Performance Tweaks#195
Merged
Merged
Conversation
After a bit of profiling I noticed that creating TaintOperations is very expensive. In hindsight this is obvious, as we have to walk the stack to get the location etc. We did however create a lot of TaintOperation objects without ever using them, i.e., due to the string itself being untainted. This change does add explicit checks where possible (in some cases we create the TaintOperation object in advance due to GC issues, this is not easily hidden behind a check) so TaintOperations are only created if the string is actually tainted, i.e., we use the TaintOperation to extend the TaintFlow.
Contributor
|
Hi @leeN, there are two failing tests, could you take a look at them please? |
Collaborator
Author
|
Ah, yes, sorry, I already had a look there and will fix them tomorrow. They used to fail with foxhound in the past, so we changed them to make them pass. This PR unbreaks them again. :) |
These tests used to be broken in Foxhound and were fixed in SAP#151. This fix was intended to mask the fact that these tests were picking up and failing due to foxhound induced side effects on untainted strings. As we call toString() on objects during TaintOperation creation, the usage of Foxhound can have visible side effects. This is shown here as these tests count how often toString() is called on an object. Now, in this branch we do not create any TaintOperation objects for these specific cases of untainted strings anymore. Consequently, the tests are now passing in their original form.
Collaborator
Author
|
Fixed. |
tmbrbr
reviewed
Jan 31, 2024
tmbrbr
approved these changes
Jan 31, 2024
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
After a bit of profiling, I noticed that creating TaintOperations is very expensive. In hindsight, this is obvious, as we have to walk the stack to get the location, etc.
We did, however, create a lot of TaintOperation objects without ever using them, i.e., due to the string itself being untainted.
This change does add explicit checks where possible (in some cases, we create the TaintOperation object in advance due to GC issues; this is not easily hidden behind a check), so TaintOperations are only created if the string is actually tainted, i.e., we use the TaintOperation to extend the TaintFlow.
I tested this on Ares6, and with the proposed changes, the average runtime goes from 70.04ms to 66.9ms, mainly due to differences in the ML benchmark.