=== mgalvin [n=mgalvin@ubuntu/member/mgalvin] has joined #ubuntu-kernel [12:05] down to 233, better than I thought [12:05] grats [12:05] almost 150 bugs [12:06] well, most were already fixed in breezy, a lot of others were fixed simply because we have 2.6.15 (containing patches needed) [12:06] I did fix 30 or so in -8.10 though [12:09] i would be happy if could fix this stupid segfault in my own code :( === zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [12:20] heylo [12:20] mjg59: i dont get a blacklight [12:28] bah [12:28] the sunfire i sent davem got banged up pretty good in transit [12:28] maybe i shouldn't have done UPS ground :/ [12:30] dilinger: Ugh. [12:39] zul: Ok, cool [12:39] zul: What BIOS version do you have? [12:40] uh...good question im on my laptop right now === doko [n=doko@dslb-084-059-073-197.pools.arcor-ip.net] has joined #ubuntu-kernel [12:48] zul: Heh :) dmidecode might tell you, if you're lucky [12:49] A03 [12:59] Ah, ok [12:59] A02 seems to work, weirdly [01:02] bah.. === zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [02:59] oh hockey game.. === mgalvin [n=mgalvin@cpe-69-205-38-37.nycap.res.rr.com] has joined #ubuntu-kernel === BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-kernel [03:17] BenC : No kernel yet? [03:17] doing the dpkg-buildpackage now [03:18] See, I completely ignored my alarm clock and slept in, just to give you more time. :) [03:18] thanks, you're so considerate :) [03:19] I expect payment in cows. [03:19] one steer, bull, cow or heffer? [03:21] As long as it's dead and tastes good. [03:24] ah, a managable bug list [03:24] active bugs, 68 [03:26] Wow. Do mine next! === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | New git tree for dapper: https://wiki.ubuntu.com/KernelGitGuide | 2.6.15-8.10uploaded (The "Fabio does Dallas" release) === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | New git tree for dapper: https://wiki.ubuntu.com/KernelGitGuide | 2.6.15-8.10 uploaded (The "Fabio does Dallas" release) [03:44] sorry to be annoying, but I am starting to write a review of flight 2, do we have an ETA of some kind? (today || end of the week)? [03:45] "Soon" [03:48] infinity: So, uh [03:48] infinity: Is our ltmodem linked on boot like our fglrx and nvidia? [03:52] fabio does dallas? ewww.. [03:53] go rent the full length video and enjoy! [03:56] mjg59 : Yes. [03:56] infinity: That's a relief [03:56] BenC : -server-bigiron? [03:56] Obvious GPL violation otherwise [03:56] (They include Linux serial code) [03:56] infinity: -server and -server-bigiron [03:57] yes. Bug. -bigiron? :) [03:57] mjg59 : Yeah, they have it in a seperate object, and we link them on boot. [03:57] BenC: i laid the groundwork for dbg in my development tree [03:58] cool [03:58] just have to figure out the post-install crap [03:58] s/Bug/But/ [03:58] infinity: bug? [03:58] oh, heh [03:59] was agreed upon by several in #ubuntu-devel [03:59] elmo, Kamion, and a few others [03:59] bigiron being a word that only people who actually know what it is would consider installing the kernel [03:59] Oh, sure, I don't get credit for the vga16fb change. :) [03:59] shit, I meant to change that before finalizing [04:00] sorry [04:00] (Like I care) [04:00] I did an identical change [04:00] I claim copyright [04:00] well, if it's broken, I can't put blame on you now :) === infinity removes your credit for the acx firmware changes out of spite [04:00] lol [04:01] infinity: I think Suse might be shipping GPL-violating ltmodem [04:01] I'm SHOCKED. [04:03] ;ater === BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-kernel [05:17] mjg59: FYI I did a hibernate on my P4 with intel IDE, and it worked fine [05:17] usb and and wireless card came back up [05:28] BenC : -8.10 is >= -rc5, right? [05:32] Oh, rc5 is old, we must be caught up to that. [05:37] no, -8.10 is -rc5 [05:37] linus was gone for a week, and just got back [05:37] he did a lot of merging yesterday, but it was too much for me to