dtchen | macli: that's one way, yes | 00:10 |
---|---|---|
macli | dtchen: I ran git clone git://kernel.ubuntu.com/ubuntu/ubuntu-jaunty.git , but got fatal: serious inflate inconsistency1) | 00:17 |
=== SEJeff is now known as SEJeff_away | ||
=== k-s is now known as k-s[AWAY] | ||
=== k-s[AWAY] is now known as k-s | ||
=== k-s is now known as k-s[AWAY] | ||
maxb | window level all | 07:50 |
=== _Purple_ is now known as _Purple_away | ||
=== _Purple_away is now known as _Purple_ | ||
=== k-s[WORK] is now known as k-s | ||
=== k-s is now known as k-s[AWAY] | ||
=== k-s[AWAY] is now known as k-s | ||
=== k-s is now known as k-s[AWAY] | ||
setuid | Question: I'm working on some firewire drivers, and noticed that the kernel I'm using doesn't support large volumes on the target device. I've seen some of the recent commits and wanted to roll a new kernel... so I built one using make-kpkg, after oldconfig and friends... | 15:46 |
setuid | ..and it seems the new kernel (2.6.29+) has renamed the SATA devices back to /dev/hdX. Is there some way I can get an *EXACT* match of the working 2.6.18 kernel's config into a working 2.6.29 kernel? It seems there's some subtle difference I can't nail down. | 15:46 |
amitk | setuid: options get moved around or renamed or removed quite a bit. It is very hard to get an exact match on two kernels that are this far apart. | 15:56 |
setuid | I would think that an oldconfig, making sure .config is the same (modulo additions in the .29 kernel), that it would work | 15:56 |
setuid | Why /dev/sdX got renamed to /dev/hdX and now refuses to boot from SATA drives, boggles me | 15:56 |
setuid | Even passing root=/dev/sda2 fails, because there _are_ no /dev/sdX devices. It's weird. | 15:57 |
rtg | its not altogether surprising. you are straddling the whole libata transition which happened about 2.6.24 | 15:58 |
setuid | Ok, is there a way to get firewire devices to recognize an 8TB volume with .18? Or is that simply not possible? | 15:59 |
trichobezoar | What has to be done to get the fix mentioned in https://bugs.launchpad.net/ubuntu/+source/linux/+bug/275120/comments/5 | 15:59 |
setuid | Everything after .18 seems to be a mess, with anything I've thrown at it. Some supports booting, others fail on some devices, it's all random and de-evolved after .18 | 15:59 |
ubot3 | Malone bug 275120 in linux "Atheros WLAN AR242x LED doesn't work" [Undecided,Fix released] | 15:59 |
trichobezoar | part of the ubuntu kernel? I upgraded to the beta but the rfkill/led dont work as specified | 15:59 |
setuid | ahhh, I think I see the issue... .18 has a bunch of lib/modules/drivers/scsi/sata*ko modules that aren't in .29 | 16:03 |
setuid | hrm, I don't remember SATA being called out specifically in the config... maybe that's the issue. | 16:04 |
setuid | amitk, Where would those be stashed? | 16:05 |
setuid | ah... here we go | 16:07 |
=== k-s[AWAY] is now known as k-s | ||
levonshe | Hello all, I am trying to use vanilla 2.6.26 kernel instead of ubuntu's 2.6.24-23-generic, but sd module does not use ide modules, any clues? | 16:20 |
setuid | modules don't use modules | 16:21 |
amitk | setuid: sure they do, they depend on other modules | 16:24 |
levonshe | _setuid_, you wrong. Modules just expose API for any interested party, look for example modules.dep or lsmod - the count more than 0 shows module used. | 16:24 |
setuid | modules rely on interfaces provided by other modules, of course, but they do not "use" them. Semantics, I know. What is your question? Because sd no longer uses the ide modules? | 16:25 |
levonshe | _setuid_, the boot process does not detect IDE disk, giving me "waiting for a root file system process" and then resolving to Busybox, I saw also that udev cannot do pre-mount scripts snf indeed, nor /dev/hda1. nor /dev/sda1 is created although I stated all possible ide and scsi modules while cre4ating initramfs | 16:44 |
setuid | levonshe, Looks like a similar problem I'm dealing with | 16:47 |
setuid | sata moved to libata, and it's disabled by default in the default kernel source | 16:47 |
=== k-s is now known as k-s[AWAY] | ||
levonshe | _setuid_, the question is general one - could smb replace ubuntu kernel and stay alive ? what is a procedure ? | 16:55 |
infinity | levonshe: Enabling libata in your kernel build should make it happy. | 17:04 |
setuid | here goes... | 17:08 |
levonshe | _infinity_ Thanks, I am dowing it right now but still I am not conviced that I can replace ubuntu kernel with vanilla kernel | 17:09 |
smb_tp | generally I would say one should try to use the ubuntu config as starting point. there probably are not all modules present but it should work | 17:09 |
setuid | yep, looks like Linux still can't handle large volumes (8TB in my case) | 17:12 |
setuid | http://rafb.net/p/aTdy0G86.html | 17:13 |
setuid | That's what I get with 2.6.29.1 | 17:13 |
setuid | This seems to be related: http://lkml.org/lkml/2009/4/10/361 | 17:14 |
setuid | It works fine over USB2 (at 600k/sec. speeds, ugh!), but not Firewire | 17:15 |
