| === bladernr_ is now known as bladernr_afk | ||
| === ericm|ubuntu is now known as ericm-afk | ||
| jussi | pgraner: lamont et al, are you happy with jsalisbury being added to the ops team in here? | 08:00 |
|---|---|---|
| apw | jussi, (depending on whether i count as et al) jsalisbury gets my vote | 08:36 |
| jussi | apw: you do count :) Ill get it done at some point in the near future. if anyone has an issue with it, PM me. | 08:53 |
| apw | jussi, thanks :) | 08:53 |
| === smb` is now known as smb | ||
| stephan | hello. | 09:13 |
| stephan | on ubuntu oneiric ocelot with kernel 3.0.0-r13-generic i had kworker issue (cpu load: 80% or more). | 09:13 |
| stephan | i installed a kernel 3.2 ... then after installation if rebootet to old kernel 3.0.0-r13. now no problems with kworker. | 09:14 |
| stephan | someone knows how to find out? | 09:14 |
| jussi | [11:16:22] [ChanServ] Flags +votiA were set on jsalisbury in #ubuntu-kernel. | 09:16 |
| jussi | apw: pgraner jsalisbury ^^^ | 09:16 |
| apw | jussi, thanks, will poke him to test when he is about | 09:17 |
| jussi | apw: this is a useful doc if you need it: https://wiki.ubuntu.com/IRC/IrcTeam/OperatorGuide | 09:18 |
| jsalisbury | jussi, apw channel op works for me. Thanks, jussi! | 14:03 |
| lamont | jussi: I get no vote here | 14:04 |
| === bladernr_afk is now known as bladernr_ | ||
| * herton -> lunch | 15:07 | |
| bjf | brendand, bug 897773 | 15:15 |
| ubot2 | Launchpad bug 897773 in linux "[Acer AR320 F1 and Acer AR160 F1] Halting shortly after booting with Lucid (2.6.32-36.79) -proposed kernel" [Undecided,Incomplete] https://launchpad.net/bugs/897773 | 15:15 |
| bjf | brendand, doesn't look like he did anything i asked him to do, however it now looks like the issue has "gone away", comments ? | 15:16 |
| brendand | bjf - well, the issue hasn't 'gone away' but it turns out it does affect the release kernel | 15:18 |
| bjf | brendand, i don't see any comments like that on the bug | 15:19 |
| bjf | brendand, i see the comment on the tracking bug | 15:20 |
| brendand | bjf - sorry, i'll update it in the bug itself | 15:22 |
| bjf | brendand, is that system supposed to have been certified for lucid ? | 15:23 |
| brendand | bjf - it is | 15:23 |
| brendand | bjf - it was certified just fine with 10.04.2 | 15:24 |
| bjf | brendand, so the regression is between 10.04.2 and 10.04.3 ? | 15:25 |
| brendand | bjf - that's what it looks like at the moment. he hasn't got around to trying the 10.04.2 image yet. i'll try and do that today, at least for the release kernel | 15:30 |
| * ogasawara back in 20 | 15:30 | |
| bjf | brendand, does that mean you have access to the HW now and can test kernels ? | 15:31 |
| brendand | bjf - supposedly. i haven't tried in anger yet | 15:32 |
| skaet | ogasawara, could you tweak and improve the kernel overview comments I put in the https://wiki.ubuntu.com/PrecisePangolin/TechnicalOverview? (when you get back) | 15:34 |
| bjf | skaet, you may want to glance over the scrollback (the interaction between myself and brendand) | 15:34 |
| arges | bryceh, hello, got a better log from our gpu hangs on the x220 machines here http://paste.ubuntu.com/755004/ this is with a 64-bit kernel, 32-bit seems to behave differently | 15:44 |
| ogasawara | skaet: ack, will do | 15:55 |
| skaet | thanks ogasawara :) | 15:57 |
| stephan | on ubuntu oneiric ocelot with kernel 3.0.0-r13-generic i had kworker issue (cpu load: 80% or more). i installed a kernel 3.2 ... then after installation if rebootet to old kernel 3.0.0-r13. now no problems with kworker. | 16:09 |
| stephan | after rebooting the old kernel 3.0.0-r13 there wasnt kworker issue anymore. any suggestions? | 16:09 |
| tgardner | meh. 'Sorry, something just went wrong in Launchpad.' | 16:12 |
| Razec | Where Can I find kernel's bug from ubuntu 11.10? | 16:57 |
| bjf | Razec: all bugs are filed in launchpad | 17:00 |
| Razec | bjf, ok, I asked because I not found a specifically option to ubuntu 11.10. But I'll take a look. Thanks so much. | 17:03 |