pull and be able to get -8.10 out for flight 2 on time [05:38] Well, I did say ">=", not ">" [05:38] So, we're on the same page. [05:38] oh, missed that [05:41] Woo, first attempt at compiling the new ATI driver fails MISERABLY. [05:41] Go ATI. [05:47] ... because they repacked the panel sources to exclude a header it needs... [05:47] W... T... F... [05:47] No other changes, except removing the header. [05:47] HATE. [05:55] lol [05:56] morning guys [05:56] hey fabbione === fabbione yawns [05:57] OO, amd64 kernels are ready. [05:57] woo new crack [05:58] amd64 always wins the first-build race [05:58] BenC: not always :) [05:58] there were times where sparc was winning :) [05:59] Well, i386 has too many flavours. [05:59] sparc only builds 2 kernels, it isn't allowed to compete yet :) [05:59] BenC: nah.. i was prebuilding the pkgs and made them enter the archive 5 minutes after the source :) [05:59] lol [06:00] CHEATING. [06:00] infinity: no.. they were properly sbuilded [06:01] you know.. that thing called pre-upload build test [06:01] they had premature access to the source though :) [06:01] CHEATING. [06:02] And powerpc and i386 are in a dead heat for second place, both building packages right now.. [06:02] ia64, as always, will be dead last (well, for the DC arches... I'm not sure where sparc and hppa fit in) [06:02] infinity: i would have done the same if we had m68k :) [06:03] no, hppa is dead last [06:03] m68k has too many subarches to play this game fairly. [06:03] Slowest hardware in Debian, but we build more kernels than anyone else. [06:04] infinity: distcc :) [06:04] If it were up to me, we'd drop all "old m68k" support, except for Amiga and Atari, and concentrate on ColdFire targets. [06:04] let's make an m68k cluster [06:04] m68k is worse than i386 cpu's...I bet we could push min cpu for ia32 to 586 before debian got rid of m68k :) [06:04] Who seriously wants to run a modern OS on a 68k Mac anyway? [06:04] BenC : MY Amiga goes toe-to-toe with My Pentium 233MMX, so that seems fair. [06:05] s/My/my/ ... Wow, I've been Microsofted. [06:05] ahaa [06:06] Anyhow, if I can shoehorn ColdFire support into Debian's toolchain, m68k could become a reality for Ubuntu in microbuntu. [06:06] Lots (and LOTS) of Linux/Coldfire projects out there right now in the embedded world. [06:07] infinity: what's Coldfire? [06:07] ColdFire == Newage m68k. [06:07] yeah, I have some Crestron equipment that I'd love to hack on [06:07] Dropped a few "useless instructions", added a few new ones (hardware RNG and crypto accel, mostly), and bumped the speed a bit. [06:08] does coldfire have mmu? [06:08] Yeahp. [06:08] Well, some do. Same at the oldskool m68k line, you can get them with and without MMU. [06:08] I'd love to have ubuntu boot on my crestron touchscreen RF remove :) [06:08] Obviously, we'd target the MMU varienty. The 68knommu target is much less interesting to Ubuntu. [06:09] *remote [06:09] The fastest ColdFires look to be on par with mid-range ARM, so it's no slouch. [06:09] Not "fast", but "fast enough". [06:10] it's got 32Megs of ram and 8megs of flash, plenty enough for OOo, firefox, and all of gnome [06:10] (And very, very low power, hence why it's being used places where people are saying no to ARM) [06:10] Oh, hey, LRM built. That can't be right. [06:10] It surely MUST be broken. === infinity oozes confidence. [06:17] so lrm is uploaded? [06:18] Nah. Testing on machines here to make sure I didn't bugger it. [06:18] But I already uploaded the .orig.tar.gz to chinstrap, so I can fire off the rest rapidly when the kernel headers get NEWed. [06:19] That 80 meg upload is hell on my connection. [06:20] I have linux-meta ready [06:21] Did you actually bump the ABI this time? :P [06:21] (Last time it was a changelog entry with no debian/rules change..) [06:23] yes :P [06:24] Hey, someone has to keep you kernel genius types from getting a big head. :) [06:24] my wife is good at stiffling my ego [06:25] If by "stifling your