[04:48] <claydoh> BenC: referred to you by hobbsee re: bug 85488
[04:49] <claydoh> what are the possible consequences of disabling 'USB selective suspend/resume and wakeup'?
[04:50] <BenC> claydoh: Breaking suspend/resume and random other USB devices
[04:50] <BenC> claydoh: It's a no-win situation...I have to break some usb host/dev setups in order to make suspend/resume work for a good portion of other people
[04:50] <claydoh> as I thought, thanks
[04:51] <BenC> claydoh: Fixing the drivers it the correct solution (to work with suspend/resume), but I don't have any of that hw, so it's hard for me to do, not to mention having the time
[04:53] <claydoh> but as some (most) sane scanner backends do work it seems easier to take it to the broken backends' maintainers for fixes 
[04:53] <BenC> claydoh: I've updated the bug report with what I just said to you
[04:54] <mjg59> BenC: Isn't USB_SUSPEND about runtime suspend?
[04:55] <BenC> mjg59: Ah right, not sure what I was thinking
[04:55] <BenC> but it is still needed, because the signaling was required for some devices to work
[04:56] <BenC> there's probably some bug references in dapper/edgy changelogs
[04:56] <BenC> Been enabled since 2.6.15-24.41
[04:56] <BenC> no bug reference, shitty
[04:57] <mjg59> Strange that it seems to have worked on edgy
[04:57] <BenC> question is why does it break sane scanning
[04:57] <BenC> that too
[10:59] <raf> goodmorning. i have compiled and installed 2.6.20.3 on feisty buntu this morning..and im unable to install the nvidia driver. i have all the neccesary programs installed gcc, kern. source. has anyone come across this problem?
[02:45] <fabbione> BenC: hey dude
[02:45] <BenC> fabbione: Hey
[02:45] <fabbione> BenC: we got a patch to test to fix DLM in the kernel, but i need to be able to build ia64.. anything we can about it?
[02:45] <fabbione> BenC: it would be really good to have it both in the kernel (the fix) and the workaround in userland)
[02:46] <BenC> fabbione: Yeah, sorry it's taken so long, my i2k doesn't like to boot very often
[02:46] <grazieno> hi people, can someone help me? https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2007-March/000553.html
[02:47] <grazieno> I'm tryng to solve the bug 75935
[02:47] <grazieno> Via VT6410 ide raid
[02:47] <mjg59> See https://wiki.ubuntu.com/KernelTeam
[02:49] <grazieno> mjg59: Did you speak to me?
[02:50] <mjg59> Yes
[02:50] <grazieno> ok, I'm looking
[02:51] <fabbione> BenC: ok.. can you try again today please so that i can test that patch? i don't have another x86 to slam in the cluster and sparc is no go on RHCS
[02:51] <BenC> fabbione: Will do
[02:51] <fabbione> thanks
[03:16] <Keybuk> err
[03:16] <Keybuk> so vmware-config.pl doesn't work on feisty
[03:16] <Keybuk> I'm certain this is known by now :p so what's the workaround?
[03:17] <fabbione> Keybuk: i build the modules manually and install them in /lib../misc/
[03:17] <fabbione> depmod -a
[03:17] <fabbione> modprobe them
[03:17] <fabbione> run vmware-config again
[03:17] <Keybuk> how do you build the modules?
[03:17] <Keybuk> that's what's failing
[03:17] <fabbione> you will need a one liner patch for vmmon
[03:17] <fabbione> the sources are in /usr/lib/vmware/...
[03:17] <fabbione> just untar them and run make
[03:17] <fabbione> one of them will fail
[03:18] <Keybuk> what's the patch?
[03:18] <fabbione> just comment out the line where it fails (duplicate definition of something)
[03:18] <fabbione> it's a .. _syscall1 IIRC
[03:18] <Keybuk> I don't get a duplicate definition error, I get a expected declaration specifiers error
[03:19] <fabbione> yeah that one.. 
