KVM: Document weakly ordered architecture requirements for dirty ring

Now that the kernel can expose to userspace that its dirty ring
management relies on explicit ordering, document these new requirements
for VMMs to do the right thing.

Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Gavin Shan <gshan@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Link: https://lore.kernel.org/r/20220926145120.27974-5-maz@kernel.org

Documentation/virt/kvm/api.rst

@@ -8019,8 +8019,8 @@ guest according to the bits in the KVM_CPUID_FEATURES CPUID leaf
 (0x40000001). Otherwise, a guest may use the paravirtual features
 regardless of what has actually been exposed through the CPUID leaf.
 
-8.29 KVM_CAP_DIRTY_LOG_RING
----------------------------
+8.29 KVM_CAP_DIRTY_LOG_RING/KVM_CAP_DIRTY_LOG_RING_ACQ_REL
+----------------------------------------------------------
 
 :Architectures: x86
 :Parameters: args[0] - size of the dirty log ring
@@ -8078,6 +8078,11 @@ on to the next GFN.  The userspace should continue to do this until the
 flags of a GFN have the DIRTY bit cleared, meaning that it has harvested
 all the dirty GFNs that were available.
 
+Note that on weakly ordered architectures, userspace accesses to the
+ring buffer (and more specifically the 'flags' field) must be ordered,
+using load-acquire/store-release accessors when available, or any
+other memory barrier that will ensure this ordering.
+
 It's not necessary for userspace to harvest the all dirty GFNs at once.
 However it must collect the dirty GFNs in sequence, i.e., the userspace
 program cannot skip one dirty GFN to collect the one next to it.
@@ -8106,6 +8111,14 @@ KVM_CAP_DIRTY_LOG_RING with an acceptable dirty ring size, the virtual
 machine will switch to ring-buffer dirty page tracking and further
 KVM_GET_DIRTY_LOG or KVM_CLEAR_DIRTY_LOG ioctls will fail.
 
+NOTE: KVM_CAP_DIRTY_LOG_RING_ACQ_REL is the only capability that
+should be exposed by weakly ordered architecture, in order to indicate
+the additional memory ordering requirements imposed on userspace when
+reading the state of an entry and mutating it from DIRTY to HARVESTED.
+Architecture with TSO-like ordering (such as x86) are allowed to
+expose both KVM_CAP_DIRTY_LOG_RING and KVM_CAP_DIRTY_LOG_RING_ACQ_REL
+to userspace.
+
 8.30 KVM_CAP_XEN_HVM
 --------------------