[erlang-questions] Segmentation fault with lcnt compiled beam and +A

[Björn-Egil Dahlberg]

On 2012-08-30 16:42, Björn-Egil Dahlberg wrote:
> Hi!
>
> Thanks for reporting this! I'll try to look into it.
I couldn't reproduce the error on current OTP_R15B02-maint branch but 
not I could reproduce the error in OTP_R15B01. Hopefully the problem is 
already fixed =)

Give me a shout if you see it in current maint branch as well.

// Björn-Egil
>
> // Björn-Egil
>
> On 2012-08-30 11:41, Gleb Peregud wrote:
>> Hello folks
>>
>> I have encountered the following segmentation fault. It happens when I
>> combine lock-counter-compiled binary and "+A N" where N >= 64. But it
>> works with up to 63 async threads. Erlang compiled without lcnt works
>> on the same machine without a problem. The stacktrace is the
>> following:
>>
>> $ gdb /opt/erlang/r15b01_lcnt/erts-5.9.1/bin/beam.smp core
>> Copyright stuff ... snip...
>> This GDB was configured as "x86_64-linux-gnu".
>> For bug reporting instructions, please see:
>> <http://bugs.launchpad.net/gdb-linaro/>...
>> Reading symbols from 
>> /opt/erlang/r15b01_lcnt/erts-5.9.1/bin/beam.smp...done.
>> [New LWP 25517]
>> [New LWP 25483]
>> ...snip...
>> [New LWP 25463]
>> [New LWP 25466]
>>
>> warning: Can't read pathname for load map: Input/output error.
>> [Thread debugging using libthread_db enabled]
>> Using host libthread_db library 
>> "/lib/x86_64-linux-gnu/libthread_db.so.1".
>> Core was generated by `/opt/erlang/r15b01_lcnt/erts-5.9.1/bin/beam.smp
>> -A 64 -- -root /opt/erlang/r15b'.
>> Program terminated with signal 11, Segmentation fault.
>> #0  erts_lcnt_lock (lock=<optimized out>) at beam/erl_lock_count.c:445
>> 445        eltd->lock_in_conflict = 0;
>> (gdb) bt
>> #0  erts_lcnt_lock (lock=<optimized out>) at beam/erl_lock_count.c:445
>> #1  0x0000000000454096 in erts_mtx_lock_x (line=3480, mtx=0x87b9d8,
>> file=<optimized out>)
>>      at beam/erl_threads.h:1342
>> #2  erts_alcu_alloc_thr_pref (type=92, extra=<optimized out>,
>> size=1664) at beam/erl_alloc_util.c:3480
>> #3  0x00000000005c81f4 in ts_event_pool (size=25,
>> endpp=0x7f00217f9e88) at common/ethr_aux.c:351
>> #4  0x00000000005c8306 in ts_event_alloc () at common/ethr_aux.c:394
>> #5  0x00000000005c86a5 in ethr_make_ts_event__ (tsepp=0x7f00217f9ec0)
>> at common/ethr_aux.c:421
>> #6  0x00000000005cd9ce in thr_wrapper (vtwd=<optimized out>) at
>> pthread/ethread.c:91
>> #7  0x00007f01f69c7e9a in start_thread () from
>> /lib/x86_64-linux-gnu/libpthread.so.0
>> #8  0x00007f01f64ed4bd in clone () from /lib/x86_64-linux-gnu/libc.so.6
>> #9  0x0000000000000000 in ?? ()
>> (gdb)
>>
>> $ uname -a
>> Linux lb4 3.2.0-24-generic #37-Ubuntu SMP Wed Apr 25 08:43:22 UTC 2012
>> x86_64 x86_64 x86_64 GNU/Linux
>> $ lsb_release -a
>> No LSB modules are available.
>> Distributor ID:    Ubuntu
>> Description:    Ubuntu 12.04 LTS
>> Release:    12.04
>> Codename:    precise
>>
>> Erlang was configured via Kerl using the following configure options:
>> "--enable-hipe --enable-smp-support --enable-threads
>> --enable-kernel-poll --enable-lock-counter".
>>
>> I can reproduce this bug on 2-core a 32 bit machine with the same
>> Ubuntu, but on it it fails if there are more than 93 async threads:
>>
>> gleber@REDACTED:~$ erl -lcnt +A 94
>> Segmentation fault
>> gleber@REDACTED:~$ erl -lcnt +A 93
>> Erlang R15B01 (erts-5.9.1) [source] [smp:2:2] [async-threads:93]
>> [hipe] [kernel-poll:false] [lock-counting]
>>
>> Eshell V5.9.1  (abort with ^G)
>> 1>
>>
>> Have anyone encountered similar problem? Can I provide any more
>> information helpful for pinpointing the issue?
>>
>> Best regards,
>> Gleb Peregud
>> _______________________________________________
>> erlang-questions mailing list
>> erlang-questions@REDACTED
>> http://erlang.org/mailman/listinfo/erlang-questions
>>
>>
>