RFR(XS): 8219214: Infinite Loop in CodeSection::dump()
lutz.schmidt at sap.com
Wed Feb 20 21:19:10 UTC 2019
Thank you, Vladimir,
for your opinion. Let's wait and hope for comments from others.
On 20.02.19, 19:18, "Vladimir Kozlov" <vladimir.kozlov at oracle.com> wrote:
I always used debugger's builtin disassembler to look on generated code. Also decode() needs up-to-date hsdis lib.
And dumping CONST section is also could be done by debugger in more flexible ways.
I vote for removing this code (including decode_all) as it duplicates debugger functionality.
On 2/20/19 5:58 AM, Schmidt, Lutz wrote:
> Dear All,
> I would like to propose the following change which fixes an infinite loop.
> Bug: https://bugs.openjdk.java.net/browse/JDK-8219214
> Webrev: http://cr.openjdk.java.net/~lucy/webrevs/8219214.00/
> The method affected is called from nowhere inside hotspot code, but could be called via a debugger for diagnostic purposes. BUT: even that doesn’t seem to have happened in the past.
> As an alternative, I would suggest to delete the method. Please voice your opinion. I am unbiased to either solution.
More information about the hotspot-compiler-dev