2009/12/19 Jiří Paleček <[email protected]>:
> On Sat, 19 Dec 2009 05:02:45 +0100, Garrett Cooper <[email protected]>
> wrote:
>
>> On Fri, Dec 18, 2009 at 6:16 PM, Jiri Palecek <[email protected]> wrote:
>>>
>>> Hello,
>>>
>>> I've found some problems with the sendfile01 testcase (the one under
>>> network/tcp_cmds). This patch fixes them:
>>>
>>> Pass the -6 parameter to gethost if we are to use ipv6 and run the
>>> ipv6 server in that case. Also, fix the awk code for getting the PID:
>>> The PID is the second column, which is $2, not $1.
>>>
>>> Also:
>>>
>>> - initialize addrlen for accept syscall in the server (=> avoid EINVAL)
>>>
>>> - use tst_resm in the client instead of tst_res, to avoid an error
>>> message interpreted as a filename
>>>
>>> Regards
>>> Jiri Palecek
>>>
>>> Signed-off-by: Jiri Palecek <[email protected]>
>>> ---
>>> testcases/network/tcp_cmds/sendfile/SF_Server | 2 +-
>>> testcases/network/tcp_cmds/sendfile/sendfile01 | 10 +++++-----
>>> testcases/network/tcp_cmds/sendfile/testsf_c.c | 2 +-
>>> testcases/network/tcp_cmds/sendfile/testsf_s.c | 2 +-
>>> 4 files changed, 8 insertions(+), 8 deletions(-)
>>>
>>> diff --git a/testcases/network/tcp_cmds/sendfile/SF_Server
>>> b/testcases/network/tcp_cmds/sendfile/SF_Server
>>> index 2383ce4..9933db4 100755
>>> --- a/testcases/network/tcp_cmds/sendfile/SF_Server
>>> +++ b/testcases/network/tcp_cmds/sendfile/SF_Server
>>> @@ -3,4 +3,4 @@
>>> cd $3
>>> exec 1< /dev/null # open descriptor 1
>>> exec 2< /dev/null # open descriptor 1
>>> -nohup ./testsf_s $1 $2 &
>>> +nohup ./$4 $1 $2 &
>>> diff --git a/testcases/network/tcp_cmds/sendfile/sendfile01
>>> b/testcases/network/tcp_cmds/sendfile/sendfile01
>>> index 4e93447..7dd4fee 100755
>>> --- a/testcases/network/tcp_cmds/sendfile/sendfile01
>>> +++ b/testcases/network/tcp_cmds/sendfile/sendfile01
>>> @@ -61,7 +61,7 @@ do_setup()
>>>
>>> exists awk diff gethost grep rsh stat
>>>
>>> - IPADDR=$(gethost $RHOST | awk '/addresses:/ {print $2}')
>>> + IPADDR=$(gethost ${EXEC_SUFFIX:+-6} $RHOST | awk '/addresses:/
>>> {print $2}')
>>>
>>> }
>>>
>>> @@ -81,17 +81,17 @@ do_test()
>>>
>>> mkdir -p $TCtmp
>>> PORT=$$
>>> - if ! rsh -l root $IPADDR "$LTPROOT/testcases/bin/SF_Server $IPADDR
>>> $PORT $LTPROOT/testcases/bin"; then
>>> + if ! rsh -l root $IPADDR "$LTPROOT/testcases/bin/SF_Server $IPADDR
>>> $PORT $LTPROOT/testcases/bin $SERVER"; then
>>> end_testcase "rsh failed to $IPADDR as root failed"
>>> fi
>>> sleep 10
>>> - PID=$(rsh -l root $IPADDR "ps -ef" | awk "\$0 !~ /awk/ && /$SERVER/
>>> && /$PORT/ {print \$1}")
>>> + PID=$(rsh -l root $IPADDR "ps -ef" | awk "\$0 !~ /awk/ && /$SERVER/
>>> && /$PORT/ {print \$2}")
>>> [ -n "$PID" ] || end_testcase "Could not start server"
>>>
>>> for clnt_fname in $FILES; do
>>> serv_fname=$TCdat/$clnt_fname
>>> SIZE=`stat -c '%s' $serv_fname`
>>> - tst_resm TINFO "Starting $SERVER $IPADDR Client_filename
>>> Server_filename Size "
>>> + tst_resm TINFO "Starting $CLIENT $IPADDR Client_filename
>>> Server_filename Size "
>>>
>>> $CLIENT $IPADDR $PORT "$TCtmp/$clnt_fname" $serv_fname $SIZE
>>> RetVal=$?