| Razec | bjf, I got it | 17:08 |
| Razec | https://launchpad.net/ubuntu/+milestone/ubuntu-11.10 | 17:08 |
| Razec | tks | 17:08 |
| === bigbash is now known as zz_bigbash | ||
| Laibsch | Is bugzilla.kernel.org not going to come back? It's been more than two months since the outage and there's no information on what the plan is. I'd like to offload kernel bugs upstream but mailing list? dunno. | 18:47 |
| arges | tgardner, howdy. looking at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/769927 . There is a cherry picked patch that fixes this issues for natty, and I found that that patch is already in oneiric. What should I change the status of the Affects linux->Oneriric field? thanks | 18:48 |
| ubot2 | Launchpad bug 769927 in linux "Kernel Oops : Dentry still in use (1) [unmount of nfs4 0:1d]" [Undecided,Confirmed] | 18:48 |
| tgardner | arges, I set it to fix committed | 18:53 |
| arges | tgardner, ok thanks | 18:53 |
| tgardner | checking to see if its been released | 18:53 |
| tgardner | arges, yep, looks like it came with 3.0 | 18:55 |
| tgardner | arges, I suppose you'll be wanting a cherry-pick next ? | 18:56 |
| arges | tgardner, got the cherry pick, typing up the SRU email | 18:57 |
| tgardner | arges, ack | 18:57 |
| * tgardner -> lunch | 19:01 | |
| patrickmw | jsalisbury, hey I updated bug 897866. | 20:01 |
| ubot2 | Launchpad bug 897866 in linux "NFS install fails with a kernel panic" [High,Confirmed] https://launchpad.net/bugs/897866 | 20:01 |
| patrickmw | bjf, ^ this bug is stopping me from running boot speed tests | 20:02 |
| bjf | tgardner: ^ | 20:03 |
| tgardner | patrickmw, do you know that the configuration is correct? i.e., does it work with natty or lucid? I would expect it fall back to busybox instead of panicing. | 20:08 |
| patrickmw | tgardner, it worked with isos before 20111129.1 | 20:12 |
| tgardner | patrickmw, thats about the time we uploaded the kernel meta package that would have referenced the 3.2 kernel | 20:17 |
| * herton -> eod | 20:21 | |
| rik_ | sforshee: can I bug you about bug 606238 again? | 20:54 |
| ubot2 | Launchpad bug 606238 in linux "synaptic touchpad not recognized on dell latitude e6510 and others" [Undecided,Confirmed] https://launchpad.net/bugs/606238 | 20:54 |
| sforshee | rik_, sure | 20:54 |
| rik_ | I've attached two logs as described on your blog to the bugreport. | 20:54 |
| rik_ | but I'm not sure they contain what is needed to further debug this | 20:54 |
| rik_ | what should I do in the vm once the driver is installed? simply move around and try scrolling etc? | 20:55 |
| * sforshee looks at the logs | 20:55 | |
| sforshee | rik_, once you have the windows driver installed in the VM you want to make sure it really is working | 20:56 |
| sforshee | so you should try to do whatever works under bare-metal windows -- scrolling, etc | 20:57 |
| rik_ | I checked the device manager and it said it was a dell touchpad. I was also able to scroll using the right side of the touchpad | 20:57 |
| sforshee | good, so the driver works | 20:57 |
| rik_ | I can also enable more options using a control panel applet, such as two-finger scrolling etc but all the options are disabled in the log | 20:58 |
| sforshee | do either of the logs you attached represent a trace from when you know the windows driver is working? | 20:58 |
| sforshee | because the data looks like normal PS/2 to me | 20:58 |
| rik_ | the 'without' is without the drive rinstalled. the 'with' one is with me scrolling | 20:58 |
| rik_ | hmm | 20:59 |
| * rik_ will create another dump | 20:59 | |
| sforshee | rik_, perhaps the "with" dump is different | 21:03 |
| rik_ | yes, "with" is with the driver installed | 21:03 |
| rik_ | "without" is without the driver installed | 21:03 |
| sforshee | it actually looks like a 6-byte protocol, I just didn't look close enough | 21:03 |
| sforshee | so you've got a good start! | 21:04 |
