832026 - Measure JSRuntime::bumpAlloc_ in the JS memory reporter.

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Nicholas Nethercote [inactive]]

We don't measure the mmaps done by JSRuntime::bumpAlloc_.

Comment 1

[Nicholas Nethercote [inactive]]

Attachment: Measure JSRuntime::bumpAlloc_ in the JS memory reporter.

This patch measures JSRuntime::bumpAlloc_.  I haven't seen it go higher than 4
KiB, but we want to have complete coverage of all the mmap calls done by
Firefox, so it's worth having just for that reason.  Also, it's possible that
it could go higher in rare cases.

Comment 2

[Sean Stangl [:sstangl]]

Comment on attachment 703602 [details] [diff] [review]
Measure JSRuntime::bumpAlloc_ in the JS memory reporter.

Review of attachment 703602 [details] [diff] [review]:
-----------------------------------------------------------------

::: js/src/jscntxt.cpp
@@ +155,5 @@
>      }
>  
> +    if (bumpAlloc_) {
> +        n += bumpAlloc_->sizeOfNonHeapData();
> +    }

uber-nit: no need for braces in single-line if()

Comment 3

[Nicholas Nethercote [inactive]]

Thanks for the fast review.


> uber-nit: no need for braces in single-line if()

Going back and forth between JS and Gecko styles is such a delight.

Comment 4

[Nicholas Nethercote [inactive]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/4f0b7a4daacf

Comment 5

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

https://hg.mozilla.org/mozilla-central/rev/4f0b7a4daacf