RFR (S): 8000313: C2 should use jlong for 64bit values

Vladimir Ivanov vladimir.x.ivanov at oracle.com
Mon Oct 8 13:06:58 PDT 2012


Thanks for the review.

As Vladimir K. already noted, JPRT does some amount of testing on all 
platforms. Regarding CTW, I decided to run it on a single 64-bit 
platform, because the changes are in shared code.

Best regards,
Vladimir Ivanov

On 10/8/12 10:07 PM, Azeem Jiva wrote:
> Looks good.  Did you run on 64bit SPARC as well?
> On 10/08/2012 12:33 PM, Vladimir Ivanov wrote:
>> http://cr.openjdk.java.net/~vlivanov/8000313/webrev.00/
>> 25 lines changed: 0 ins; 0 del; 25 mod
>> This is a code cleanup. It replaces all occurrences of long with jlong
>> in C2 code to guarantee that 64-bit value is always used (it's not the
>> case for type long) and no overflow in subsequent arithmetic operations
>> occurs.
>> Testing: JPRT, full CTW on solaris_x64
>> Best regards,
>> Vladimir Ivanov

More information about the hotspot-compiler-dev mailing list