Bug#381297: [Pkg-utopia-maintainers] Bug#381297: dbus: X keyboard interference debuging - small progress
On 11/17/07, Sjoerd Simons [EMAIL PROTECTED] wrote: On Wed, Aug 09, 2006 at 07:04:13PM +0300, Martin-Éric Racine wrote: ke, 2006-08-09 kello 18:34 +0300, Martin-Éric Racine kirjoitti: ke, 2006-08-09 kello 00:31 +0300, Martin-Éric Racine kirjoitti: ti, 2006-08-08 kello 22:33 +0200, Sjoerd Simons kirjoitti: On Thu, Aug 03, 2006 at 06:16:29PM +0300, Martin-Éric Racine wrote: I welcome further step-by-step instructions to perform, towards finding the source of this keyboard input disruption by dbus-enabled apps in X. I doubt it's dbus (thatis the session bus) itself but instead something on the session bus. Which dies when you kill the session bus. Try looking at the difference between your procress listing before and after killing the session bus. Probably same apps have quit (among which the one guilty for your slugish keyboard) As mentioned in the original bug report, been there done that. NO SINGLE APPLICATION died as the result of killing dbug daemons. Hmm. It seems I forgot to actually mention it. Anyhow, NO, there is no other application that dies in the process of killing D-Bus components. However, as suggested over IRC, I straced this and it indeed taps into the keyboard: Here's a shorter one, until the keystroke that I pressed right when infinite scrolling kicked-in: me2165 0.0 0.2 2648 648 ?S18:50 0:00 /usr/bin/dbus-launch --exit-with-session x-session-manager me2168 0.0 0.1 2052 424 ?Ss 18:50 0:00 dbus-daemon --fork --print-pid 8 --print-address 6 --session root 3953 0.0 0.1 2052 424 ?Ss 18:54 0:00 dbus-daemon --fork --print-pid 8 --print-address 6 --session nino:/root# strace -p 2165 60 Process 2165 attached - interrupt to quit select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \234, 512)= 1 select(5, [0 4], NULL, [0 4], NULL Sorry for not getting back to you earlier. But do you still see this issue with recent versions of dbus? Yes, I do. -- Martin-Éric Racine http://q-funk.iki.fi
Bug#381297: [Pkg-utopia-maintainers] Bug#381297: dbus: X keyboard interference debuging - small progress
On Sat, Nov 17, 2007 at 02:51:54PM +0200, Martin-Éric Racine wrote: Sorry for not getting back to you earlier. But do you still see this issue with recent versions of dbus? Yes, I do. So what versions of dbus and X11 are you running, current sid? I just checked the code again and dbus-launch only opens the X11 display and then does _nothing_ with it except for waiting for a disconnection signal. And any events it might receive are popped and ignore right away. So i'm still completely in the dark here. Ofcourse on none of my machine it even gets any events except for the disconnection event. Sjoerd -- Do your part to help preserve life on Earth -- by trying to preserve your own.
Bug#381297: [Pkg-utopia-maintainers] Bug#381297: dbus: X keyboard interference debuging - small progress
On 11/17/07, Sjoerd Simons [EMAIL PROTECTED] wrote: On Sat, Nov 17, 2007 at 02:51:54PM +0200, Martin-Éric Racine wrote: Sorry for not getting back to you earlier. But do you still see this issue with recent versions of dbus? Yes, I do. So what versions of dbus and X11 are you running, current sid? Etch. -- Martin-Éric Racine http://q-funk.iki.fi
Bug#381297: [Pkg-utopia-maintainers] Bug#381297: dbus: X keyboard interference debuging - small progress
ke, 2006-08-09 kello 00:31 +0300, Martin-Éric Racine kirjoitti: ti, 2006-08-08 kello 22:33 +0200, Sjoerd Simons kirjoitti: On Thu, Aug 03, 2006 at 06:16:29PM +0300, Martin-Éric Racine wrote: I welcome further step-by-step instructions to perform, towards finding the source of this keyboard input disruption by dbus-enabled apps in X. I doubt it's dbus (thatis the session bus) itself but instead something on the session bus. Which dies when you kill the session bus. Try looking at the difference between your procress listing before and after killing the session bus. Probably same apps have quit (among which the one guilty for your slugish keyboard) As mentioned in the original bug report, been there done that. NO SINGLE APPLICATION died as the result of killing dbug daemons. Hmm. It seems I forgot to actually mention it. Anyhow, NO, there is no other application that dies in the process of killing D-Bus components. However, as suggested over IRC, I straced this and it indeed taps into the keyboard: me2116 0.1 0.7 4452 1868 ?Ss 18:24 0:00 -bash startx -- vt1 -s 5 -quiet -p 5 -nolisten tcp dpms -dpime2140 0.0 0.2 2428 684 ?S18:24 0:00 \_ xinit /etc/X11/xinit/xinitrc -- /usr/bin/X vt1 -s 5 -quiroot 2141 4.0 5.0 18368 12400 tty1 Ss+ 18:24 0:04 \_ /usr/bin/X :0 vt1 -s 5 -quiet -p 5 -nolisten tcp dpm me2144 0.0 0.4 4776 1036 ?S18:24 0:00 \_ x-session-manager me2182 0.0 0.2 4400 704 ?Ss 18:24 0:00 \_ /usr/bin/ssh-agent /usr/bin/dbus-launch --exit-wme2187 0.9 0.7 4852 1900 ?SNs 18:24 0:00 \_ icewmbg me2188 0.7 1.6 8384 