>>> @@ -118,7 +118,7 @@ do_test()
>>>
>>> do_cleanup()
>>> {
>>> - PID=$(rsh -n -l root $RHOST "ps -ef" | awk "\$0 !~ /awk/ &&
>>> /$SERVER/ && /$PORT/ {print \$1}")
>>> + PID=$(rsh -n -l root $RHOST "ps -ef" | awk "\$0 !~ /awk/ &&
>>> /$SERVER/ && /$PORT/ {print \$2}")
>>> [ -n "$PID" ] && rsh -n -l root $RHOST kill -9 $PID
>>> tst_cleanup
>>> }
>>> diff --git a/testcases/network/tcp_cmds/sendfile/testsf_c.c
>>> b/testcases/network/tcp_cmds/sendfile/testsf_c.c
>>> index 52b9ecf..7683927 100644
>>> --- a/testcases/network/tcp_cmds/sendfile/testsf_c.c
>>> +++ b/testcases/network/tcp_cmds/sendfile/testsf_c.c
>>> @@ -120,7 +120,7 @@ main (int argc, char *argv[])
>>> }
>>> }
>>>
>>> - tst_res(TINFO, "Asking for remote file: %s", serv_fname);
>>> + tst_resm(TINFO, "Asking for remote file: %s", serv_fname);
>>>
>>> tst_resm(TINFO, "File %s received\n", argv[4]);
>>>
>>> diff --git a/testcases/network/tcp_cmds/sendfile/testsf_s.c
>>> b/testcases/network/tcp_cmds/sendfile/testsf_s.c
>>> index bc77712..38baf8c 100644
>>> --- a/testcases/network/tcp_cmds/sendfile/testsf_s.c
>>> +++ b/testcases/network/tcp_cmds/sendfile/testsf_s.c
>>> @@ -92,7 +92,7 @@ main(int argc, char *argv[])
>>> tst_exit();
>>> }
>>>
>>> - socklen_t fromlen;
>>> + socklen_t fromlen = sizeof(from);
>>>
>>> /* process connections */
>>> while (1) {
>>
>> Hi Jiri,
>> I agree with you on all points, but the ps column:
>>
>> gcoo...@orangebox ~ $ ps ef
>> PID TTY STAT TIME COMMAND
>> 4631 tty1 S 0:00 -bash TERM=linux HOME=/home/gcooper
>> SHELL=/bin/bash U
>> 4636 tty1 S+ 0:00 \_ /bin/sh /usr/bin/startx
>> MANPATH=/etc/java-config-
>> 4652 tty1 S+ 0:00 \_ xinit /etc/X11/xinit/xinitrc --
>> -nolisten tcp
>> 4667 tty1 S 0:00 \_ /bin/sh /etc/X11/xinit/xinitrc
>> DISPLAY=:0
>> 4682 tty1 S 0:00 \_ /bin/sh /etc/xdg/xfce4/xinitrc
>> MANPAT
>> 4699 tty1 S 0:00 \_ /usr/bin/xfce4-session
>> MANPATH=/e
>> 4705 tty1 S 0:02 \_ xfwm4 --sm-client-id
>> 2d6757e3
>> 4707 tty1 Sl 0:03 \_ Thunar --sm-client-id
>> 276f4b6
>> 4711 tty1 S 0:06 \_ xfce4-panel -r
>> --sm-client-id
>> 4716 tty1 Sl 0:00 | \_
>> /usr/libexec/xfce4/panel-
>> 4724 tty1 Sl 0:00 | \_
>> /usr/libexec/xfce4/panel-
