[erlang-questions] Erlang segfault ?? OMG

[Lukas Larsson]

On Mon, Sep 22, 2014 at 10:19 AM, 卓一抗 <zhuoyikang@REDACTED> wrote:

>
> * From the place where the error happens I assume that this is during
> start up. Is that a correct assumption? =>
>
> it's not start up , my program run few hours and get the segment fault.
>

Then the error did not happen in erts_bfalc_init as that is only called
during startup before any schedulers are even started.


>
> * Can you get a full core file?
>
> => the core file is too large , more 2G , it not write successfully..
>
>
It should be possible to get quite a lot of information out of a partial
core. Have you tried just setting ulimit -c to 1 GB and then check if you
can get any info from that core?


>
>
> * Are you running any non-Erlang/OTP nifs/linked-in drivers?
>
> not nifs.... competely erlang/otp
>

Are you 100% sure? No third party libs use nifs/linked-in drivers? This
type of error almost always turn out to be nif/linked-in driver related.


>
> 2014-09-22 16:06 GMT+08:00 Lukas Larsson <lukas@REDACTED>:
>
>> Hello,
>>
>> This is not a known issue.
>>
>> You will have to provide a little more information in order for me to
>> understand what is going on.
>>
>> * From the place where the error happens I assume that this is during
>> start up. Is that a correct assumption?
>> * Is it possible to reproduce?
>> * Can you get a full core file?
>> * What OS are you using?
>> * How did you install Erlang/OTP?
>> * Are you running any non-Erlang/OTP nifs/linked-in drivers?
>>
>> Lukas
>>
>> On Sun, Sep 21, 2014 at 4:38 PM, 卓一抗 <zhuoyikang@REDACTED> wrote:
>>
>>> hello，everybody :
>>>
>>> my erlang node crash and dmesg report:
>>>
>>> segfault at 18 ip 00000000004565f4 sp 00007fffe9178698 error 4 in
>>> beam[400000+1e9000]
>>>
>>> and i objdump -d beam to track this code :
>>>
>>>   456597:       74 5b                   je     4565f4
>>> <erts_bfalc_init+0x4a4>
>>>   456599:       48 8b 71 08             mov    0x8(%rcx),%rsi
>>>   45659d:       40 f6 c6 02             test   $0x2,%sil
>>>   4565a1:       74 51                   je     4565f4
>>> <erts_bfalc_init+0x4a4>
>>>   4565f4:       48 8b 41 18             mov    0x18(%rcx),%rax
>>>
>>>  so simple code =>
>>>
>>>   void
>>>   erts_bfalc_init(void)
>>>   {
>>>   atoms_initialized = 0;
>>>   }
>>>   .//erts/emulator/beam/erl_bestfit_alloc.h:39:void
>>> erts_bfalc_init(void);
>>>
>>>
>>> does it a bug on erlang rb15b02?  and  anyone occur the same question
>>> like me ?
>>>
>>> I need some help.....
>>>
>>>
>>> _______________________________________________
>>> erlang-questions mailing list
>>> erlang-questions@REDACTED
>>> http://erlang.org/mailman/listinfo/erlang-questions
>>>
>>>
>>
>
> _______________________________________________
> erlang-questions mailing list
> erlang-questions@REDACTED
> http://erlang.org/mailman/listinfo/erlang-questions
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20140922/f45f508b/attachment.htm>