firefox main menù is unusable because off a long delay. all other firefox button work well. the problem arise after emerging xorg-1.8.0. Reproducible: Always Steps to Reproduce: 1. open firefox 2. click on file menu or any other main menu element 3. wait for a long time for the nexted menu list
Created attachment 229771 [details] emerge Info
running firefox from a terminal, I read on standard error: cloc3@s939 ~ $ firefox Fontconfig error: "local.conf", line 73: not well-formed (invalid token) socket(): Famiglia dell'indirizzo non gestita dal protocollo socket(): Famiglia dell'indirizzo non gestita dal protocollo using strace, I see this: output from: strace -fe open firefox >output.txt 2>&1 ... [pid 20754] open("/usr/lib64/gconv/ISO8859-1.so", O_RDONLY) = 59 Process 20758 attached Process 20753 detached socket(): Famiglia dell'indirizzo non gestita dal protocollo Process 21823 attached (waiting for parent) Process 21823 resumed (parent 20754 ready) [pid 21823] open("/proc/self/fd", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 54 Process 20754 suspended [pid 21823] open("/etc/ld.so.cache", O_RDONLY) = 3 ... strace -fe socket firefox >output.txt 2>&1 [pid 8559] socket(PF_NETLINK, SOCK_RAW, 0) = 54 [pid 8554] socket(PF_INET6, SOCK_STREAM, IPPROTO_IP) = -1 EAFNOSUPPORT (Address family not supported by protocol) socket(): Famiglia dell'indirizzo non gestita dal protocollo Process 8674 attached Process 8554 suspended Process 8677 attached Process 8679 attached (waiting for parent) Process 8677 detached Process 8679 resumed (parent 8674 ready) [pid 8679] socket(PF_FILE, SOCK_STREAM, 0) = 3 ...
I am unable to duplicate this, if you wuold issue a LC_ALL="C" firefox for the strace and terminal.
cloc3@s939 ~ $ LC_ALL="C" firefox Fontconfig error: "local.conf", line 73: not well-formed (invalid token) socket(): Address family not supported by protocol socket(): Address family not supported by protocol cloc3@s939 ~ $
Created attachment 229933 [details] strace output sorry for the bzip2 output, but the text file was long too much for bugzilla filters. LC_ALL="C" strace -f firefox >strace.out 2>&1 for both output I do: 1) start command 2) click on file menu 2') wait for a long time 3) click on quit 3') wait for a long time
Ahi, downgrading to xorg-1.7* seems not enough to avoid the issue. so, I'm not able to suggest you the real cause. anyway, the problem affects only mozilla-firefox (not seamonky, for example) and is undipendent from the desktop envinronment. it happens also runnig firefox from a new user, in a nake X environment.
(In reply to comment #6) > Ahi, downgrading to xorg-1.7* seems not enough to avoid the issue. > so, I'm not able to suggest you the real cause. > > anyway, the problem affects only mozilla-firefox (not seamonky, for example) > and is undipendent from the desktop envinronment. > > it happens also runnig firefox from a new user, in a nake X environment. > Did you remember to run "emerge `qlist -I -C x11-drivers/`" after upgrading and downgrading xorg server? I am still very much unable to reproduce the problem you describe.
(In reply to comment #7) > (In reply to comment #6) > > Ahi, downgrading to xorg-1.7* seems not enough to avoid the issue. > > so, I'm not able to suggest you the real cause. > > > > anyway, the problem affects only mozilla-firefox (not seamonky, for example) > > and is undipendent from the desktop envinronment. > > > > it happens also runnig firefox from a new user, in a nake X environment. > > > > Did you remember to run "emerge `qlist -I -C x11-drivers/`" after upgrading and > downgrading xorg server? I am still very much unable to reproduce the problem > you describe. > yes, off course. I've rebuild x11-drivers/ and a lot of other (pseudo-)related packages. but not one of them was good to solve my issue. I've also run revdep-rebuild.