[03:19] <fabbione> i need to get off this office.. the Niagara is melting my head
[03:19] <fabbione> -ETOONOISE
[03:19] <crimsun> (http://icanthack.com/?p=53  , for instance; see the comments)
[03:19] <Keybuk> commenting out that line seems to solve it, thanks
[03:19] <Keybuk> then I get
[03:19] <Keybuk> /tmp/vmware-config1/vmnet-only/userif.c:629: error: CHECKSUM_HW undeclared (first use in this function)
[03:19] <crimsun> http://icanthack.com/wp-content/uploads/2007/01/vmnet.patch
[03:19] <fabbione> uh?
[03:20] <fabbione> what version of vmware are you installing?
[03:20] <fabbione> i didn't get that one with 5.5.x
[03:20] <fabbione> 5.5.3-34685
[03:21] <Keybuk> 5.5.2
[03:21] <BenC> Keybuk: check the patches in lrm for the modules there
[03:21] <BenC> same stuff
[03:21] <Keybuk> BenC: it never seems to want to use those modules though
[03:21] <BenC> Keybuk: tar xf, patch, tar cf, let vmware-config do it's thing
[03:21] <BenC> Keybuk: Patch the source, and let vmware-config.pl build from it
[03:22] <BenC> Keybuk: Or get ws6 beta and it will build without problems
[03:57] <Mithrandir> \o/
[03:57] <BenC> w00t
[03:58] <BenC> "Assigned to me (64)"
[03:59] <Nafallo> rock on!
[04:58] <Mithrandir> BenC: can you tell me anything about the scope of the bug?  Is it as bad as I was afraid of on #-devel?
[04:58] <BenC> Mithrandir: it's so bad we need it in beta
[04:58] <Mithrandir> ok
[04:58] <BenC> Mithrandir: but we already have a fix, and I am in the process of uploading the new tarball, and a test package for sladen to try
[04:59] <Mithrandir> ok, goodie.
[04:59] <BenC> I'll hold the tarball on rookery till he confirms the fix
[04:59] <BenC> Mithrandir: it's non abi bump, I just patches 2.6.20-12.19
[04:59] <BenC> my stuff in git is too big and untested right now to upload
[04:59] <BenC> plus it's an abi bump
[04:59] <Mithrandir> ok, so it's localised to just the IDE driver?
[05:00] <BenC> it could affect a lot of things actually
[05:00] <Mithrandir> ugh.
[05:00] <BenC> but I think libata is the only thing using devres that we pulled in for last upload
[05:00] <Mithrandir> do you have the diff somewhere?  Not that I know much of the kernel, but more eyes can't hurt.
[05:00] <BenC> the bug was actually in lib/iomap.c
[05:01] <BenC> the diff was pulled from an upstream fix for the bug
[05:01] <BenC> I just cherry picked it, and it mentions the exact trace back we have
[05:01] <Mithrandir> ok, that sounds good then.
[05:02] <kylem> iz pretty critical fix
[05:02] <BenC> <paste, sue me>
[05:02] <BenC>     [PATCH]  pci_iomap_regions() error handling fix
[05:02] <BenC> 
[05:02] <BenC>     It appears that the pcim_iomap_regions() function doesn't get the error
[05:02] <BenC>     handling right. It BUGs early at boot with a backtrace along the lines of:
[05:02] <BenC> 
[05:02] <BenC>     ahci_init
[05:02] <BenC>     pci_register_driver
[05:02] <BenC>     driver_register
[05:02] <BenC>     [...] 
[05:02] <BenC>     ahci_init_one
[05:02] <BenC>     pcim_iomap_region
[05:02] <BenC>     pcim_iounmap
[05:02] <BenC> so it's dead on
[05:02] <mjg59> Ooooooooooooooooh
[05:02] <mjg59> It's the case where ahci tries to load, fails and then ata_piix takes over
[05:02] <BenC> right
[05:03] <mjg59> That makes sense
[05:03] <mjg59> And would explain why it only seems to be hitting ICH6
[05:04] <kylem> right
[05:04] <kylem> in hindsight pulling in this devres stuff may not have been such a hot idea...
