[00:47:54] <SeligArkin> I've been poking around chrome os's source trying to find how it reconnects so fast, and how hard it would be to pull over, one place I was looking was dhcpcd -- am I to understand you guys made your own dbus fork from the core, or am I mis-understanding this diff?
[02:35:04] *** lipsinV2 has joined #chromium-os
[02:43:25] *** powderluv has joined #chromium-os
[02:43:25] *** ChanServ sets mode: +v powderluv
[03:05:58] *** yusukes has quit IRC
[03:18:58] *** jglasgow_ has joined #chromium-os
[03:19:24] *** jglasgow_ has quit IRC
[03:19:58] *** jglasgow_ has joined #chromium-os
[03:20:05] *** ChanServ sets mode: +v jglasgow_
[03:30:09] *** jglasgow_ has quit IRC
[03:32:41] *** rbyers has joined #chromium-os
[03:47:35] *** powderluv_ has joined #chromium-os
[03:47:35] *** ChanServ sets mode: +v powderluv_
[03:49:47] *** powderluv has quit IRC
[03:52:08] *** yusukes has joined #chromium-os
[03:54:17] *** powderluv_ has quit IRC
[03:55:53] *** perezd has joined #chromium-os
[04:01:01] *** powderluv has joined #chromium-os
[04:01:01] *** ChanServ sets mode: +v powderluv
[04:01:16] *** powderluv has quit IRC
[04:26:31] *** behdad has joined #chromium-os
[04:43:58] *** BladeFreak has quit IRC
[04:53:53] *** robe_ has joined #chromium-os
[05:06:09] *** perezd has left #chromium-os
[05:09:44] *** oc80z has quit IRC
[05:10:30] *** oc80z has joined #chromium-os
[05:13:53] *** oc80z has quit IRC
[05:13:53] *** oc80z has joined #chromium-os
[05:20:33] *** oc80z has quit IRC
[05:21:32] *** oc80z has joined #chromium-os
[05:31:40] *** rbyers has quit IRC
[05:34:39] *** behdad has quit IRC
[05:54:51] *** behdad has joined #chromium-os
[06:16:54] *** SoulRaven has joined #chromium-os
[06:17:05] *** SoulRaven has joined #chromium-os
[06:17:13] *** SoulRaven has joined #chromium-os
[06:17:22] *** SoulRaven has joined #chromium-os
[06:17:33] *** SoulRaven has joined #chromium-os
[06:17:43] *** SoulRaven has joined #chromium-os
[06:19:36] *** SoulRaven has joined #chromium-os
[06:22:22] *** powderluv has joined #chromium-os
[06:22:22] *** ChanServ sets mode: +v powderluv
[06:26:50] *** powderluv has quit IRC
[06:48:31] *** behdad has quit IRC
[07:08:41] *** powderluv has joined #chromium-os
[07:08:41] *** ChanServ sets mode: +v powderluv
[07:12:55] *** powderluv has quit IRC
[07:33:47] *** powderluv has joined #chromium-os
[07:33:47] *** ChanServ sets mode: +v powderluv
[07:52:36] *** Fractal has quit IRC
[07:58:33] *** powderluv has quit IRC
[09:22:47] *** vmil86 has joined #chromium-os
[09:32:17] *** mascondante has quit IRC
[09:41:44] *** mascondante has joined #chromium-os
[10:17:30] *** SoulRaven has quit IRC
[10:18:14] *** SoulRaven has joined #chromium-os
[10:29:02] *** jochen__ has quit IRC
[10:29:08] *** jochen__ has joined #chromium-os
[10:29:09] *** ChanServ sets mode: +v jochen__
[10:56:22] *** SoulRaven has quit IRC
[10:56:31] *** SoulRaven has joined #chromium-os
[11:00:46] *** SoulRaven has quit IRC
[11:02:38] *** SoulRaven has joined #chromium-os
[11:05:06] *** [Sou|Raven] has joined #chromium-os
[11:07:06] *** SoulRaven has quit IRC
[11:09:30] *** [Sou|Raven] has quit IRC
[11:09:52] *** SoulRaven has joined #chromium-os
[11:11:50] *** e_t_ has joined #chromium-os
[11:13:14] *** patcito has quit IRC
[11:14:04] *** e_t_ has joined #chromium-os
[11:14:20] *** SoulRaven has quit IRC
[11:14:22] *** e_t_ has quit IRC
[11:18:14] *** SoulRaven has joined #chromium-os
[11:19:14] *** e_t_ has joined #chromium-os
[11:22:40] *** SoulRaven has quit IRC
[11:27:17] *** e_t_ has quit IRC
[11:37:34] *** SoulRaven has joined #chromium-os
[11:40:18] *** [Sou|Raven] has joined #chromium-os
[11:42:13] *** SoulRaven has quit IRC
[11:43:32] *** SoulRaven has joined #chromium-os
[11:44:44] *** [Sou|Raven] has quit IRC
[11:48:35] *** [Sou|Raven] has joined #chromium-os
[11:48:39] *** SoulRaven has quit IRC
[11:53:10] *** [Sou|Raven] has quit IRC
[11:57:40] *** borios has joined #chromium-os
[13:49:39] *** KannanPK has joined #chromium-os
[13:50:52] <BigWookie> Just if somebody is interrested in the intel graphics bug, here is my report http://pastebin.com/kpVfTrAM
[13:52:29] <KannanPK> anyone can tell me about any beginners resources for chromium OS
[13:52:31] <KannanPK> ?