(In reply to comment #7) > Did you remember to run "emerge `qlist -I -C x11-drivers/`" after upgrading and > downgrading xorg server? FTR, X won't load drivers if they weren't (re)built with the same ABI, so Marco would have had a non-working X if he hadn't done that :) > I am still very much unable to reproduce the problem > you describe. I'd bet on an another component rather than Xorg proper: kernel, X driver, pixman, cairo, nss? Anyhow, there's not much I can say about this, Firefox works great on all the boxes I have, with and without Xorg 1.8. Don't hesitate to CC us back if one of the X components are to blame. Cheers
(In reply to comment #9) > > I'd bet on an another component rather than Xorg proper: kernel, X driver, > pixman, cairo, nss? > this is my conf: cloc3@s939 ~ $ uname -rm 2.6.31-gentoo-r10-p5q x86_64 cloc3@s939 ~ $ emerge -pv cairo dev-libs/nss `qlist -IC x11-drivers/` pixman xulrunner mozilla-firefox These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild R ] dev-libs/nss-3.12.6-r1 USE="utils" 0 kB [ebuild R ] x11-libs/pixman-0.18.0 USE="sse2 (-altivec) -mmx -static-libs" 0 kB [ebuild R ] x11-drivers/nvidia-drivers-195.36.24 USE="acpi gtk (multilib) -custom-cflags" 0 kB [ebuild R ] x11-drivers/xf86-input-evdev-2.3.2 USE="-debug" 0 kB [ebuild R ] x11-drivers/xf86-input-keyboard-1.4.0 USE="-debug" 0 kB [ebuild R ] x11-drivers/xf86-input-mouse-1.5.0 USE="-debug" 0 kB [ebuild R ] x11-drivers/xf86-input-void-1.3.0 USE="-debug" 0 kB [ebuild R ] x11-drivers/xf86-video-nouveau-0.0.15_pre20100329 USE="-debug" 0 kB [ebuild R ] x11-drivers/xf86-video-nv-2.1.17 USE="-debug" 0 kB [ebuild R ] x11-drivers/xf86-video-vesa-2.3.0 USE="-debug" 0 kB [ebuild R ] x11-drivers/xf86-video-vmware-11.0.1 USE="-debug" 0 kB [ebuild R ] x11-libs/cairo-1.8.8-r1 USE="X lcdfilter opengl svg xcb (-aqua) -cleartype -debug -directfb -doc -glitz" 0 kB [ebuild R ] net-libs/xulrunner-1.9.2.3-r1 USE="alsa dbus gnome java libnotify -custom-optimization -debug -startup-notification -system-sqlite -wifi" 0 kB [ebuild R ] www-client/mozilla-firefox-3.6.3 USE="alsa dbus gnome java libnotify -bindist -custom-optimization -startup-notification -system-sqlite -wifi" LINGUAS="en_GB it -af -ar -as -be -bg -bn -bn_BD -bn_IN -ca -cs -cy -da -de -el -en -en_US -eo -es -es_AR -es_CL -es_ES -es_MX -et -eu -fa -fi -fr -fy -fy_NL -ga -ga_IE -gl -gu -gu_IN -he -hi -hi_IN -hr -hu -id -is -ja -ka -kk -kn -ko -ku -lt -lv -mk -ml -mr -nb -nb_NO -nl -nn -nn_NO -oc -or -pa -pa_IN -pl -pt -pt_BR -pt_PT -rm -ro -ru -si -sk -sl -sq -sr -sv -sv_SE -ta -ta_LK -te -th -tr -uk -vi -zh_CN -zh_TW" 0 kB
Created attachment 230191 [details] strace output for firefox-bin another hint. firefox-bin works fine. this is a strace log for comparison, if it makes help you. I'm not able to contrive better schemes...
maybe i'm wrong but from this > [pid 8559] socket(PF_NETLINK, SOCK_RAW, 0) = 54 > [pid 8554] socket(PF_INET6, SOCK_STREAM, IPPROTO_IP) = -1 EAFNOSUPPORT i'd say that the not supported address family error is related to ipv6 support lacking (which should not be an issue)
(In reply to comment #12) > maybe i'm wrong but from this > > [pid 8559] socket(PF_NETLINK, SOCK_RAW, 0) = 54 > > [pid 8554] socket(PF_INET6, SOCK_STREAM, IPPROTO_IP) = -1 EAFNOSUPPORT > i'd say that the not supported address family error is related to ipv6 support > lacking (which should not be an issue) > thanks. anyway, my system is built with -ipv6 useflag. what can I do to correctly setup my conf?
s939 linux # zcat /proc/config.gz |grep IPV # CONFIG_IPV6 is not set CONFIG_NF_DEFRAG_IPV4=m CONFIG_NF_CONNTRACK_IPV4=m
setting true network.dns.disableIPv6 in about:config does not help. :cry:
(In reply to comment #13) > (In reply to comment #12) > > maybe i'm wrong but from this > > > [pid 8559] socket(PF_NETLINK, SOCK_RAW, 0) = 54 > > > [pid 8554] socket(PF_INET6, SOCK_STREAM, IPPROTO_IP) = -1 EAFNOSUPPORT > > i'd say that the not supported address family error is related to ipv6 support > > lacking (which should not be an issue) > > > thanks. > anyway, my system is built with -ipv6 useflag. > > what can I do to correctly setup my conf? > that is an nss call which returned false nothing to worry about. As to your issue I have no clue. As I and others have been unable to duplicate this I am not sure which direction to point you in.
actually run 'firefox -safe-mode' this will disable all addons. Let us know your results.
