Call for testing: OpenSSH-6.2

Tom Christensen tgc at
Wed Mar 13 07:14:18 EST 2013

On 03/06/2013 10:07 PM, Tom Christensen wrote:
> First problem is that configure hangs in the select+rlimit test.
I see that a timeout was added and the test now passes on IRIX 5.3.

> A generic IRIX issue is the use of killpg in sshd.c.
> sshd.c: In function `grace_alarm_handler':
> sshd.c:368: warning: implicit declaration of function `killpg'
> On IRIX using this function requires _BSD_SIGNALS to be defined
> otherwise results are unpredictable.
> See the manpage here for more details:
> Looking at the kill() manpage it seems to me that killpg could be
> replaced with kill(0, SIGTERM) to achieve the same thing.
Any opinion on this one?

Building openssh with -D_BSD_SIGNALS forced on causes numerous warnings 
like the below to appear:
gcc -D_BSD_SIGNALS -g -O2 -Wall -Wpointer-arith -Wuninitialized 
-Wsign-compare -Wformat-security -fno-strict-aliasing 
-D_FORTIFY_SOURCE=2 -fno-builtin-memset -std=gnu99  -I. -I.. -I. -I./.. 
-I/usr/tgcware/include/openssl -I/usr/tgcware/include -DHAVE_CONFIG_H -c 
In file included from ../openbsd-compat/openbsd-compat.h:151,
                  from ../includes.h:172,
                  from bsd-arc4random.c:17:
../openbsd-compat/bsd-misc.h:99:1: warning: "signal" redefined
In file included from 
                  from ../openbsd-compat/getrrsetbyname.h:57,
                  from ../openbsd-compat/openbsd-compat.h:45,
                  from ../includes.h:172,
                  from bsd-arc4random.c:17:
warning: this is the location of the previous definition

This is because the system <signal.h> does a #define signal BSDsignal 
when _BSD_SIGNALS is defined.


More information about the openssh-unix-dev mailing list