[13:55:52] *** KannanPK has quit IRC
[14:43:50] *** behdad has joined #chromium-os
[14:45:09] *** rbyers has joined #chromium-os
[14:54:39] *** sadrul has quit IRC
[14:55:42] *** rbyers has quit IRC
[14:57:04] *** rbyers has joined #chromium-os
[15:34:18] *** rharrison_chrome has joined #chromium-os
[15:34:19] *** ChanServ sets mode: +v rharrison_chrome
[15:46:21] *** SoulRaven has joined #chromium-os
[15:49:22] <ellyjones> crosbot lives :O
[15:55:32] *** rbyers has quit IRC
[15:56:35] *** sadrul has joined #chromium-os
[15:56:35] *** ChanServ sets mode: +v sadrul
[15:57:13] *** behdad has quit IRC
[15:57:44] *** rbyers has joined #chromium-os
[15:57:44] *** ChanServ sets mode: +v rbyers
[16:02:07] <ellyjones> crosbot: stat tree
[16:02:07] <crosbot> ellyjones: tree: Tree is open (cycled green)
[16:02:10] *** behdad has joined #chromium-os
[16:02:11] *** ChanServ sets mode: +v behdad
[16:10:41] *** petermayo has joined #chromium-os
[16:15:32] *** petermayo has left #chromium-os
[16:24:55] <ellyjones> "It's okay, I'm an electrical engineer."
[16:30:41] <ellyjones> crosbot: stat tree
[16:30:41] <crosbot> ellyjones: tree: Tree is open (cycled green)
[16:30:43] <ellyjones> wat
[16:30:46] <ellyjones> why you lie, crosbot
[16:31:02] <ellyjones> oh, no, the build lamp lies
[17:38:06] *** cmasone has joined #chromium-os
[17:38:06] *** ChanServ sets mode: +v cmasone
[18:07:20] *** jrbarnette has joined #chromium-os
[18:15:03] *** jrbarnette_ has joined #chromium-os
[18:15:45] *** jrbarnette has quit IRC
[18:15:46] *** jrbarnette_ is now known as jrbarnette
[18:15:58] *** BladeFreak has joined #chromium-os
[18:16:45] *** dubroy has quit IRC
[18:18:09] <cmasone> so...the repeatedly failing PFQ is no concern?
[18:18:31] <cmasone> ellyjones: ^^^
[18:19:14] *** petermayo has joined #chromium-os
[18:19:15] *** ChanServ sets mode: +v petermayo
[18:23:08] *** oc80z has quit IRC
[18:23:55] *** oc80z has joined #chromium-os
[18:24:05] *** xiyuan has quit IRC
[18:24:34] *** Keybuk has joined #chromium-os
[18:24:35] *** ChanServ sets mode: +v Keybuk
[18:26:52] *** saintlou has joined #chromium-os
[18:26:52] *** ChanServ sets mode: +v saintlou
[18:29:32] *** jrbarnette_ has joined #chromium-os
[18:33:25] *** jrbarnette has quit IRC
[18:33:26] *** jrbarnette_ is now known as jrbarnette
[18:33:30] *** powderluv has joined #chromium-os
[18:33:32] *** ChanServ sets mode: +v powderluv
[18:35:51] *** perezd has joined #chromium-os
[18:41:41] <ellyjones> cmasone: hrm
[18:42:02] <ellyjones> cmasone: you raise a good point
[18:42:34] *** powderluv has quit IRC
[18:42:41] *** powderluv has joined #chromium-os
[18:42:41] *** ChanServ sets mode: +v powderluv
[18:44:10] <ellyjones> okay, yeah, I'm going to revert josephsih's change
[18:46:06] *** Inumedia has quit IRC
[18:46:41] *** Inumedia has joined #chromium-os
[18:47:56] * ellyjones taps her foot while 'repo sync' pulls down the change...
