[12:58] <Toroman> hi!  I have an usb wifi dongle (usr805422), when i try to boot up ubuntu with my usr805422, I receive host system error and I cant use my dongle until next reboot....   it only works when i boot up my ubuntu without usb dongle and wait to load desktop manager and ndiswrapper :(    http://pastebin.com/me5a7918   here is more inf. please help me
[12:59] <defcon> belkin ralink rt73 wifi dongle causes a kernel panic when I plug it in
[01:00] <defcon> Toroman, same problem here, kernel panic my caps/scroll lock blink
[01:00] <defcon> everything worked fine in feisty
[01:01] <Toroman> with usr805422 no kernel panic problem but only my usb host dies :D
[01:04] <Toroman> i try my dongle with other distros such as pardus, no problem occured...
[01:06] <JanC> many ralink drivers heave issues with multiprocessing
[01:07] <JanC> so if you have a multi-core/multithreading CPU...
[01:08] <Toroman> I'v very very single and old core :D
[01:09] <JanC> so then there is another issue...  ;)
[01:12] <Toroman> i try this dongle with 3 different computers but they all deliver same error
[01:13] <mjg59> amitk_: http://www.codon.org.uk/~mjg59/tmp/usb_autosuspend
[01:15] <Toroman> I have to unplug my dongle and boot my comp wait  ndiswrapper loads then plug my dongle
[03:25] <defcon> ikm having an issue with gutsy and the 2.6.22-9 kernel, when I plug in my belkins usb wifi dongle I get an instant kernel panic
[03:29] <lamont`> defcon: neato.  what exactly is the panic?>
[03:30] <defcon> how do I debug this and properly report it
[03:30] <\x6e\x65\x72\x64> Hi, I'm trying to compile a dapper kernel on a feisty machine, because the feisty machine is faster, but I'm getting some errors, do I have to make any special changes for this to work?
[03:30] <defcon> it instantly freezes my box and blinks caps lock/scroll lock
[03:32] <defcon> lamont, I have tried -rt and it doesnt panic but I am unable to modprobe the module rt73.ko from serialmonkey, and cannot use the supplied drivers with ubuntu
[03:32] <defcon> 2.6.22-9-386 got the same issue
[03:35] <lamont`> defcon: and the actual panic message is????
[03:35] <lamont`> ah - it freezes the machine, rather than panicing... got it
[03:35] <defcon> i didnt write it down, should there be a log of it
[03:36] <defcon> apon boot it does the same
[03:36] <lamont`> given that panic doesn't trust anything, it tends to only record the info on the console....
[03:36] <defcon> wont boot when initializing interfaces
[03:36] <lamont`> that makes sense
[03:36] <defcon> capslock/scroll lock blink
[03:37] <defcon> what does that mean
[03:37] <lamont`> OTOH, without knowing what the panic message was, life sucks.  If there's no panic message, then life gets more interesting for tracking things down
[03:37] <lamont`> defcon: dunno what the blinking stuff means
[03:38] <mjg59> Blinking means that the kernel has oopsed
[03:38] <defcon> ok so I'll reinstall the kernel and try again supplying u with an error message, I only get it during recovery mode
[03:38] <lamont`> mjg59: cool.  I'll remember that
[03:39] <defcon> am i in the right place for this to be resolved? 
[03:39] <lamont`> defcon: I'm close to running out the door for home, and will be gone for at least an hour at that point.  I'll read scrollback when I get there.
[03:39] <lamont`> defcon: either here, or (better yet), all the detail you can provide, into a bug filed in launchpad
[03:39] <defcon> lamont, ok
[03:39] <lamont`> that way it gets some tracking.
[03:40] <lamont`> and actually, this is the right channel to be discussing your fix for the bug, to make sure that it's implemented correctly, etc.  rather than a support channel....
[03:40] <lamont`> likewise, discussions on figuring out how to best debug the problem are on topic.
[03:40] <lamont`> "fix my bug please, kthxbye" would be off topic. :-)
[03:41] <lamont`> at least, that's how i read things...
