RFR (XS): 8171155: Scanning method file for initialized final field updates can fail for non-existent fields
zoltan.majo at oracle.com
Tue Dec 13 14:02:37 UTC 2016
Thank you, Vladimir, for the review!
On 12/13/2016 02:47 PM, Vladimir Ivanov wrote:
> Looks good.
> Best regards,
> Vladimir Ivanov
> On 12/13/16 2:35 PM, Zoltán Majó wrote:
>> please review the fix for 8171155.
>> The JVM assumes the field it's looking for always exists. As a result,
>> the JVM processes uninitialized data if the field does not exists, which
>> results in a crash. The suggested fix is to check if the field exists
>> before processing it. Please find more details in the bug description.
>> I found the bug while JPRT-testing the 8u-backport of
>> 8157181: Compilers accept modification of final fields outside
>> initializer methods
>> I was not able to reproduce the issue with 9, but (judging from the
>> code) JDK 9 is likely to be affected as well. So I'd prefer to fix this
>> problem in JDK 9 as well (and then backport the fix to 8u) instead of
>> pushing the fix only into 8u.
>> Thank you!
>> Best regards,
More information about the hotspot-dev