SIGABORT on linux-sparc

[Sébastien Bernard]

Gary Benson a écrit :
> Dennis Gilmore wrote:
>   
>> On Tuesday 08 April 2008, S�bastien Bernard wrote:
>>     
>>> Gary Benson a �crit :
>>>       
>>>> Bernard S�bastien wrote:
>>>>         
>>>>> 2nd: the sparc32 build goes all the way but any run led to a
>>>>> SIGABORT.  As show at the end.
>>>>>
>>>>> I'm putting the hs_err_pid1983.log in attachment produced by
>>>>> the latest build of the jdk.
>>>>>           
>>>> Ah, the code that produces the hs_err_pid*.log files calls abort(),
>>>> so the SIGABORT is possibly a side-effect of that.
>>>>
>>>>         
>>>>> #  Internal Error (os_linux_zero.cpp:144), pid=1983, tid=1887509392
>>>>> #  Error: caught unhandled signal 10
>>>>>           
>>>> Is signal 10 USR1 on your machine?
>>>>         
>>> No, signal 10 is SIGBUS.
>>>       
>> Which means that memory access in not aligned
>>     
>
> Interesting.  Do you know what it was trying to access?
>
> Cheers,
> Gary
>
>   
See my previous mail.
The sigbus happens in 
icedtea6/openjdk-ecj/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp 
line 356.

The line is :

    case T_LONG: 
      SET_LOCALS_LONG(*(jlong *) result, 0);
      break;

The SET_LOCALS_LONG is breaking the access.
The value of result according to gdb is  (long long int *) 0xf781c420