levonshe | One more problem puzzles me. I made installation from alternative cd (kernel 2.6.24-23-i368) and several very basic symbols (like irq_restore) were absent comparing to generic kernel for x86 ?why .. | 17:15 |
=== SEJeff_away is now known as SEJeff | ||
setuid | hrm, sg_readcap indicates that this device DOES support RC16... something is fishy here. | 17:32 |
setuid | I wonder if I pull Linus' tree if the results will differ | 17:32 |
setuid | read capacity (16): transport: Host_status=0x05 [DID_ABORT] | 17:33 |
setuid | Driver_status=0x00 [DRIVER_OK, SUGGEST_OK] | 17:33 |
=== k-s[AWAY] is now known as k-s | ||
Fundlefroop | ? | 18:17 |
macli | hi, I git clone git://kernel.ubuntu.com/ubuntu/ubuntu-karmic.git, cd ubuntu-karmic and git tag | grep Ubuntu*, but it shows nothing | 18:41 |
* setuid boots 2.6.30 and hopes for the best | 18:43 | |
rtg | macli: none too surprising since I've have not created any tags yet | 18:43 |
macli | rtg: thanks, I thought I could try the latest kernel like 2.3.30 for ubuntu using ubuntu-karmic.git | 18:46 |
rtg | macli: its a bit steamy yet, so beware. your best bet is one of the kernels in http://kernel.ubuntu.com/~kernel-ppa/mainline | 18:47 |
macli | rtg: so I could simply download the kernel image from the link and boot from it, no need to compile | 18:51 |
rtg | macli: yep. wget http://kernel.ubuntu.com/~kernel-ppa/mainline/SOMEKERNEL.deb; sudo dpkg -i SOMEKERNEL.deb | 18:52 |
z_existence | anyone know about rfkill handling ? i get XF86WLAN key pressed when I press the switch, but nothing is actually DONE | 18:56 |
setuid | Can someone tell me why Linux is de-evolving so fast? It's becoming more and more unstable as the months and years go on. | 19:07 |
setuid | Stuff that used to work, now no longer does, and when someone fixes that, they break something else, and it just cascades. | 19:08 |
z_existence | anyone working on linux-backports-modules-jaunty ? | 19:08 |
macli | I git cloned ubuntu-jaunty.git and ran AUTOBUILD=1 NOEXTRAS=1 fakeroot debian/rules binary-generic | 19:43 |
macli | but get ld: BFD (GNU Binutils for Ubuntu) 2.18.0.20080103 internal error, aborting at ../../bfd/merge.c line 865 in _bfd_merged_section_offset | 19:43 |
rtg | macli: why don't you just build it normally the first time, e.g., 'debuild -b' | 19:44 |
macli | rtg: I followed https://help.ubuntu.com/community/Kernel/Compile#Install%20the%20new%20kernel, it does not mention 'debuild -b' and I typed the debuild -b , it found so such command | 19:47 |
smb_tp_ | macli, probably "debuild -b -uc -us" otherwise it wan't to sign the packages | 19:47 |
rtg | macli: apt-get install devscripts build-essential fakeroot | 19:47 |
* setuid has a howto on my blog for this | 19:49 | |
setuid | http://blog.gnu-designs.com/building-custom-kernels-for-ubuntu | 19:49 |
rtg | this is a good page for vanilla Ubuntu builds: https://wiki.ubuntu.com/KernelTeam/KernelMaintenance#Performing builds | 19:50 |
setuid | Anyone know of a good > 4 bay SATA enclosure? I'm trying to replace my Drobo2 with something that'll work with Linux/LVM | 19:57 |
popey | setuid: edge10 | 19:57 |
setuid | No kernels support the Drobo over 1394b and the Linux kernel that is on the firmware of the Drobo itself, doesn't support RC16 | 19:57 |
alex_joni | setuid: there's a lan thingie for drobo | 19:58 |
popey | droboshare | 19:58 |
setuid | alex_joni, Right, but it too, is just a Linux embedded target, doesn't support RC16 either | 19:58 |
alex_joni | setuid: bummer | 19:58 |
setuid | Drobo is limited to 2TB partitions on Linux (mine is 8TB, works with Linux, but not over 1394) | 19:58 |
popey | setuid: edge10 DAS or NAS devices are v nice | 19:58 |
setuid | And I'm limited to ext3, no optimizations... can't use XFS, geli, or cryptfs either | 19:59 |
setuid | popey, Looking now | 19:59 |
setuid | The DAS801t looks nice | 19:59 |
setuid | The TS-809 Pro is nice too, but cost-prohibitive (126MB/sec. reads and 111MB/s. writes) | 20:00 |
setuid | $2,300+ for that little device | 20:00 |
setuid | Anyone happen to know if I can LVM multiple external eSATA enclosures together into one logical volume? | 20:04 |
dtchen | rtg: i've been waiting for the pcm_lib and mid-layer changes to settle a bit. looks like a should be able to push a patch in a day or so. | 21:24 |
dtchen | looks like i* | 21:24 |
erle- | the topic is not up to date? | 21:41 |
erle- | i would like to discuss a bug i filed a few days ago: | 21:41 |
erle- | https://bugs.launchpad.net/ubuntu/+source/linux-ports-meta/+bug/361222 | 21:41 |
ubot3 | Malone bug 361222 in linux-ports-meta "kernel 2.6.28-11 boot fail with dm-crypt" [Undecided,New] | 21:41 |
dtchen | erle-: well, 2.6.28-11.42 has been tagged | 21:46 |
=== mcasadevall is now known as NCommander | ||
=== NCommander is now known as mcasadevall | ||
z_existence | anyone familiar with acer_wmi | 22:54 |
=== mcasadevall is now known as NCommander | ||
=== Daviey_ is now known as Daviey |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!