Andrey,
>> How to have sure gcc take the option via configure : configure --disable-optimizations because it's not work for me with gdb?
>No options involved, --disable-optimizations simply doesn't strip debug info (as gcc usually does with -O3). Unless gcc on aix does something weird, that should just happen by itself.
But i have symbols in the binary.
# file fio
fio: executable (RISC System/6000) or object module not stripped
#
if i check with nm:
# nm fio
:T1000=eIO_LOG_TYPE_LAT:1,IO_LOG_TYPE_CLAT:2,IO_LOG_TYPE_SLAT:3,IO_LOG_TYPE_BW:4,IO_LOG_TYPE_IOPS:5,; - 0
:T1001=eIO_LOG_TYPE_LAT:1,IO_LOG_TYPE_CLAT:2,IO_LOG_TYPE_SLAT:3,IO_LOG_TYPE_BW:4,IO_LOG_TYPE_IOPS:5,; - 0
:T1001=eIO_LOG_TYPE_LAT:1,IO_LOG_TYPE_CLAT:2,IO_LOG_TYPE_SLAT:3,IO_LOG_TYPE_BW:4,IO_LOG_TYPE_IOPS:5,; - 0
:T1002=eIP_F_ONRB:1,IP_F_ONLIST:2,IP_F_TRIMMED:4,IP_F_IN_FLIGHT:8,; - 0
...
> You reported 128 threads in the original problem report. Is the below a differnt one?
When i launch fio with: HDISK=hdisk6 RWTYPE=rw RRATIO=0 BS=65536 QD=128 NJ=128 /home/padmin/fio/fio-5f3bd0f/fio myfio.fio
Form unix with ps, i have 129 fio process running. When fio is stucking, it staying alive only 1 the main process that fork the 128 process.
>> (dbx) thread
>> thread state-k wchan state-u k-tid mode held scope function
>> >$t1 run running 84803767 k no sys _p_nsleep
>> $t2 run running 110231691 u no sys __ioctl
> Thread 2 sleeps in the ioctl that the trace reports as failed. May be this is the root cause behind your issue?
Perhaps but i don't know how to progress againt it?
Regards,
Thierry--
To unsubscribe from this list: send the line "unsubscribe fio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