[05:05] <Mithrandir> ok, so we get the fix confirmed for a couple of boxes, then upload, then roll isos then test.
[05:05] <BenC> well, we needed some of the new libata drivers and libata-acpi fixes...was easier to pull 3 commits for devres than backport the stuff
[05:05] <BenC> Mithrandir: sounds good
[05:07] <BenC> test deb is almost done
[05:07] <BenC> tarball is on rookery already
[05:07] <kylem> BenC, 8-way not fast enough? :P
[05:07] <BenC> hehe
[05:08] <BenC> it's built, but kernel-package is just too stupid to not try to run the build 5 more times during the packaging process
[05:12] <BenC> deb is heading to rookery
[05:14] <crimsun> I can try, too (I'm affected)
[05:14] <crimsun> s/can/would like to/
[05:16] <BenC> http://people.ubuntu.com/~bcollins/kernels/bug-93648/linux-image-2.6.20-12-generic_2.6.20-12.20_i386.deb
[05:16] <BenC> 8172368b874be9ed6777c818052e08a4  linux-image-2.6.20-12-generic_2.6.20-12.20_i386.deb
[05:24] <crimsun> BenC: -12.20 confirmed working on my hardware.
[05:25] <crimsun> thanks much
[05:25] <BenC> crimsun: great, thanks
[05:25] <\sh> BenC: where can I download your test package? I would like to test it on my t43
[05:25] <BenC> BenC http://people.ubuntu.com/~bcollins/kernels/bug-93648/linux-image-2.6.20-12-generic_2.6.20-12.20_i386.deb
[05:25] <BenC>  8172368b874be9ed6777c818052e08a4  linux-image-2.6.20-12-generic_2.6.20-12.20_i386.deb
[05:26] <\sh> give me a few
[05:28] <\sh> dpkg: Fehler beim Bearbeiten von linux-image-2.6.20-12-generic_2.6.20-12.20_i386.deb (--install):
[05:28] <\sh>  nicht ganz gelesen in buffer_copy (Backend dpkg-deb whrend ./boot/vmlinuz-2.6.20-12-generic)
[05:28] <\sh> Running postrm hook script /sbin/update-grub.
[05:28] <\sh> You shouldn't call /sbin/update-grub. Please call /usr/sbin/update-grub instead!
[05:28] <\sh> rebooting now
[05:28] <BenC> ignore that, it's just older kernel-package on my build box
[05:28] <BenC> or old stuff on your system :)
[05:36] <\sh> BenC: works like a charme
[05:39] <BenC> \sh: Thanks
[05:40] <Nafallo> 3/3 :-)
[05:44] <\sh> BenC: np :)
[08:33] <Lure> BenC: re bug 63123> so you need the actual commit in acpi-test where the fix was pulled in/merged into linus tree?
[08:34] <BenC> Lure: No, I need the commit SHA id from linux-2.6.git that I can use with git-cherry-pick
[08:34] <Lure> BenC: I am still looking into polling issue - not sure if this is required or not
[08:34] <BenC> Lure: One of them was ok, but it didn't pick cleanly, looks like it depends on other commits not listed
[08:34] <dade`> BenC: 
[08:34] <dade`> r u alive
[08:34] <BenC> dade`: barely
[08:34] <dade`> I'm using 2.6.20-12-generic
[08:35] <dade`> and psmouse does not work
[08:35] <Lure> BenC: ok, I have manually applied them here, will try git-cherry-pick and find you workable set
[08:35] <BenC> Lure: Ok, please order them as well, so I get the right ones in dependency order
[08:35] <BenC> Lure: Thanks
[08:35] <Lure> BenC: ok, thank you
[08:35] <dade`> cat /dev/input/eventN where N is the right one :P does not give output
[08:36] <BenC> dade`: Does dmesg show anything interesting?