| rik_ | I'm making another dump with two finger scrolling options, and rotating enabled | 21:04 |
| sforshee | rik_, what you really need from the first dump is the initialization, so what you have is good for a starting point | 21:05 |
| rik_ | ok | 21:05 |
| rik_ | my last dump has me two-finger scrolling, two-finger rotating and two-finger zooming | 21:05 |
| rik_ | I will also upload it | 21:06 |
| rik_ | is this enough information to determine what the correct settings for the alps.c file should be? | 21:07 |
| rik_ | the latest dump is attached | 21:07 |
| sforshee | it's certainly enough to get started | 21:07 |
| rik_ | to be honest I have no idea where to go from here, so I hope you can find some time to have a look at it in the near future | 21:08 |
| rik_ | the 6-byte protocol is not one of the current V2,3,4 protocols? | 21:09 |
| sforshee | i can't tell, but you have to get past the initialization before any of that matters | 21:09 |
| rik_ | ah ok. I assume the init is in the beginning of the file and are the three values in the alps.c file that should respond in some way? | 21:10 |
| sforshee | it's more than that. look at the hw_init functions | 21:11 |
| sforshee | it's definitely not the same as v3 or v4 | 21:11 |
| rik_ | you're right, it seems a lot more complex than a few values | 21:14 |
| rik_ | and out of my field of expertise | 21:15 |
| yofel | hey, Is there a way to freeze the system during boot? 3.2.0-2 in precise doesn't boot on my thinkpad t510, and I can't read the error as it's thrown so often that I can't read it | 21:16 |
| sforshee | rik_, I don't anticipate being able to spend a lot of time on this in the immediate future. I'll try to look at it deeper when I have some time. If you want to dig in further in the meantime I can give you some assistance | 21:19 |
| rik_ | sforshee: OK, what can I do? | 21:20 |
| sforshee | you might want to start by taking a look at this as an general PS/2 mouse introduction: http://www.computer-engineering.org/ps2mouse/ | 21:21 |
| arges | yofel, you can boot with a working kernel and read through an older /var/log/dmesg.0, /var/log/dmesg.1.gz, etc | 21:21 |
| yofel | arges: problem is that I would need a mounted rootfs for the logs, but the kernel doesn't get past initrd from what I see | 21:22 |
| yofel | I'll try 3.2.0-999.201111300416 for now | 21:22 |
| sforshee | rik_, in your dumps, you'll notice blocks that start with the driver sending a reset (ff) and end with it enabling data reporting (f4) | 21:23 |
| arges | yofel, you deleted the older kernels? | 21:23 |
| yofel | nope, 3.1.0-2 boots fine | 21:23 |
| yofel | but 3.2.0 didn't write anything to the logs | 21:23 |
| rik_ | sforshee: yes, I scrolled down on the page you sent and noticed the flags in the log | 21:23 |
| sforshee | rik_, the parts in between are initializations. it would be good to try to figure out what the initialization is doing, but that's also difficult. you can simply recreate the stream of commands in the alps driver though to get started | 21:24 |
| rik_ | sforshee: where exactly in the alps file should I do that? | 21:24 |
| rik_ | sforshee: so I should focus on everything between the ff reset and f4? | 21:25 |
| sforshee | rik_, just make your model use a new protocol version, e.g. v5. add a v5 initialization function to send the initialization you need | 21:25 |
| sforshee | then I'd make alps_process_byte dump the data and exit | 21:26 |
| sforshee | you'll want a USB mouse to use while you're doing this | 21:26 |
| rik_ | can you elaborate on "make alps_process_byte dump the data and exit"? | 21:27 |
| sforshee | if you send the initialization correctly the data you see should look similar to what you see in the dumps you took | 21:27 |
| sforshee | are you using the code from the dkms packages I made? | 21:27 |
| rik_ | I believe it's based on a version 10 | 21:28 |
| rik_ | is that possible? | 21:28 |
| sforshee | yes, that's possible | 21:28 |
