[07:23] is there an ETA for the .27-3 kernel? [14:32] sorry for the newbie question. Where can some one get info about become a kernel developer? [14:50] smb_tp: amd64 - if i installed just that (so not apps/modules) would it help any? [14:50] i wouldn't have network, and maybe even no shell, so I wouldn't be able to upload logs [14:51] but I would be able to see if it pauses at the same point [14:51] CarlFK, It is just the same code just amd64 version for some others to test. So I don't expect a different behaviour. [14:51] figured as much [14:52] CarlFK, Sure you can try. If that works, then maybe there is a 32/64 bit problem lurking [14:56] smb_tp: cat /proc/cpuinfo model name : AMD Turion(tm) 64 X2 Mobile Technology TL-60 [14:56] dpkg: error processing linux-image-2.6.27-3-generic_2.6.27-3.3smb5_amd64.deb (--install): [14:56] package architecture (amd64) does not match system (i386) [14:56] guessing it is looking at the current .. um... os [14:58] CarlFK, Oh yeah. It looks at you installed architecture. You can force it but all your binaries are 32 bit so this probably won't work [14:59] all we need is for it to boot, not ... um; be usefull, right ? [15:01] CarlFK, At some point the ramdisk takes over and that would be 32bit stuff (i don't think there is biarch support), so it might not get very far... [15:04] um.. magic happend [15:05] carl@dv67:~$ uname -a [15:05] Linux dv67 2.6.27-3-generic #1 SMP Mon Sep 8 22:51:51 UTC 2008 x86_64 GNU/Linux [15:05] shell, network... [15:05] CarlFK, Oh right, should probably have more confidence in biarch support then. :) [15:06] [ 0.000000] Command line: root=UUID=270934eb-b4c5-4f17-87e4-8ebea9d98ed3 ro highres=off nohz=off vga=6 [15:06] no pauses - what params do you want ? [15:06] CarlFK, ok, that combination has worked with i386 kernel too [15:07] yes [15:07] CarlFK, without "highres=off nohz=off" with "apic=debug debug" [15:08] CarlFK, Only need to know whether this stops or not. Logs/dmesg only if it happens to _not_ pause [15:08] ok [15:08] what is this biarch support? [15:09] it paused. at least it isn't a surprise :) [15:10] Ok, so in this case i don't need anything more at the moment [15:22] rtg, BenC: apparently neither gcc-4.1 and gcc-4.2 is needed to build kernels in intrepid on community ports. is this correct? [15:24] doko: dunno. I've yet to look at the community ports packages. [15:24] doko: Ben is on vacation this week. [15:33] Hello. I have Ubuntu Hardy kernel 2.6.24-21. This kernel is buggy and I know that > 2.6.25 kernels work with my machine. How can I get newer kernel to Hardy and all necessary packages (linux-ubuntu-modules, linux-headers and linux-image)? Some packages are missing from Intrepid and PPA repositories.. [15:34] timbba: not 'best' but good enough: wget http://people.ubuntu.com/~smb/bug254668/linux-image-2.6.27-3-generic_2.6.27-3.3smb5_i386.deb dpkg -i [15:34] timbba: try the kernel PPA at http://ppa.launchpad.net/kernel-ppa/ubuntu [15:35] I have tried already PPA kernels.. there are no ubuntu-modules packages, so my wifi isn't working [15:36] timbba: 2.6.27 has LUM folded into the package, so LUM is no longer needed. [15:36] So I can get kernel packages, but how about those modules for my wifi card (Zydas based) [15:36] timbba: if they aren't in the 2.6.27 package, then they are not supported. [15:38] oh god.. so nobody can't get Zydas wifi to work? [15:38] I can't remember. you'll have to look in the kernel source for 2.6.27-rc5 [15:39] timbba: what card is it? [15:40] edup something :) [15:41] timbba@media-timbuktu ~ $ dmesg | grep zd [15:41] [ 32.583130] zd1211rw 2-2:1.0: eth0 [15:41] [ 32.583145] usbcore: registered new interface driver zd1211rw [15:41] [ 165.221220] zd1211rw 2-2:1.0: firmware version 4605 [15:41] [ 165.261183] zd1211rw 2-2:1.0: zd1211 chip 079b:004a v4330 high 00-60-b3 