Created attachment 230563 [details] LC_ALL="C" DISPLAY=":0" strace firefox -safe-mode >strace.out 2>&1 this comes from a new user, with an empty homeDir, running a simple X command from a character console. I'm afraid: no improvements. :)
possible workaround: simple add ipv6 module on the kernel. s939 ~ # zcat /proc/config.gz |grep IPV CONFIG_IPV6=m # CONFIG_IPV6_PRIVACY is not set # CONFIG_IPV6_ROUTER_PREF is not set # CONFIG_IPV6_OPTIMISTIC_DAD is not set # CONFIG_IPV6_MIP6 is not set CONFIG_IPV6_SIT=m CONFIG_IPV6_NDISC_NODETYPE=y # CONFIG_IPV6_TUNNEL is not set # CONFIG_IPV6_MULTIPLE_TABLES is not set # CONFIG_IPV6_MROUTE is not set CONFIG_NF_DEFRAG_IPV4=m CONFIG_NF_CONNTRACK_IPV4=m # CONFIG_NF_CONNTRACK_IPV6 is not set simple way to reproduce the issue: type from root with firefox running: sysctl -w net.ipv6.conf.all.disable_ipv6=1 (put 0 to re-fix) this user on italian forum has the same problem, with the same error: http://forums.gentoo.org/viewtopic-p-6272005.html#6272005
I can't repreduce the problem. Using "sysctl -w net.ipv6.conf.all.disable_ipv6=1" makes no difference for me. I tried with currently running firefox and tried to restart it too... I'm on vanilla-sources-2.6.33.1 Anyway... ------------- ariel ~ # zcat /proc/config.gz |grep IPV CONFIG_IPV6=y # CONFIG_IPV6_PRIVACY is not set # CONFIG_IPV6_ROUTER_PREF is not set # CONFIG_IPV6_OPTIMISTIC_DAD is not set # CONFIG_IPV6_MIP6 is not set CONFIG_IPV6_SIT=y # CONFIG_IPV6_SIT_6RD is not set CONFIG_IPV6_NDISC_NODETYPE=y # CONFIG_IPV6_TUNNEL is not set # CONFIG_IPV6_MULTIPLE_TABLES is not set # CONFIG_IPV6_MROUTE is not set CONFIG_NF_DEFRAG_IPV4=m CONFIG_NF_CONNTRACK_IPV4=m CONFIG_NF_CONNTRACK_IPV6=m CONFIG_IP6_NF_MATCH_IPV6HEADER=m ariel ~ #
the behaviour descripted by Apetrini depends from the useflag ipv6. my system is completly built with the useflag -ipv6. in this moment, to completly fix the firefox bad behaviour, I'm forced to set the ipv6 useflag and to add the ipv6 module in the kernel.
(In reply to comment #21) > the behaviour descripted by Apetrini depends from the useflag ipv6. > my system is completly built with the useflag -ipv6. > > in this moment, to completly fix the firefox bad behaviour, I'm forced to set > the ipv6 useflag and to add the ipv6 module in the kernel. > which version of nss do you all have installed? If it is less then 3.12.6 please update to testing nss and report. If this resolves it we will go ahead with stabilizing nss-3.12.6.
3.12.6-r1 :cry:
(In reply to comment #22) > > If it is less then 3.12.6 > please update to testing nss and report. If this resolves it we will go ahead > with stabilizing nss-3.12.6. > if this bug blocks nss stabilization, please change the title.
(In reply to comment #22) > (In reply to comment #21) > > the behaviour descripted by Apetrini depends from the useflag ipv6. > > my system is completly built with the useflag -ipv6. > > > > in this moment, to completly fix the firefox bad behaviour, I'm forced to set > > the ipv6 useflag and to add the ipv6 module in the kernel. > > > > which version of nss do you all have installed? If it is less then 3.12.6 > please update to testing nss and report. If this resolves it we will go ahead > with stabilizing nss-3.12.6. > I have 3.12.6-r1 which is already stable on amd64 and x86.
I tried everything. I also removed ipv6 support from my kernel. strace shows something like " ... socket(PF_INET6, SOCK_STREAM, IPPROTO_IP) = -1 EAFNOSUPPORT (Address family not supported by protocol)..." and there isn't any row with "socket(): Address family not supported by protocol". but there are no slowdowns in the menu; firefox works fine. Anyway I take a wild guess and I say it's glibc related, since I have sys-libs/glibc-2.11.1 and Marco Clocchiatti has 2.10.1-r1.(both on amd64) Anyway... I don't have any other idea on how to face that, sorry.
(In reply to comment #26) > > Anyway I take a wild guess and I say it's glibc related, since I have > sys-libs/glibc-2.11.1 and Marco Clocchiatti has 2.10.1-r1.(both on amd64) > may it be. upgrading to sys-libs/glibc-2.11.1 I also have no issue.
