[12:26] <BenC> mjg59: should we consider the ide suspend/resume patch?
[12:29] <mjg59> BenC: Which one?
[12:29] <mjg59> Oh, the one that hit LKML? We already have it
[12:29] <mjg59> If it gets merged, drop our one
[12:37] <dilinger> oh
[12:37] <dilinger> if this doesn't work, i won't be able to reboot back into 2.6.12, will i?
[12:37] <dilinger> thanks to udev
[12:39] <jbailey> You just won't have dynamic devices.
[12:40] <dilinger> ok
[12:40] <jbailey> Hmm
[12:40] <jbailey> And nothing that got modprobed by hotplug.
[12:40] <jbailey> So do an lsmod first. =)
[12:41] <dilinger> ok
[12:45] <BenC> mjg59: ok
[12:46] <jbailey> dilinger: You still have keyboard access to this box, yes?
[12:46] <jbailey> dilinger: Networking doesn't come up automatically with the new udev, known bug.
[12:46] <dilinger> jbailey: there is a keyboard hooked up to it; however, it's a few hundred miles away
[12:46] <jbailey> dilinger: Err.
[12:46] <jbailey> Don't do this then.
[12:46] <dilinger> jbailey: what needs to be done for that?
[12:47] <dilinger> well,i've got someone there who can fix it
[12:47] <BenC> add "auto eth0" to /etc/network/interfaces
[12:47] <jbailey> dilinger: No idea, I haven't tracked it closely enough to know.
[12:47] <dilinger> oh right, since by default it uses a hotplug mapping
[12:47] <BenC> right
[12:47] <dilinger> i can just kill that mapping as well now, right?
[12:48] <BenC> yeah
[12:48] <dilinger> thanks
[12:48] <BenC> auto works for both kernels/udev
[12:49] <dilinger> yep
[12:49] <dilinger> just wanted to make sure nothing else might be relying on the mapping
[12:50] <BenC> anyone know off hand, the url for the sparc/hppa build logs?
[12:55] <makx> jbailey: you are only on freenode?
[12:57] <makx> aah you left oftc..
[12:58] <jbailey> BenC: http://bld-3.mmjgroup.com/buildLogs/
[12:58] <BenC> thanks
[12:58] <jbailey> makx: I don't always log in everywhere.
[12:58] <jbailey> makx: I have different accounts / different IRC networks depending on what I'm doing atm.
[12:59] <makx> hehe, ok jbailey. was quite surprised, how easy Md was to the initramfs-tools hooks. seems to work out nicely :)
[01:00] <jbailey> Cool.
[01:00] <BenC> lamont-away: ping
[01:02] <BenC> lamont-away: any idea why dh_clean failed on hppa and causes linux-source-2.6.15 build failure?
[01:15] <dilinger> woo!
[01:15] <dilinger> it worked
[01:15] <dilinger> that's w/ SMP enabled, too (which didn't work w/ breezy's 2.6.12)
[01:19] <BenC> sweet
[01:20] <dilinger> i'll do some stress testing over the weekend, see if i can't get it to fall over
[01:20] <dilinger> oh great, now i can use smartd, too
[01:25] <jbailey> dilinger: Nice! =)
[01:25] <dilinger> hm.  actually, no i can't :/
[01:28] <dilinger> dilinger@throat:~$ sudo smartctl -s on /dev/sda
[01:28] <dilinger> Informational Exceptions (SMART) disabled
[01:30] <dilinger> i guess i'll be looking at that awful code again, this needs to work for my day job as well
[05:02] <jbailey_> BenC: FYI - new kernel doesn't fix the X hangs on ppc64. =)
[05:02] <jbailey_> (It was time to go to bed anyway)
[05:03] <infinity> Who uses ppc64 anyway?  No one /I/ like.
[05:03] <jbailey_> But...
[05:03] <jbailey_> But...
[05:03] <jbailey_> Are you sure?
[05:04] <infinity> Oh, now that makx has gotten udev torn out of Debian's initramfs, I'm going to look at the remaining merge candidates.
[05:05] <jbailey_> Well fine.
[05:05] <jbailey_> Sleep for me!
[09:24] <doko> BenC, fabbione: did you plan an k8 flavour for the i386 arch?
[09:24] <doko> for dapper ...
[09:25] <Earthpig> Good question: I have to use k8 in i386 mode at work due to legacy apps. :-/
[09:25] <infinity> An i386 chroot, or ia32-libs aren't good enough for you?
[09:26] <Earthpig> ia32-libs weren't enough; sun-jre-1.3 wouldn't work with it.
[09:26] <Earthpig> didn't get around to a chroot.
[09:27] <fabbione> doko: didn't we agree on amd64-generic?
[09:27] <fabbione> k8 is already specific
[09:29] <infinity> Earthpig ...?  Sun provides a linux-amd64 build of their JRE too...
[09:29] <Earthpig> infinity: only 1.5, from memory.
[09:30] <doko> fabbione: yes
[09:30] <infinity> Earthpig : And 1.4... You really need 1.3?... Ugh.
