Surgical fix for bad assertion generation #55626
Merged
Conversation
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
Say we have a cast like this: CAST(uint <- long). What value does this tree compute? [int.MinValue..int.MaxValue] - IR operates on signed TYP_INTs. But assertion prop generated [0..uint.MaxValue] for it. The confusion created by this "how to interpret TYP_UINT" question caused a bug where for the assertion generated for the above cast, in the form of [0..uint.MaxValue], propagation could remove a checked cast in the form of CAST_OVF(uint < int). The proper fix is to generate proper ranges for such casts. The surgical fix proposed here is to always treat casts to TYP_UINT as if they were to TYP_INT. This is conservative, but always correct. The generated assertion is useless of course, but that makes this a zero-diff change.
ghost
added
the
area-CodeGen-coreclr
|
Hello @kunalspathak! Because this pull request has the p.s. you can customize the way I help with merging this pull request, such as holding this pull request until a specific person approves. Simply @mention me (
|
|
Failures seems to be infra issues. |
jakobbotsch
reviewed
Jul 20, 2021
ghost
locked as resolved and limited conversation to collaborators
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Say we have a cast like this:
CAST(uint <- long).What value does this tree compute?
[int.MinValue..int.MaxValue]- IR operates on signedTYP_INTs.But assertion prop generated
[0..uint.MaxValue]for it.The confusion created by this "how to interpret
TYP_UINT" questioncaused a bug where for the assertion generated for the above cast,
in the form of
[0..uint.MaxValue], propagation could removea checked cast in the form of
CAST_OVF(uint < int).The proper fix is to generate proper ranges for such casts, it is in #55186 and waiting for .NET 7.
The surgical fix proposed here, for .NET 6, is to always treat casts to
TYP_UINTas if they were to
TYP_INT. This is conservative, but always correct.The generated assertion is useless of course, but that makes this a
zero-diff change (verified via replaying all
win-x64&win-x86collections).Fixes #54842
cc @kunalspathak