| sforshee | just a second, let me get that code | 21:28 |
| sforshee | so you should see a function named alps_packet_dump, right? | 21:30 |
| rik_ | sforshee: yes | 21:30 |
| sforshee | okay, where that's being called in alps_process_packet -- change alps_packet_dump to be called unconditionally, then return | 21:31 |
| sforshee | that will stop the driver from trying to process the data | 21:31 |
| yofel | k, 3.2.0-999.201111300416 boots, but has no iwlagn module, 3.2.0-2 fails pretty fast with what seems to be a DRAM error (like 100 errors / sec) | 21:32 |
| sforshee | rik_, but you need to be sure that you set psmouse->pktsize to 6 | 21:32 |
| rik_ | in the alps_init function for my V5? | 21:34 |
| rik_ | sforshee: would you mind adding this info to the bug report and/or blog about it? | 21:34 |
| rik_ | other can help out as well then | 21:35 |
| rik_ | I hope my chat client is logging this :-) | 21:35 |
| sforshee | rik_, I'll see if I can think of something fairly generic to post, but it's really rather ad-hoc. I'm just making it up as I go :) | 21:36 |
| rik_ | sforshee: do you have a dump for the v3 and/or v4 you added to the file? That way I can compare how you do the init there? | 21:37 |
| rik_ | if I change the routines for v3 with the unconditional dump, would that work? v3 has 6 for pktsize? | 21:40 |
| rik_ | only v4 has 8, or is it the other way around? | 21:40 |
| sforshee | rik_, I don't think I have any posted anywhere, but I also don't think they would prove very helpful. Yours looks substantially different | 21:40 |
| sforshee | only v4 has 8, so if you add a v5 version it should end up with 6 as it is now | 21:40 |
| rik_ | sforshee: I would like to see their init and see how you have translated that into code | 21:41 |
| rik_ | ah ok | 21:41 |
| sforshee | rik_, http://people.canonical.com/~sforshee/alps-touchpad/latitude-E6520-sniff.txt | 21:42 |
| sforshee | that's a v3 | 21:42 |
| sforshee | but it will be hard to follow. The driver doesn't just replicate the sequence directly | 21:43 |
| rik_ | sforshee: ok, I'll have a look at this. But it's getting late now. I'll read the docs you mentioned and will come back here when I find some time to look at this | 21:44 |
| rik_ | sforshee: thanks for the help so far! | 21:45 |
| sforshee | rik_, np. Feel free to ask more questions as the come up. | 21:45 |
| rik_ | sforshee: don't worry, they will come up | 21:46 |
| rik_ | sforshee: the file you gave has blocks of sequences that do specific things like set the sensitivity? They are not a simple replay of everything? | 21:47 |
| rik_ | what does an e6 and e7 do? | 21:48 |
| sforshee | rik_, the v3 and v4 initialization uses a register access protocol that involves switching modes, reading and writing to registers | 21:48 |
| sforshee | it's fairly convoluted | 21:49 |
| rik_ | does my dump look like something similar? How can I know/"see" this? | 21:49 |
| sforshee | there's a command sequence to enter "command mode" where the registers can be read and written -- ec ec ec e9 | 21:50 |
| sforshee | then ea exits command mode | 21:50 |
| sforshee | in your dump i see it enter and exit command mode, but pretty much everything happens when a v3/v4 touchpad would not be in command mode | 21:50 |
| rik_ | I guess my last dump is more interesting then, as it has more features enabled? | 21:51 |
| sforshee | i can't tell. All i really know is that it's different | 21:51 |
| sforshee | most of your initialization doesn't happen in command mode | 21:52 |
| rik_ | in your dump it always responds with fa to a successfull command? | 21:52 |
| sforshee | or else the way to getting to command mode is different, or it doesn't have command mode at all | 21:52 |
| sforshee | the fa response is standard ps/2 | 21:52 |
| rik_ | sforshee: I have to go now. Thanks again for your time. If you find time to look at this, please do. I'm sure you'll beat me to it ;-) | 21:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!