[RESEND,v2] kvm: x86: Keep the lock order consistent

Commit Message

Hao Peng Oct. 31, 2022, 7:20 a.m. UTC

From: Peng Hao <[e-mail]>

Acquire SRCU before taking the gpc spinlock in wait_pending_event() so as
 to be consistent with all other functions that acquire both locks.  It's
 not illegal to acquire SRCU inside a spinlock, nor is there deadlock
 potential, but in general it's preferable to order locks from least
 restrictive to most restrictive, e.g. if wait_pending_event() needed to
 sleep for whatever reason, it could do so while holding SRCU, but would
 need to drop the spinlock.

Thanks Sean Christopherson for the comment.

Signed-off-by: Peng Hao <[e-mail]>
---
 arch/x86/kvm/xen.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)


@@ -986,8 +986,8 @@ static bool wait_pending_event(struct kvm_vcpu
*vcpu, int nr_ports,
        }

  out_rcu:
-       srcu_read_unlock(&kvm->srcu, idx);
        read_unlock_irqrestore(&gpc->lock, flags);
+       srcu_read_unlock(&kvm->srcu, idx);

        return ret;
 }
--
2.27.0

Comments

Sean Christopherson Nov. 3, 2022, 11:15 p.m. UTC | #1

Nit, capitalize KVM in the shortlog, i.e. "KVM: x86:".

On Mon, Oct 31, 2022, Hao Peng wrote:
> From: Peng Hao <[e-mail]>
> 
> Acquire SRCU before taking the gpc spinlock in wait_pending_event() so as
>  to be consistent with all other functions that acquire both locks.  It's
>  not illegal to acquire SRCU inside a spinlock, nor is there deadlock
>  potential, but in general it's preferable to order locks from least
>  restrictive to most restrictive, e.g. if wait_pending_event() needed to
>  sleep for whatever reason, it could do so while holding SRCU, but would
>  need to drop the spinlock.

Extra whitespace at the beginning of each line should be deleted.

> Thanks Sean Christopherson for the comment.
> 
> Signed-off-by: Peng Hao <[e-mail]>
> ---

Nits aside,

Reviewed-by: Sean Christopherson <[e-mail]>

Patch

diff --git a/arch/x86/kvm/xen.c b/arch/x86/kvm/xen.c
index 2dae413bd62a..766e8a4ca3ea 100644
--- a/arch/x86/kvm/xen.c
+++ b/arch/x86/kvm/xen.c
@@ -964,8 +964,8 @@  static bool wait_pending_event(struct kvm_vcpu
*vcpu, int nr_ports,
        bool ret = true;
        int idx, i;

-       read_lock_irqsave(&gpc->lock, flags);
        idx = srcu_read_lock(&kvm->srcu);
+       read_lock_irqsave(&gpc->lock, flags);
        if (!kvm_gfn_to_pfn_cache_check(kvm, gpc, gpc->gpa, PAGE_SIZE))
                goto out_rcu;