Обсуждение: BUG #11397: enable-profiling hangs configure at thread safety checking

Поиск
Список
Период
Сортировка

BUG #11397: enable-profiling hangs configure at thread safety checking

От
amutu@amutu.com
Дата:
The following bug has been logged on the website:

Bug reference:      11397
Logged by:          Jov
Email address:      amutu@amutu.com
PostgreSQL version: 9.3.5
Operating system:   FreeBSD r10 amd64
Description:

I configure PG with this cmd,and configure hangs:
./configure --enable-debug --enable-profiling --enable-cassert
--able-dtrace

....
checking for collateindex.pl... no
checking for xsltproc... no
checking for osx... no
checking for sgml2xml... no
checking for sx... no
checking thread safety of required library functions...
load: 0.58  cmd: conftest 25565 [umtxn] 34.63r 0.00u 0.00s 0% 3244k
load: 0.42  cmd: conftest 25565 [umtxn] 58.06r 0.00u 0.00s 0% 3244k
^[[B
load: 0.23  cmd: conftest 25565 [umtxn] 89.09r 0.00u 0.00s 0% 3244k

If I remove the enable-profiling option,the configure goes well.

some more info:
jovz ~]$ ps auxwww | grep conftest
jovz 30276   0.0  0.2 30804 3260  0  I+    9:14下午    0:00.01 ./conftest
jovz 30302   0.0  0.1 18724 2348  1  S+    9:16下午    0:00.01 grep conftest
[jovz ~]$ procstat -kk 30276
  PID    TID COMM             TDNAME           KSTACK

30276 100087 conftest         -                mi_switch+0xde
sleepq_catch_signals+0xb2 sleepq_wait_sig+0xf _sleep+0x265 umtxq_sleep+0x129
do_lock_umutex+0x64f __umtx_op_wait_umutex+0x78 amd64_syscall+0x357
Xfast_syscall+0xfb

Re: BUG #11397: enable-profiling hangs configure at thread safety checking

От
Alvaro Herrera
Дата:
amutu@amutu.com wrote:

> If I remove the enable-profiling option,the configure goes well.
>
> some more info:
> jovz ~]$ ps auxwww | grep conftest
> jovz 30276   0.0  0.2 30804 3260  0  I+    9:14下午    0:00.01 ./conftest
> jovz 30302   0.0  0.1 18724 2348  1  S+    9:16下午    0:00.01 grep conftest
> [jovz ~]$ procstat -kk 30276
>   PID    TID COMM             TDNAME           KSTACK
>
> 30276 100087 conftest         -                mi_switch+0xde
> sleepq_catch_signals+0xb2 sleepq_wait_sig+0xf _sleep+0x265 umtxq_sleep+0x129
> do_lock_umutex+0x64f __umtx_op_wait_umutex+0x78 amd64_syscall+0x357
> Xfast_syscall+0xfb

Isn't this a kernel bug, or a libc bug, or a combination thereof?

--
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services