Fix compiler crash in memset optimisation #64084
Closed
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
NOT converts 0 to 0xFFFFFFFF, which is not what we want. Also this commit removes setge and setuge optimisation which is never actually selected.
Note that halfword load and store are not being selected at the moment.
- Fix instructions conflicts between NanoMips and other Mips targets that caused tests from the MC/Mips group to fail - Fix several tests - Fix debug location of SP adjustment instruction that caused DebugInfo/Mips/delay-slot.ll test to fail - Add 'nanomips' feature to llvm-lit config script so that tests can be marked as UNSUPPORTED for NanoMips target
- Fix expected output for jumptable.ll and mips_generated_funcs.ll tests
- No longer need to mark a test ExtDebugInfo.cpp that needs object emitter as UNSUPPORTED - Delete line left behind by a merge conflict in sanitize-coverage-old-pm.c test
This allows the INS instruction to be used to set the value of one field of the bitfields struct.
* Add LTO section predicting & reporting tool.
The macro nature of these instructions makes them slower than their individual instruction equivalents, so only use them when optimising for size.
Merging volatile accesses into lwm/swm can change the order of accesses and may have other undesirable effects.
Avoid swm/lwm of volatile accesses.
Only use LWM/SWM when optimising for size.
Make sure that we do not use unsupported features by NanoMips
…euristics Improve MemCopyOpt heuristics to account for combinable stores
Added isBarier flag to definition of BRSC_NM. Suppressed warnings during build: - Reordered NanoMipsJumpTableInfo constructor parameters. - Added override identifier to use useIPRA.
Allow OR to INS combine without second register masking only if maximum possible value given by KnownBits for the second register can fit inside SMSize0.
NanoMips: OR combine optimization
Fix brsc and suppress warnings
|
This repository does not accept pull requests. Please follow http://llvm.org/docs/Contributing.html#how-to-submit-a-patch for contribution to LLVM. |
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.
The legacy
getDestAlignmentmember used to get alignment for existingmemsetintrinsics can use a value of 0 to represent undefined alignment, which will cause the alignment offset calculation to fail as it uses this as the denominator of a modulo operation.