Reproducible: Unknown Steps to Reproduce: Just boot my system up :) Actual Results: No known results. Very ocassionaly the computer stops responding, XMMS stops playing, only mouse seem to work. After ~10s it wakes up and behaves normally again. I don't know if this is related to the bug. dmesg is full of the following: warning: many lost ticks. Your time source seems to be instable or some driver is hogging interupts Losing some ticks... checking if CPU frequency changed. warning: many lost ticks. Your time source seems to be instable or some driver is hogging interupts Losing some ticks... checking if CPU frequency changed. etc etc, and they are added randomly: no message in one minute, then 30 lines in one second, etc. I thought that it is because of cpudyn, but when I stopped it the messages just continued to roll. At http://lkml.org/lkml/2004/6/28/208 a guy had the same problem (however I am not getting the dma-related stuff). He thinks that it is the APIC bug. I didn't tried the noacpi switch yet. $ cat .config |grep APIC CONFIG_X86_GOOD_APIC=y CONFIG_X86_IO_APIC=y CONFIG_X86_LOCAL_APIC=y I can't find them in the kernel's menuconfig to turn them off. uhost: Linux moto 2.6.7-gentoo-r8 #4 Sun Jul 4 13:39:41 CEST 2004 x86_64 4 GNU/Linux system: Acer Aspire 1501LMi CFLAGS="-O2 -pipe -fomit-frame-pointer" CHOST="x86_64-pc-linux-gnu" CXXFLAGS="${CFLAGS}" GENTOO_MIRRORS="http://gentoo.ynet.sk/pub http://gentoo.inode.at" MAKEOPTS="-j2" USE="-oss acpi alsa -arts canna cdr cjk cups dvd dvdr gtk gtk2 icq java ldap -kde kerberos -motif nls xvid mozilla moznoirc -mysql pam -qt readline samba -svga tiff unicode usb X" FEATURES="sandbox"
ok, I finally tried the noapic. The result: a) xmms took 90% of cpu time (of amd64 3000 :-) b) the bug was not fixed
I can confirm everything that was written in the original description, I'm having exactly the same problem here (gentoo-dev-sources-2.6.7-r8). I tried pci=noacpi but this didn't help either. I also tried unloading all ALSA modules since someone mentioned that this worked for him, but this didn't solve it either.
Check the manufacturers' website for a BIOS update.
Yeah, I'm checking for new BIOS periodically, unfortunately there is none available for download at the moment. I'm curious what might cause the error - is it bios, hw or amd64 linux port (or all ;)? I don't know.
I'm also getting this problem - I don't get it with the 2.6.5-r1 kernel through...
We are discussing the matter also at http://forums.gentoo.org/viewtopic.php?t=191716&sid=04cb95582e49ca7a91591309bc6fdb3a so you may check there for more info. I have found more interesting issues. I don't know if that relates to the bug, but just to be sure I'll post them here. I am sure that I'm using the one (and only) correct IDE kernel module, however sometimes whole system stops and after ~10s it wakes up. In dmesg I have found something like this: "DMA uncatched interrupt, recovering" (I'll post the correct msg when it happens again ;). Moreover, my keyboard ocassionaly 'lets the key pressed' (I release the key but the system thinks that it is still pressed. Quite annoying when it happens with Backspace :-), so I must hit it one more time to actually release it. I'm sure that the keyboard is ok (it's keyboard on notebook, bought 2 weeks ago). I'm using the synaptics driver for touchpad. Sometimes (when I do the drag operation - doublehit the touchpad and move) the cursor goes in the direction even if I stop moving the finger on the touchpad. In dmesg the following occurs: psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 - driver resynched. warning: many lost ticks. Your time source seems to be instable or some driver is hogging interupts Losing some ticks... checking if CPU frequency changed. warning: many lost ticks. Your time source seems to be instable or some driver is hogging interupts Losing some ticks... checking if CPU frequency changed. psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 4 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 4 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 lost sync at byte 1 psmouse.c: TouchPad at isa0060/serio4/input0 - driver resynched. However this never happens with my USB mouse. I finally found how to get the config info, so I'll post it here one more time :-) Portage 2.0.50-r9 (default-amd64-2004.0, gcc-3.3.3, glibc-2.3.4.20040605-r0, 2.6.7-gentoo-r9) ================================================================= System uname: 2.6.7-gentoo-r9 x86_64 4 Gentoo Base System version 1.4.16 Autoconf: sys-devel/autoconf-2.59-r3 Automake: sys-devel/automake-1.8.3 ACCEPT_KEYWORDS="amd64" AUTOCLEAN="yes" CFLAGS="-O2 -pipe -fomit-frame-pointer" CHOST="x86_64-pc-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3/share/config /usr/lib/mozilla/defaults/pref /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-O2 -pipe -fomit-frame-pointer" DISTDIR="/usr/portage/distfiles" FEATURES="autoaddcvs ccache sandbox" GENTOO_MIRRORS="http://gentoo.ynet.sk/pub http://gentoo.inode.at" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="X acpi alsa amd64 apm avi berkdb canna cdr cjk crypt cups dvd dvdr encode esd foomaticdb gdbm gif gnome gpm gtk gtk2 icq imlib java jpeg kerberos ldap libg++ libwww mikmod mozilla moznoirc mpeg ncurses nls nogcj oggvorbis opengl pam pdflib perl png python quicktime readline samba sdl slang spell ssl tcpd tiff truetype unicode usb xml2 xmms xv xvid zlib"
Hi again, I have moved to 2-6-7-r11 kernel and nothing had changed at all. The 'DMA bug' text is as follows: hda: DMA interrupt recovery hda: lost interrupt It seems to me that the VIA interrupt controller is not working properly, but honestly I don't know.
