Fix DFS tree construction #31129
Merged
Fix DFS tree construction #31129
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
It's called a Depth-First tree after all, not a dept-first-but-feel-free-to-record-any-of-the-parents tree. In particular, in order for the semi-dominator condition to hold, the parent in the DFS tree needs to be the predecessor with the highest DFS number. Here we were accidentally doing the opposite casuing us to look at the wrong node in case the sdom and the idom are not the same. To understand #31121, consider the following CFG (minimized from the bug report to show the issue), as well as the corresponding DFS numbering and sdom assignment ``` CFG DFS sdom A 1 0 | \ | \ |\ B C 2 5 1 1 /|/ /|/ /|/ | D | 3 | 1 \| \| \| E 4 2 ``` This bug caused us to record the parent of `E` as `B`, when it should have been `D` (the relevant invariant here is that the parent in the DFS tree is the predecessor with the highest DFS number). As a result, when computing idoms from the sdoms, we were incorrectly looking at `B`, seeing that the sdom matched the ancestor in the DFS tree and thus concluding that `E`'s idom was `B` rather than `A`. As a result, we neglected to insert a phi node in `E`. Fixes #31121
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Outdated suggestions cannot be applied.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.
It's called a Depth-First tree after all, not a
dept-first-but-feel-free-to-record-any-of-the-parents tree. In
particular, in order for the semi-dominator condition to hold,
the parent in the DFS tree needs to be the predecessor with the
highest DFS number. Here we were accidentally doing the opposite
casuing us to look at the wrong node in case the sdom and the idom
are not the same. To understand #31121, consider the following
CFG (minimized from the bug report to show the issue), as well as
the corresponding DFS numbering and sdom assignment
This bug caused us to record the parent of
EasB, when it shouldhave been
D(the relevant invariant here is that the parent in the DFStree is the predecessor with the highest DFS number).
As a result, when computing idoms from the sdoms, we
were incorrectly looking at
B, seeing that the sdom matched theancestor in the DFS tree and thus concluding that
E's idom wasBrather than
A. As a result, we neglected to insert a phi node inE.Fixes #31121