ego", you mean "crushing your soul", mine's pretty good at that too. [06:25] Is hppa still gcc-3.4 in this upload (and still the only gcc-3.4 arch)? [06:26] yeah [06:26] Hrm. [06:27] Can I reliably extract the compiler from /usr/src/linux-headers-2.6.15-$(ABI)/Makefile or something, rather than hardcoding it in each LRM upload? [06:27] nah [06:28] but if you used the kernel build system like a good module build, it would automatically do all this for you :) [06:28] Some LRM modules do, not all. [06:30] madwifi, for instance, is well-behaved. [06:30] If it wasn't for that little ath_hal bit, I'm sure it would find its way into the main tree. === infinity gets out and pushes his amd64 to build faster. [06:34] BenC : -7.9 was also -rc5, yes? [06:34] (or close enough)? [06:34] close [06:35] Kay. I'm testing some patches that were for "-rc5 and over", but I can't be bothered snagging the new kernels from the buildds to test... [06:35] sweet, the big three are done [06:36] Specifically, stuff targetted at the new vm_insert_page() stuff. [06:36] anything new from nvidia for that? [06:36] Yeah. Which is what I'm testing. :) [06:36] I keep seeing WARN_ON()'s [06:36] ah, cool [06:38] Sometime, I need to sit down and backport some of these changes to the -legacy driver... [06:38] Not that -legacy doesn't work, but I suspect it's getting flakier and flakier with each kernel release. [06:39] infinity: i assume we will have to drop it sooner or later [06:39] at a certain point it won't build anymore [06:40] Upstream has committed to supporting it. [06:40] zander keeps releasing patches for it. [06:40] Just not as many as he does for the new driver. [06:41] So, it builds, it works, it's just not as loved. [06:41] ok [06:45] hmm Kamion was mentioning a udeb pkg that does the default kernel selection in d-i [06:45] but i can't remember which one it is [06:45] ah there it is [06:45] base-installer [06:46] Okay, NVIDIA stuff (old and new) works, now to test fglrx... Kiss the laptop goodbye. [06:47] (fglrx will porbably still OOPS, from what I understand... ATI sucks) [06:47] BenC : When do you expect 2.6.15 to be final? [06:47] BenC : ATI "doesn't support unreleased kernels" like NVIDIA does. [06:47] infinity: i am pretty sure .15 is not too far from now [06:48] Good. Then maybe the next fglrx release will stop sucking. [06:48] Oh well, off to test this one anyway. [06:48] Linux v2.6.15-rc5 [06:48] [06:48] Hey, for no other reason than the fact that I'll be off-line for a [06:48] week. [06:48] [06:48] Of course, I could force everybody to just use git (and when I'm emperor [06:48] of the world, don't think I won't!), but it seems some people want to [06:48] just test official releases. Even if they are just -rc's. [06:48] [06:48] By the time I'm back, Andrew will have fixed all my bugs, and I'll [06:48] release it as 2.6.15 and take all the credit. [06:48] [06:48] Mwahahahaaa [06:48] [06:48] Signed-off-by: Linus Torvalds [06:49] What day was that? [06:49] Date: Sat Dec 3 21:10:42 2005 -0800 [06:49] Okay, so triple his "one week" estimate, and that gives us a final release in 11 days or so. [06:49] Not too bad. [06:50] I think it will be a new years release [06:50] christmas coming up, he wont be getting a lot done [06:51] Just cause Linux like flash and show? [06:51] s/Linux/Linus/ === infinity remembers the "greased turkey" release. [06:51] Which conjured up all sorts of images I didn't want. [06:51] THANKS LINUS. [06:51] yeah, holiday releases are always good :) === JaneW [n=JaneW@196.36.161.235] has joined #ubuntu-kernel === BenC goes to bed [07:03] night Ben [07:36] BenC : Wait! [07:36] Meh, too late, I guess. [08:00] infinity: anything i can do? === lbm [n=lbm@cpe.atm4-0-1301006.0x50a0824e.vgnxx6.customer.tele.dk] has joined #ubuntu-kernel === JaneW [n=JaneW@196.36.161.235] has joined #ubuntu-kernel === chmj [n=chmj@196.36.161.235] has joined #ubuntu-kernel [08:30] fabbione : No, I just wanted him to put his linux-meta changes