[03:41] <defcon> cool
[03:41] <defcon> I appreciate it lamont 
[03:41] <defcon> i'll work on this
[03:41] <defcon> dling/installing the kernel now
[03:42] <defcon> going to reboot now and write this down
[03:42] <defcon> bbl
[03:52] <defcon_> lamont,  lamont` gotta type this out now
[03:54] <defcon_> [39.928296]  kobject_add failed for :d-0000008 with -EEXIST dont try to register things with the same name in the same directory
[03:54] <defcon_> [39.928660]  Kernel Panic - not syncing Creation of kmalloc slab kmalloc_dma-2 size=2 failed
[03:55] <defcon_> where can I post this on launchpad
[03:55] <mjg59> Sounds like a bug in p54
[03:55] <mjg59> defcon_: There should be several other lines as well
[03:56] <defcon_> keep in mind this is a belkins F5D7050 USB wifi card
[03:56] <defcon_> if that matters
[04:08] <defcon_> new bug https://bugs.launchpad.net/ubuntu/+bug/130078
[04:11] <mjg59> defcon_: We need the rest of the message
[04:13] <defcon_> thats where it froze
[04:13] <defcon_> it didnt give any more
[04:14] <mjg59> defcon_: What are the lines before that?
[04:16] <defcon_> I dont remember, let me try and think, it all looked normal
[04:17] <defcon_> something about b**** doesnt have a config file, I know its an app for security policies
[04:17] <defcon_> but I didnt enable it
[04:19] <defcon_> I know
[04:19] <defcon_> bastille does not have a configuration file
[04:19] <defcon_> thats it
[04:20] <defcon_> I did not set anything for bastille, so it shouldnt effect anything I dont think
[04:20] <defcon_> I just removed bastille
[05:17] <bullgard4> kern.log prints: "pnp: Failed to activate device 00:06." How can I determine what '00:06' stands for?
[05:19] <johanbr> lspci
[05:22] <mjg59> No
[05:25] <bullgard4> johanbr: My lspci -vv outputs in the format xxxx:yy:zz.z. How does this format map to the pnp format aa::bb?
[05:25] <bullgard4> aa:bb
[05:26] <mjg59> It doesn't
[05:26] <johanbr> x=a and y=b, I think.
[05:26] <kylem> no.
[05:26] <kylem> god.
[05:26] <mjg59> That output isn't from the pci system
[05:27] <kylem> bullgard4, install pnputils and run lspnp.
[05:27] <johanbr> Oh... sorry for spreading misinformation.
[05:58] <bullgard4> kylem: lspnp prints: "00:06 PNP0303 IBM enhanced keyboard (101/102-key, PS/2 mouse support)." --  Thank you very much.
[05:58] <kylem> bullgard4, no problemo.
[05:58] <lamont> EE: Missing modules (start begging for mercy)
[06:02] <lamont> kylem: so are we already committed to -9.26?
[06:02] <kylem> pardon?
[06:02] <lamont> hppa64.modules was also incorrect.
[06:02] <lamont> mind you, I'm happy just pushing -9.25hppa1 for the moment, in my archive
[06:03] <lamont> as long as we make sure things get right in -9.26 when it does hit...
[06:03] <kylem> i have a few patches i want in tribe4, but they're likely abi-breaking, so i'll have to schedule them with Ben.
[06:04] <lamont> build takes a couple hours, so we'll know in the morning if it's any good.
[06:05] <kylem> if it's just a simple patch, please just post it to kernel-team (instead of asking for a pull for just a single commit)
[06:05] <lamont> I'll post the diff, although it's basically a "replace hppa64.modules with the good one"
[06:05] <kylem> ok.
[06:06] <lamont> I wonder... if it's -9.25hppa1 do I want it to be abi/2.6.22-9.25 or abi/2.6.22-9.24?
[06:07] <lamont> 25
[06:27] <defcon_> lamont, https://bugs.launchpad.net/ubuntu/+bug/130078
[06:28] <lamont> defcon_: and what did it say _before_ those two lines?
[06:28] <lamont> full context is kinda important...
