RFR: 8262017: C2: assert(n != __null) failed: Bad immediate dominator info. [v2]

[Christian Hagedorn]

On Wed, 9 Jun 2021 07:54:14 GMT, Roland Westrelin <roland at openjdk.org> wrote:

>> Christian Hagedorn has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Add case for positive stride*scale
>
> Roughly, the problem as I understand it from your description/patch is that: new bounds are computed using long arithmetic which causes bounds to not fit in an int and to be incorrect once the bounds are casted to an int. Is that it? Can you give some more details?
> 
> So the previous patch dropped some overflow checks under the assumption, I suppose, that there was no need for them because of the move to long arithmetic and now we're finding that it's doesn't really work that way. Was long arithmetic the right move then? With the previous code added back, does it still solve problems? Or was that approach fundamentally flawed?

Thanks @rwestrel and @vnkozlov for your reviews! I'm closing this PR and will reopen a new one for JDK 17.

-------------

PR: https://git.openjdk.java.net/jdk/pull/4362