[9] RFR (XS): 8075270: Print locals & stack slots location for PcDescs (original) (raw)

Vladimir Kozlov vladimir.kozlov at oracle.com
Tue Mar 17 17:11:12 UTC 2015

Looks good. Thank you for testing it.

Thanks, Vladimir

On 3/17/15 9:31 AM, Vladimir Ivanov wrote:

Thanks, Vladimir.

Can you look how scalarized (EA) objects are displayed? It looks fine. Simple test: static int f(int a) { Integer i = new Integer(a), j = new Integer(10); nonInlined(); return i.intValue() + j.intValue(); } PcDesc(pc=0x105a08274 offset=14 bits=0): EA::f at 19 (line 10) Locals - l0: empty - l1: obj[14] - l2: obj[15] Objects - 14: reg rbp [10],int - 15: 10 While experimenting with that, I found that: (1) "Objects" section indentation is broken; (2) tty is used where st is expected. Updated webrev in place: http://cr.openjdk.java.net/~vlivanov/8075270/webrev.00/ Best regards, Vladimir Ivanov

Thanks, Vladimir K On 3/16/15 12:26 PM, Vladimir Ivanov wrote: http://cr.openjdk.java.net/~vlivanov/8075270/webrev.00/ https://bugs.openjdk.java.net/browse/JDK-8075270

Enhance -XX:+PrintNMethods output with additional info about locals & stack slots location (if present): PcDesc(pc=0x110b71b5c offset=17c bits=0): java.lang.String::indexOf at 43 (line 1562) Locals - l0: empty - l1: stack[0] - l2: empty - l3: reg rbp [10],int - l4: stack[8],narrowoop - l5: stack[4] Right now it looks like: PcDesc(pc=0x110b71b5c offset=17c bits=0): java.lang.String::indexOf at 43 (line 1562) Testing: manual. Thanks! Best regards, Vladimir Ivanov

More information about the hotspot-compiler-dev mailing list