[06:28] <defcon_> before it said that it said cannot load bastille config, I never configured it so I dont believe thats a problem
[06:29] <defcon_> I had to write down then type that out whats on there hehe
[06:29] <defcon_> before that it said loading network interfaces
[06:29] <defcon_> and thats it
[06:29] <defcon_> brb
[06:32] <lamont> ok
[07:35] <defcon_> lamont, is there a way to get the log without writing it all down
[07:35] <defcon_> or a script or something to log it all
[07:36] <lamont> defcon_: personally, I tend to use a camera
[07:36] <defcon_> hehe
[07:36] <lamont> serial consoles are also kinda cool
[07:38] <defcon_> i could enable bootlogd
[07:38] <defcon_> ./etc/default/bootlogd
[07:38] <defcon_> im suprised it is disabled by default
[07:42] <lamont> kylem: fwiw, the diff is 5kb, 500 lines, all of it hppa64.modules
[07:42] <lamont> short enough to just email to the list
[07:42] <lamont> once the build succeeds
[07:42] <infinity> s/once/if/
[07:43] <infinity> Although I have every confidence in your test building process, LaMont. :P
[07:43] <defcon_> brb reboot
[07:43] <lamont> infinity: heh./
[07:43] <lamont> the only oddity is that it's test-building -9.25hppa1 instead of -9.26
[07:45] <lamont> it's doing hppa64/modules build right now... I wish it'd hurry up
[07:46] <lamont> hrm.. should be OTOO 40 minutes
[08:15] <defcon> I see bootlogd is disabled, and broken
[08:15] <defcon> any other app to log the boot seq so I can get a kernel log of it Panicing
[08:15] <defcon> I dont have a digital camera to snap it
[08:33] <defcon> new kernel update still bombs when it see's my belkin usb dongle
[08:44] <defcon> will a kern.log help with a kernel panic
[11:04] <Nafallo> mjg59: http://paste.ubuntu-nl.org/32346/ <-- that's what I could save from the serial window booting with apic=debug
[11:05] <Nafallo> mjg59: http://home.nafallo.info/bugs/sata_nv_take2.txt <-- and that's with the drives at ports 3 and 4.
[01:15] <mjg59> Nafallo: Does acpi_use_timer_override work?
[01:16] <Nafallo> mjg59: instead of noacpi?
[01:16] <Nafallo> apic even
[01:18] <mjg59> Yes
[01:20] <Nafallo> we shall see. brb.
[01:28] <Nafallo> http://paste.ubuntu-nl.org/32360/
[01:28] <Nafallo> http://paste.ubuntu-nl.org/32361/
[01:29] <Nafallo> the first one without apic=debug
[01:35] <Nafallo> http://home.nafallo.info/bugs/sata_nv_take3.txt
[01:35] <Nafallo> mjg59: any ideas? :-)
[01:38] <Nafallo> putting current gutsy -server on there should work, right?
[01:53] <Nafallo> well, that didn't work... :-P
[02:04] <mjg59> Wow. Your hardware is impressively broken.
[02:05] <Nafallo> ASA M2N-E
[02:05] <Nafallo> ASUS even
[02:06] <Nafallo> mjg59: so. what can I do? :-)
[02:07] <mjg59> Get a new motherboard?
[02:07] <Nafallo> ouch
[02:07] <mjg59> Hm. apic=debug doesn't seem to have helped much.
[02:08] <Nafallo> indeed it didn't :-/
[02:09] <mjg59> There ought to have been a dump of the apic mappings
[02:11] <Nafallo> can I force it to do that on the running system or somethiing? :-)
[02:12] <mjg59> Oh, wait, I see the output now
[02:16] <Nafallo> mjg59: I found logcheck have nice logs. do you want all the latest reboots? ;-)
[02:23] <mjg59> No, I don't think that would really help
[02:24] <Nafallo> ehrm. I think you might want to see this one... what's you mailaddress?
[02:27] <Nafallo> ah, no. seems the .txt on the server has the info.
[02:29] <mjg59> Nafallo: It, uh, seems to have the same problem?
[02:30] <Nafallo> they all have :-P
[02:31] <Nafallo> hmm. a friend told me to remove the disabling code from spurious.c :-P
[02:33] <mjg59> Nafallo: So why would I want to look at that one especially?
[02:33] <mjg59> I suspect your machine just has wrong IRQ routing for the noapic case
[02:33] <mjg59> You could try pci=noacpi
[02:34] <mjg59> But it would be more interesting to fix the actual bug (the fact that it doesn't seem willing to boot with the apic)
[02:34] <mjg59> Have you tried the 64-bit version?
[02:34] <Nafallo> I agree. but what more can I do to nail down the cause?
[02:35] <Nafallo> yes. when I removed all drives except the raptors and installed gutsy tribe-3 x86_64 server
[02:35] <ivoks> mjg59: (sorry for distraction) could you take a look at bug 119130 when you have time; it has a solution, so it should be an easy one...
[02:37] <mjg59> Nafallo: And?
[02:39] <Nafallo> mjg59: same thing.
[02:39] <mjg59> Nafallo: Which same thing?
[02:40] <Nafallo> the irq 7: nobody cared thing
[02:40] <mjg59> Nafallo: And without noapic?
[02:40] <Nafallo> and had to boot with noapic
[02:40] <mjg59> ivoks: I don't really have time to deal with acpi-support right now
[02:40] <mjg59> With the same failure message?
