native_safe_halt enables interrupts, and you just shouldn't
call rcu_irq_enter() with interrupts enabled. Reorder the
call with the following local_irq_disable() to respect the
invariant.
Reported-by: Ross Zwisler <[email protected]>
Cc: [email protected]
Signed-off-by: Paolo Bonzini <[email protected]>
---
arch/x86/kernel/kvm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index 14f65a5f938e..2a7835932b71 100644
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -161,8 +161,8 @@ void kvm_async_pf_task_wait(u32 token)
*/
rcu_irq_exit();
native_safe_halt();
- rcu_irq_enter();
local_irq_disable();
+ rcu_irq_enter();
}
}
if (!n.halted)
--
2.9.3
On Wed, Apr 26, 2017 at 04:56:26PM +0200, Paolo Bonzini wrote:
> native_safe_halt enables interrupts, and you just shouldn't
> call rcu_irq_enter() with interrupts enabled. Reorder the
> call with the following local_irq_disable() to respect the
> invariant.
>
> Reported-by: Ross Zwisler <[email protected]>
> Cc: [email protected]
> Signed-off-by: Paolo Bonzini <[email protected]>
Acked-by: Paul E. McKenney <[email protected]>
> ---
> arch/x86/kernel/kvm.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
> index 14f65a5f938e..2a7835932b71 100644
> --- a/arch/x86/kernel/kvm.c
> +++ b/arch/x86/kernel/kvm.c
> @@ -161,8 +161,8 @@ void kvm_async_pf_task_wait(u32 token)
> */
> rcu_irq_exit();
> native_safe_halt();
> - rcu_irq_enter();
> local_irq_disable();
> + rcu_irq_enter();
> }
> }
> if (!n.halted)
> --
> 2.9.3
>
2017-04-26 22:56 GMT+08:00 Paolo Bonzini <[email protected]>:
> native_safe_halt enables interrupts, and you just shouldn't
> call rcu_irq_enter() with interrupts enabled. Reorder the
> call with the following local_irq_disable() to respect the
> invariant.
>
> Reported-by: Ross Zwisler <[email protected]>
> Cc: [email protected]
> Signed-off-by: Paolo Bonzini <[email protected]>
I'm afraid we miss this commit. This commit can fix the below warning
against my testing, so feel free to add:
Tested-by: Wanpeng Li <[email protected]>
=============================
WARNING: suspicious RCU usage
4.12.0-rc3+ #23 Tainted: G O
-----------------------------
kernel/rcu/tree.c:1040 rcu_irq_enter() invoked with irqs enabled!!!!
other info that might help us debug this:
rcu_scheduler_active = 2, debug_locks = 0
1 lock held by update-notifier/1661:
#0: (&mm->mmap_sem){++++++}, at: [<ffffffff94061bfc>]
__do_page_fault+0x13c/0x5a0
stack backtrace:
CPU: 3 PID: 1661 Comm: update-notifier Tainted: G O
4.12.0-rc3+ #23
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
rel-1.10.2-0-g5f4c7b1-prebuilt.qemu-project.org 04/01/2014
Call Trace:
dump_stack+0x99/0xce
lockdep_rcu_suspicious+0xd7/0x110
rcu_irq_enter+0xa0/0xb0
kvm_async_pf_task_wait+0x210/0x260
do_async_page_fault+0x77/0xb0
? do_async_page_fault+0x77/0xb0
async_page_fault+0x28/0x30
RIP: 0010:copy_page+0x7/0x10
RSP: 0000:ffffaa3d834d7da0 EFLAGS: 00010286
RAX: 0000000000000002 RBX: 0000000000eb3f00 RCX: 0000000000000200
RDX: 0000000000000000 RSI: ffff93f6bf36f000 RDI: ffff93f33acfc000
RBP: ffffaa3d834d7e58 R08: ffff93f6db998cb0 R09: 0000000000000000
R10: 00000000fef20a58 R11: b143443d00000000 R12: 0000000000000200
R13: ffffaa3d834d7dc0 R14: ffff93f6bf36f000 R15: ffff93f6db998000
? __handle_mm_fault+0xed1/0x11f0
handle_mm_fault+0x18a/0x350
__do_page_fault+0x29f/0x5a0
trace_do_page_fault+0x6b/0x2f0
do_async_page_fault+0x53/0xb0
async_page_fault+0x28/0x30
RIP: 0033:0x7f7d448b2a84
RSP: 002b:00007ffc129bead8 EFLAGS: 00010202
RAX: 00007f7d42a063c0 RBX: 0000000000000003 RCX: 00007f7d42a06400
RDX: 0000000000000080 RSI: 0000000000000000 RDI: 00007f7d42a063c0
RBP: 00007ffc129bed80 R08: 00007f7d42a06440 R09: 000000000000b000
R10: 00007ffc129beb10 R11: 00007f7d42a06440 R12: 00007f7d44a9f500
R13: 00007ffc129bee68 R14: 0000000000000002 R15: 000000000000fd01
> ---
> arch/x86/kernel/kvm.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
> index 14f65a5f938e..2a7835932b71 100644
> --- a/arch/x86/kernel/kvm.c
> +++ b/arch/x86/kernel/kvm.c
> @@ -161,8 +161,8 @@ void kvm_async_pf_task_wait(u32 token)
> */
> rcu_irq_exit();
> native_safe_halt();
> - rcu_irq_enter();
> local_irq_disable();
> + rcu_irq_enter();
> }
> }
> if (!n.halted)
> --
> 2.9.3
>
2017-06-05 12:21+0800, Wanpeng Li:
> 2017-04-26 22:56 GMT+08:00 Paolo Bonzini <[email protected]>:
> > native_safe_halt enables interrupts, and you just shouldn't
> > call rcu_irq_enter() with interrupts enabled. Reorder the
> > call with the following local_irq_disable() to respect the
> > invariant.
> >
> > Reported-by: Ross Zwisler <[email protected]>
> > Cc: [email protected]
> > Signed-off-by: Paolo Bonzini <[email protected]>
>
> I'm afraid we miss this commit. This commit can fix the below warning
Applied to kvm/master,
> Tested-by: Wanpeng Li <[email protected]>
thanks for the testing and ping!