lkml.org 
[lkml]   [2021]   [Mar]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH] x86/kvm: Fix broken irq restoration in kvm_wait
From
Date
On 11/03/21 16:54, Sean Christopherson wrote:
> On Tue, Feb 23, 2021, Wanpeng Li wrote:
>> On Tue, 23 Feb 2021 at 13:25, Wanpeng Li <kernellwp@gmail.com> wrote:
>>>
>>> From: Wanpeng Li <wanpengli@tencent.com>
>>>
>>> After commit 997acaf6b4b59c (lockdep: report broken irq restoration), the guest
>>> splatting below during boot:
>>>
>>> raw_local_irq_restore() called with IRQs enabled
>>> WARNING: CPU: 1 PID: 169 at kernel/locking/irqflag-debug.c:10 warn_bogus_irq_restore+0x26/0x30
>>> Modules linked in: hid_generic usbhid hid
>>> CPU: 1 PID: 169 Comm: systemd-udevd Not tainted 5.11.0+ #25
>>> RIP: 0010:warn_bogus_irq_restore+0x26/0x30
>>> Call Trace:
>>> kvm_wait+0x76/0x90
>>> __pv_queued_spin_lock_slowpath+0x285/0x2e0
>>> do_raw_spin_lock+0xc9/0xd0
>>> _raw_spin_lock+0x59/0x70
>>> lockref_get_not_dead+0xf/0x50
>>> __legitimize_path+0x31/0x60
>>> legitimize_root+0x37/0x50
>>> try_to_unlazy_next+0x7f/0x1d0
>>> lookup_fast+0xb0/0x170
>>> path_openat+0x165/0x9b0
>>> do_filp_open+0x99/0x110
>>> do_sys_openat2+0x1f1/0x2e0
>>> do_sys_open+0x5c/0x80
>>> __x64_sys_open+0x21/0x30
>>> do_syscall_64+0x32/0x50
>>> entry_SYSCALL_64_after_hwframe+0x44/0xae
>>>
>>> The irqflags handling in kvm_wait() which ends up doing:
>>>
>>> local_irq_save(flags);
>>> safe_halt();
>>> local_irq_restore(flags);
>>>
>>> which triggered a new consistency checking, we generally expect
>>> local_irq_save() and local_irq_restore() to be pared and sanely
>>> nested, and so local_irq_restore() expects to be called with
>>> irqs disabled.
>>>
>>> This patch fixes it by adding a local_irq_disable() after safe_halt()
>>> to avoid this warning.
>>>
>>> Cc: Mark Rutland <mark.rutland@arm.com>
>>> Cc: Thomas Gleixner <tglx@linutronix.de>
>>> Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
>>> ---
>>> arch/x86/kernel/kvm.c | 4 +++-
>>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
>>> index 5e78e01..688c84a 100644
>>> --- a/arch/x86/kernel/kvm.c
>>> +++ b/arch/x86/kernel/kvm.c
>>> @@ -853,8 +853,10 @@ static void kvm_wait(u8 *ptr, u8 val)
>>> */
>>> if (arch_irqs_disabled_flags(flags))
>>> halt();
>>> - else
>>> + else {
>>> safe_halt();
>>> + local_irq_disable();
>>> + }
>>
>> An alternative fix:
>>
>> diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
>> index 5e78e01..7127aef 100644
>> --- a/arch/x86/kernel/kvm.c
>> +++ b/arch/x86/kernel/kvm.c
>> @@ -836,12 +836,13 @@ static void kvm_kick_cpu(int cpu)
>>
>> static void kvm_wait(u8 *ptr, u8 val)
>> {
>> - unsigned long flags;
>> + bool disabled = irqs_disabled();
>>
>> if (in_nmi())
>> return;
>>
>> - local_irq_save(flags);
>> + if (!disabled)
>> + local_irq_disable();
>>
>> if (READ_ONCE(*ptr) != val)
>> goto out;
>> @@ -851,13 +852,14 @@ static void kvm_wait(u8 *ptr, u8 val)
>> * for irq enabled case to avoid hang when lock info is overwritten
>> * in irq spinlock slowpath and no spurious interrupt occur to save us.
>> */
>> - if (arch_irqs_disabled_flags(flags))
>> + if (disabled)
>> halt();
>> else
>> safe_halt();
>>
>> out:
>> - local_irq_restore(flags);
>> + if (!disabled)
>> + local_irq_enable();
>> }
>>
>> #ifdef CONFIG_X86_32
>
> A third option would be to split the paths. In the end, it's only the ptr/val
> line that's shared.
>
> ---
> arch/x86/kernel/kvm.c | 23 ++++++++++-------------
> 1 file changed, 10 insertions(+), 13 deletions(-)
>
> diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
> index 5e78e01ca3b4..78bb0fae3982 100644
> --- a/arch/x86/kernel/kvm.c
> +++ b/arch/x86/kernel/kvm.c
> @@ -836,28 +836,25 @@ static void kvm_kick_cpu(int cpu)
>
> static void kvm_wait(u8 *ptr, u8 val)
> {
> - unsigned long flags;
> -
> if (in_nmi())
> return;
>
> - local_irq_save(flags);
> -
> - if (READ_ONCE(*ptr) != val)
> - goto out;
> -
> /*
> * halt until it's our turn and kicked. Note that we do safe halt
> * for irq enabled case to avoid hang when lock info is overwritten
> * in irq spinlock slowpath and no spurious interrupt occur to save us.
> */
> - if (arch_irqs_disabled_flags(flags))
> - halt();
> - else
> - safe_halt();
> + if (irqs_disabled()) {
> + if (READ_ONCE(*ptr) == val)
> + halt();
> + } else {
> + local_irq_disable();
>
> -out:
> - local_irq_restore(flags);
> + if (READ_ONCE(*ptr) == val)
> + safe_halt();
> +
> + local_irq_enable();
> + }
> }
>
> #ifdef CONFIG_X86_32
> --
>

I'll send this one tomorrow.

Paolo

\
 
 \ /
  Last update: 2021-03-11 19:10    [W:0.079 / U:0.220 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site