somewhere, so I could upload them when the kernels are NEWed. [08:30] I assume he changed stuff for the -server rename, etc, and didn't want to duplicate the effort. [08:31] infinity: ah ok === jane_ [n=JaneW@196.36.161.235] has joined #ubuntu-kernel === rhubear [n=rhubear@217.24.128.152] has joined #ubuntu-kernel === rhubear [n=rhubear@217.24.128.152] has left #ubuntu-kernel ["Leaving"] === mjg59 [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel === CataEnry [n=Enrico@host149-23.pool8250.interbusiness.it] has joined #ubuntu-kernel [12:46] hi all === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel === mjg59 [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel === mjg59 [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel [01:58] linux-meta uploaded, for the interested [01:59] fabbione: ping [01:59] BenC: Hey - saw the unix.ko builtin in the changelog. I'm curious why you built it in, we loaded it in the initramfs. [02:00] BenC: pong [02:00] BenC: Adam did upload linux-meta too [02:00] jbailey: Keybuk/Kamion said to build it in [02:00] he did? [02:00] yes [02:00] oh well, collision :) [02:00] jbailey: everything was loading it unconditionally - good sign that it should be built-in [02:01] BenC: what can i do for you? [02:01] Kamion: Except for larger kernels making building floppies and such suck a little bit more. [02:01] fabbione: vtun is up if you want to push a kernel build [02:01] yeah, well I didn't really mind either way; I certainly see your argument [02:01] BenC: i am build gcj-4.1 now to track down the gcc (as) segfault [02:02] BenC: i noticed that the tunnel problems happen my side when the tunnel dies... [02:02] I didn't have any objections to building it in, but I don't think I said to build it in :) [02:02] so i could restart it easily [02:02] fabbione: ah, didn't notice you were on there :) [02:02] BenC: eheheh i can sneak everywhere :P === BenC feels violated [02:02] but vtun has a bug on the server side and doesn't ifconfig down / up when the tunnel restarts.. so it hangs [02:03] if i stop the server, ifconfig down, start the server, everything is cool [02:03] Kamion: slight exageration on my part to avoid getting blamed :) [02:03] BenC: eheh :) [02:03] Kamion, BenC: It's not a big deal, since it's small. I'm occasionally hacking on floppy support for d-i in Debian. Having small kernels is really important. =) [02:03] jbailey: does d-i itself need unix.ko? [02:04] BenC: did you see the mail from Claire? [02:04] not yet [02:04] BenC: ok [02:04] let me know if you want to answer or should I [02:04] both ways work for me [02:04] BenC: Probably. I don't think I need it for the bit where I get the user to put in a new floppy into the drive and hit enter to load stuff off of it, though. [02:04] fabbione: oh yeah, got that [02:04] was going to email him this morning [02:04] BenC: That's Debian stuff mind you - I don't know that we care about floppy installs in Ubuntu. [02:05] BenC: ok perfect [02:05] BenC: yes, we need unix.ko, but having it built-in is fine too [02:05] i.e. we need the feature, not necessarily the module [02:05] floppy installs in Ubuntu have never worked yet, but they'd be really nice to do at *some* point [02:05] well, I _had_ to switch to ide-core being built-in, so adding unix was a small increase compared [02:06] Kamion: Okay. Honestly, I had assumed we would never care here. [02:06] otherwise it would have been a big todo exporting the kernel command_line out to the ide-core modules [02:06] Kamion: I still maintain that as long as we include openoffice in the default install, we clearly don't care about any machine older than about 5 years. =) [02:07] (And at that, they were high end machines 5 years ago) [02:07] I will never ever ever try to do a floppy install for as long as I live [02:09] how long has x86 been able to boot from floppy? [02:09] err, cdrom === BenC remembers doing nt4 installs that had to start with a 3 floppy set [02:09] at least 5 years [02:09] actually... 