I have the same problem. I have these packages installed: www-client/mozilla-firefox-3.6.7 dev-libs/nss-3.12.6-r1 net-libs/xulrunner-1.9.2.7 sys-libs/glibc-2.11.2 I didn't try it on a native X server. Only through freenx-server on both the windows NX client, and on another linux box with the opennx client. For my kernel config: # CONFIG_IPV6 is not set I do however have the ipv6 use flag set. This does not seam to match. Reading the comments, setting IPV6 in the kernel will probably fix this. But I do think it's a bug nonetheless.
this thread was very helpful for me (same problem on two different gentoo machines): http://forums.gentoo.org/viewtopic-t-828606-start-0.html i've done the mentioned workaround in ~/.gtk-2.0 and it works for me: gtk-enable-event-sounds=0 gtk-enable-input-feedback-sounds=0
Same problem for me. Work-around works. This is the interesting part of the stacktrace for when it's blocked for a few seconds: #0 pthread_cond_wait () at ../nptl/sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S:140 #1 0x00007fffdbd96aff in pa_threaded_mainloop_wait (m=0x23caa30) at pulse/thread-mainloop.c:212 #2 0x00007fffdbfafc75 in pulse_driver_open (c=0x23ca3e0) at pulse.c:429 #3 0x00007fffdc9ffcb0 in driver_open (c=0x23ca3e0) at dso.c:273 #4 0x00007fffdc9f82d8 in context_open_unlocked (c=0x23ca3e0) at common.c:293 #5 0x00007fffdc9f8af8 in ca_context_play_full (c=0x23ca3e0, id=0, p=0x23ca450, cb=0, userdata=0x0) at common.c:517 #6 0x00007fffdc9f8eb0 in ca_context_play (c=0x23ca3e0, id=0) at common.c:462 #7 0x00007ffff5ce2067 in nsSound::PlayEventSound (this=<value optimized out>, aEventId=6) at nsSound.cpp:501 #8 0x00007ffff57c1948 in nsMenuPopupFrame::ShowPopup (this=0x13db140, aIsContextMenu=<value optimized out>, aSelectFirstItem=0) at nsMenuPopupFrame.cpp:642 #9 0x00007ffff57cce2b in nsXULPopupManager::ShowPopupCallback (this=0x1613f80, aPopup=0x18354d0, aPopupFrame=0x13db140, aIsContextMenu=<value optimized out>, aSelectFirstItem=<value optimized out>) at nsXULPopupManager.cpp:601 #10 0x00007ffff57cd613 in nsXULPopupManager::FirePopupShowingEvent (this=0x1613f80, aPopup=0x18354d0, aMenu=<value optimized out>, aPresContext=<value optimized out>, aPopupType=ePopupTypeMenu, aIsContextMenu=0, aSelectFirstItem=0) at nsXULPopupManager.cpp:1085 #11 0x00007ffff57b85e6 in nsPopupBoxObject::OpenPopup (this=0x238cc10, aAnchorElement=<value optimized out>, aPosition=<value optimized out>, aXPos=0, aYPos=0, aIsContextMenu=0, aAttributesOverride=0) at nsPopupBoxObject.cpp:133
I also tried a lot of things firefox was working again by unmerging media-sound/pulseaudio (0.9.21.2-r1) have not try to rebuild it yet
(In reply to comment #29) > this thread was very helpful for me (same problem on two different gentoo > machines): > http://forums.gentoo.org/viewtopic-t-828606-start-0.html > > i've done the mentioned workaround in ~/.gtk-2.0 and it works for me: > gtk-enable-event-sounds=0 > gtk-enable-input-feedback-sounds=0 Thanks for this, it helped me too, it has been really frustrating that the few gtk2 based applications has been slow and seem to wait for something. I really look forward for the QT version of firefox.
The workaround does not work for me, but I can observer the blocking behaviour only when a second user logs in to X. I have USE="-ipv6 pulseaudio" set, and can't get sound in pulseaudio to work on the second X session, which is probably related.
(In reply to comment #33) > The workaround does not work for me, but I can observer the blocking behaviour > only when a second user logs in to X. I have USE="-ipv6 pulseaudio" set, and > can't get sound in pulseaudio to work on the second X session, which is > probably related. > Is anyone still having issues with this?
Me, confirmed again that the bug still exists.
Please test if the issue persists with stable xorg and firefox-9.0. Also make sure to test with a blank profile. If the problem persists, reopen this bug. Thanks!
I can no longer replicate this bug using stable xorg and Firefox. Works for me.
(In reply to comment #37) > I can no longer replicate this bug using stable xorg and Firefox. Works for me. Thanks for getting back to us.