>> 4712 tty1 S 0:01 \_ xfdesktop
>> --sm-client-id 293b
>> 21073 tty1 S 0:00 /usr/bin/xterm
>> MANPATH=/etc/java-config-2/current-sys
>> 21075 pts/1 Ss 0:00 \_ bash
>> MANPATH=/etc/java-config-2/current-system-vm
>> 21080 pts/1 R+ 0:00 \_ ps ef
>> MANPATH=/etc/java-config-2/current-syst
>> 3396 tty1 S 0:00 /usr/bin/xterm
>> MANPATH=/etc/java-config-2/current-sys
>> 3398 pts/0 Ss 0:00 \_ bash
>> MANPATH=/etc/java-config-2/current-system-vm
>> 21065 pts/0 S+ 0:00 \_ vim sendfile01
>> MANPATH=/etc/java-config-2/cur
>> 3350 tty1 Sl 1:02 /usr/bin/firefox
>> MANPATH=/etc/java-config-2/current-s
>> 30806 tty1 Sl 5:26 /usr/lib/jvm/sun-jdk-1.6/bin/java -classpath
>> /usr/sha
>> 4791 tty1 S 0:00 /usr/libexec/uim-helper-server
>> MANPATH=/etc/java-conf
>> 4714 tty1 S 0:00 xfce4-settings-helper --display :0.0
>> --sm-client-id 2
>> 4706 tty1 S 0:00 xfsettingsd
>> MANPATH=/etc/java-config-2/current-system
>> 4696 tty1 S 0:00 /usr/bin/dbus-launch --sh-syntax
>> --exit-with-session
>> gcoo...@orangebox ~ $ ps --version
>> procps version 3.2.8
>>
>> So to bypass this issue I used ps eopid,cmd:
>>
>> 400 [crypto/2]
>> 401 [crypto/3]
>> 587 [iscsi_eh]
>> 593 [scsi_eh_0]
>> 596 [scsi_eh_1]
>> 599 [scsi_eh_2]
>> 602 [scsi_eh_3]
>> 605 [scsi_eh_4]
>> 608 [scsi_eh_5]
>> 692 [kpsmoused]
>> 700 [kstriped]
>> 703 [kondemand/0]
>> 704 [kondemand/1]
>> 705 [kondemand/2]
>> 706 [kondemand/3]
>> 729 [usbhid_resumer]
>> 809 [reiserfs/0]
>> 810 [reiserfs/1]
>> 811 [reiserfs/2]
>> 812 [reiserfs/3]
>> 1906 /sbin/udevd --daemon
>> 2303 /usr/sbin/acpid
>> 3350 /usr/bin/firefox
>> 3396 /usr/bin/xterm
>> 3398 bash
>> 3535 supervising syslog-ng
>> 3536 /usr/sbin/syslog-ng
>> 3596 /usr/bin/dbus-daemon --system
>> 3653 /usr/sbin/crond
>> 3707 /usr/sbin/gpm -m /dev/input/mice -t ps2
>> 3768 /usr/sbin/hald --use-syslog --verbose=no
>> 3771 hald-runner
>> 3774 hald-addon-input: Listening on /dev/input/event4 /dev/input/event3
>> /dev/in
>> 3787 hald-addon-storage: polling /dev/sr0 (every 2 sec)
>> 3790 /usr/libexec/hald-addon-cpufreq
>> 3791 hald-addon-acpi: listening on acpid socket /var/run/acpid.socket
>> 4250 /sbin/dhcpcd -h orangebox eth0
