[02:43] <Kano> hi, some things about your current kernel
[02:44] <Kano> a) ndiswrapper is too old, remove or update - with 1.30 and installed usb wlan drivers you get a system lock
[02:44] <Kano> b) pata-sis is borken in git, -8- kernel worked
[02:45] <Kano> if you want to fix some little things like hostap + via intel southbridge bugfix take a look here
[02:45] <Kano> http://kanotix.com/files/kernel/kernel-update-pack-noide/source/
[02:45] <Kano> ndiswrapper 1.37 is not fully stable with 2.6.20 for me, will try to find a stable svn snapshot
[02:46] <Kano> bye
[03:43] <mgalvin> crimsun: any idea if/when the patch mentioned in #
[03:43] <mgalvin> https://launchpad.net/bugs/87253 will make it into feisty?
[05:44] <crimsun> mgalvin: it's already in feisty's ubuntu-2.6.git; expect it in the next kernel upload (you'd have to ask Ben for an ETA)
[05:46] <mgalvin> crimsun: ok cool, thanks, as long as it make it's way in eventually I am happy :)
[05:51] <mgalvin> crimsun: i asked earlier and got an "i think so" but... do you know for sure if these are in or getting in... http://svn.sourceforge.net/viewvc/mactel-linux/trunk/kernel/mactel-patches-2.6.20/
[05:52] <mgalvin> i only ask again because if they are not I am interested in looking into it, it just don't want to waste time if they are already in or planned to be in
[05:57] <crimsun> I can only speak for audio [which I've addressed] 
[05:58] <crimsun> I'm afraid I don't have any additional time atm to check
[06:00] <mgalvin> ok, np, thanks
[06:01] <mgalvin> i can dig in myself and check
[06:07] <crimsun> be aware that some of that code is already outdated
[06:07] <crimsun> certainly the sigmatel diff is incomplete
[06:10] <mgalvin> crimsun: thanks for the heads up
[12:25] <Whoopie> BenC: Hi, will tp_smapi added to ubuntu kernel 2.6.20?
[02:05] <fforw> hi
[02:05] <fforw> I have a Dell Latitude D820 laptop and have been trying to get feisty fawn running on it. works fine so far but teh wireless drivers don't work.. found the driver (ipw3945) and followed the build instructions (involving replacing a module from the kernel source)
[02:05] <fforw> but now I get some "no rule for target ubuntu/net/d80211/ieee80211.o" error.. (translated back from German)
[02:05] <fforw> any ideas?
[02:10] <fforw> the module is ieeee80211, a prerequisite for ipw3945
[02:14] <neuralis> er
[02:14] <fforw> yes?
[02:14] <neuralis> ubuntu ships ipw3945 in l-r-m; why did you build it manually?
[02:14] <fforw> because it did not work..
[02:15] <neuralis> you'll have to be a bit more specific.
[02:15] <fforw> looked around and found instructions to install a newer version of it including the new ieee80211 
[02:16] <fforw> lspci showed the device, the module was loaded but no activity.. everything seemed fine, it just can't find anything or connect to anything
[02:18] <neuralis> strange. well, as long as you have the kernel headers and general build tools installed, compiling the new driver should simply work.
[02:19] <fforw> the ieee80211 build process complains about a conflicting kernel version and wants to disable/outcomment it and add its own version
[02:19] <fforw> conflicting module version
[02:19] <fforw> in the kernel
[02:28] <fforw> bbl
[07:38] <[g2] > BenC do you still need a dmes/lspci for the Kernel SATA boot issue ?
[07:39] <BenC> which boot issue?
[07:39] <BenC> bug number would be helpful
[07:45] <[g2] > BenC one sec
[07:45] <[g2] > https://launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/84964
[07:46] <[g2] > from my limited searching it still appears to be a critical bug
[07:47] <[g2] > basically the box doesn't boot without "irqpoll" on herd 4
[07:57] <kylem> i think my crestline reproduces that.
[07:57] <kylem> [g2] , can you try booting with acpi=off or whatever and tell me if that helps? (it's what i did)
[07:58] <kylem> (booting with irqpoll didn't fix things for me, which is why i ask)
[08:05] <[g2] > kylem thx. I'm trying that now
[08:06] <[g2] > it's good to know that someone has reproduced the issue also
[08:07] <kylem> i just tried without the piix ide driver built into the kernel at all, and it's still an issue, so i'm kind of confused.
[08:07] <kylem> livecd boots are fine, after install boots are boned, which confuses the hell out of me. :\
[08:08] <kylem> [g2] , what is the symptom you see? i get "link down" on all sata links from ahci.
[08:09] <kylem> hmm, i should try with i386.
[08:40] <[g2] > kylem without irqpoll or acpi=off it won't boot from cdrom
[08:40] <[g2] > that's i386 on a P4 with an intel 865 mobo
[08:42] <kylem> and edgy works flawlessly for you?