[08:36] <dade`> I tried reloading the module and I get
[08:36] <dade`> [  586.136000]  input: SynPS/2 Synaptics TouchPad as /class/input/input9
[08:36] <dade`> [  586.336000]  psmouse.c: Failed to enable mouse on isa0060/serio1
[08:36] <dade`> BenC: this--^
[08:38] <BenC> dade`: Looks like it is failing to send a command to the ps2 device
[08:39] <BenC> dade`: Does going back to -11 fix it?
[08:39] <BenC> Nothing related to this changed between -11 and -12, so I'm a bit confused
[08:39] <dade`> BenC: don't know if -11 worked.. I user edgy before
[08:39] <dade`> used
[08:40] <BenC> Does booting edgy kernel fix it?
[08:40] <dade`> Should I try now ? :P
[08:40] <BenC> sure :)
[08:41] <dade`> ok.. wait -.- bbl
[08:49] <BenC> crimsun: FYI, all your patches from last night/this morning, applied. Thanks
[08:49] <bdmurray> BenC: Do you have a second?
[08:49] <BenC> bdmurray: Sure
[08:50] <bdmurray> it seems suspend and hibernate are confused on my laptop
[08:50] <bdmurray> hibernate writes to swap and suspend goes into low power
[08:50] <BenC> Sounds like a gnome-power-manager issue
[08:50] <BenC> bdmurray: Does it do that when you write mem/disk to /sys/power/state ?
[08:50] <bdmurray> I tried sudo pmi action and the behavior was the same
[08:51] <BenC> oh way, no that's right
[08:51] <BenC> suspend is low power, and hibernate is suspend2disk
[08:51] <bdmurray> really? that seems backwards to me
[08:51] <BenC> hibernate requires no power, suspend is low power sleep mode
[08:52] <BenC> well, the naming is just a userspace convention...the kernel just knows suspen2mem and suspend2disk
[08:52] <Mithrandir> bdmurray: hibernate is what bears do during winter.  Suspend is what you do when you delay something for a bit.  Makes lots of sense to me.
[08:52] <BenC> userspace seems to call s2mem Suspend/Stand-by, and s2disk Hibernate
[08:53] <dade`> ..
[08:53] <bdmurray> Mithrandir: that might help me remember, thanks
[08:53] <dade`> the edgy kernel does not boot anymore, is that right ?
[08:54] <dade`> BenC: 
[08:54] <BenC> dade`: Possible that it doesn't agree with feisty userspace
[08:54] <BenC> dade`: Do you have a edgy livecd you can boot?
[08:54] <dade`> heh, so I could try the -11, but you just told me that there is no difference in psmouse
[08:55] <dade`> BenC: I don't but.. I always used the psmouse touchpad on edgy
[08:55] <dade`> I'm sure it works
[08:56] <BenC> dade`: I just need to alleviate any possibility that somehow the hw has gone bad...not likely but it has happened before
[08:56] <BenC> I've chased a bug for 2 days just to have the person tell me the cable they used for a USB device was just too long :/
[08:57] <psusi> lol
[08:57] <dade`> it's a lappie, internal touchpad.. 15 hours ago I was using edgy and my sweet touchpad
[08:58] <BenC> dade`: Well, it's an isolated problem. I've been through almost the entire bug list yesterday and I didn't see anyone else mentioning anything like this
[08:59] <BenC> dade`: If you could try to test it under the edgy kernel somehow (just boot edgy kernel to initramfs, and check dmesg for the mouse mesg's)
[09:00] <dade`> hm
[09:20] <dade`> BenC: -11 has the same problem. I'm going to download an edgy live cd :(
[09:32] <crimsun> BenC: thanks
[09:39] <BenC> dade`: Ok
[11:06] <pepsiman> Hi, I've upgraded to feisty and the cx88-dvb module isn't loaded at boot, it was in edgy.  Any idea why it has changed?  The other modules for this card are loaded correctly