Tom Marchant wrote:

>Why not make your STC run AMODE 64, at least part of the time? It is easy
>to switch modes, but you do have to be careful that those programs that run
>AMODE 64 save and restore the 64-bit registers. The same with your JNI. It
>doesn't have to be "Full 64-bit".

 

Yeah, that's an option, but it's a lot more work and has other possible
impacts. Given the relatively low uptake of Java on z, we'd rather isolate
the changes to Java-land.

 

Andrew Rowley suggested:

>NewDirectByteBuffer looks like it would allow you to allocate buffers from
the JNI side (and therefore control the location in storage) and use them
from Java. I don't know what the performance implications are.

Not sure this helps, as the Java side has no relationship with the JNI other
than calling it atomically. This seems more like it's for use in a
cooperative/persistent environment, where the Java side is going to know
more about the JNI side.

 

But I'm no Java programmer (I don't even drink coffee), so maybe? Thanks.


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to