[02:40] <Nafallo> yes.
[02:40] <Nafallo> dapper worked though :-)
[02:40] <ivoks> mjg59: ok
[02:41] <Nafallo> with noapic...
[02:41] <Nafallo> so "worked"
[02:42] <Nafallo> atleast I didn't get the irq7 on 2.6.15 :-)
[03:52] <BenC> Anyone else have a Turion x2 laptop that they are running or have tried to run gutsy on?
[03:53] <BenC> rtg: I'm going to start a new thread on lkml about this problem...it's pretty serious regression for us
[03:54] <rtg> BenC: There are several LP reports that I think are related.
[03:57] <BenC> rtg: any pointers?
[03:59] <rtg> I'll see if I can round 'em up.
[04:11] <lamont> BenC: btw, more hppa tweaks for .26 :-(
[04:14] <BenC> rtg: Cc'd you
[04:15] <lamont> imagelist=$(cat kernel-versions | grep ^hppa | awk '{print $4}') && \
[04:15] <lamont>         for i in $imagelist; do \
[04:15] <lamont>           dpkg -x $(ls ../linux-image-$i\_*hppa.deb) \
[04:15] <lamont>                 debian/d-i-hppa; \
[04:15] <lamont>         done
[04:15] <lamont> dpkg-deb: --extract takes at most two arguments (.deb and directory)
[04:24] <BenC> Nafallo: override the PCI slot IRQ's in the BIOS?
[04:25] <BenC> Nafallo: in a lot of cases, the "irqX and nobody care" is a symptom of a broken BIOS, so check for an update on that as well
[04:25] <Nafallo> BenC: thanks :-)
[04:26] <Nafallo> a slightly more expensive solution is PCI-E 3ware RAID1 ;-)
[04:26] <Nafallo> but I'll check this one out first :-)
[04:26] <Nafallo> I should have PNPOS installed in BIOS, right? :-)
[04:27] <BenC> should, yes
[04:30] <Nafallo> well, I'll look soon enough then.
[04:30] <Nafallo> need to put VGA in the box again :-P
[04:33] <Nafallo> hmm. beta bios? ;-)
[04:33] <Nafallo> http://support.asus.com/download/download.aspx?SLanguage=en-us&model=M2N-E <-- /me is confused
[05:12] <abogani> BenC: Could you enable linux-image-debug build for rt? Thanks.
[05:12] <BenC> abogani: can't, the custom-binary builds don't support that
[05:13] <abogani> BenC: Uh, sorry!
[05:29] <BenC> kylem: Do you have a Turion laptop?
[05:29] <kylem> nope.
[05:29] <kylem> i only have one amd box these days.
[05:32] <mjg59> BenC: There's a discussion about missing timer interrupts on mobile AMDs on lkml right now
[05:42] <BenC> mjg59: ah, that's exactly what I need, thanks
[05:43] <BenC> mjg59: subject?
[05:46] <mjg59>  ACPI on Averatec 2370
[05:46] <mjg59> nolapic_timer ought to work if it's the same issue
[05:50] <kylem> Regarding the issue described above that Frank and I are having, I've
[05:50] <kylem> narrowed it down to commit e9e2cdb4[1] : [PATCH]  clockevents: i386 drivers
[05:50] <kylem> heh, must be the same problem.
[05:52] <BenC> mjg59: It didn't work for me
[05:52] <BenC> kylem: same commit that Tim and I came up with
[05:52] <BenC> so three of us separately bisected to the same one
[05:53] <kylem> check out the message from Linus there.
[05:53] <BenC> yeah, I'm about to test that
[05:53] <BenC> kylem: data point, x86_64 kernel boots fine for me
[05:53] <kylem> BenC, it would, it uses completely different code.
[05:54] <BenC> another reason to merge i386/x86_64 :/
[05:54] <mjg59> Andi insists on maintaining different timer code because, erm, well.
[05:55] <kylem> BenC, well, x86_64 hasn't got the new time code yet, because it's still buggy. (clearly. :)
[06:13] <amitk_> mjg59: Patch looks good. Can you send it with a sign-off?
[06:13] <mjg59> amitk_: Didn't it have one?