[18:48:09] <sosa> you can do it from the Gerrit interface
[18:48:26] <ellyjones> oh, wow, you can
[18:48:53] <ellyjones> reverted
[18:49:18] <sosa> actually the revert button creates a Gerrit review for yourself which you have to TBR a push to
[18:49:25] <sosa> it is misleading it is
[18:49:42] <sosa> http://gerrit.chromium.org/gerrit/4660
[18:50:09] <ellyjones> landed the revert for real
[18:50:18] <ellyjones> why was the tree green this morning? the pfq must have been failing for days
[18:50:27] <sosa> looked at the status msg
[18:50:33] <sosa> apparently chromium status thought it was red
[18:50:45] <sosa> but for some reason it was showing up green
[18:51:15] <ellyjones> ah, no, it started failing at pfq build 6571, which was this morning, with only josephsih in the blamelist
[18:51:30] <ellyjones> sosa: huh, weird. You want to file it? :P
[18:51:41] <sosa> sure, though i'm not sure if someone will be able to reproduce it
[18:52:07] <ellyjones> I am going to declare that the tree should now go green
[18:52:17] <sosa> i sadly disagree
[18:52:25] <sosa> check out tegra2 binary
[18:52:49] <sosa> looks like Randall's change redified it again
[18:52:51] <ellyjones> ooh
[18:53:41] <sosa> it is extra confusing because the buildbot looks like it took Randall's change before he committed it but the first red tegra2 binary does show his commit in the manifest it built
[18:54:19] <ellyjones> yeah, I was just noticing that
[18:54:26] <ellyjones> 6300 has his commit, but does not blame him
[18:54:28] <ellyjones> that is a bad race
[18:54:32] <sosa> yah
[18:54:35] <sosa> i'll file a bug on it as well
[18:54:37] <sosa> prob to myself
[18:54:51] <ellyjones> cool
[18:54:54] <sosa> the issue is the http repo mirrors getting much behind the git mirrors
[18:55:25] <sosa> but i can file a bug to try to avoid that issue ... i'm not sure I can actually fix the fact that the http mirrors get behind or there is a "real" fix as any mirroring can be behind
[18:56:50] <ellyjones> I dunno where PRIu32 is from, so I'm going to revert rspangler's change instead of trying to fix it
[18:57:19] <ellyjones> some internal canary bots are sick - do we care?
[18:58:30] *** arkaitzr has joined #chromium-os
[18:58:31] *** ChanServ sets mode: +v arkaitzr
[18:59:17] <ellyjones> sosa: ack reverting rspangler's change?
[18:59:23] <sosa> ellyjones:  you would have to ask a TPM I believe responsible for TOT.  I believe we're moving the default such that TPM's care about tree status's and will elevate to sheriffs when they want you to take action.
[18:59:28] <sosa> ellyjones: acked
[18:59:46] <ellyjones> reverted, so that should fix tegra2-binary
[18:59:52] <sosa> agreed
[18:59:57] <ellyjones> so which tpm? should I ask trond? (he's in town right now)
[19:00:05] <sosa> ellyjones: all is good again in the world
[19:00:11] <ellyjones> metaphorically
[19:00:21] <sosa> maybe check with brad
[19:00:27] <sosa> i don't actually know if those builds have ever been green
[19:01:00] <sosa> the no-gobi is brand new and scottz set it up on Friday, so I don't expect much from it
[19:01:12] <ellyjones> no-gobi is really sad, yes
[19:01:23] <ellyjones> the alex bsp is not designed to build with no gobi right now :P
[19:02:10] <ellyjones> crosbot: stat tree
[19:02:10] <crosbot> ellyjones: tree: Tree is open (pfq going green after revert of 4565, tegra2-bin going green after revert of 4611)
[19:02:13] <ellyjones> crosbot: announce +tree
[19:02:13] <crosbot> ellyjones: announcing tree
[19:02:16] <ellyjones> okay
[19:05:17] <ellyjones> sosa: you see the tangent canary? it looks like it has git problems
[19:05:26] <sosa> it does.
[19:05:30] <ellyjones> and has for quite some time
[19:05:45] <ellyjones> can I foist that one off on you?