AL2230_RF pa0 g---- [15:41] timbba: that doesn't help a lot [15:41] timbba@media-timbuktu ~ $ lsusb [15:41] Bus 002 Device 002: ID 079b:004a Sagem [15:42] timbba: is it in http://linuxwireless.org/en/users/Drivers/zd1211rw/devices? [15:43] yes it is: Sagem XG760A zd1211 079b 004a AL2230 [15:47] timbba: then please file a bug in launchpad describing how the device is not working. From a quick glance, it looks like intrepid's 2.6.27-based kernel should have support for it. No separate modules needed. [15:48] it doesn't find the firmware.. that is the problem. The card itself is recognized [15:49] timbba: that would be the "describing how the device is not working" [15:49] ok :) [15:49] timbba: 'ls -l /lib/firmware' [15:49] how about copying firmware from previous kernel? [15:50] timbba: exactly... :) [15:50] oh.. I have already removed 2.6.27 kernel so there is just... [15:50] ls -la /lib/firmware/ [15:50] total 12 [15:50] drwxr-xr-x 3 root root 4096 2008-09-09 17:21 . [15:50] drwxr-xr-x 15 root root 4096 2008-09-09 17:20 .. [15:50] drwxr-xr-x 3 root root 4096 2008-09-02 05:51 2.6.24-21-generic [15:51] but I will give it a try.. I'll copy the firmware after I have reinstalled the kernel... Let's see what happens [15:52] rtg: I forget why we moved from /lib/firmware//* to /lib/firmware/*. And do our packages handle that transition? [15:55] amitk: Ben's contention was that firmware didn't change enough to bother keeping separate copies per ABI. For most drivers that's probably true. I assume upgrades are handled, but I've not a dist-upgrade in a bit (since FireFox didn't work on some web sites, I went back to Hardy) [16:03] probably a good idea [17:01] Hello Everyone... I'm standing in for BenC today (he is on holiday) for the Weekly Kernel IRC Meeting. [17:02] ogasawara: Can you tell us how we are looking wrt to 2.6.27? [17:03] [Link Regressions https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bugs?field.tag=regression-2.6.27 ] [17:03] pgraner: from my point of view I think we are looking quite good with the move to 2.6.27 [17:04] pgraner: there are the noted list of regressions (link above) but I have gotten good feedback from a handful of those bugs that they will be resolved with the next kernel we release [17:04] Looks like we are sitting at about 30 regressions, with most of them having fixes upstream or plans to get them fixed soon. [17:04] pgraner: for example the suspend/resume issues should hopefully go away [17:04] ogasawara: great [17:05] pgraner: some of the boot failures will look to also be resolved [17:05] ogasawara: we are sitting at about 70 bugs fixed by 2.6.27 [17:05] [LINK https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bugs?field.tag=fixed-2.6.27] [17:05] pgraner: yup, so if we stuck with 2.6.26 we'd have to consider most of those unresolved then [17:06] So far the feeling of the kernel team is that there are no real show stoppers and that we are on track for GO decision on the 14th per [LINK https://wiki.ubuntu.com/KernelTeam/2.6.27-kernel-plan] [17:07] rtg: is looking at RaLink RT2860 LP# 210725 issues [17:07] s/rtg:/rtg/ [17:07] when will -3 be uploaded? [17:08] tjaalton: as soon as -rc6 is in the can. [17:08] rtg: ah, thanks [17:09] anyone from the kernel team have anything to add? [17:09] pgraner: I think it's Stefan thats looking inti the RaLink driver stuff. He built a DKMS package IIRC. [17:10] rtg: ah, correct for some reason I had you down for it. I guess I'm trying to give you too much to do... [17:10] I'm going to start putting some of the DKMS packages on the kernel PPA, so that they are semi-official [17:10] correction smb_tp will be looking at the RaLink RT2860 LP# 210725 issues [17:11] Any questions? [17:12] Ok I guess we'll call this meeting over for today. If anyone has questions or follow up questions you can post them