[06:14] <mjg59> amitk_: Oh - I tried to mail you, but amitk@u.c doesn't work
[06:14] <mjg59> It's on lkml and usb-dev
[06:14] <kylem> mjg59, it's just amit@
[06:17] <amitk_> mjg59: thanks
[06:22] <BenC> rebooting to test the amd broken timer thing
[06:29] <BenC> sweet, it worked
[06:32] <kylem> yeah, i figured it would. the probelm seems to be with nohz, you can't turn off the lapic timer
[06:33] <BenC> I wonder if "nolapic nohz=on" would work
[06:33] <BenC> err nohz=off
[06:34] <BenC> kylem: do you think forcing "return 1" is a good work around for us to use?
[06:34] <BenC> in production I mean
[06:35] <BenC> kylem: does a broken lapic timer mean that kernel disabled nohz?
[06:35] <kylem> i say we add the proper define in now (test the cpuid) and reevaluate the decision after before -rc?
[06:35] <kylem> a lot of testing should happen for beta, so we can see if the issue reoccurs now that we know what to do...
[06:36] <BenC> kylem: you sounds like you're on top of this...can you get a patch in, and a new kernel today (or before Monday)?
[06:37] <kylem> BenC, er, well, need info from you or tim first.
[06:37] <BenC> let me know what you need and I can get it
[06:37] <kylem>  (b) make that function print out the values it uses for debugging (ie the
[06:37] <kylem>      xtended family and model numbers, and the MSR_K8_ENABLE_C1E MSR
[06:37] <BenC> I assume cpuid
[06:37] <kylem>      values)?
[06:37] <kylem> right.
[06:37] <BenC> ok
[08:00] <Nafallo> BenC, mjg59: I think the damn upgrade fixed it.
[08:00] <Nafallo> BIOS
[08:00] <mjg59> Nafallo: Hurrah. Does it work without noapic now?
[08:00] <Nafallo> mjg59: I think so. I'll check my commandline :-)
[08:01] <Nafallo> [   64.748254]  Kernel command line: root=/dev/md1 ro console=tty0 console=ttyS0,38400n8 break=premount
[08:01] <Nafallo> :-D
[08:01] <mjg59> Excellent.
[08:01] <Nafallo> too easy :-P
[08:01] <Nafallo> why didn't I think about this before? :-)
[08:12] <rtg> BenC or mjg59: Am I smoking crack on this one: https://bugs.launchpad.net/ubuntu/+bug/129477 ? What does a Synaptics touchpad do under windows that these guys expect?
[08:13] <mjg59> rtg: It's a new-model ALPS pad. The kernel doesn't recognise it.
[08:13] <rtg> mjg59: So PS/2 emulation is the best we can expect w/Feisty, right?
[08:13] <mjg59> With Feisty, yes
[08:14] <rtg> mjg59: Thanks.
[08:15] <mjg59> rtg: In PS/2 emulation mode it only reports relative motion. In native mode it gives absolute coordinates which give the synaptics userspace driver the ability to implement additional functionality (like scrolling)
[08:15] <rtg> I guess I've never used a touchpad like taht.
[08:16] <mjg59> Basically everything (other than very old stuff and bleeding-edge alps) works like that in Linux
[08:19] <Nafallo> http://home.nafallo.info/bugs/sata_nv.txt btw :-)
[08:19] <Nafallo> I can't see anything weird :-)
[08:21] <mjg59> Yeah, looks fine
[08:21] <mjg59> Note how stuff is all on different interrupts now
[08:26] <gnomefreak> we are tptally against putting 2.6.23 in multiverse right?
[08:26] <gnomefreak> totally even
[08:26] <mjg59> Do you mean universe?
[08:27] <gnomefreak> yeah sorry
[08:27] <mjg59> And I suspect the answer is yes
[08:27] <gnomefreak> someone asked this this morning im getting around to it now
[08:27] <zul> i am against it
[08:27] <gnomefreak> once gutsy stable we cant really touch it
[08:28] <infinity> We've had newer kernels in universe before, and it was a complete disaster.
[08:29] <infinity> So, yeah, we're pretty firmly against repeating that fiasco.
[08:30] <mjg59> Oh, yes, I'd forgotten that
[08:30] <mjg59> 2.6.11, wasn't it?
[08:30] <mjg59> Back in Hoary
[08:30] <infinity> Yeah.
[09:24] <rtg> mjg59: I'm about to instrument a Feisty kernel to extract the Dell 1420 ALPS touchpad E7 report. Is there an easier way? Do you think simply adding alps_model_info will enable the scroll bars?
[09:25] <mjg59> rtg: The issue came up on lkml - simply adding the device to the table didn't seem to work, so I suspect the protocol has changed slightly
[09:25] <rtg> mjg59: drat.