LLVM Bugzilla is read-only and represents the historical archive of all LLVM issues filled before November 26, 2021. Use github to submit LLVM bugs

Bug 48362 - opt -jump-threading hangs
Summary: opt -jump-threading hangs
Status: RESOLVED FIXED
Alias: None
Product: new-bugs
Classification: Unclassified
Component: new bugs (show other bugs)
Version: unspecified
Hardware: PC Linux
: P enhancement
Assignee: Unassigned LLVM Bugs
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2020-12-02 06:42 PST by Mikael Holmén
Modified: 2020-12-02 23:50 PST (History)
3 users (show)

See Also:
Fixed By Commit(s): 9d6d24c25

Attachments
bbi-50589_reduced.ll reproducer (3.25 KB, text/plain)
2020-12-02 06:42 PST, Mikael Holmén 		Details
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Mikael Holmén 2020-12-02 06:42:25 PST 
Created attachment 24227 [details]
bbi-50589_reduced.ll reproducer

Reproduce with:
 opt -o - bbi-50589_reduced.ll -S -jump-threading 

It doesn't seem to terminate.

This starts happening after 5486e00dc3e3:

    [InstSimplify] remove poison-unsafe insertelement of undef value
    
    PR45481:
    https://bugs.llvm.org/show_bug.cgi?id=45481
    
    SDAG has an identical transform to this, so there's little
    chance of any real-world impact. OTOH, that means we are
    effectively sweeping the bug out of sight because poison
    exists in codegen too.
Comment 1 Sanjay Patel 2020-12-02 08:39:21 PST 
In the debugger, this is getting stuck in SimplifyExtractElementInst() -> llvm::findScalarElement().
Comment 2 Sanjay Patel 2020-12-02 10:48:56 PST 
I suspect that we could have hit this bug with a less-undef test case even before 5486e00dc3e3.

I don't know much about jumpthreading, so I put in a lower-level exit in the analysis that was getting tripped up:
https://reviews.llvm.org/rG9d6d24c25056
Comment 3 Mikael Holmén 2020-12-02 23:50:54 PST 
Thanks!