1215337 - Cache slotSpan()

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Steve Fink [:sfink] [:s:]]

While single-stepping in gdb, I found a place where slotSpan() gets called for every iteration through a loop.

Comment 1

[Steve Fink [:sfink] [:s:]]

Attachment: Cache spotSpan()

Yes, I should time this and things. But I don't want to spend the time, and the resulting code is at least as easy to follow.

Comment 2

[Terrence Cole [:terrence]]

Comment on attachment 8674579 [details] [diff] [review]
Cache spotSpan()

Review of attachment 8674579 [details] [diff] [review]:
-----------------------------------------------------------------

::: js/src/jsobj.cpp
@@ +3703,5 @@
>          {
>              GetObjectSlotNameFunctor func(nobj);
>              JS::AutoTracingDetails ctx(trc, func);
>              JS::AutoTracingIndex index(trc);
> +            uint32_t nslots = nobj->slotSpan(); // cache

Might as well use |const uint32_t|, since that's kinda what you're aiming to show the compiler.

Also, "cache" isn't really that clear. How about a full comment like: // Since tracing can mutate the target, the compiler will not generally be able to prove that slotSpan is constant across the loop.

Comment 3

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/f41fd6c7d522

Comment 4

[Phil Ringnalda (:philor)]

Backed out in https://hg.mozilla.org/integration/mozilla-inbound/rev/1266f78145c4 along with bug 1215336.

Comment 5

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/0fbd23348147

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/0fbd23348147

Comment 7

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/0fbd23348147