>> 4339 /sbin/portmap
>> 4400 /sbin/rpc.statd --no-notify
>> 4512 /usr/sbin/sshd
>> 4580 /sbin/agetty 38400 tty2 linux
>> 4581 /sbin/agetty 38400 tty3 linux
>> 4582 /sbin/agetty 38400 tty4 linux
>> 4583 /sbin/agetty 38400 tty5 linux
>> 4584 /sbin/agetty 38400 tty6 linux
>> 4613 [rpciod/0]
>> 4614 [rpciod/1]
>> 4615 [rpciod/2]
>> 4616 [rpciod/3]
>> 4622 [nfsiod]
>> 4627 [lockd]
>> 4629 /bin/login --
>> 4631 -bash
>> 4636 /bin/sh /usr/bin/startx
>> 4652 xinit /etc/X11/xinit/xinitrc -- -nolisten tcp -br -auth
>> /home/gcooper/.ser
>> 4653 X :0 -nolisten tcp -br -auth /home/gcooper/.serverauth.4636
>> -deferglyphs 1
>> 4667 /bin/sh /etc/X11/xinit/xinitrc
>> 4682 /bin/sh /etc/xdg/xfce4/xinitrc
>> 4691 /usr/bin/ssh-agent -s
>> 4696 /usr/bin/dbus-launch --sh-syntax --exit-with-session
>> 4697 /usr/bin/dbus-daemon --fork --print-pid 5 --print-address 7
>> --session
>> 4699 /usr/bin/xfce4-session
>> 4701 /usr/libexec/xfconfd
>> 4705 xfwm4 --sm-client-id 2d6757e33-4d33-453a-ba49-da83d783acde --display
>> :0.0
>> 4706 xfsettingsd
>> 4707 Thunar --sm-client-id 276f4b61e-2bba-4207-8e5e-feed55cbe9b4 --daemon
>> 4709 /usr/libexec/gam_server
>> 4711 xfce4-panel -r --sm-client-id 260365c00-2a57-4640-9170-8e74ce91bf1e
>> 4712 xfdesktop --sm-client-id 293b54ac0-804e-45e2-9ae9-2b9adbaeddde
>> --display :
>> 4714 xfce4-settings-helper --display :0.0 --sm-client-id
>> 20a17a11f-321d-40f4-87
>> 4716 /usr/libexec/xfce4/panel-plugins/xfce4-menu-plugin socket_id
>> 18874394 name
>> 4723 /usr/libexec/gvfsd
>> 4724 /usr/libexec/xfce4/panel-plugins/xfce4-mixer-plugin socket_id
>> 18874431 nam
>> 4791 /usr/libexec/uim-helper-server
>> 21065 vim sendfile01
>> 21073 /usr/bin/xterm
>> 21075 bash
>> 21113 ps -eopid,cmd
>> 28404 /usr/libexec/gvfs-hal-volume-monitor
>> 28689 /sbin/udevd --daemon
>> 28690 /sbin/udevd --daemon
>> 30806 /usr/lib/jvm/sun-jdk-1.6/bin/java -classpath
>> /usr/share/bcprov/lib/bcprov.
>> 30860 xchat
>> 31904 /usr/libexec/gconfd-2
>> gcoo...@orangebox ~ $ ps -eopid,cmd
>>
>> Do you agree that this is an ok choice?
>
> Yes, it is arguably more robust and readable (although I don't think your
> example shows the need for it).