[19:05:50] <sosa> yeah that's fine
[19:06:59] <ellyjones> and stumpy had a chrome segfault during ChromeFirstRender
[19:07:03] <ellyjones> and that's all the bots accounted for :) woo
[19:07:27] *** lipsinV2 has quit IRC
[19:07:48] *** wfrichar has joined #chromium-os
[19:07:48] *** ChanServ sets mode: +v wfrichar
[19:14:40] *** perezd has left #chromium-os
[19:23:39] *** vlaviano has joined #chromium-os
[19:23:39] *** ChanServ sets mode: +v vlaviano
[19:29:56] <kliegs> sosa: It looks like tangent canary just needs a clobber
[19:30:18] <kliegs> sosa: Also - I think clobber may be broken in cbuildbot - I had some trouble clobbering a different bot last week (one off the tree) and ended up just doing a rm -rf on /b/cbuild
[19:35:09] *** fischman has joined #chromium-os
[19:35:16] *** ChanServ sets mode: +v fischman
[19:36:37] *** dubroy has joined #chromium-os
[19:50:04] *** dhendrix has quit IRC
[19:50:13] *** powderluv has quit IRC
[19:50:18] *** dhendrix has joined #chromium-os
[19:50:19] *** ChanServ sets mode: +v dhendrix
[19:57:35] *** thieule has joined #chromium-os
[19:57:35] *** ChanServ sets mode: +v thieule
[20:07:55] *** achuith has quit IRC
[20:13:30] *** jrbarnette_ has joined #chromium-os
[20:16:28] *** achuith has joined #chromium-os
[20:16:29] *** ChanServ sets mode: +v achuith
[20:16:43] *** achuith has quit IRC
[20:17:17] *** achuith has joined #chromium-os
[20:17:18] *** ChanServ sets mode: +v achuith
[20:17:32] *** jrbarnette has quit IRC
[20:19:19] *** gspencer has quit IRC
[20:20:30] *** jrbarnette_ has quit IRC
[20:28:50] <mononoke> ellyjones: I just noticed you mention TPM problems
[20:29:01] <mononoke> Are those build or run-time problems?
[20:30:41] <ellyjones> mononoke: haha, no - tpm = technical program manager :)
[20:30:59] <mononoke> ah, phew
[20:31:15] <mononoke> sorry, I was reading quickly.
[20:31:22] <mononoke> gotta go to a meeting
[20:36:29] *** jrbarnette has joined #chromium-os
[20:36:30] *** ChanServ sets mode: +v jrbarnette
[21:08:10] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from None: )'
[21:14:04] <wfrichar> sigh.
[21:14:27] <wfrichar> Can anyone interpret http://chromeos-botmaster.mtv.corp.google.com:8026/builders/TOT%20Pre-Flight%20Queue/builds/4008/steps/Report/logs/stdio for me?
[21:15:01] <ellyjones> what does it mean when u-boot prints 'cmdline after update:' and then the kernel never boots?
[21:16:55] <ellyjones> (that's the last console output I ever get)
[21:17:35] <wfrichar> ellyjones: I don't know for certain. What's supposed to happen (I think) is that U-boot expands the built-in kernel commandline, replacing any %D, %P, %U with the drive, partition, and UUID of the selected kernel partitions. The next step is to invoke the kernel with that commandline.
[21:17:50] <wfrichar> so if nothing happens after that, it's generally because the kernel fell over.
[21:18:22] <wfrichar> That's how the x86 bootstub works. I haven't looked at ARM, but it should do the same thing, although in a different way.
[21:18:29] <ellyjones> fell over before printing _anything_? :(
[21:18:49] <ellyjones> do you know who a good person to ask about this on Arm would be?
[21:18:50] <wfrichar> Yeah. I see that in x86 occasionally.
[21:19:07] <ellyjones> how does one diagnose falling over without printing anything?
[21:19:14] <wfrichar> Anton?
[21:19:28] <wfrichar> Gabe?
[21:19:37] <wfrichar> Benson?
[21:19:45] <wfrichar> dianders?
[21:19:55] <wfrichar> All these people are at lunch right now, of course.
[21:20:06] <ellyjones> rats
[21:20:43] <ellyjones> I would think that anything that could go wrong in early kernel would cause it to scream on its serial console
[21:20:56] <wfrichar> For x86, I used to craft a special kernel command line that has things like earlyprintk=serial,ttyS0,115200 console=ttyS0,115200 loglevel=7 in it.
[21:21:31] <ellyjones> I already have earlyprintk console=ttyS0,115200
[21:21:46] <wfrichar> Bleah. Then I got nuthin. Sorry.
