[04:34] i heard the intel eeprom corruption issue has been fixed, has that been put into synaptic yet? === mdz_ is now known as mdz [09:09] what does the "PIIX3...enabling passive release" message mean which comes up in KVM? [09:23] mdz: It means that the bios in kvm fails to enable passive release for the piix3 pci bridge, and the kernel fixes this. [09:24] I don't think it should be a KERN_ERR thing. [09:26] Seems I'm not alone: http://lkml.org/lkml/2008/8/8/198 [09:27] mdz: Oh, or are you asking what passive release is? [09:31] soren: no, you answered my question, thanks [09:31] mdz: Any time. [11:09] umm... we have a serious regression in intrepid's kernel [11:10] drbd isn't functional - while the drbd works on linus's kernel, it doesn't on ubuntu's, even tough it's the same patch/module [11:12] could someone assist me in detecting where is the bug? [11:17] i think that's a known issue [11:17] check the meeting minutes of the server team [11:18] thanks [11:36] laga: can't find anything... do you recall which meeting that was? [11:41] doh. it references your nick name... [11:41] http://ubuntuserver.wordpress.com/2008/09/30/server-team-20080930-meeting-minutes/ [11:42] eh... :) [11:43] bottom line: drbd 8.2.7~rc1 works with vanilla kernel and doesn't with ubuntu's latest git [11:43] and vanilla kernel is the latest 2.6.27 RC? [11:43] no, rc7 [11:44] the same one which is base for ubuntu's kernel [11:44] ah, okay [11:44] my usual answer to such problems is git-bisect, but maybe one of the kernel developers already has an idea what's going on [11:45] ivoks: kernel config differences/ [11:45] ? [11:45] IntuitiveNipple: i used ubuntu servers config for vanilla [11:46] there are differences, cause vanilla doesn't have all that stuff ubuntu has [11:49] The Kconfig looks pretty straightforward [11:49] anybody know (or can guess) when the ubuntu kernel will be rebased against -rc8? I have a machine that needs a fix from there :) [11:49] IntuitiveNipple: everything looks fine [11:49] IntuitiveNipple: it just doesn't work - i have a strace output from userland tools [11:49] ivoks: Is the module loading and not working - any log clues? [11:50] module is loading, but when userspace want's to connect to it, it can't [11:50] but cn is loaded [11:51] :q [11:51] ups :) [11:54] i could upload strace output if that could help [11:55] with ubuntu kernel, userspace tool timeouts (strace show that it reads eth0 and never finishes) [11:56] while on the vanilla kernel, same tool works (strace shows on that same point read(5, "Unconfigured\n", 4096) ) [11:59] paran: it has already been rebased against -rc8. Upload will probably happen on Monday. [12:00] IntuitiveNipple: is it normal to have something like 'ESPIPE (Illegal seek)'? :) [12:01] ivoks: I've not seen that much so, no, I doubt it. [12:01] ivoks: Is it using a single ethernet port, or the bonding driver? [12:02] single [12:02] amitk_: sounds great, thanks. [12:03] IntuitiveNipple: if you are interested in problem, i could provide you access to machine (both i386 and amd64) with full setup [12:07] ivoks: I doubt I know enough of the specifics to be of any help [12:07] IntuitiveNipple: ok, thanks for assisting [12:08] ivoks: Is the *only* difference between the working and non-working configuration, the kernel the system boots with? [12:09] IntuitiveNipple: yes [12:11] IntuitiveNipple: i don't think that ESPIPE is problem, since seeking socket isn't possible at all :) [12:12] ivoks: possibly a symptom of the root cause [12:13] could be... [12:13] i've asked drbd devs, so i'm waiting for an answer [13:56] hi! any change getting the HSO driver to ship with intrepid? http://tinyurl.com/4jynn6 [13:56] there are many option 3g devices that need this driver [13:57] and it's actively developed [13:59] I honestly thought it was included in vanilla 2.6.27, but obviously it's not [13:59] Wellark^: hso is already merged into the mainline kernel in 2.6.26, so we carry it already. Is the pharscape driver a different version? [13:59] see drivers/net/usb/hso.c [14:00] hmm.. didn't spot that in gitweb.. [14:01] Wellark^: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-intrepid.git;a=tree;f=drivers/net/usb;h=704c3d6c5ecce21316f65a33d802a1ce54ed6b8a;hb=HEAD [14:02] ah, ok. I checked drivers/usb/serial where option.c is :) [14:03] hmm.. option forum states that 1.6 is the latest.. intrepid kernel has 1.2.. [14:12] BenC: can you think of a good reason _not_ to go with upstream e1000e? I've just pulled the NVM protection patch from Linus, so I'm thinking about dropping the SourceForge version of e1000e. [14:13] rtg: none that I'm aware of...what does Intel say about it? [14:13] do they have any requirements for us to use the sf one? [14:13] BenC: it seems upstream is where all of their development effort has gone lately. I'm wondering why they have a duplicate project in the first place. [14:14] BenC: I'm not aware of any requirements from Intel. [14:15] BenC: I think I'll just swizzle the configs for now and see if anyone squawks. [14:16] rtg: ok [14:17] BenC: I'm alos gonna upload -rc8 plus that one e1000e patch. [14:21] rtg: ok [14:23] BenC: is there a trick to setting these tags after a rebase so that the next time you 'insertchanges' you don't get all of the UBUNTU history repeated? [14:25] rtg: debian/scripts/misc/retag [14:26] rtg: ./debian/scripts/misc/retag [14:26] huh [14:27] fairly new script [14:27] I was just gonna ask when that little tidbit appeared :) === ivoks_ is now known as ivoks === BenC1 is now known as BenC === asac_ is now known as asac [18:07] ibex 2.6.27-3-generic clonezilla output "Try to turn on the harddisk "/dev/sda" DMA... HDIO_GET_DMA failed: Inappropriate ioctl for device" [18:08] what command is it probably running? [18:13] hdparm [18:24] # hdparm -d /dev/sda " HDIO_GET_DMA failed: Inappropriate ioctl for device" [18:24] is this a problem? [18:29] CarlFK, No, just a result from moving to lib(s)ata. IIRC this automatically wants to set the right dma mode [18:30] thanks. I'll pass that on to clonezilla [18:31] can someone help me out ? [18:31] i have some strange feelings about virtualization [18:32] what will ubuntu be supporting with the intro of intrepid? [18:33] i'm using xen but i read at a few places intrepid will only support xen domu? [18:36] can someone confirm that? [19:12] [ 10.725634] ata1.00: limited to UDMA/33 due to 40-wire cable [19:12] it's an 80 [19:12] who's problem? [19:18] you could try a knoppix cd [19:18] see if thats giving the same output? [19:18] it could be a hardware problem? [19:19] what does the bios say [19:19] CarlFK, It's from libata-core, which kernel version? [19:19] Linux dhcp91 2.6.27-3-generic #1 SMP Wed Sep 10 16:02:00 UTC 2008 i686 GNU/Linux [19:20] I was about to install daily u-server, so should have -4 in about 30 min [19:24] CarlFK, Doesn't look to me like there was much activity in that area since then... [19:27] cobradevil: bios - hardly says anything - I am lucky to see a model number [19:30] CarlFK, From cable_is_40wire() in drivers/ata/libata-core.c it looks either the controller thinks it is 40wires or it doesn't know and none of the drives on that controller can tell... [19:30] smb_tp: might be older drives - 40 and 80 gig [19:34] CarlFK, There is some description to force cable type in Documentation/kernel-parameters libata-force.. Maybe that could help [19:34] CarlFK, libata.force= (to be correct) [19:34] ill give it a shot and see what happens [19:34] thanks === asac_ is now known as asac-the-nm-rock === asac-the-nm-rock is now known as nm-rocker === nm-rocker is now known as asac [22:16] where are docs for append ... libata.force=2:1.5g,2.15:3.0g,2.03:noncq,udma4 [22:53] smb_tp: 2.6.27-4-server works fine: [ 4.510417] ata1.00: configured for UDMA/100 [ 4.551749] ata1.01: configured for UDMA/66 [23:19] rtg: http://launchpadlibrarian.net/17919677/syslog (from bug 273633) ... any ideas? what does "disassociating by local choice" means? [23:20] rtg: also this guy has a bogus ethernet driver (eth0: driver is 'NULL(info.linux.driver)'. [23:21] this means that hal didnt provide any driver name (i added that NULL (info...) thing as a hack at some point for broadcom wireless) [23:44] CarlFK, Hm, since the libata code did change, maybe it was the code for the controller. Well, it works now... === Ng_ is now known as Ng