From: Paolo Bonzini pbonzini@redhat.com
commit 250715a6171a076748be8ab88b274e72f0cfb435 upstream.
The syzkaller folks reported a NULL pointer dereference that seems to be cause by a race between KVM_CREATE_IRQCHIP and KVM_CREATE_PIT2. The former takes kvm->lock (except when registering the devices, which needs kvm->slots_lock); the latter takes kvm->slots_lock only. Change KVM_CREATE_PIT2 to follow the same model as KVM_CREATE_IRQCHIP.
Testcase:
#include <pthread.h> #include <linux/kvm.h> #include <fcntl.h> #include <sys/ioctl.h> #include <stdint.h> #include <string.h> #include <stdlib.h> #include <sys/syscall.h> #include <unistd.h>
long r[23];
void* thr1(void* arg) { struct kvm_pit_config pitcfg = { .flags = 4 }; switch ((long)arg) { case 0: r[2] = open("/dev/kvm", O_RDONLY|O_ASYNC); break; case 1: r[3] = ioctl(r[2], KVM_CREATE_VM, 0); break; case 2: r[4] = ioctl(r[3], KVM_CREATE_IRQCHIP, 0); break; case 3: r[22] = ioctl(r[3], KVM_CREATE_PIT2, &pitcfg); break; } return 0; }
int main(int argc, char **argv) { long i; pthread_t th[4];
memset(r, -1, sizeof(r)); for (i = 0; i < 4; i++) { pthread_create(&th[i], 0, thr, (void*)i); if (argc > 1 && rand()%2) usleep(rand()%1000); } usleep(20000); return 0; }
Reported-by: Dmitry Vyukov dvyukov@google.com Signed-off-by: Paolo Bonzini pbonzini@redhat.com Signed-off-by: Radim Krčmář rkrcmar@redhat.com Signed-off-by: Zubin Mithra zsm@google.com --- Notes: * Syzkaller triggered a GPF in kvm_pic_clear_all with the following stacktrace when fuzzing a 4.4 kernel. Call Trace: [<ffffffff81270a07>] lock_acquire+0x294/0x2f3 kernel/locking/lockdep.c:3592 [<ffffffff832bd6f7>] __raw_spin_lock include/linux/spinlock_api_smp.h:144 [inline] [<ffffffff832bd6f7>] _raw_spin_lock+0x2f/0x3e kernel/locking/spinlock.c:151 [<ffffffff81084dce>] spin_lock include/linux/spinlock.h:302 [inline] [<ffffffff81084dce>] pic_lock arch/x86/kvm/i8259.c:45 [inline] [<ffffffff81084dce>] kvm_pic_clear_all+0x33/0x5b arch/x86/kvm/i8259.c:213 [<ffffffff810968fb>] kvm_free_irq_source_id+0xc8/0xde arch/x86/kvm/irq_comm.c:197 [<ffffffff810926bb>] kvm_create_pit+0x24e/0x5c5 arch/x86/kvm/i8254.c:713 [<ffffffff8103aabf>] kvm_arch_vm_ioctl+0x592/0x17db arch/x86/kvm/x86.c:3858 [<ffffffff81013cb4>] kvm_vm_ioctl+0xb7d/0xbfa arch/x86/kvm/../../../virt/kvm/kvm_main.c:2959 [<ffffffff8149dc2e>] vfs_ioctl fs/ioctl.c:43 [inline] [<ffffffff8149dc2e>] do_vfs_ioctl+0xcb0/0xd0f fs/ioctl.c:630 [<ffffffff8149dcfe>] SYSC_ioctl fs/ioctl.c:645 [inline] [<ffffffff8149dcfe>] SyS_ioctl+0x71/0xad fs/ioctl.c:636 [<ffffffff832be2fa>] entry_SYSCALL_64_fastpath+0x31/0xb3
* This patch resolves a conflict that arises in arch/x86/kvm/i8254.c:kvm_create_pit()
* This commit is present in linux-4.9.y
* Tests run: Chrome OS tryjobs, Syzkaller reproducer
arch/x86/kvm/i8254.c | 5 +++-- arch/x86/kvm/x86.c | 4 ++-- 2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/arch/x86/kvm/i8254.c b/arch/x86/kvm/i8254.c index ab53187275794..ef56d70d10264 100644 --- a/arch/x86/kvm/i8254.c +++ b/arch/x86/kvm/i8254.c @@ -678,7 +678,6 @@ static const struct kvm_io_device_ops speaker_dev_ops = { .write = speaker_ioport_write, };
-/* Caller must hold slots_lock */ struct kvm_pit *kvm_create_pit(struct kvm *kvm, u32 flags) { struct kvm_pit *pit; @@ -733,6 +732,7 @@ struct kvm_pit *kvm_create_pit(struct kvm *kvm, u32 flags) pit->mask_notifier.func = pit_mask_notifer; kvm_register_irq_mask_notifier(kvm, 0, &pit->mask_notifier);
+ mutex_lock(&kvm->slots_lock); kvm_iodevice_init(&pit->dev, &pit_dev_ops); ret = kvm_io_bus_register_dev(kvm, KVM_PIO_BUS, KVM_PIT_BASE_ADDRESS, KVM_PIT_MEM_LENGTH, &pit->dev); @@ -747,13 +747,14 @@ struct kvm_pit *kvm_create_pit(struct kvm *kvm, u32 flags) if (ret < 0) goto fail_unregister; } + mutex_unlock(&kvm->slots_lock);
return pit;
fail_unregister: kvm_io_bus_unregister_dev(kvm, KVM_PIO_BUS, &pit->dev); - fail: + mutex_unlock(&kvm->slots_lock); kvm_unregister_irq_mask_notifier(kvm, 0, &pit->mask_notifier); kvm_unregister_irq_ack_notifier(kvm, &pit_state->irq_ack_notifier); kvm_free_irq_source_id(kvm, pit->irq_source_id); diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 66adcd00b3ddd..ef95cb80cadc2 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -3867,7 +3867,7 @@ long kvm_arch_vm_ioctl(struct file *filp, sizeof(struct kvm_pit_config))) goto out; create_pit: - mutex_lock(&kvm->slots_lock); + mutex_lock(&kvm->lock); r = -EEXIST; if (kvm->arch.vpit) goto create_pit_unlock; @@ -3876,7 +3876,7 @@ long kvm_arch_vm_ioctl(struct file *filp, if (kvm->arch.vpit) r = 0; create_pit_unlock: - mutex_unlock(&kvm->slots_lock); + mutex_unlock(&kvm->lock); break; case KVM_GET_IRQCHIP: { /* 0: PIC master, 1: PIC slave, 2: IOAPIC */
On Wed, Jul 17, 2019 at 01:35:01PM -0700, Zubin Mithra wrote:
From: Paolo Bonzini pbonzini@redhat.com
commit 250715a6171a076748be8ab88b274e72f0cfb435 upstream.