[21:22:12] <ellyjones> alright
[21:22:15] <wfrichar> At that point I start looking for GPIOs attached to LEDs that I can turn off & on.
[21:22:16] <ellyjones> well, I'll try bisecting
[21:23:31] <wfrichar> In other news, the build fell over. I'm supposed to be sheriff, but I don't know how to interpret the failure. Anyone? Anyone?
[21:24:46] <cjb> wfrichar: any idea what the error is?
[21:25:19] <cjb> reading scrollback, there are some interesting ways to debug early boot failures
[21:25:20] <ellyjones> I am also sheriffing
[21:25:38] <ellyjones> wfrichar: the 'report' step is never interesting
[21:25:45] *** nick761 has quit IRC
[21:25:56] <wfrichar> bin/ctest fails in one case, bin/cros_au_test_harness in another.
[21:26:09] <ellyjones> chrome sig 11
[21:26:10] <ellyjones>   suite_Smoke/login_CryptohomeMounted
[21:26:19] <wfrichar> where do you find that?
[21:26:34] <cjb> we use this patch at OLPC, for early boot debugging:  http://dev.laptop.org/git/xo1.75-kernel-bringup/commit/?id=3d252d891def274e294520480a8860b315678060
[21:26:35] <ellyjones> http://chromeos-botmaster.mtv.corp.google.com:8026/builders/TOT%20Pre-Flight%20Queue/builds/4008/steps/cbuildbot/logs/stdio
[21:26:49] <cjb> it allows you to see printks from being the serial port init happens, which is pretty late during boot
[21:28:14] *** nick761 has joined #chromium-os
[21:28:20] *** dtu has joined #chromium-os
[21:28:20] *** ChanServ sets mode: +v dtu
[21:29:14] <wfrichar> ellyjones: that link is FULL of errors. How do we tell which is a real error?
[21:29:55] <ellyjones> wfrichar: long and painful experience? :(
[21:30:34] <wfrichar> ugh.
[21:30:58] <wfrichar> So, are you going to look at it, or should I?
[21:31:26] <wfrichar> (warning: If I look at it, I'm just going to be asking you a bunch of idiot questions)
[21:31:44] <wfrichar> (but I'm okay with doing that)
[21:31:46] <cjb> I don't mind taking a look if one of you wants to repaste the link somewhere
[21:31:55] * cjb just a random hacker, though
[21:32:56] <wfrichar> Thanks, cjb, but I'm not sure I can do that. Might be some partner-sensitive info in the logs.
[21:33:05] <cjb> Ah, okay, didn't occur to me.  No problem.
[21:33:25] <ellyjones> wfrichar: it might provide more total learning if you look, but I am happy to do it instead if you are busy
[21:33:54] <wfrichar> I'll do it. Vince is on it with me.
[21:35:39] <ellyjones> alright, cool
[21:38:01] <ellyjones> nope, still hangs
[21:41:47] <ellyjones> wfrichar: you recommended people before, but I cannot map those names to LDAP names
[21:41:54] <ellyjones> what are their LDAP names?
[21:47:59] *** benchan has joined #chromium-os
[21:48:00] *** ChanServ sets mode: +v benchan
[21:48:25] *** benchan is now known as Guest54974
[21:48:31] <dianders> ellyjones: you still getting hangs on boot of ARM?  Every time, or just sometimes?
[21:48:59] <ellyjones> every time, before any output is printed
[21:49:08] <ellyjones> just after "cmdline after update:"
[21:49:25] <ellyjones> HEAD is 40080245260a65403fb0ef107582193f392459bb
[21:49:30] <dianders> Hmmm.  Booting from what device?  ...and what u-boot version?
[21:49:49] <ellyjones> kaen, 2010.09, built with USE=-device_tree
[21:50:13] <dianders> Hmmm.  Possible that USE=-device_tree hasn't been tested w/ latest u-boot.
[21:50:37] <dianders> This is with the old u-boot, right?  You were going to do that to try to upgrade things...
[21:50:37] <ellyjones> I'm not using latest u-boot
[21:50:44] <ellyjones> 2010.09 is old, yes
[21:51:04] <ellyjones> and yes, I'd like to upgrade it, except I do not yet have an A-A cable and now I cannot boot my kaen any more :\
[21:51:17] <dianders> Hmmm.  I guess I could believe that wouldn't work.  Ick.  Nobody has A-A cables there?
