[12:07] jbailey: is there any docs for hooks and scripts, or am i reading the shell? :) [12:07] s/for/for creating/ [12:09] There's a HACKING file in there. [12:09] It doesn't tell you much, though. Feel free to improve it. =0 [12:10] hehe, ok === doko_ [n=doko@dslb-084-059-078-087.pools.arcor-ip.net] has joined #ubuntu-kernel === BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-kernel === mgalvin [n=mgalvin@ubuntu/member/mgalvin] has joined #ubuntu-kernel === zul [n=chuck@ubuntu/member/zul] has joined #ubuntu-kernel [01:00] heylo [01:07] Heya Chuck [01:16] hey jeff how is it going? [01:17] In a conforming program, this register contains a function pointer [01:17] %edx [01:17] that the application should register with atexit(BA_OS). This [01:17] function is used for shared object termination code [see Dynamic [01:17] Linking in Chapter 5 of the System V ABI] . [01:17] zul: This is what I'm reading a tm. [01:17] I'm not sure if I'm doing well or doing poorly. =) [01:17] sounds exciting...im going crazy the project i was working on went live yesterday [01:17] 4 hospitals...fun fun...but its calming down [01:18] re-installing ubuntu on my server [01:18] another 14 hospitals will be joining in 3 months [01:47] Nice! [01:52] stressful though [01:53] but more time for ubuntu now [01:59] brb...need to check on the servers though === zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [02:13] meh.. [02:17] i love it when a standard is not followed [02:31] standards are for pussies [02:31] thanks.. [02:33] BenC: get_wireless_stats have changed so alot of external network cards are going to get the a message like in bug #20498 [02:33] im going to see if upstream has the changes if not make a patch.. [02:33] ok [02:35] if i could only spell [02:35] that message is harmless [02:35] I get it with my rt2500 driver aswell, and it is working fine [02:36] hmm...then it must be something else [02:36] im going to fix it anyways [02:36] wish hostap had a usb interface [02:37] hostap_{cs,pci} seem to be doing well replacing prism2 versions [02:37] i swear to god everytime i open something the guinea pig thinks its going to be feed [02:37] lol [02:41] it even has a big fat ass carrot that its chewing on.. [02:44] fabbione: you are up early [02:45] zul: yeah we have a meeting in 15 minutes [02:45] oh goody [02:49] BenC: did anyone try the linux-image-server-lowend yet? [02:52] not that I know of [02:52] heh lets try it [02:53] BenC: are you planning to rename them, don't you? [02:53] yeah, already done in git [02:54] ok [02:54] i will start testing them on the next upload [02:54] not that i have -bigiron [02:54] but well [02:54] it should boot on common 686 hardware, but it's not optimal [02:54] I need to do comparable amd64 images too [02:55] we will need server images for all arches mostlikely [02:55] i mean.. if there is a reason to have them [02:55] BenC: it boots at least [02:56] ppc doesn't need them, and ppc64 supports everything it can [02:56] sparc64 doesn't need it, and neither does ia64 or hppa, that I know of [02:56] perfect [02:56] maybe a -server for each one, but that's about it [02:56] that's great [02:57] just with HZ=100, and all preempt disabled [02:57] yeah [02:57] that's what i was thinking about [02:57] we need to talk with Benh soon [02:57] he is preparing some patches for quad-g5 [02:57] that's not actually supported in the standard kernels [02:57] that's something people will want to use with dapper === calc [n=ccheney@ip70-185-4-246.ma.dl.cox.net] has joined #ubuntu-kernel [03:14] fabbione: wrt udev and sysfs root block detection is there a place to note which drivers are broken? [03:15] fabbione: i noticed you mentioned something about udev in your report [03:15] calc: ???? [03:16] no i think you are confusing things around a bit [03:16] yaird builds its initramfs and includes drivers based on the data in sysfs [03:16] ah ok [03:16] no it's a completely differnt thing [03:16] ok [03:16] in short [03:16] was just noting that some drivers don't work with