The syzkaller folks reported a NULL pointer dereference that seems to be cause by a race between KVM_CREATE_IRQCHIP and KVM_CREATE_PIT2. The former takes kvm->lock (except when registering the devices, which needs kvm->slots_lock); the latter takes kvm->slots_lock only. Change KVM_CREATE_PIT2 to follow the same model as KVM_CREATE_IRQCHIP.
Testcase:
#include <pthread.h> #include <linux/kvm.h> #include <fcntl.h> #include <sys/ioctl.h> #include <stdint.h> #include <string.h> #include <stdlib.h> #include <sys/syscall.h> #include <unistd.h> long r[23]; void* thr1(void* arg) { struct kvm_pit_config pitcfg = { .flags = 4 }; switch ((long)arg) { case 0: r[2] = open("/dev/kvm", O_RDONLY|O_ASYNC); break; case 1: r[3] = ioctl(r[2], KVM_CREATE_VM, 0); break; case 2: r[4] = ioctl(r[3], KVM_CREATE_IRQCHIP, 0); break; case 3: r[22] = ioctl(r[3], KVM_CREATE_PIT2, &pitcfg); break; } return 0; } int main(int argc, char **argv) { long i; pthread_t th[4]; memset(r, -1, sizeof(r)); for (i = 0; i < 4; i++) { pthread_create(&th[i], 0, thr, (void*)i); if (argc > 1 && rand()%2) usleep(rand()%1000); } usleep(20000); return 0; }
Reported-by: Dmitry Vyukov dvyukov@google.com Signed-off-by: Paolo Bonzini pbonzini@redhat.com Signed-off-by: Radim Krčmář rkrcmar@redhat.com Signed-off-by: Zubin Mithra zsm@google.com
Notes:
- Syzkaller triggered a GPF in kvm_pic_clear_all with the following stacktrace
when fuzzing a 4.4 kernel. Call Trace: [<ffffffff81270a07>] lock_acquire+0x294/0x2f3 kernel/locking/lockdep.c:3592 [<ffffffff832bd6f7>] __raw_spin_lock include/linux/spinlock_api_smp.h:144 [inline] [<ffffffff832bd6f7>] _raw_spin_lock+0x2f/0x3e kernel/locking/spinlock.c:151 [<ffffffff81084dce>] spin_lock include/linux/spinlock.h:302 [inline] [<ffffffff81084dce>] pic_lock arch/x86/kvm/i8259.c:45 [inline] [<ffffffff81084dce>] kvm_pic_clear_all+0x33/0x5b arch/x86/kvm/i8259.c:213 [<ffffffff810968fb>] kvm_free_irq_source_id+0xc8/0xde arch/x86/kvm/irq_comm.c:197 [<ffffffff810926bb>] kvm_create_pit+0x24e/0x5c5 arch/x86/kvm/i8254.c:713 [<ffffffff8103aabf>] kvm_arch_vm_ioctl+0x592/0x17db arch/x86/kvm/x86.c:3858 [<ffffffff81013cb4>] kvm_vm_ioctl+0xb7d/0xbfa arch/x86/kvm/../../../virt/kvm/kvm_main.c:2959 [<ffffffff8149dc2e>] vfs_ioctl fs/ioctl.c:43 [inline] [<ffffffff8149dc2e>] do_vfs_ioctl+0xcb0/0xd0f fs/ioctl.c:630 [<ffffffff8149dcfe>] SYSC_ioctl fs/ioctl.c:645 [inline] [<ffffffff8149dcfe>] SyS_ioctl+0x71/0xad fs/ioctl.c:636 [<ffffffff832be2fa>] entry_SYSCALL_64_fastpath+0x31/0xb3
This patch resolves a conflict that arises in arch/x86/kvm/i8254.c:kvm_create_pit()
This commit is present in linux-4.9.y
Tests run: Chrome OS tryjobs, Syzkaller reproducer
Now queued up, thanks.
greg k-h
linux-stable-mirror@lists.linaro.org