[21:51:37] <ellyjones> no; I have one being shipped, but it's not here yet
[21:52:11] <dianders> Sigh.  Let me think if I can figure out any other good way to upgrade your u-boot...
[21:52:17] <ellyjones> alright
[21:52:25] *** Guest54974 has quit IRC
[21:52:31] *** benchan_ has joined #chromium-os
[21:52:32] *** ChanServ sets mode: +v benchan_
[21:52:35] <ellyjones> I am doing this: USE=-device_tree ./build_image --boot_args "noinitrd no_console_suspend"
[21:53:02] * ellyjones tries rebuilding kernel from clean with USE=-device_tree
[21:54:07] *** SoulRaven has quit IRC
[21:54:13] *** ihf has joined #chromium-os
[21:56:57] <dianders> ellyjones: There ought to be some way to get u-boot to load a new version over TFTP, then flash that.  ...but I don't think anyone has actually tried it.  I've got to run to another meeting now, unfortunately.  I guess the other route is like you're doing: trying to see if USE=-device_tree can be figured out...
[21:57:06] <ellyjones> yeah
[21:57:15] <ellyjones> we'll see, I guess
[21:57:20] *** davidjames has joined #chromium-os
[21:57:20] *** ChanServ sets mode: +v davidjames
[21:57:41] <dianders> OK.  I'll check back after my meeting and see if I can come up with a quick set of steps for you...  Good luck!
[21:58:08] *** patcito has joined #chromium-os
[21:58:18] <sosa> nirnimesh committing on closed tree :p
[21:58:42] <davidjames> wfrichar, vlaviano: TOT preflight queue cycled green, time to open?
[21:59:04] *** rbyers has quit IRC
[21:59:58] *** rbyers has joined #chromium-os
[22:00:36] *** vmil86 has quit IRC
[22:01:06] <crosbot> tree became 'Tree is open (chrome crash on logout issue: 18158)'
[22:03:41] <wfrichar> just opened it.
[22:08:59] <crosbot> tree became 'Tree is open (chrome crash on logout issue: 17899)'
[22:09:18] *** DeezeNuts has joined #chromium-os
[22:09:22] <DeezeNuts> Hello
[22:10:40] <ellyjones> hi
[22:10:55] <ellyjones> ok
[22:11:04] <ellyjones> dianders: a discovery was made: cros_workon_make and USE flags do not go together
[22:11:16] <ellyjones> or at least, cros_workon_make and a *change* of USE flags
[22:11:46] <cmasone> ellyjones: yeah, cros_workon_make is intended for iterative building.  If you wish to force new USE flags and stuff, pass --reconf
[22:11:47] <DeezeNuts> I have a CR48, is it possible to remove chrome OS completely and reinstall it? I currently have Ubuntu 10.10 installed
[22:11:52] <ellyjones> cmasone: excellent
[22:12:01] <ellyjones> DeezeNuts: recovery mode will let you reinstall it
[22:12:04] <cmasone> ellyjones: at least, I'm pretty sure that works :-)
[22:12:19] <DeezeNuts> ellyjones, will it remove my ubuntu?
[22:12:28] <DeezeNuts> I want to install chromium os, not chrome
[22:13:10] <ellyjones> oh
[22:13:17] <ellyjones> just build a chromium OS stick and install it
[22:17:11] <cjb> hmph, my latest build doesn't work.  Don't have any output past "Freeing init memory" at all, although I can chroot into C-ROOT and see that the binaries are runnable.
[22:17:33] <cjb> so I'm looking for a way to get early init debugging.  just passing console=ttyS2,115200 isn't good enough to get init debugging over there.
[22:28:45] <cjb> I eventually got a console, and X is running and I can run /opt/google/chrome/chrome manually and have it come up, so that's progress.  But I still need to debug what's going wrong in early boot somehow.  Does anyone know which script kicks off launching the X session/window manager/etc?
[22:29:25] <marcheu> /sbin/session_manager_setup.sh
[22:29:53] <cjb> cool, thanks
[22:31:44] <cjb> woot!
[22:34:19] *** mascond0 has joined #chromium-os
[22:35:05] *** mascondante has quit IRC
[22:37:09] <cjb> http://dev.laptop.org/~cjb/chromiumos-xeno-boot.jpg :)
[22:37:12] <wfrichar> DeezeNuts: if you use the stock installer, it will repartition your SSD. If you build your own image, you should be able to just use dd to copy the kernel and rootfs partition.
[22:37:34] <DeezeNuts> I see
[22:37:52] <DeezeNuts> If i uninstall chrome though will my ubuntu be intact?