sysfs detection since they don't register with sysfs properly [03:16] assume you install on a USB stick [03:16] that you move around [03:17] the device changes name [03:17] ah cool [03:17] so be able to mount root and the rest [03:17] without binding to a specifc device name [03:17] so it uses a magic id on the root block that is contained in the initramfs? [03:18] no, we will use new udev properties [03:18] like /dev/disc/by-uuid/ [03:18] or /dev/disc/by-lable [03:18] label [03:18] that are symlinks to the real device name [03:19] and they are generated at boot [03:19] so you have a unique way to identify them [03:19] even if the device changes names [03:19] cool [03:20] label sometimes could break if not unique, eg putting another / in to copy old data from [03:20] iirc redhat used to just use "/" or something for root which caused issues [03:21] no we are not going to use LABELS everywhere [03:21] we are targetting uuid [03:22] if uuid is not available, then generate a pseudo uuid strlen(LABEL) [03:22] and use that one [03:23] ok [03:24] sounds good [03:29] BenC: if you have time can you gimme access to your sparc? [03:29] if you can just enable the account and sudo i will take care of the rest myself [03:29] I'm having problems punching a hole through this damn satellite modem :/ [03:29] ah [03:29] if you have a linux gw in the middle, just use redir and tell me the non standard port [03:30] it doesn't need to be 22 ;) [03:30] I have a static IP, but It's all private network inside (NAT), and there doesn't seem to be a way to tell the modem to port forward [03:31] BenC: ok.. we can work on that easily [03:31] i can run a vtun server instance here and we can play private tunnels [03:31] don't understand the point of allowing me to get a static IP if it wont let me get connections [03:31] not sure how well that will work over this hellish latency [03:31] hold on [03:32] it works pretty well.. latency isn't an issue and vtun does his job of re-establishing connections if it dies [03:32] sure [03:33] ok, how do we setup vtun? [03:34] i will prepare the configs and send it to you... [03:34] installing vtun package now [03:34] it's very simple [03:34] i will be server and you client [03:34] ok [03:34] since you are the one that needs to make holes in the modem [03:34] right [03:34] i don't remember all the details right now. [03:34] i will have to check them again [03:34] but that's basically how it works: [03:34] server listen on tcp port [03:35] client connect to server, auth, etc. [03:35] if everything matches they run an ifconfig on a tun device [03:35] that's it [03:35] pvt networks [03:35] what address space do you use at home? [03:35] because we need to avoid clashing [03:36] 192.168.1.0/24 [03:36] ergh.. ok [03:36] i have the same net here [03:36] so if i use a 192.168.2.0 would be ok for you? [03:37] make my network 192.168.1.13/32 :) [03:37] yeah [03:37] that's just for the p2p [03:38] once we can ping the tunnel, i can jump via it to the sparc directly or something [03:59] staying up to get this configured? [04:00] yeah [04:00] why not [04:00] let me file a bug and let's do it [04:00] ok [04:06] Oh rock === mjg59 has working hotswap on a SATA laptop [04:09] BenC: After getting comments from -ide, I may have some patches for you... [04:09] sweet [04:09] BenC: Also, this should work on any of the libata PATA drivers [04:09] Hence my interest in them :) [04:09] Anyone here know where the userspace setup code is for a new process on powerpc? [04:10] crt0.o, or do you mean task creation in the kernel? [04:10] Task creation in the kernel. On ppc, crt0.o's first function seems to get started with %r7 set incorrectly. [04:11] Troubleshooting a klibc segfault. [04:11] let me look... [04:11] Thanks. [04:12] BenC: 192.168.0.something would be ok for you? [04:12] To make life more fun, I'm on a ppc64 kernel with a 32 bit userspace klibc. [04:12] instead of 192.168.2.0 === jbailey wonders if he should try a 32bit kernel, too, just to prove it. [04:12] i forgot i have