7 [02:10] win95 install could be booted from cdrom, couldn't it? [02:10] yup [02:10] BenC: nt4 could boot off CD [02:10] I still get requests *shrug* [02:10] Mithrandir: SCSI Cdrom, yes. [02:10] the original nt4 installs I did couldn't [02:10] Mithrandir: Not the first IDEs [02:10] from owners of laptops that don't boot properly off CD-ROM [02:11] jbailey: sure, but SCSI is teh way. [02:11] the laptop can run Ubuntu fine once it's installed, it's just that the BIOS is fucked [02:11] Kamion: point them to netboot :) [02:11] I don't think I own a machine with a floppy anymore [02:11] fabbione: I do, but seriously dude, that's stupid and tedious for laptops [02:11] Mithrandir: SCSI is an indiciation that you came from the sabdfl "I have too much money" school of business. =) [02:11] except my i2k, but I'd rip it out if I could [02:11] I think I first booted off CDROMs back before there were IDE cdroms. [02:11] or USB sticks, but they can't all boot from USB either [02:11] jbailey: my father worked at the university and brought all the new and shiny stuff home to play with^W^Wtest. [02:12] at the moment they have to use Smart Boot Manager which apparently is some kind of floppy->CD bootstrap weirdness [02:12] Kamion: i install my laptops with netboot... what's wrong with it? [02:12] Mithrandir: Yeah, I remember that from when my father owned the first computer store in the lowermainland [02:12] fabbione: not everybody is you [02:12] Kamion: i know.. and nobody will ever wish that :) [02:12] Mithrandir: All sorts of crazy new shit. Like 300 baud modems when all the cool kids had 110. =) [02:12] fabbione: not everybody has a convenient TFTP/DHCP setup on another machine, for one ... [02:13] fabbione: Or like me, who does, but can't be bothered to tweak his linksys to ignore a particular host. [02:13] jbailey: that's because you are lazy :) [02:13] my ssh sessions on this satellite connection feel like 300 baud [02:13] fabbione: I'm not lazy enough, clearly. [02:13] fabbione: This is something I'm working on. [02:13] jbailey: you should just plug 2 ethernets in one machine and use it as a gw [02:14] BenC: i feel your pain! [02:14] fabbione: Nah. What I should do is hack the wireless box and run my tftp server on it directly. [02:14] fabbione: But since even the new linksys "L" boxes are crippled somewhat, I think I'll wait for another hobbiest hacking box. [02:14] I wonder if I could turn dialup back on and have the linksys route my ssh traffic over that [02:14] dialup was 5 times less lag than this dialup [02:16] jbailey: get wrt54g(s) [02:16] They don't sell the s's anymore, apparently. [02:16] I have my two running dd-wrt, and 7db atennas, and they work great [02:16] yeah, they do actually, at walmart [02:16] Eh, really? [02:16] just don't get v5 [02:16] v4 is ok [02:17] BenC: I'll go look at my local store, otherwise can I get you to pick one up next time you're at walmart and I'll buy it off of you? [02:17] sure [02:17] walmart probably has such a long stock line that they just haven't run out yet. [02:17] yeah, I think that is the case [02:17] I need to head downtown today anyway at some point. I'll check future shop. [02:17] g or gs, it's $50 and $80 [02:18] ? [02:18] 'cause yeah The gs had 32mb of ram, IIRC? [02:18] BenC: http://www.futureshop.ca/ is a large computer chain here. [02:18] don't think so [02:18] just had the speed boost [02:18] v5 only has 4megs [02:18] Mind you, I should probably go support the only union walmart in qubec. =) [02:18] that's why they can't be hacked on [02:18] lol [02:21] Hmm [02:22] I forgot that it's decemeber. Can't buy anything for myself - house rules. [02:27] speaking of which [02:27] i need a Gigabit switch [02:31] who is going to be required to go to the distro sprint, all of the distro team? [02:31] I was already planning a World Series of Power Circuit tournament that was on Feb 7 :/ [02:32] s/Power/Poker/ === zul [n=chuck@ubuntu/member/zul] has joined #ubuntu-kernel [02:34] heylo [02:34] BenC: yes.. all of the distro team [02:34] that would include you as well [02:37] Note to BenC: If you have something important to you to do. Book it early. =) [02:37] That way at least you can fall back on the "But I paid for this already" argument. I don't know how well it works, mind you, but... [02:37] it's already booked, I have a reservation with 3 other guys who are going to [02:38] tournament fee paid [02:41] When/where is the distro sprint? [02:41] Yeah, make sure you note these in your staffcalendar. [02:41] mjg59: End of January, in the UK. [02:41] mjg59: I think we didn't want you to have to leave your army or something. =) [02:42] Haha [02:42] Except I'm in New Zealand [02:42] which tournament BenC? [02:42] You guys all suck [02:42] zul: World Series of Poker Circuit even in Harrah's in Atlantic City [02:43] cool...im totally in the wrong area [02:43] mjg59: you looked at #18164 at all? [02:43] mjg59: New Zealand? WTH are you down there? [02:43] jbailey: better class of sheep [02:43] mjg59: Haven't you seen from the Canonical folks what happens when you stand upside down for too long? [02:43] LCA [02:44] BenC: Yeah, Linux hates HPs [02:44] It's worse on the nx6125 for some reason [02:44] Haven't figured out if it's Linux or the BIOS [02:44] I'll give it a kick at some point [02:45] (Can somebody please check whether suspend is broken for them on x86 and work out why? I still don't have time, and won't until some time next year) [02:45] the kernel.org bug report shows they are getting acpi thermal events from hw, just that they are getting out to userspace [02:45] mjg59: I did /etc/acpi/hibernate.sh last night on my P4, and it worked [02:46] BenC: Hibernate is fine, it's suspend that's broken [02:47] BenC: Rather, the kernel bug shows that the DSDT is being executed, but not that stuff is getting to the higher kernel levels. Sigh. [02:48] That'll be fun to track down [02:49] Actually, it might not be too bad [02:49] Sprinkle some printks through it, see where it's dropping them [02:50] mjg59: how do I force suspend? [02:51] BenC: /etc/acpi/sleep.sh [02:51] Might need to enable it in /etc/default/acpi-support first [02:59] how are others seeing it fail? [03:00] mine didn't really do much, except corrupt my vga, and if up/down my network interface [03:02] mjg59: the odd thing about that bug is that if they run acpi -t, it gets events, so to me it sounds like a kernel bug [03:03] BenC: Machine hangs when it should go to sleep [03:03] BenC: acpi -t executes further DSDT code [03:04] setting acpi sleep to ram didn't work for me it seems [03:04] acpi standby worked [03:04] neither froze the machine though [03:07] suspend to mem goes down (disables everything, I can see in dmesg), but comes right back up [03:08] how do I enable suspend to disk? [03:08] That's what hibernate does [03:09] ok [03:09] well, I didn't freeze up, -8.10-686 kernel [03:10] standby fixed my vga console too [03:10] Well, if it didn't actually sleep, then it's not going through the codepath that's failing [03:10] how can I debug why it didn't sleep? [03:10] maybe I need to build a acpi debug enabled kernel [03:11] dmesg ought to give a hint [03:12] says nothing [03:12] after the last IRQ disable, it starts enabling irqs/devices again [03:13] Right [03:13] What does /sys/power/state look like? [03:13] "standby disk", so I guess I don't have suspend to mem support [03:13] my G4 does have suspend to mem [03:13] guess I can give it a try [03:14] ah, but that's isn't acpi, so probably pointles [03:19] BenC : Sorry for the collision, I wasn't sure when you were coming back. [03:19] no problem === mgalvin [n=mgalvin@ubuntu/member/mgalvin] has joined #ubuntu-kernel [03:22] mjg59 : Say, when do I get my screen-lock (screen close) button working again? [03:24] infinity: Oh, soon [04:06] hi [04:06] finally my acx module is working, but i dont know why :o [04:06] did you upgrade to -8? [04:07] nope :o [04:07]