[22:37:58] <DeezeNuts> or will that mess up something with the bootlader
[22:37:59] <ellyjones> hrm
[22:38:01] <DeezeNuts> *loader
[22:38:07] <wfrichar> Depends on how you uninstall it.
[22:38:34] <ellyjones> dianders: I have noticed something that may be of interest: when the kernel command line is <512 bytes, the system boots ok...
[22:38:55] <wfrichar> The BIOS just looks for kernel partitions, validates them in priority order (bigger numbers are higher priority, 0 means 'ignore'). In dev-mode it allows self-signed kernels, in normal mode it doesn't.
[22:39:32] <wfrichar> Assuming the kernel signature is valid, the BIOS just launches it. What happens next is up to the kernel.
[22:39:41] <ellyjones> or rather, if I change the boot opts such that the kernel commandline is > 512 bytes, it no longer boots
[22:39:45] <wfrichar> (or, more specifically, to the bootstub which is attached to the kernel)
[22:40:06] <wfrichar> ellyjones: that is suspicious!
[22:40:10] <ellyjones> yes
[22:40:11] <ellyjones> it is
[22:40:24] <ellyjones> in fact, it suddenly stops booting in a "no console output at all after u-boot" kind of way
[22:41:07] <ellyjones> I can't think of what could possibly cause that though
[22:41:36] <nirnimesh> sosa: sorry, I didn't notice
[22:42:21] <wfrichar> ellyjones: All the obvious suspects spring to mind: the command-line modifying part of U-Boot runs off the end, the kernel's command-line parser runs off the end, the command-line is being truncated by bits of the kernel image (or vice-versa).
[22:42:37] <ellyjones> wfrichar: yeah... hrm
[22:43:11] <wfrichar> there was discussion here last week that some kernel command-line args are not recognized if they're the last term in the string. I can't recall the details.
[22:43:25] <ellyjones> asm-generic/setup.h:4:#define COMMAND_LINE_SIZE 512
[22:43:30] <ellyjones> Fffffff.
[22:43:34] <ellyjones> I have been had.
[22:43:38] <wfrichar> yup.
[22:43:57] <cjb> yeah, I was gonna say this problem sounded familiar to me :)
[22:44:04] <wfrichar> x86 is 4096 chars, not actually #defined AFAIK, but implicit in the bootstub layout.
[22:44:12] <cjb> we hit it on OLPC, and we moved some options into CONFIG_CMDLINE
[22:44:16] <cjb> which lets you compile some options into the kernel
[22:44:24] <ellyjones> wfrichar: this is arm :P
[22:44:38] <wfrichar> I know, I'm just offering what little context I have.
[22:44:41] <cmasone> THIS
[22:44:42] <ellyjones> o*ho*!
[22:44:43] <cmasone> IS
[22:44:44] <gauravsh> 512 bytes oughta be enough be for everyone.
[22:44:47] <cmasone> AAARRRRRMMMMM!!!
[22:44:48] <ellyjones> 418         saved_command_line = alloc_bootmem(strlen (boot_command_line)+1);
[22:44:52] <ellyjones> 420         strcpy (saved_command_line, boot_command_line);
[22:44:58] <ellyjones> "boom!"
[22:45:27] <wfrichar> Yes, ARM, but we'd really like as much of everything to be as identical as possible in every way.
[22:46:19] <ellyjones> okay, the arm setup_arch() strlcpy()s it into a smaller buffer
[22:46:27] <wfrichar> Ideally we'd use U-Boot and FDT for x86 too instead of the backward-compatible-to-8088 mode.
[22:46:28] <ellyjones> so it's not running off the end, but we do lose some commandline options
[22:46:40] <ellyjones> that seems bad
[22:46:50] <ellyjones> will anyone complain if I bump COMMAND_LINE_SIZE up? :)
[22:50:18] <cmasone> wfrichar: vlaviano ellyjones I have a sneaking suspicion the x86 pfq is hung while testing
[22:50:43] <wfrichar> hmph. poke it.
[22:51:26] <ellyjones> "Hey, kid, wanna see a hung vm?"
[22:53:53] <gauravsh> anyone know our process for getting a third party package source into src/third_party?
[22:55:01] <cmasone> gauravsh: ask davidjames, brecause I always forget
[22:55:52] <cmasone> that is my process
[22:56:03] <gauravsh> all roads lead to davidjames, apparently. :)
[22:56:40] <gauravsh> cmasone: I will update our developer documentation appropriately.
