musl mips 64 help (fwd)

musl mips 64 help (fwd)

[Rich Felker]

I'm forwarding this to the list in case anyone else has ideas:

----- Forwarded message from Mahesh Bodapati <maheshbodapati90@gmail.com> -----

Date: Tue, 16 Feb 2016 18:40:23 +0530
Message-ID: <CAKd0kDLPFnFVO=y1yoHcBZVyrycbS1H4YvxuVswEMp7zhaLGMw@mail.gmail.com>
Subject: musl mips 64 help
From: Mahesh Bodapati <maheshbodapati90@gmail.com>
To: Rich Felker <dalias@libc.org>

Hi Rich,

We are getting FAIL with musl mips n64 dynamic build libraries and we
didn’t see issue with musl mips n64 static build libraries.



FAIL ./src/regression/pthread_cond-smasher.exe [status 1]



$ gdb ./src/regression/pthread_cond-smasher.exe





   x136
settimeout(&ts);


B+ x137           int ret = condition_timedwait(&cond_main, &mut[phase],
&ts);


   x138           tell("main seeing %u threads in phase %u, %s\n",
inside[phase], phase,
errorstring(ret));


B+>x139           if (ret) {


   x140             t_error("main thread in phase %u (%u threads inside),
finished waiting: %s\n", phase, inside[phase],
errorstring(ret));


   x141             return t_status;


   x142
}


   x143         }





(gdb) p ret

*$1 = 145*

(gdb) n

*(gdb) n*

*src/regression/pthread_cond-smasher.c:140: main thread in phase 0 (0
threads inside), finished waiting: Operation timed out*

(gdb)





What may be the cause for getting timeout with dynamic build which is
passing with static build.

----- End forwarded message -----

Re: musl mips 64 help (fwd)

[Szabolcs Nagy]

* Rich Felker <dalias@libc.org> [2016-02-16 15:32:34 -0500]:
> I'm forwarding this to the list in case anyone else has ideas:
> 
> ----- Forwarded message from Mahesh Bodapati <maheshbodapati90@gmail.com> -----
> 
> Date: Tue, 16 Feb 2016 18:40:23 +0530
> Message-ID: <CAKd0kDLPFnFVO=y1yoHcBZVyrycbS1H4YvxuVswEMp7zhaLGMw@mail.gmail.com>
> Subject: musl mips 64 help
> From: Mahesh Bodapati <maheshbodapati90@gmail.com>
> To: Rich Felker <dalias@libc.org>
> 
> Hi Rich,
> 
> We are getting FAIL with musl mips n64 dynamic build libraries and we
> didn???t see issue with musl mips n64 static build libraries.
> 

i don't know what might be the cause
(this test was written by Jens Gustedt, 10 child
threads and one main thread communicate with
condvars for 10 iterations)

i changed the test to use a 0.5s timeout (so in case
of failure the test finishes instead of deadlock),
if you run it in an emulator that may not be enough.
(i should probably revert that)

the test can print trace output if you compile
it with -DDEBUG, you may also try to increase
the timeout.