RFR (XXS): 8027364: PSScavenge accounts too large code section to StringTable unlink [Was: Re: Trivial bug in psScavenge.cpp?]

[Srinivas Ramakrishna]

Looks good; thx for reviewing other uses as well (and for finding the
parallel compact issue of not splitting the components apart).

-- ramki


On Mon, Oct 28, 2013 at 2:37 AM, Thomas Schatzl
<thomas.schatzl at oracle.com>wrote:

> Hi all,
>
>   can I have reviews for the following small change: it fixes the scope
> of some GCTraceTime timer, i.e. the braces that limited the scope to
> string table processing were missing.
>
> The change adds them.
>
> Bug:
> https://bugs.openjdk.java.net/browse/JDK-8027364
>
> Webrev:
> http://cr.openjdk.java.net/~tschatzl/8027364/webrev/
>
> Testing:
> jprt, compilation
>
> I had a look at other GCTraceTime uses in the code, and this seems to be
> the only case where the scope seems wrong.
>
> There is another place which could be considered buggy, that is in
> psParallelCompact.cpp, line 2114 where the GCTraceTime("class
> unloading"...) covers unloading, string table and symbol table
> unlinking. Other collectors seperate them out. I will probably create a
> separate CR for that if others agree that this is an issue.
>
> @Jeremy: We decided to try to get this fix in asap, possibly before we
> need to ask for exceptions to push, so I took over creating the patch.
> Jeremy, I hope you don't mind :) I will add the proper contribution.
>
> Thomas
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20131028/7b0f7519/attachment.htm>