[09:30] <Earthpig> infinity: oracle installer :)
[09:31] <Earthpig> as of November there's an amd64 build, but back in June there wasn't.
[09:46] <CataEnry> hi all
[02:19] <CataEnry> hi all
[04:15] <BenC> lamont: ping
[04:21] <lamont> ack
[04:22] <zul> heylo
[04:23] <BenC> lamont: the hppa build failure, should I worry about it?
[04:39] <lamont> it's retrying now
[04:45] <BenC> ok, so it wasn't my fault? :)
[05:48] <zul> dude its *always* your fault ;)
[05:49] <BenC> some cases of "fault" allow me to delegate responsibility though :)
[06:21] <fabbione> BenC: sparc is uploading 7.9 now
[06:22] <BenC> cool
[06:51] <infinity> BenC : Do we know the cause of these messages?
[06:51] <infinity> Assertion failed! qc->n_elem > 0,drivers/scsi/libata-core.c,ata_fill_sg,line=2504
[06:51] <BenC> nope, never seen/heard of it
[06:52] <infinity> Joy.  I get them every once in a while.  Oddly enough, I can reliably reproduce it when running update-initramfs -u, of all things.
[06:53] <BenC> something is calling ata_fill_sg() with a zero sized sg list
[06:53] <BenC> well, zero sized queue command list
[06:54] <BenC> looks harmless, since it will just not do anything
[06:54] <infinity> Harmless, as long as it wasn't meant to do anything. :)
[06:54] <infinity> But if it IS harmless, we'll want to shut it up before release.  A dmesg with a mess of those is scary.
[06:54] <BenC> well, true, if it had commands, but didn't update the list elem count, then that's bad :)
[06:55] <BenC> what ata driver are you using?
[06:55] <infinity> ata_piix
[06:56] <mjg59> infinity: ahci loads, finds it can't claim your device and then fails. ata_piix then gets loaded.
[06:57] <mjg59> Some PCI devices can be used in either mode, but it depends on BIOS setup
[06:57] <infinity> mjg59 : Sure, but typically when things fail to bind to a device, they gracefully unload again.
[06:57] <mjg59> infinity: Not nowadays, no
[06:57] <mjg59> PCI drivers stick around, in case you hotplug something they can bind to
[06:58] <mjg59> And so you can echo stuff into new_id
[06:58] <infinity> But if I hotplug something, it would get modprobed again anyway, wouldn't it?
[06:58] <mjg59> Well, yes
[06:58] <mjg59> But the latter argument still stands
[06:58] <infinity> The New World Order is too confusing for me. :)
[06:59] <mjg59> Ah!
[06:59] <infinity> Perhaps this is why I still run monolithic kernels on my servers.
[06:59] <mjg59> That would be why my battery life is sucking
[06:59] <mjg59> Every time ipw2100 gets a corrupt packet, it disables C3 and C4
[06:59] <mjg59> Cunting thing
[06:59] <CataEnry> bye :)
[07:14] <siretart> hi
[07:14] <siretart> AcidPils1: acx111 working for you?
[07:15] <AcidPils1> not really, it loads but the firmware is missing
[07:15] <AcidPils1> ERROR 2 trying to open firmware image file 'RADIO16.BIN': file not found - make sure this EXACT filename is in eXaCtLy this directory!
[07:15] <siretart> hm. 
[07:16] <siretart> lets see what firmware is in the package
[07:18] <AcidPils1> there is a RADIO15.BIN but no RADIO16.BIN in /usr/lib/firmware
[07:18] <siretart> interesting
[07:19] <infinity> File a bug on linux-restricted-modules for me to include the newer firmware.
[07:19] <AcidPils1> http://paste.ubuntulinux.nl/5485 <-- that are the files
[07:19] <infinity> Bonus points if you can point me to where to download it from.
[07:19] <AcidPils1> infinity: ok 
[07:19] <siretart> AcidPils1: do you have perhaps a RADIO16.BIN on your drivers cd (the windows ones)
[07:20] <AcidPils1> wrong question, do you perhaps have a windows driver cd ;)
[07:20] <siretart> ah, I see
[07:21] <infinity> Okay, I found a copy of that firmware.
[07:21] <infinity> File the bug, please, so I remember in the morning.
[07:21] <infinity> It's about 6 hours past my bedtime now. :)
[07:21] <AcidPils1> infinity: ok 
[07:22] <siretart> :)
[07:23] <infinity> Can you try the firmware here and confirm it works?
[07:23] <infinity> http://195.66.192.167/linux/acx_patches/acx_fw/acx111_xterasys_xn2522g/fw3/
[07:30] <BenC> infinity: any luck with vga16fb yet?
[07:36] <AcidPils1> infinity: seems to work, but i dont get a a link to my ap, strange
[07:36] <AcidPils1> the error is gone
[07:38] <siretart> AcidPils1: ifconfig up the interface first
[07:38] <BenC> ifup eth0 actually
[07:39] <BenC> or whatever your interface is
[07:39] <AcidPils1> ifup wlan0 i know
[07:39] <AcidPils1> dhcp client comes up but gets no ip
[07:39] <BenC> iwconfig wlan0 show anything?