allocated .0.0/24 for p2p and tunnels :) [04:12] nah, my network looks like internet->192.168.0.0/24->192.168.1.0/24 [04:13] ok [04:13] no problem [04:13] between my sat modem and AP is 192.168.0.0/24 [04:13] 192.168.2. it is [04:13] I'd change, but I have 12 systems on 192.168.1.0/24 [04:13] that's all right [04:13] don't worry [04:22] night guys [04:25] good night zul [04:25] BenC: Setting up VPN to your farm? [04:25] yeah, fabbione wants to reach out and touch a cow [04:26] zul: night [04:26] BenC: yeah i am almost done [04:26] You should remind him that your cows are better poker players than he is. You had to practice with *someone* out there. [04:26] g'n zul =) [04:26] fabbione: even logged into a machine that was colo'd in a barn? :) [04:26] lol [04:26] ahahha [04:27] just 30 feet from your shell, is a steaming cow pie [04:27] lovely [04:27] i love cows :) [04:28] BenC: mail with config is on the way [04:28] ok [04:29] time for a smoke :) [04:30] BenC: I need to go pass out. Can I catch up with you tomorrow on the %r7 weirdness? [04:30] yeah [04:30] Cool, thanks. [04:30] good night [04:30] g'n all. =) [04:34] night jeff [04:36] BenC: got the mail? [04:38] yup [04:38] the tunnel is up [04:38] tag, you're it :) [04:39] sparcbuildd is the user [04:39] ping 192.168.2.2 [04:39] PING 192.168.2.2 (192.168.2.2) 56(84) bytes of data. [04:39] 64 bytes from 192.168.2.2: icmp_seq=1 ttl=64 time=1127 ms [04:39] 64 bytes from 192.168.2.2: icmp_seq=2 ttl=64 time=993 ms [04:39] is the tunnel terminated on the e3k? [04:39] yeah [04:39] ssh directly to iy [04:39] perfect [04:39] testing now [04:40] are you running iptables on that machine? [04:40] no [04:41] I can't ping you anymore [04:41] neither can i [04:41] hmm [04:42] there it is [04:42] connection is shaky [04:42] ok, restarted it [04:42] seems to work now [04:42] latency really sucks on this sat [04:42] no problem.. i can live with that [04:43] yes i am in [04:43] 1500ms ave [04:43] cool, your chroot is thee [04:43] create chroots in /org/chroots/ [04:43] ok [04:43] 222G free, so plenty of room [04:43] do i have sudo access? [04:43] perfect [04:43] way too much [04:43] yeah [04:44] if the vtun is down, it will mean I am debugging :) [04:44] thanks dude [04:44] sure [04:44] that works for me [04:44] np [04:45] i feel your IRC pain ;) [04:46] good thing is I started using xchat instead of ssh'ing to a shell and using bitchx [04:46] that did suck [04:46] oh yeah [04:46] i remember when ops.l.o was constantly under attack [04:46] i could barely ssh to it [04:47] yeah, that got ugly for awhile [04:47] i know.. [04:47] our ISP was cool tho [04:47] they managed to block the attack at core routers level [04:47] nice [04:49] let's start to test the toy with gcc-3.4 :) [04:49] burn it up :) [04:49] oh i will [04:49] eheh [04:50] ccache is installed, so try to bind mount it [04:50] oh nice [04:50] /org/ccache is my cache [04:50] where do you store the ccache? [04:50] hmm [04:50] there might permission problems [04:51] anyway to share it? [04:51] ah, just make your own and bind mount it :) [04:51] right [04:52] but i am sure how that can help you [04:53] there is quite high pkt loss [04:53] vtun is starving again [04:55] let see if it can come up again by itself [04:56] may need to put a ping check script on it to restart when it dies [04:56] vtun should be able to notice that itself [04:56] that's why i want to wait a bit [04:56] it has a 60 sec timeout [04:57] actually [04:57] i forgot to add it to the configs :/ [04:58] can you add: [04:58] timeout 60; [04:58] in the session part? [04:58] and restart vtun please? [05:03] interesting [05:03] it comes up.. [05:03] but it looks like it dies very very fast [05:04] ok [05:06] here is up again now [05:06] did you touch anything? [05:06] restarted it [05:06] ok [05:06] added the timeout too [05:06] let see if that helps [05:25] it seems more stable now === trevilor [n=gnu@p549CD1C6.dip.t-dialin.net] has joined #ubuntu-kernel === trevilor [n=gnu@p549CD1C6.dip.t-dialin.net] has joined #ubuntu-kernel === chmj [n=chmj@dsl-165-121-252.telkomadsl.co.za] has joined #ubuntu-kernel === Elektrochelovek [i=vasiliev@62.212.201.69] has joined #ubuntu-kernel === CataEnry [n=cataenry@host149-23.pool8250.interbusiness.it] has joined #ubuntu-kernel [10:26] hi :) [11:43] bye === jane_ [n=JaneW@dsl-165-120-130.telkomadsl.co.za] has joined #ubuntu-kernel === nomed [n=daniele@host102-127.pool8254.interbusiness.it] has joined #ubuntu-kernel [01:28] hi all [01:28] i would post an issue about tg3 module [01:28] there are some cards that don't work .. [01:29] what's strange is that using an hoary kernel this issue is not present [01:29] it really seems a breezy-dapper issue .. [01:29] even using a vanilla kernel there aren't problems [01:30] you'll need to file a proper bug report with a lot more detailed info [01:31] what info will you need .. [01:31] ? [01:31] do you apply any patch to that module ? [01:32] and what can be the reason for which using hoary kernel this issue is not present and even using a vanilla kernel? [01:32] it happens just with breezy and dapper kernels .. === nomed [n=daniele@host102-127.pool8254.interbusiness.it] has left #ubuntu-kernel ["/etc/init.d/nomed] === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel === CataEnry [n=Enrico@host149-23.pool8250.interbusiness.it] has joined #ubuntu-kernel [02:46] hi :) === slushpupie [i=jay@slushpupie.com] has joined #ubuntu-kernel === zul [n=chuck@ubuntu/member/zul] has joined #ubuntu-kernel [03:03] morning [03:14] Heya Chuck! === zul__ [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [03:47] jbailey: do i get whipped with some form of wet pasta for using _log_msg in an initramfs script? [03:48] i just want to spit out some information to the user, but w/out any warning/being/done/etc prefix [03:48] Eh, depends if you'd like the whipping or not, I guess. [03:48] I don't see any real problem with it. If it's in the initramfs it should be generally safe to use. === infinity whips dilinger. [03:52] jbailey: cool, ok [03:52] dilinger : Other than the fact that users think is "icky" or even "scary" to see a _msg without an _end_msg, I don't care. [03:52] infinity: please use clean pasta, not the stuff i've already doused w/ cheese or spaghetti sauce [03:53] What about pesto? [03:53] Spaghetti sauce stings when it gets in the cuts. [03:53] that green stuff never comes out of your skin [03:54] dilinger : I'd be inclined to say that if it's informational, but the user doesn't really NNED to know it's happening (with a success or failure), it should be guarded in a VERBOSE check. [03:54] s/NNED/NEED/ [03:55] Same goes for all init scripts, really, nothing special about initramfs. [03:55] doesn't _log_msg do that implicitly? [03:55] if [ "$quiet" = "y" ] ; then return; fi [03:58] Oh, so it does. [03:59] I really should rewrite that stuff to look like the LSB things in Debian/Ubuntu. [03:59] I only boot with usplash these days, so I forgot the tty output is different. [04:00] jbailey: btw, i dunno if you've already seen it, but makx has some good documentation in his initramfs tree [04:01] dilinger: Ah, handy. Perhaps the new initramfs-tools maintainer will see fit to merge those in. [04:01] *ahem* === jbailey hides. [04:01] jbailey: i dunno, i hear he's real negligent about updating his packages. you should take them over. [04:01] ;p [04:01] ACtually, hmm. [04:01] Don't you have main upload rights? [04:01] I'm planning on making a date with Max next week to run down our diffs. [04:02] But I'd be happy to have dilinger in on it. [04:02] We've worked passably well together in the past. :) [04:02] :) [04:03] infinity: VGA16FB? [04:03] jbailey: me? depends on elmo [04:03] mjg59 : How do you feel about trying to squeeze the world into 350px vertical? [04:03] dilinger: No, it depends on the techboard. [04:03] jbailey: i revoked my old key and had Clint ask to get my new one in the keyring (for debian, not ubuntu) [04:04] for ubuntu, no [04:04] mjg59 : Either he's insane, or I'm insane, but Kamion's informed me