x86/kvm: Alloc dummy async #PF token outside of raw spinlock

Drop the raw spinlock in kvm_async_pf_task_wake() before allocating the
the dummy async #PF token, the allocator is preemptible on PREEMPT_RT
kernels and must not be called from truly atomic contexts.

Opportunistically document why it's ok to loop on allocation failure,
i.e. why the function won't get stuck in an infinite loop.

Reported-by: Yajun Deng <yajun.deng@linux.dev>
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index b48ce07756ca83a5ab05e99d8e996c91b270d095..35b3c5836703bbf63d1c6683e7411a332c17cafa 100644 (file)
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -191,7 +191,7 @@ void kvm_async_pf_task_wake(u32 token)
 {
        u32 key = hash_32(token, KVM_TASK_SLEEP_HASHBITS);
        struct kvm_task_sleep_head *b = &async_pf_sleepers[key];
-       struct kvm_task_sleep_node *n;
+       struct kvm_task_sleep_node *n, *dummy = NULL;
 
        if (token == ~0) {
                apf_task_wake_all();
@@ -203,28 +203,41 @@ again:
        n = _find_apf_task(b, token);
        if (!n) {
                /*
-                * async PF was not yet handled.
-                * Add dummy entry for the token.
+                * Async #PF not yet handled, add a dummy entry for the token.
+                * Allocating the token must be down outside of the raw lock
+                * as the allocator is preemptible on PREEMPT_RT kernels.
                 */
-               n = kzalloc(sizeof(*n), GFP_ATOMIC);
-               if (!n) {
+               if (!dummy) {
+                       raw_spin_unlock(&b->lock);
+                       dummy = kzalloc(sizeof(*dummy), GFP_KERNEL);
+
                        /*
-                        * Allocation failed! Busy wait while other cpu
-                        * handles async PF.
+                        * Continue looping on allocation failure, eventually
+                        * the async #PF will be handled and allocating a new
+                        * node will be unnecessary.
+                        */
+                       if (!dummy)
+                               cpu_relax();
+
+                       /*
+                        * Recheck for async #PF completion before enqueueing
+                        * the dummy token to avoid duplicate list entries.
                         */
-                       raw_spin_unlock(&b->lock);
-                       cpu_relax();
                        goto again;
                }
-               n->token = token;
-               n->cpu = smp_processor_id();
-               init_swait_queue_head(&n->wq);
-               hlist_add_head(&n->link, &b->list);
+               dummy->token = token;
+               dummy->cpu = smp_processor_id();
+               init_swait_queue_head(&dummy->wq);
+               hlist_add_head(&dummy->link, &b->list);
+               dummy = NULL;
        } else {
                apf_task_wake_one(n);
        }
        raw_spin_unlock(&b->lock);
-       return;
+
+       /* A dummy token might be allocated and ultimately not used.  */
+       if (dummy)
+               kfree(dummy);
 }
 EXPORT_SYMBOL_GPL(kvm_async_pf_task_wake);