Fwd: Review request for JDK-8017177: more explicit code location information in hs_err crash log

[Vladimir Kozlov]

Hi Doug,

I think this is good. One thing I would change is offset value to 
hexadecimal value instead of "+%d]".

And you need review and sponsor from runtime group.

thanks,
Vladimir

On 6/20/13 7:31 AM, Albert Noll wrote:
> Begin forwarded message:
>
>> From: Doug Simon <doug.simon at oracle.com>
>> Subject: Review request for JDK-8017177: more explicit code location
>> information in hs_err crash log
>> Date: June 20, 2013 4:12:09 PM GMT+02:00
>> To: "hotspot-dev at openjdk.java.net developers"
>> <hotspot-dev at openjdk.java.net>
>>
>> When debugging compiler issues, it would be useful to know where
>> exactly in compiled code a crash happened. Currently, only the name
>> and signature of a compiled frame is shown in a crash stack trace. If
>> would be useful to know the address and offset of the crashing
>> instruction.
>>
>> For example, instead of:
>>
>> Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
>> v ~DeoptimizationBlob
>> J
>> java.util.concurrent.ConcurrentHashMap$Segment.remove(Ljava/lang/Object;ILjava/lang/Object;)Ljava/lang/Object;
>>
>>
>> one would get something like:
>>
>> Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
>> v ~DeoptimizationBlob
>> J
>> java.util.concurrent.ConcurrentHashMap$Segment.remove(Ljava/lang/Object;ILjava/lang/Object;)Ljava/lang/Object;
>> @ 0x7f502892e095 [0x7f502892e000+149]
>>
>> Potential issue: may break hs_err crash log parsers
>>
>> open webrev at http://cr.openjdk.java.net/~dnsimon/JDK-8017177/
>
>
>