assert disabling and enabling network card on linux

[mdionisio]

I have the following stack trace:

Program received signal SIGABRT, Aborted.
[Switching to Thread 0x44670450 (LWP 243)]
0x40406460 in __GI_raise (sig=sig@entry=6)
at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
56 return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig);
(gdb) bt
#0 0x40406460 in __GI_raise (sig=sig@entry=6)

at ../nptl/sysdeps/unix/sysv/linux/raise.c:56

#1 0x40409e88 in GI_abort () at abort.c:89
#2 0x403ad728 in zmq::zmq_abort (errmsg=) at err.cpp:76
#3 0x403c0000 in read (this=, data=,

size_=<optimized out>) at stream_engine.cpp:566

#4 zmq::stream_engine_t::read (this=, data_=,

size_=<optimized out>) at stream_engine.cpp:523

#5 0x403c070c in zmq::stream_engine_t::in_event (this=0x41b14000)

at stream_engine.cpp:197

#6 0x403ad1fc in zmq::epoll_t::loop (this=0x166370) at epoll.cpp:154
#7 0x403c41ec in thread_routine (arg_=0x1663b8) at thread.cpp:83
#8 0x4004efc4 in start_thread (arg=0x44670450) at pthread_create.c:314
#9 0x404a7b30 in ?? () at ../ports/sysdeps/unix/sysv/linux/arm/clone.S:97

from /lib/libc.so.6
#10 0x404a7b30 in ?? () at ../ports/sysdeps/unix/sysv/linux/arm/clone.S:97

from /lib/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Simply removing the errno_assert in

int zmq::stream_engine_t::read (void *data_, size_t size_)

the problem is solved.

[hintjens]

It'd be better to catch specific errors. Can you find out what the errno is
that that point?

On Thu, May 14, 2015 at 12:45 PM, Michele Dionisio <notifications@github.com

wrote:

I have the following stack trace:

Program received signal SIGABRT, Aborted.
[Switching to Thread 0x44670450 (LWP 243)]
0x40406460 in GI_raise (sig=sig@entry=6)
at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
56 return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig);
(gdb) bt
#0 0x40406460 in GI_raise (sig=sig@entry=6)
at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
#1 #1 0x40409e88 in GI_abort
() at abort.c:89
#2 #2 0x403ad728 in
zmq::zmq_abort (errmsg=) at err.cpp:76
#3 #3 0x403c0000 in read (this=,
data=,
size=) at stream_engine.cpp:566
#4 #4 zmq::stream_engine_t::read
(this=, data=,
size=) at stream_engine.cpp:523
#5 #5 0x403c070c in
zmq::stream_engine_t::in_event (this=0x41b14000)
at stream_engine.cpp:197
#6 #6 0x403ad1fc in
zmq::epoll_t::loop (this=0x166370) at epoll.cpp:154
#7 #7 0x403c41ec in
thread_routine (arg=0x1663b8) at thread.cpp:83
#8 #8 0x4004efc4 in start_thread
(arg=0x44670450) at pthread_create.c:314
#9 #9 0x404a7b30 in ?? () at
../ports/sysdeps/unix/sysv/linux/arm/clone.S:97
from /lib/libc.so.6
#10 #10 0x404a7b30 in ?? () at
../ports/sysdeps/unix/sysv/linux/arm/clone.S:97
from /lib/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Simply removing the errno_assert in

int zmq::stream_engine_t::read (void *data_, size_t size_)

the problem is solved.

—
Reply to this email directly or view it on GitHub
#1399.

[mdionisio]

I'm testing it on an embedded device with linux 2.6.37. I have no access to stderr, so I have to build a special version of zmq that log that error on file, so I need few time to answer to your question.

It is possible to insert a compile option to:

1. add compile option to add syslog print on any assert ?
2. add a compile option to remove assert on any error check ?

[hintjens]

You can't remove asserts, these are always for cases where the code does
not know how to continue. What you're hitting here is an unknown errno that
we have to deal with... simplest is to add a printf of errno before the
assert.

On Sat, May 16, 2015 at 11:09 AM, Michele Dionisio <notifications@github.com

wrote:

I'm testing it on an embedded device with linux 2.6.37. I have no access
to stderr, so I have to build a special version of zmq that log that error
on file, so I need few time to answer to your question.

It is possible to insert a compile option to:

1. add compile option to add syslog print on any assert ?
2. add a compile option to remove assert on any error check ?

—
Reply to this email directly or view it on GitHub
#1399 (comment).

[mdionisio]

the errno is EINVAL = 22

[hintjens]

Michele,

Could you retest on libzmq master and tell us the stack backtrace if it's
still crashing?

Thanks

On Fri, Oct 30, 2015 at 2:06 PM, Michele Dionisio notifications@github.com
wrote:

the errno is EINVAL = 22

—
Reply to this email directly or view it on GitHub
#1399 (comment).

[mdionisio]

On libzmq 4.1.3 there is the same issue and the same fix (but on file src/tcp.cpp) solve the problem.

patch:

--- a/src/tcp.cpp
+++ b/src/tcp.cpp
@@ -192,7 +192,7 @@ int zmq::tcp_write (fd_t s_, const void *data_, size_t size_)
                    && errno != EBADF
                    && errno != EDESTADDRREQ
                    && errno != EFAULT
-                   && errno != EINVAL
+//                   && errno != EINVAL    
                    && errno != EISCONN
                    && errno != EMSGSIZE
                    && errno != ENOMEM
@@ -241,7 +241,7 @@ int zmq::tcp_read (fd_t s_, void *data_, size_t size_)
    if (rc == -1) {
        errno_assert (errno != EBADF
                    && errno != EFAULT
-                   && errno != EINVAL
+//                 && errno != EINVAL     
                    && errno != ENOMEM
                    && errno != ENOTSOCK);
        if (errno == EWOULDBLOCK || errno == EINTR)

[mdionisio]

I'm reading old isses and I think my one is duplicated of #829. I'm using PUB-SUB pattern but the result is the same.

[hintjens]

OK, fixed on master and backported to 4-1 stable. Thanks!

On Sat, Oct 31, 2015 at 6:19 PM, Michele Dionisio notifications@github.com
wrote:

I'm reading old isses and I think my one is duplicated of #829
#829. I'm using PUB-SUB pattern
but the result is the same.

—
Reply to this email directly or view it on GitHub
#1399 (comment).