prevent duplication of nodeRef on deserialization #355
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mpollmeier
reviewed
Feb 20, 2023
| if (node == null) throw new IllegalStateException("unable to read node from disk; id=" + id); | ||
| this.node = node; | ||
| assert(this.node == node); |
There was a problem hiding this comment.
Java's asserts are not being checked at runtime by default 😢
I tend to throw new AssertionError in this case, but I guess there's other options.
Also: the error text in this case wouldn't be very helpful.
|
I take that back, this works, I didn't upgrade the joern version in the odbv2 branch. |
Co-authored-by: Michael Pollmeier <michael@michaelpollmeier.com>
mpollmeier
approved these changes
Feb 20, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
As we saw from joernio/flatgraph#11 , we have a problem when deserializing nodes: There are too many and too heavy (ShiftLeftSecurity/overflowdb-codegen#205) NodeRefs.
This addresses the "too many" part. The issue is the following: When we deserialize a node, then we ask NodeFactory to please give us a new NodeDb object with the data. But the API helpfully also allocates a NodeRef object that is kept alive by the NodeDb!
So the pointer structure when deserializing looks like this:
This is of course nonsense. We fix this by passing pre-existing NodeRef pointers into the
createNodefunction and only allocate a newNodeRefwhen that isnull.This should save one NodeRef per node, which weighs 32 bytes (with ShiftLeftSecurity/overflowdb-codegen#205; before it was 40 bytes) .
PS. can you add some more eyes that I got the referenceManager queue / registerRef right? We never want to add a nodeRef twice, nor do we want to forget it.