Re: [PATCH v3 6/7] memory: Attach MemoryAttributeManager to guest_memfd-backed RAMBlocks

[David Hildenbrand]

On 14.03.25 09:21, Chenyi Qiang wrote:

Hi David & Alexey,

Hi,

To keep the bitmap aligned, I add the undo operation for
set_memory_attributes() and use the bitmap + replay callback to do
set_memory_attributes(). Does this change make sense?

I assume you mean this hunk:

+    ret = 
memory_attribute_manager_state_change(MEMORY_ATTRIBUTE_MANAGER(mr->rdm),
+                                                offset, size, to_private);
+    if (ret) {
+        warn_report("Failed to notify the listener the state change of "
+                    "(0x%"HWADDR_PRIx" + 0x%"HWADDR_PRIx") to %s",
+                    start, size, to_private ? "private" : "shared");
+        args.to_private = !to_private;
+        if (to_private) {
+            ret = ram_discard_manager_replay_populated(mr->rdm, &section,
+                                                       
kvm_set_memory_attributes_cb,
+                                                       &args);
+        } else {
+            ret = ram_discard_manager_replay_discarded(mr->rdm, &section,
+                                                       
kvm_set_memory_attributes_cb,
+                                                       &args);
+        }
+        if (ret) {
+            goto out_unref;
+        }


Why is that undo necessary? The bitmap + listeners should be held in sync 
inside of
memory_attribute_manager_state_change(). Handling this in the caller looks 
wrong.

I thought the current implementation properly handles that internally? In which 
scenario is that not the case?

--
Cheers,

David / dhildenb