[08:42] <[g2] > kylem yeah
[08:42] <kylem> ok.
[08:42] <[g2] > edgy's installed on the hd
[08:42] <[g2] > I'm really looking forward to feisty as I want the updated ivtv drivers
[08:45] <kylem> can you cat /proc/interrupts from your working system and /msg it to me?
[08:45] <[g2] > kylem sure is pastebin ok ?
[08:46] <kylem> i think i just found something in common between working and non-working on my system.
[08:46] <kylem> absolutely.
[08:46] <kylem> i figured my acpi bios was just broken, but if other people are seeing this, something more insidious might be happening (the box i'm seeing it on isn't production so.. :/)
[08:47] <[g2] > http://pastebin.ca/370648
[08:48] <kylem> ah, no sata?
[08:48] <[g2] > yeah it's a sata drive
[08:48] <[g2] > but the sata driver isn't recognized
[08:48] <[g2] > with acpi=off
[08:49] <[g2] > well at least I didn't see it recognized
[08:49] <kylem> hrm
[08:49] <[g2] > I could reboot with irqpoll
[08:50] <kylem> ahci uses msi for me, so even with pci=nomsi irqpoll i am still getting no luck.
[08:50] <[g2] > irqpoll allows me to boot
[08:51] <[g2] > boot from cdrom which is on ide
[08:51] <[g2] > I've got a sata dvd drive and I don't remember if irqpoll worked with that
[08:52] <kylem> hrm. ok.
[08:52] <[g2] > kylem did you want to see the edgy working system ?
[08:52] <kylem> /proc/interrupts and dmesg on pastebin would be fantastic.
[08:52] <[g2] > from edgy ?
[08:52] <kylem> yeah.
[08:52] <[g2] > ok. sure
[08:53] <kylem> i'm adding some crap to my kernel to see if i actually get any interrupts from libata (it shares the irq with uhci, so it's hard to tell.)
[08:53] <kylem> for a while i thought maybe uhci was stealing the interrupts but it seems not.
[08:57] <[g2] > http://pastebin.ca/370676
[08:57] <kylem> thanks
[08:58] <kylem> yeah, definitely looks like there's no interrupt.
[08:59] <[g2] > libata
[09:07] <BenC> so is piix conflicting with ata_piix?
[09:08] <BenC> if that's the issue, I can fix it real quick
[09:09] <kylem> no.
[09:09] <kylem> not for me at least.
[09:09] <BenC> the last comment is a different bug then?
[09:10] <kylem> it's entirely possible intel managed to fuck the acpi bios completely on this sdv.
[09:10] <BenC> I thought I had commented out the entries in piix that matched in ata_piix
[09:10] <kylem> but it's pretty similar, edgy/feisty cds work, first boot into installed system fails miserably and no disks get see.
[09:10] <kylem> n
[09:13] <BenC> a conflict between piix and ata_piix seems impossible
[09:14] <[g2] > kylem / BenC there's an oops in usb during the failed boot
[09:14] <BenC> oh wait
[09:14] <BenC> CONFIG_ATA_PIIX isn't defined, CONFIG_ATA_PIIX_MODULE is
[09:14] <BenC> retarded
[09:15] <kylem> [g2] , oh?
[09:15] <[g2] > kylem yeah.
[09:16] <kylem> interesting.
[09:17] <kylem> i don't think i see that. can you try booting, wait for the initramfs, cat /proc/kmsg and take a picture of the oops?
[09:17] <[g2] > kylem I can capture the boot on serial
[09:18] <[g2] > I'll boot with console=ttyS0
[09:18] <kylem> ok.
[09:18] <[g2] > and capture it on my laptop
[09:18] <[g2] > i'll pastebin in a minute or two
[09:19] <BenC> kylem: This *_MODULE thing could be the cause of a lot of problems, so keep it in mind when you check bug reports
[09:19] <BenC> I fixed all the ones I had
[09:19] <kylem> _MODULE?
[09:19] <BenC> kylem: The idea was that some of the pata drivers only had some of the matching PCI id's for the same IDE driver, so I had to keep the IDE driver enabled, but make it disable the matching PCI id's (so it would use pata)
[09:20] <kylem> i see. hmm.
[09:20] <BenC> but I used "#ifndef CONFIG_ATA_PIIX", and it wasn't matching
[09:20] <BenC> it could account for this sort of thing (CD working, but normal boot isn't)
[09:21] <BenC> piix, generic and siimage were the three drivers affected
[09:22] <kylem> hmm.
[09:23] <BenC> piix should only be used on a hand full of devices
[09:24] <[g2] > http://pastebin.ca/370701
[09:26] <BenC> g2: Can you pastebin "lspci -vv"?
[09:26] <BenC> and /proc/interrupts
[09:26] <zul_> heylo
[09:26] <[g2] > BenC sure. Do you want me to boot edgy ant past that ?
[09:26] <[g2] > or feisty with irqpoll ?