that the VGA 25-line screen is 640x350, not 640x400. [04:06] He's pretty much right, but is there any compelling reason to do that? === rhubear [n=rhubear@217.24.128.152] has joined #ubuntu-kernel [04:07] If we're looking for maximum compatibility... [04:07] The problem isn't that we're using a non-standard screenmode, the problem is that we're overflowing 128K [04:07] I'm fairly certain 640x400 should work all over, but if it doesn't, I don't want to do this switch twice. [04:08] All the symptoms (bottom 80 lines missing, bottom 80 lines overlap top 80 lines, so on) seem to match the 128K issue [04:08] I agree on that point. [04:08] And I /think/ we'll solve all that with 400px. === rhubear [n=rhubear@217.24.128.152] has left #ubuntu-kernel ["Leaving"] [04:08] But, I really, really hate being wrong. [04:09] A whole lot. [04:09] Shame I do it so often. [04:09] Anyhow, I'll just bite the bullet and patch shit together tomorrow and see if it at least boots on the one VGA card I have here in the house. [04:09] (Yes, a total of one, WHAT A GREAT TEST) [04:10] Not counting the laptop, since laptop displays will force pretty much anything on your screen without complaint. [04:10] It's genuinely a 3 line patch [04:10] You can just grab the timing data out of modedb.c [04:10] Yeah, I know. [04:10] I've been suffering overload. [04:10] But I also have mdz breathing down my neck about this one, so I need to just do it and get it tested. [04:10] Heh, ok [04:12] Meh. === infinity puts another sticky on his desktop and WRITES ON IT IN ALLCAPS. === lamont [n=lamont@mib.fc.hp.com] has joined #ubuntu-kernel === infinity -> bed. [04:15] BenC: ping? [04:18] BenC: unping.. ttyl [04:18] ok [04:44] Hngh. [04:45] (Also: hngh) [04:45] Consensus seems to be that I should put this in ACPI rather than in SCSI [04:46] So, is there any way that I can write a module that will be (a) loaded after scsi, and (b) loaded before any scsi hosts? [04:46] It seems that it's not acceptable for scsi to have any dependency on the module concerned [04:57] sounds like a job for initramfs [05:10] mjg59: have it depend on scsi and load it in the initramfs, yes. [05:31] Mithrandir: That works for our case, but it doesn't work in general [05:31] Which means it wouldn't go upstream [05:32] true [05:51] Hmm. I guess I can do /this/... [06:07] hey, neat. My external cdrw no longer floods printks [06:07] ARGH. [06:07] No I can't. === g47o [n=g47o@pc-145-4-239-201.cm.vtr.net] has joined #ubuntu-kernel === BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-kernel [07:02] BenC: new gnu-efi uploaded to debian last night, new elilo imminent. [07:02] once those hit ubuntu, you should be good to go for playing with initramfs [07:02] new elilo means initramfs support? [07:02] so given a git tree, how do I roll it back to an older revision? [07:02] sweet [07:02] mdz: git-checkout [07:03] lamont: why does elilo need to know the difference between initrd and initramfs? [07:03] where commitish is a commit sha1 id [07:03] mdz: elilo had a bug in how it handled the image. didn't matter for initrd, fatal for initramfs [07:03] mdz: apt-get install git- git-core [07:03] BenC: mjg59 has asked me to try an older 2.6.15 with HOTPLUG_CPU and see how long ago suspend broke [07:04] specifically, it passed in NUMPAGES*sizeof(PAGE) instead of the true byte-length of the image. [07:04] what would be a good value for commitish to start with? [07:04] mdz: hold a sec [07:05] mdz: git-log [07:05] and look for "Linux v2.6.14" [07:05] use that commit to start [07:05] if it works, then you can use git-bisect to find the commit that broke it [07:06] if you need help with git-bisect, let me know, it's pretty easy once you start it === zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [07:10] heylo [07:11] mizar:[...ce/tmp/mdz/linux/ubuntu-2.6] git-checkout 741b2252a5e14d6c60a913c77a6099abe73a854a [07:11] git checkout: you need to specify a new branch name [07:11] hey chuck [07:11] did you get the git-core? [07:11] hey ben [07:11] yep [07:11]