3964 ?Ss 18:24 0:00 \_ icewm me2189 0.1 0.8 6200 2056 ?Ss 18:24 0:00 \_ icewmtray root 2117 0.0 0.1 1592 488 tty7 Ss+ 18:24 0:00 /sbin/getty 38400 tty7 me2176 6.1 10.4 108172 25552 ?Sl 18:24 0:05 /usr/lib/firefox/firefox-bin me2185 0.0 0.2 2648 632 ?S18:24 0:00 /usr/bin/dbus-launch --exit-with-session x-session-manager me2186 0.0 0.1 2052 424 ?Ss 18:24 0:00 dbus-daemon --fork --print-pid 8 --print-address 6 --sessionme2193 0.3 0.9 4852 2324 ?S18:24 0:00 /usr/lib/libgconf2-4/gconfd-2 10 nino:/root# strace -p 2185 Process 2185 attached - interrupt to quit select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \234, 512)= 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340\310, 512)= 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340\310, 512)= 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340\310, 512)= 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340\310, 512)= 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \35, 512) = 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, ., 512) = 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \256, 512)= 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, ., 512) = 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \256, 512)= 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, ., 512) = 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \256, 512)= 1 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \235, 512)
Bug#381297: [Pkg-utopia-maintainers] Bug#381297: dbus: X keyboard interference debuging - small progress
ke, 2006-08-09 kello 18:34 +0300, Martin-Éric Racine kirjoitti: ke, 2006-08-09 kello 00:31 +0300, Martin-Éric Racine kirjoitti: ti, 2006-08-08 kello 22:33 +0200, Sjoerd Simons kirjoitti: On Thu, Aug 03, 2006 at 06:16:29PM +0300, Martin-Éric Racine wrote: I welcome further step-by-step instructions to perform, towards finding the source of this keyboard input disruption by dbus-enabled apps in X. I doubt it's dbus (thatis the session bus) itself but instead something on the session bus. Which dies when you kill the session bus. Try looking at the difference between your procress listing before and after killing the session bus. Probably same apps have quit (among which the one guilty for your slugish keyboard) As mentioned in the original bug report, been there done that. NO SINGLE APPLICATION died as the result of killing dbug daemons. Hmm. It seems I forgot to actually mention it. Anyhow, NO, there is no other application that dies in the process of killing D-Bus components. However, as suggested over IRC, I straced this and it indeed taps into the keyboard: Here's a shorter one, until the keystroke that I pressed right when infinite scrolling kicked-in: me2165 0.0 0.2 2648 648 ?S18:50 0:00 /usr/bin/dbus-launch --exit-with-session x-session-manager me2168 0.0 0.1 2052 424 ?Ss 18:50 0:00 dbus-daemon --fork --print-pid 8 --print-address 6 --session root 3953 0.0 0.1 2052 424 ?Ss 18:54 0:00 dbus-daemon --fork --print-pid 8 --print-address 6 --session nino:/root# strace -p 2165 60 Process 2165 attached - interrupt to quit select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \340H, 512) = 2 select(5, [0 4], NULL, [0 4], NULL) = 1 (in [0]) ioctl(4, FIONREAD, [0]) = 0 read(0, \234, 512)= 1 select(5, [0 4], NULL, [0 4], NULL -- Martin-Éric Racine http://q-funk.iki.fi
Bug#381297: [Pkg-utopia-maintainers] Bug#381297: dbus: X keyboard interference debuging - small progress
On Thu, Aug 03, 2006 at 06:16:29PM +0300, Martin-Éric Racine wrote: I welcome further step-by-step instructions to perform, towards finding the source of this keyboard input disruption by dbus-enabled apps in X. I doubt it's dbus (thatis the session bus) itself but instead something on the session bus. Which dies when you kill the session bus. Try looking at the difference between your procress listing before and after killing the session bus. Probably same apps have quit (among which the one guilty for your slugish keyboard) Sjoerd -- Biology is the only science in which multiplication means the same thing as division.
Bug#381297: [Pkg-utopia-maintainers] Bug#381297: dbus: X keyboard interference debuging - small progress
ti, 2006-08-08 kello 22:33 +0200, Sjoerd Simons kirjoitti: On Thu, Aug 03, 2006 at 06:16:29PM +0300, Martin-Éric Racine wrote: I welcome further step-by-step instructions to perform, towards finding the source of this keyboard input disruption by dbus-enabled apps in X. I doubt it's dbus (thatis the session bus) itself but instead something on the session bus. Which dies when you kill the session bus. Try looking at the difference between your procress listing before and after killing the session bus. Probably same apps have quit (among which the one guilty for your slugish keyboard) As mentioned in the original bug report, been there done that. NO SINGLE APPLICATION died as the result of killing dbug daemons. -- Martin-Éric Racine http://q-funk.iki.fi