[09:26] <BenC> either way should be fine
[09:27] <[g2] > both are fine with me
[09:28] <BenC> looks like your ICH5 and UHCI are sharing IRQ's, just wondering if anything else is on it
[09:32] <[g2] > http://pastebin.ca/370715
[09:36] <BenC> wtf
[09:36] <[g2] > the UHCI, IDE and SATA are all on irq 169
[09:36] <BenC> is this on edgy?
[09:36] <[g2] > yeah
[09:37] <BenC> ok, irq's are routed to different numbers...
[09:37] <[g2] > 2.6.17-11-generic
[09:37] <BenC> I wish libata would use the driver name instead of just libata for request_irq() :/
[09:38] <BenC> weird, on this one ide, sata and uhci show same irq in lspci, but it's not showing that in proc/interrupts
[09:39] <BenC> ah, 201 is the audio
[09:39] <[g2] > I'm guessing the ide is getting remapped
[09:39] <BenC> so libata is taking taking 169 twice, and uhci_hcd is taking it once
[09:41] <BenC> Can you do /proc/interrupts on feisty?
[09:41] <[g2] > sure
[09:42] <[g2] > BenC do you want me to boot with irqpoll or acpi=off  ?  Does it matter at all ?
[09:42] <BenC> try without
[09:42] <[g2] > it doesn't boot without that's the original problem
[09:43] <BenC> be better to see what it's doing with acpi irq routing
[09:43] <[g2] > ok so irqpoll ?
[09:46] <BenC> yeah
[09:50] <[g2] > BenC ok... I'm winding up in BusyBox in the initramfs (I'm guessing there's a race condition)
[09:51] <[g2] > in the initramfs /proc/interrupts has usb 18 with uhci_hcd:usb3, ide2 with 100K ints
[09:52] <BenC> ah, ide2
[09:52] <BenC> ok, are you in busybox?
[09:52] <[g2] > ide1 is mapped to int 15
[09:52] <[g2] > yeah
[09:52] <BenC> rmmod piix
[09:52] <BenC> modprobe ata_piix
[09:52] <BenC> see if that does anything
[09:52] <[g2] > device or resource busy ?  -f ?
[09:52] <BenC> hmm
[09:53] <BenC> ide modules are prone to being load-only
[09:53] <[g2] > device or resource busy.
[09:53] <BenC> ok, boot with whatever options you need to get to 2.6.20 being up
[09:53] <BenC> rm -f /lib/modules/`uname -r`/kernel/drivers/ide/piix.ko
[09:53] <BenC> update-initramfs -u
[09:53] <BenC> then reboot without any options
[09:56] <[g2] > there is no piix in drivers/ide
[10:00] <[g2] > BenC this is from the LiveCD
[10:00] <BenC> Ah, it's drivers/ide/pci/piix.ko
[10:01] <BenC> But you can't do this from a livecd
[10:02] <[g2] > nod
[10:03] <[g2] > an yeah it is in ide/pci
[10:11] <[g2] > BenC ok. I powered off for a few minutes and booted feisty with irqpoll
[10:13] <[g2] > http://pastebin.ca/370766
[10:17] <BenC> [g2] : Ok, can you pastebin lsmod output?
[10:22] <[g2] > http://pastebin.ca/370780
[10:29] <BenC> piix is definitely loaded, and shouldn't be
[10:30] <BenC> do you have feisty installed or are you doing this from livecd alone?
[10:31] <[g2] > BenC livecd, I'd like to get feisty installed
[10:31] <BenC> You should be ok if you just rm the piix.ko driver
[10:31] <BenC> next kernel upload will have that fixed
[10:31] <BenC> I think piix and ata_piix don't play well together
[10:31] <[g2] > ok
[11:02] <[g2] > BenC / kylem thx -- Feisty's installed now
[11:02] <BenC> getting rid of piix fixed it up?
[11:02] <BenC> or still using irqpoll?
[11:03] <[g2] > BenC no I just installed and rm'd piix.ko and booted
[11:03] <BenC> sounds good, thanks for testing
[11:03] <[g2] > so to be clear, no I'm not using irqpoll
[11:03] <[g2] > and yes I am not loading piix
[11:04] <[g2] > it's a pleasure testing :)
[11:04] <[g2] > thx for the great help
[11:04] <BenC> [g2]  next kernel update should be fixed without needing any special hacks
[11:04] <BenC> np
[11:04] <[g2] > next+1 :)
[11:04] <[g2] > I'm apt upgrading now so I don't think you mean that one
[11:05] <[g2] > BenC you don't really have a hardware lab to test kernels on correct ?
[11:06] <[g2] > heh... other than "the world"
[11:28] <BenC> [g2] : Lab, no, just the machines I have, and other developers have
[11:28] <BenC> we have a support center that does testing for certain machines
[11:29] <[g2] > BenC how might one get a machine added to support center ?
[11:30] <[g2] > or do similar testing
[11:48] <BenC> [g2] : Contact jbailey@canonical.com