Just for further information: I installed a 32bit system on my machine with exactly the same kernel version/configuration - no lost ticks messages so I think it's definitely an amd64 bug.
Just tried the mm-sources (2.6.7-mm6) with the same configuration => no lost ticks messages.
While having no problems with mm-sources-2.6.7-r2 after upgrading to mm-sources-2.6.8rc3-r2 the "losing many ticks" messages are back.
I'm convinced its a gentoo problem because if you google "many lost ticks" pretty much the only thing that shows up is the thread on the gentoo forums. If it was a general kernel problem I would assume many other people from other distros would be having the problem too. And it seems that most people that switch to vanilla or mm-sources fix the problem. So maybe we should bring this to the attention of whoever tweakes the gentoo-dev-sources if it hasn't been already???
imho you should ( idon't have amd64 hardware, i am just watching this bug) start experimenting with the different mm-sources kernels. peter said that it worked with 2.6.7-r2, but is not working with 2.6.8rc3-r2. so now you should go back to 2.6.7-r2, verify if it's working. if yes, go to the NEXT mm-sources kernel version. check if it's working and so on. so you need those 2 mm-sources versions where the change occurs. after that, it's a matter of diff ;) (or to look thru the mm-patchset, and see what has changed)
I can confirm comment #10: mm-sources-2.6.7-r6 is OK, but mm-sources-2.6.8_rc2-r2 has the problem. Also, gentoo-dev-sources-2.6.5 was OK, but gentoo-dev-sources-2.6.7 introduced the bug for me.
I was getting the losing clock ticks, etc lines in my messages log. I turned of X and they stopped. So then I started back X again and as soon as I logged in they started up again. I eventually narrowed it down to gkrellm2. I tried changing themes and nothing changed. I also turned off all of the meters and it continued displaying the messages in my log. I know for a fact that it is linked to gkrellm2 for me. I will say though that the messages are ALMOST at regular intervals. They are normally 5 seconds apart (seconds that are divisable by 5), yet there are sometimes that there is one that is out of sync.
I tried to turn off gkrellm2 (2.1.28-r1) on my 2.6.7-gentoo-r12. My gkrellm2 behaves exactly as Eric has described - messages with 5 second periodicity, etc. My GKrellm2 monitors only Mem, Swap, hda temp (through the hddtemp daemon) and something called THRC/THRS - it measures the temperatures of my notebook, I believe that they are extracted via the ACPI interface - I don't have lm_sensors installed. However there are also messages with 30 second periodicity and I cannot turn them off - they are present regardless of whether gkrellm2 is running or not.
I do not have any problems with the message when gkrellm2 is off. When I turn on gkrellm2 after deleting my settings I sometimes get a single message but then I really don't have any problems. I went through and turned on my settings one by one and once it seemed that my volume plugin was doing it - however, it might have been a one-time thing. I have found that one thing that WAS causing it was my battery monitor which uses ACPI. When I turned it off I had much fewer messages appear. When I then tried changing the volume using the volume plugin and from then on I had an increased amount of messages... I am sure about effect the battery monitor had on the messages, however, the sound plugin I am less sure of because when I turn off the plugin, the messages continue.
I want to add that I have this problem as well - just not very regularly. It locks up for about 6-7 secs, and in /var/log/messages, I found a strange thing with these things (I don't remember that I had this issue ad this time, since it is some time ago, but there is some interesting info here): Jul 5 13:20:08 Metis Your time source seems to be instable or some driver is hogging interupts Jul 5 13:20:08 Metis Losing some ticks... checking if CPU frequency changed. Jul 5 13:20:08 Metis warning: many lost ticks. As of cpu-frequency, it's changing frequently because I'm using cpudynd. I don't know whether this is related. And when stepping through /var/log/messages, it looks like I don't get these messages very often (once every 2 weeks or so) but it if happens, there are just _tons_ of these messages. May be, just logging these messages hogs the cpu, but I find that pretty unlikely
I was getting this message on my amd64 machine too. I found a comment on some mailing list (I don't have a URL for it, sorry) that said there is a bug in many vendors' BIOS on amd64 relating to the BIOS emulating PS/2 kbd/mouse for their USB equivalents. It seems that it made the kernel driver think the PS/2 interface supported more features than it actually did, leading to the problem. I think it doesn't show up if you are actually using USB kbd or mouse, just if you are using PS/2. The note suggested disabling this feature in the BIOS, which I did, and the messages stopped for me.
WORKSFORME (>=gentoo-dev-sources-2.6.9-r4)
I'm using gentoo-dev-sources-2.6.9-r14 and I'm getting this message. I'm also using version 2004.3 What information can I post that would be helpful?
