KVM: x86: Add SRCU protection for KVM_GET_SREGS2
PATCH 1/1 KVM: x86: Add SRCU protection for KVM_GET_SREGS2
Add SRCU read-side protection to KVM_GET_SREGS2 ioctl handler.
__get_sregs2() may read guest memory when caching PDPTR registers:__get_sregs2() -> kvm_pdptr_read() -> svm_cache_reg() -> load_pdptrs() -> kvm_vcpu_read_guest_page() -> kvm_vcpu_gfn_to_memslot()
kvm_vcpu_gfn_to_memslot() dereferences memslots via __kvm_memslots(), which uses srcu_dereference_check() and requires either kvm->srcu or kvm->slots_lock to be held. Currently only vcpu->mutex is held, triggering lockdep warning:
=============================
WARNING: suspicious RCU usage in kvm_vcpu_gfn_to_memslot 6.12.59+ #3 Not tainted
-----------------------------include/linux/kvm_host.h:1062 suspicious rcu_dereference_check() usage!
other info that might help us debug this:
rcu_scheduler_active = 2, debug_locks = 1
1 lock held by syz.5.1717/15100:
#0: ff1100002f4b00b0 (&vcpu->mutex){+.+.}-{3:3}, at: kvm_vcpu_ioctl+0x1d5/0x1590
Call Trace: <TASK> __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0xf0/0x120 lib/dump_stack.c:120 lockdep_rcu_suspicious+0x1e3/0x270 kernel/locking/lockdep.c:6824 __kvm_memslots include/linux/kvm_host.h:1062 [inline] __kvm_memslots include/linux/kvm_host.h:1059 [inline] kvm_vcpu_memslots include/linux/kvm_host.h:1076 [inline] kvm_vcpu_gfn_to_memslot+0x518/0x5e0 virt/kvm/kvm_main.c:2617 kvm_vcpu_read_guest_page+0x27/0x50 virt/kvm/kvm_main.c:3302 load_pdptrs+0xff/0x4b0 arch/x86/kvm/x86.c:1065 svm_cache_reg+0x1c9/0x230 arch/x86/kvm/svm/svm.c:1688 kvm_pdptr_read arch/x86/kvm/kvm_cache_regs.h:141 [inline] __get_sregs2 arch/x86/kvm/x86.c:11784 [inline] kvm_arch_vcpu_ioctl+0x3e20/0x4aa0 arch/x86/kvm/x86.c:6279 kvm_vcpu_ioctl+0x856/0x1590 virt/kvm/kvm_main.c:4663 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:907 [inline] __se_sys_ioctl fs/ioctl.c:893 [inline] __x64_sys_ioctl+0x18b/0x210 fs/ioctl.c:893 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xbd/0x1d0 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f
Found by Linux Verification Center (linuxtesting.org) with Syzkaller.
---
Note 1: commit 85e5ba83c016 ("KVM: x86: Do all post-set CPUID processing during vCPU creation") in v6.14+ reduces the likelihood of hitting this path by ensuring proper MMU initialization, but does not eliminate the requirement for SRCU protection when accessing guest memory.
Note 2: KVM_SET_SREGS2 is not modified because __set_sregs_common() already acquires SRCU when update_pdptrs=true, which covers the case when PDPTRs must be loaded from guest memory.
--- arch/x86/kvm/x86.c | 2 ++ 1 file changed, 2 insertions(+)
arch/x86/kvm/x86.c +2 -0
@@ -6619,7 +6619,9 @@ long kvm_arch_vcpu_ioctl(struct file *filp, r = -ENOMEM ;if (!u.sregs2) goto out; +kvm_vcpu_srcu_read_lock (vcpu);__get_sregs2 (vcpu, u.sregs2); +kvm_vcpu_srcu_read_unlock (vcpu); r = -EFAULT ;if (copy_to_user (argp, u.sregs2,sizeof (struct kvm_sregs2 ))) goto out;
Signature · 1 line
2.50.1
Sean Christopherson
maintainerQuoted 9 lines · Vasiliy Kovalev
---
Note 1: commit85e5ba83c016("KVM: x86: Do all post-set CPUID processing during vCPU creation") in v6.14+ reduces the likelihood of hitting this path by ensuring proper MMU initialization, but does not eliminate the requirement for SRCU protection when accessing guest memory.Note 2:
KVM_SET_SREGS2is not modified because__set_sregs_common()already acquires SRCU when update_pdptrs=true, which covers the case when PDPTRs must be loaded from guest memory.
On the topic of the update_pdptrs behavior, what if we scope the fix to precisely reading the PDPTRs? Not for performance reasons, but for documentation purposes, e.g. so that future readers don't look at __get_sregs() and wonder why that call isn't wrapped with SRCU protection.
I.e.
arch/x86/kvm/x86.c +2 -0
@@ -12145,9 +12145,11 @@ static void __get_sregs2(struct kvm_vcpu *vcpu, struct kvm_sregs2 *sregs2)return ;if (is_pae_paging (vcpu)) { +kvm_vcpu_srcu_read_lock (vcpu);for (i =0 ; i <4 ; i++) sregs2->pdptrs[i] =kvm_pdptr_read (vcpu, i); sregs2->flags |=KVM_SREGS2_FLAGS_PDPTRS_VALID ; +kvm_vcpu_srcu_read_unlock (vcpu); } }
Quoted 21 lines
--- arch/x86/kvm/x86.c | 2 ++ 1 file changed, 2 insertions(+)arch/x86/kvm/x86.c +2
-0@@ -6619,7 +6619,9 @@ long kvm_arch_vcpu_ioctl(struct file *filp, r = -ENOMEM ;if (!u.sregs2) goto out; +kvm_vcpu_srcu_read_lock (vcpu);__get_sregs2 (vcpu, u.sregs2); +kvm_vcpu_srcu_read_unlock (vcpu); r = -EFAULT ;if (copy_to_user (argp, u.sregs2,sizeof (struct kvm_sregs2 ))) goto out;Signature · 1 line
2.50.1
Quoted 7 lines · Sean Christopherson
Quoted 6 lines · Vasiliy Kovalev
---
Note 1: commit85e5ba83c016("KVM: x86: Do all post-set CPUID processing during vCPU creation") in v6.14+ reduces the likelihood of hitting this path by ensuring proper MMU initialization, but does not eliminate the requirement for SRCU protection when accessing guest memory.
Quoted 8 lines
Quoted 3 lines
Note 2:
KVM_SET_SREGS2is not modified because__set_sregs_common()already acquires SRCU when update_pdptrs=true, which covers the case when PDPTRs must be loaded from guest memory.On the topic of the
update_pdptrsbehavior, what if we scope the fix to precisely reading the PDPTRs? Not for performance reasons, but for documentation purposes, e.g. so that future readers don't look at__get_sregs()and wonder why that call isn't wrapped with SRCU protection.
Agreed, moving the lock inside __get_sregs2() makes the requirements clearer. I've verified it fixes the issue and sent v2:
https://lore.kernel.org/all/[email protected]/
Signature · 2 lines
Thanks, Vasiliy