tcpbench core dumps which issued with ctrl and 4

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

tcpbench core dumps which issued with ctrl and 4

Sevan / Venture37-2
Hi,
As the subject says, if you run tcpbench on OpenBSD & press ctrl and 4
on keyboard, tcpbench with coredump. Tested on -current snapshots form
this month.
Pointing gdb to the core file reports
#0  0x07915849 in kevent () at <stdin>:2
2       <stdin>: No such file or directory.
        in <stdin>

#0  0x07915849 in kevent () at <stdin>:2
#1  0x07f3df87 in kq_dispatch (base=0x7f17a800, arg=0x7fdcc800, tv=0x0)
    at /usr/src/lib/libevent/kqueue.c:193
#2  0x07f3d843 in event_base_loop (base=0x7f17a800, flags=0)
    at /usr/src/lib/libevent/event.c:493
#3  0x07f3dd09 in event_loop (flags=0) at /usr/src/lib/libevent/event.c:426
#4  0x07f3dd2e in event_dispatch () at /usr/src/lib/libevent/event.c:364
#5  0x18ac40d6 in ?? () from /usr/bin/tcpbench
#6  0x78b42940 in ?? ()
#7  0x00000000 in ?? ()
Current language:  auto; currently asm


Sevan

Reply | Threaded
Open this post in threaded view
|

Re: tcpbench core dumps which issued with ctrl and 4

Theo de Raadt
control-4 is control-\, depending on the type of tty you are on.

This is a function of the tty subsystem; it delivers SIGQUIT to the
tty session processes.

You are going to discover almost the entire works this way, since
very few processes block SIGQUIT.

This feature is probably older than you.

> As the subject says, if you run tcpbench on OpenBSD & press ctrl and 4
> on keyboard, tcpbench with coredump. Tested on -current snapshots form
> this month.
> Pointing gdb to the core file reports
> #0  0x07915849 in kevent () at <stdin>:2
> 2       <stdin>: No such file or directory.
>         in <stdin>
>
> #0  0x07915849 in kevent () at <stdin>:2
> #1  0x07f3df87 in kq_dispatch (base=0x7f17a800, arg=0x7fdcc800, tv=0x0)
>     at /usr/src/lib/libevent/kqueue.c:193
> #2  0x07f3d843 in event_base_loop (base=0x7f17a800, flags=0)
>     at /usr/src/lib/libevent/event.c:493
> #3  0x07f3dd09 in event_loop (flags=0) at /usr/src/lib/libevent/event.c:426
> #4  0x07f3dd2e in event_dispatch () at /usr/src/lib/libevent/event.c:364
> #5  0x18ac40d6 in ?? () from /usr/bin/tcpbench
> #6  0x78b42940 in ?? ()
> #7  0x00000000 in ?? ()
> Current language:  auto; currently asm
>
>
> Sevan
>

Reply | Threaded
Open this post in threaded view
|

Re: tcpbench core dumps which issued with ctrl and 4

Sevan / Venture37-2


On 21/12/2015 18:16, Theo de Raadt wrote:
> control-4 is control-\, depending on the type of tty you are on.
>
> This is a function of the tty subsystem; it delivers SIGQUIT to the
> tty session processes.
>
> You are going to discover almost the entire works this way, since
> very few processes block SIGQUIT.
>
> This feature is probably older than you.

Undestood, the reason I sent the email was because it coredumped, should
SIGQUIT be added to signal_handler() in src/usr.bin/tcpbench/tcpbench.c
so that tcpbench exits without the coredump or actually the current
behaviour is what is intended.


Sevan

Reply | Threaded
Open this post in threaded view
|

Re: tcpbench core dumps which issued with ctrl and 4

Theo de Raadt
> On 21/12/2015 18:16, Theo de Raadt wrote:
> > control-4 is control-\, depending on the type of tty you are on.
> >
> > This is a function of the tty subsystem; it delivers SIGQUIT to the
> > tty session processes.
> >
> > You are going to discover almost the entire works this way, since
> > very few processes block SIGQUIT.
> >
> > This feature is probably older than you.
>
> Undestood, the reason I sent the email was because it coredumped, should
> SIGQUIT be added to signal_handler() in src/usr.bin/tcpbench/tcpbench.c
> so that tcpbench exits without the coredump or actually the current
> behaviour is what is intended.

it is behaving *exactly* as intended.