_flush_shadow_all() (via mmu_notifier::release ) which does nothing on arm. Later this causes us to access pages that were already released [via exit_mmap() -> unmap_vmas()] when we actually get to unmap the stage2 pagetable [via kvm_arch_destroy_vm() -> kvm_free_stage2_pgd()]. This triggers crashes with CONFIG_DEBUG_PAGEALLOC, which unmaps any free'd pages from the linear map. [ 757.644120] Unable to handle kernel paging request at virtual address ffff800661e00000 [ 757.652046] pgd = ffff20000b1a2000 [ 757.655471] [ffff800661e00000] *pgd=00000047fffe3003, *pud=00000047fcd8c003, *pmd=00000047fcc7c003, *pte=00e8004661e00712 [ 757.666492] Internal error: Oops: 96000147 [#3] PREEMPT SMP [ 757.672041] Modules linked in: [ 757.675100] CPU: 7 PID: 3630 Comm: qemu-system-aar Tainted: G D 4.8.0-rc1 #3 [ 757.683240] Hardware name: AppliedMicro X-Gene Mustang Board/X-Gene Mustang Board, BIOS 3.06.15 Aug 19 2016 [ 757.692938] task: ffff80069cdd3580 task.stack: ffff8006adb7c000 [ 757.698840] PC is at __flush_dcache_area+0x1c/0x40 [ 757.703613] LR is at kvm_flush_dcache_pmd+0x60/0x70 [ 757.708469] pc : [] lr : [] pstate: 20000145 ... [ 758.357249] [] __flush_dcache_area+0x1c/0x40 [ 758.363059] [] unmap_stage2_range+0x458/0x5f0 [ 758.368954] [] kvm_free_stage2_pgd+0x34/0x60 [ 758.374761] [] kvm_arch_destroy_vm+0x20/0x68 [ 758.380570] [] kvm_put_kvm+0x210/0x358 [ 758.385860] [] kvm_vm_release+0x2c/0x40 [ 758.391239] [] __fput+0x114/0x2e8 [ 758.396096] [] ____fput+0xc/0x18 [ 758.400869] [] task_work_run+0x108/0x138 [ 758.406332] [] do_exit+0x48c/0x10e8 [ 758.411363] [] do_group_exit+0x6c/0x130 [ 758.416739] [] get_signal+0x284/0xa18 [ 758.421943] [] do_signal+0x158/0x860 [ 758.427060] [] do_notify_resume+0x6c/0x88 [ 758.432608] [] work_pending+0x10/0x14 [ 758.437812] Code: 9ac32042 8b010001 d1000443 8a230000 (d50b7e20) This patch fixes the issue by moving the kvm_free_stage2_pgd() to kvm_arch_flush_shadow_all(). Cc: # 3.9+ Tested-by: Itaru Kitayama Reported-by: Itaru Kitayama Reported-by: James Morse Cc: Marc Zyngier Cc: Catalin Marinas Cc: Christoffer Dall Signed-off-by: Suzuki K Poulose Signed-off-by: Christoffer Dall --- arch/arm/kvm/arm.c | 2 -- arch/arm/kvm/mmu.c | 1 + 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c index 75f130ef6504..c94b90d43772 100644 --- a/arch/arm/kvm/arm.c +++ b/arch/arm/kvm/arm.c @@ -158,8 +158,6 @@ void kvm_arch_destroy_vm(struct kvm *kvm) { int i; - kvm_free_stage2_pgd(kvm); - for (i = 0; i < KVM_MAX_VCPUS; ++i) { if (kvm->vcpus[i]) { kvm_arch_vcpu_free(kvm->vcpus[i]); diff --git a/arch/arm/kvm/mmu.c b/arch/arm/kvm/mmu.c index a3faafe70081..e9a5c0e0c115 100644 --- a/arch/arm/kvm/mmu.c +++ b/arch/arm/kvm/mmu.c @@ -1894,6 +1894,7 @@ void kvm_arch_memslots_updated(struct kvm *kvm, struct kvm_memslots *slots) void kvm_arch_flush_shadow_all(struct kvm *kvm) { + kvm_free_stage2_pgd(kvm); } void kvm_arch_flush_shadow_memslot(struct kvm *kvm, -- 2.20.1