Open Bug 1483345 Opened 5 years ago Updated 1 year ago

Try and re-attach NewObject stubs after GC

Categories

(Core :: JavaScript Engine: JIT, enhancement, P3)

Product:
Core
Component:
JavaScript Engine: JIT
Type:
enhancement

Tracking

()

Project Flags:
Performance Impact medium
Tracking Flags:
Tracking Status
firefox63 --- affected

People

(Reporter: mgaudet, Unassigned)

References

(Blocks 1 open bug)

Details

In Bug 1480524, Comment 16 it's pointed out that the design of the NewObject stub is such that if the GC discards the cache stubs, we subsequently will never try to attach a new stub because we'll still have a template around.
Priority: -- → P3
Might be an easy fix for a minor QF win. Baseline optimizations of new Objects may stop working under certain GC timings. In profiles we'd be looking for DoNewObject (which should have been called DoNewObjectFallback like the rest of the IC fallbacks).
Blocks: CacheIR
Whiteboard: [qf]
Some nits to fix at the same time:
 - s/DoNewObject/DoNewObjectWithTemplate
 - Remove unused NewObject_WithTemplate
Whiteboard: [qf] → [qf:p2]
Performance Impact: --- → P2
Whiteboard: [qf:p2]
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.