[22:57:16] *** saintlou has quit IRC
[22:58:40] *** saintlou has joined #chromium-os
[22:58:40] *** ChanServ sets mode: +v saintlou
[23:02:07] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-mario-binary" from 6f6a12802e8294c709ad2e45e7367b6f4036df6f: mazda at chromium dot org <mazda at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>, tsepez at chromium dot org <tsepez at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[23:02:17] <ellyjones> hrm
[23:02:27] <ellyjones> on Arm it appears to be 1024 though... what gives?
[23:02:47] <ellyjones> the 'platform-neutral' size differs from the 'real' one?
[23:08:54] <davidjames> wfrichar: That tree closure looks like http://crosbug.com/16574
[23:08:54] <davidjames> vlaviano: ^
[23:09:17] <vlaviano> wfrichar and sosa are taking a look at it atm
[23:10:19] <cjb> hmph.  I can't get the "Continue" button on the splash screen to become clickable after associating to wifi.  I think I'm not really associating, though.  Anyone have tips for debugging networking?  (am using libertas/sd8686 driver)
[23:10:43] <vlaviano> there are some diagnostic tools in crosh (ctrl-alt-t)
[23:10:49] <vlaviano> the output of chrome://system might also help
[23:11:11] <cjb> don't think either of those work at the splash screen
[23:11:31] <vlaviano> yeah, maybe stick in usb ethernet dongle to get past that point and then debug
[23:11:44] <cjb> ah, good idea, thanks
[23:11:54] <cjb> is there any way to get past the splash without being online?
[23:12:21] *** Adys has quit IRC
[23:12:35] <vlaviano> ctrl-alt-f2?
[23:12:45] *** Adys has joined #chromium-os
[23:12:49] <marcheu> cjb: a usb to ethernet dongle is what seems easiest right now
[23:12:58] <cjb> ok
[23:13:23] <marcheu> or ethernet cable if you have rj45
[23:13:49] <cjb> nah, don't
[23:14:01] <marcheu> maybe the olpc is indeed a chromebook
[23:14:53] <superm1> ctrl-alt-L should get you past that screen too if it's a nonofficial build
[23:15:01] <superm1> cjb, ^
[23:15:14] <cjb> superm1: ooh, thank you
[23:15:52] <cjb> it did!
[23:17:28] <cjb> chrome://system is impressive
[23:18:44] *** saintlou has quit IRC
[23:26:44] *** saintlou has joined #chromium-os
[23:26:44] *** ChanServ sets mode: +v saintlou
[23:31:19] *** rharrison_chrome has quit IRC
[23:32:53] <wfrichar> The tree has been closed for 30 minutes. There have been 23 commits since then. WTF?
[23:33:18] <davidjames> wfrichar: Looking at Chrome commits?
[23:33:42] <wfrichar> http://buildbot.jail.google.com/i/chromiumos/waterfall
[23:34:00] <davidjames> wfrichar: Yeah that waterfall shows Chrome commits as well, which have a different tree status
[23:34:06] <wfrichar> Ah. That's probably it.
[23:34:43] <wfrichar> Yeah, there are only a handful that have a sha1 on the waterfall.
[23:35:16] <davidjames> wfrichar: Actually Chrome commits show up with a sha1 too now
[23:36:10] <wfrichar> I'm not seeing that.
[23:36:14] <davidjames> wfrichar: E.g. http://build.chromium.org/p/chromiumos/changes/36029 is a Chrome commit and has a sha1. But format is slightly different
[23:36:51] <wfrichar> hm.
[23:37:33] <davidjames> kmixter's commit was pushed on a closed tree though :(
[23:37:57] <redpig> good thing it only removes code :p
[23:42:21] <gauravsh> folks, I am planning to add add a new git repo under third_party which will hold the opencryptoki source + our fixes
[23:42:24] <gauravsh> any objections to that?
[23:42:51] <gauravsh> http://dev.chromium.org/chromium-os/how-tos-and-troubleshooting/git-server-side-information mentions I should poll IRC or cros-dev before going ahead with it
[23:43:08] <gauravsh> speak now or forever hold your peace
[23:45:29] *** kmixter has joined #chromium-os
[23:45:38] *** saintlou has quit IRC
[23:47:34] *** rbyers has quit IRC
[23:50:56] *** saintlou has joined #chromium-os
[23:50:57] *** ChanServ sets mode: +v saintlou
[23:53:07] *** Styx has joined #chromium-os
[23:53:20] <gauravsh> the unpacked git checkout for this package is around 12M





top