[07:40] <BenC> like the correct essid and such
[07:40] <AcidPils1> yes
[07:40] <BenC> signal?
[07:40] <AcidPils1> Link Quality:0  Signal level:0  Noise level:0
[07:40] <fabbione> re
[07:41] <BenC> fabbione: hey, there's an aweful lot of bugs concerning 2.6.12-10, regressions from 2.6.12-9
[07:41] <fabbione> BenC: what kind of regressions?
[07:41] <BenC> if you search bugzilla for 2.6.12-10, you'll find them all
[07:41] <BenC> crashes, especially acpi related
[07:41] <fabbione> binary crap -ENOCARE.. we can't fix them
[07:41] <fabbione> we didn't touch ACPI with security :/
[07:41] <BenC> but they all say "worked in 2.6.12-9"
[07:42] <BenC> most crash during boot, during hotplug
[07:42] <fabbione> i can see 3 bugs
[07:43] <fabbione> 20107: it's definetely not the kernel
[07:44] <fabbione> we call an helper script to update grbu menu stuff
[07:45] <BenC> 20366, 20391
[07:45] <BenC> those are two in my current list
[07:46] <fabbione> yes.. the same i am looking at
[07:47] <BenC> I've seen more, but I don't think they all say 2.6.12-10 (some start out, "security upgrade" or something)
[07:47] <fabbione> the only ones i saw were binary drivers related
[07:47] <BenC> if they are invalid, can you close them?
[07:48] <fabbione> i find it weird because they are all reporting problems again subsystem that we didn't touch at all
[07:48] <BenC> yeah, I knew it was odd
[07:48] <BenC> but when > 1 complain about the same thing, it tends to add some credibility
[07:49] <fabbione> yup i agree
[07:49] <AcidPils1> hmmm, wlan0 (WE) : Driver using old /proc/net/wireless support, please fix driver ! may this be the error?
[07:49] <BenC> AcidPils1: nah, my rt2500 says the same thing, and it works
[07:50] <AcidPils1> ah ok, stupid card :(
[07:51] <fabbione> BenC: there is only a scary patch in -10- that's the one that changes the ABI
[07:51] <fabbione> i wonder if that could be the culprit
[07:51] <fabbione> it's quite invasive
[07:51] <fabbione> ideally we could try to revert that and see if it fixes the problem
[07:55] <fabbione> but it's contained inside sysctl
[07:56] <fabbione> it's an internal 
[07:56] <fabbione> it's an internal function that changes
[07:56] <fabbione> but it adds stuff to a struct...
[08:03] <fabbione> i really can't see anything there
[08:03] <fabbione> if there was a remote problem with acpi and the scary patch, none of the machines would have survived longer than a few seconds
[08:04] <fabbione> because the change touches all subsystem
[08:04] <fabbione> +s
[08:06] <AcidPils1> now its getting strange, acx_pci: probe of 0000:01:07.0 failed with error -5, and then ndiswrapper version 1.5 loaded (preempt=yes,smp=yes) oO
[08:07] <AcidPils1> why ndiswrapper? i used it but deletet it from /etc/modules
[08:11] <BenC> maybe ndiswrapper is taking the resources
[08:11] <BenC> try deleting the acx entry for ndiswrapper (maybe some script is autoloading ndiswrapper because it is associated with one of your devices?)
[08:15] <mjg59> infinity: There's 640x400 timing information in modedb.c
[08:15] <mjg59> infinity: Sticking that in vga16fb.c and changing the vertical refresh seems to be good
[08:20] <mjg59> Anyone here familiar with sata?
[08:23] <jbailey> What type of familiarity are you looking for?
[08:23] <mjg59> Oh utter arse
[08:24] <mjg59> The code I've been working on for the past few days was already sent to linux-ide on Saturday
[08:24] <mjg59> (Ha ha ha)
[08:37] <BenC> lol
[09:09] <mjg59> Oh, except this one's less good than mine
[09:11] <mjg59> Massive wheel re-invention
[09:49] <dilinger> jbailey: what does initramfs-tools actually use to detect which modules to modprobe during boot?
[09:55] <makx> udev
[09:57] <makx> the initramfs-tools hooks are in the udev package
[10:08] <dilinger> thanks
[10:10] <dilinger> makx: at least w/ the udev that's in sid (backported to a sarge-like system), i get no love wrt piix and ide-generic.  i'm putting htem in /etc/mkinitramfs/modules for now
[10:10] <dilinger> oops, wrong channel
[10:27] <dilinger> makx: ping
[11:27] <jbailey> dilinger: I'ts a bit different from breezy to current dapper.
[11:27] <jbailey> dilinger: In breezy, it walked sysfs and modprobed the modalias.
[11:27] <jbailey> dilinger: In dapper (2.6.15) it walks sysfs and tickle the uevent files.  udev then sees a new uevent.