>
> Note that there are differences between ps -ef, ps ef and ps -e f:
>
> ji...@debian:~$ ps -ef | head -10
> UID PID PPID C STIME TTY TIME CMD
> root 1 0 0 21:15 ? 00:00:00 init [2]
> root 2 0 0 21:15 ? 00:00:00 [kthreadd]
> root 3 2 0 21:15 ? 00:00:00 [migration/0]
> root 4 2 0 21:15 ? 00:00:00 [ksoftirqd/0]
> root 5 2 0 21:15 ? 00:00:00 [events/0]
> root 6 2 0 21:15 ? 00:00:00 [khelper]
> root 35 2 0 21:15 ? 00:00:00 [kblockd/0]
> root 36 2 0 21:15 ? 00:00:00 [kacpid]
> root 37 2 0 21:15 ? 00:00:00 [kacpi_notify]
> ji...@debian:~$ ps ef | head -10
> PID TTY STAT TIME COMMAND
> 7465 pts/12 Ss 0:00 /bin/bash --noediting -i PWD=/home/jirka
> TERM=dumb TERMCAP= COLUMNS=136 EMACS=t INSIDE_EMACS=22.3.1,comint GT
> 7498 pts/12 R+ 0:00 \_ ps ef KDE_MULTIHEAD=false
> KDEWM=/usr/bin/openbox SHELL=/bin/bash TERM=dumb
> XDG_SESSION_COOKIE=e07c53451a9
> 7499 pts/12 D+ 0:00 \_ head -10 KDE_MULTIHEAD=false
> KDEWM=/usr/bin/openbox SHELL=/bin/bash TERM=dumb XDG_SESSION_COOKIE=e07c5345
> 6355 pts/10 Ss+ 0:00 /bin/bash GDM_XSERVER_LOCATION=local
> GDM_LANG=cs_CZ LANGUAGE= KDE_MULTIHEAD=false KDE_SESSION_VERSION=4 KONSO
> 6343 pts/9 Ss+ 0:00 /bin/bash GDM_XSERVER_LOCATION=local
> GDM_LANG=cs_CZ LANGUAGE= KDE_MULTIHEAD=false KDE_SESSION_VERSION=4 KONSO
> 6337 pts/8 Ss+ 0:00 /bin/bash GDM_XSERVER_LOCATION=local
> GDM_LANG=cs_CZ LANGUAGE= KDE_MULTIHEAD=false KDE_SESSION_VERSION=4 KONSO
> 6325 pts/7 Ss+ 0:00 /bin/bash GDM_XSERVER_LOCATION=local
> GDM_LANG=cs_CZ LANGUAGE= KDE_MULTIHEAD=false KDE_SESSION_VERSION=4 KONSO
> 6315 pts/6 Ss+ 0:00 /bin/bash GDM_XSERVER_LOCATION=local
> GDM_LANG=cs_CZ LANGUAGE= KDE_MULTIHEAD=false KDE_SESSION_VERSION=4 KONSO
> 6300 pts/5 Ss+ 0:00 /bin/bash GDM_XSERVER_LOCATION=local
> GDM_LANG=cs_CZ LANGUAGE= KDE_MULTIHEAD=false KDE_SESSION_VERSION=4 KONSO
> ji...@debian:~$ ps -e f | head -10
> PID TTY STAT TIME COMMAND
> 2 ? S< 0:00 [kthreadd]
> 3 ? S< 0:00 \_ [migration/0]
> 4 ? S< 0:00 \_ [ksoftirqd/0]
> 5 ? S< 0:00 \_ [events/0]
> 6 ? S< 0:00 \_ [khelper]
> 35 ? S< 0:00 \_ [kblockd/0]
> 36 ? S< 0:00 \_ [kacpid]
> 37 ? S< 0:00 \_ [kacpi_notify]
> 100 ? S< 0:00 \_ [kseriod]
> ji...@debian:~$
>
> So you actually used a different command than the one that's used in the
> testcase. (the first listing is consistent with POSIX specification)
Interesting -- didn't know that!
How about this instead:
gcoo...@orangebox ~ $ ps -eo pid,cmd | head -n 10
PID CMD
1 init [3]
2 [kthreadd]
3 [migration/0]
4 [ksoftirqd/0]
5 [migration/1]
6 [ksoftirqd/1]
7 [migration/2]
8 [ksoftirqd/2]
9 [migration/3]
I'm just proposing this because if --columns is set on the target
system, it's more likely to truncate output and give false positives.
Sound good :)?
Thanks,
-Garrett
Thanks,
-Garrett
------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