here and I will answer or you can email me at pgraner@canonical.com. [17:12] thanks! [17:16] I've been bitten by a usb bug which has existed upstream since 2.6.25 and which prevents me from connecting to my Motorola phone. A patch exists, but is not yet accepted upstream. Is there any chance this patch could go in? [17:16] johanbr: https://bugs.launchpad.net/ubuntu/hardy/+source/linux/+bug/263217 ? [17:19] laga: Probably not. 2.6.24 works fine for me, but later kernels don't, which is the opposite of what's stated in that bug report. [17:20] There's a discussion archived at http://permalink.gmane.org/gmane.linux.usb.general/8532 [17:21] there's a test PPA [17:22] maybe you can try that kernel [17:22] the problem sounds familiar ;) [17:25] I doubt it'll work, but I could try. [17:54] evening. is it too late to request v4l-dvb/mantis driver to be included for intrepid yet? the driver gives ~270 dvb card and stick drivers available and builds currently without issues on 8.10 [17:58] 270 *new* dvb boards? [17:58] Bor_Ed: is 270 a count or a model number? [17:59] module count [17:59] including frontend, jingles and bells [18:00] and how is it different from the v4l-dvb tree included in the current kernel? [18:00] doesn't support the dvb-c or dvb-t devices around europe that well [18:01] so you're basically suggesting that the kernel devs should include regressions? ;) [18:02] well, I was pointed to here from #ubuntu-devel since the driver level stuff belongs to you guys (I heard, no pun) [18:03] * laga is not a kernel dev, just for the record [18:03] I read that mantis is getting to v4l-dvb kernel tree but not in time for release of 8.10 [18:03] but as one of the mythbuntu guys, i'm interested in better DVB support [18:04] I had to fetch and compile the mantis tree but it worked as a charm on first trial on stock intrepid with build-essential&the rest of the deal [18:04] if it's not even in v4l-dvb yet then changes are probably slim [18:04] maybe it'd be possible to cherry-pick certain (new) device drivers [18:05] but that's just my opinion ;) [18:05] * Bor_Ed is long time linux user and spotted the lack of feat in current ubu... servers still stay on debian, sorry :D [18:06] afaik, mantis is getting merged to v4l-dvb at some point in the future, the prob is though that currently v4l-dvb kernel driver doesn't support much more than couple of dvb-s decoders [18:16] and ati binary driver still refuses to build so I could have DRI [20:09] hi... I'm trying to install the latest Linus kernel. I have the kernel compiled, but can't get an initrd generated. [20:09] I would really like to avoid make-kpkg and creating and installing deb packges. lacking another way that I knew of, I tried that, and it doesn't work.... [20:10] first, the -append-kernel-version freaks out on the -rc5 plus my -foo [20:10] so, I dropped that [20:11] second, it aborts with "arch/xen/ not found". well, yes, that dir does not exist in latest kernels. why does it have to try to build a Xen kernel? I just want an initrd! [20:11] so, I pass --arch "x86_64", but then it fails with "I do not know where the kernel image goes to [kimagedest undefined] ... Please specify a subarch, and try again.." [20:11] I am stuck [20:12] nevermind the last one. I accidentially passed "x86" as arch [20:13] it still trips over the -rc5, though, even without --append-kernel-version [20:14] I had to clean everything and rebuild from scratch. that'll take almost 1h *sigh* [22:06] is there any way to create an initrd without make-kpkg? that takes 20 minutes for me *without* any code or config changes. i.e. first make-kpkg takes 1h20m, second (directly afterwards) takes half an hour (!). creating in initrd takes seconds (e.g. on suse). [22:07] BenB: mkinitramfs [22:07] mjg59: thanks! [22:07] (I wish I'd have known that 2 hours earlier)