[15] RFR: 8237894: CTW: C1 compilation fails with assert(x->type()->tag() == f->type()->tag()) failed: should have same type

[Jamsheed C M]

Hi Tobias,

i have made conservative fix for assert failing types. it will still be 
awkward for subtypes.

C1 doesn't remove dead codes and it even inline through even constant nulls.

http://cr.openjdk.java.net/~jcm/8237894/webrev.01/

Best regards,

Jamsheed

On 14/02/2020 01:06, Jamsheed C M wrote:
> Hi Tobias,
>
> On 13/02/2020 16:32, Tobias Hartmann wrote:
>> Looks reasonable to me. Did you verify that there is no difference in 
>> generated code for cases with
>> null loads of the same type?
>
> null loads of same type will have difference as value numbering is 
> disabled.
>
> but more importantly, i missed null constants passed locally in to 
> consideration, will do type hashing as conservative fix for this 
> special scenario of asserts
>>
>> Please update the copyright dates and the affects versions in the bug.
>
> sure.
>
> Best regards,
>
> Jamsheed
>
>
>