/srv/irclogs.ubuntu.com/2011/06/20/#ubuntu-kernel.txt

bullgard4I obtained a message "panic occurred, switching back to text console". What Natty log file should reflect that?05:04
ohsixthat's a panic; it wouldn't have made it to a log file, you need to take a screenshot of it or otherwise read it to start investigating what it is05:12
ohsixer screenshot -> photo05:14
bullgard4ohsix: When having made a photo, is the proper way to report to Launchpad using the command '~$ ubuntu-bug linux-image-2.6.38-8-generic' and attaching there the photo as a file?05:27
ohsixthats what i did last time i reported a panic05:28
bullgard4Thank you.05:28
* apw yanws08:28
* apw waves to cking 08:34
jjohansenmorning apw, cking08:35
apwjjohansen, not again!  morning08:35
* apw wonders if mumble is broke, or if its just none of us is in CC08:36
ckingmorning one and all, here we go again08:36
jjohansenapw: nope I'm not here, you are still groggy got get some coffee08:36
=== smb` is now known as smb
jjohansens/got/go/08:37
jjohansenapw: so for aufs in the oneiric kernel we have commit dc7c488aeee4e3a50d6e76c5a6eba01d4ecfe432, and then b105abf699ff68fe0960a2568c14da8f82cfe4eb which is the merge from upstream which should have contained the patch in first commit mentioned, but that commit never got dropped08:45
apwjjohansen, looking08:47
apwjjohansen, ok the 'merge' from upstream is just a physical copy over of the aufs code08:48
apwso we'd never know to remove the old commit, does t08:48
jjohansenapw: ah, right08:48
apwthat matter ?08:48
jjohansenapw: nope, except it shows up as a sauce patch to check status on08:49
apwjjohansen, ahh just mark it 'merged upstream already' and i'll try and figure out how to lose it :)08:50
jjohansenapw: will do :)08:50
apw7467571 cpuidle: menu: fixed wrapping timers at 4.294 seconds09:58
ppisatilucid/fsl-imx51 have already 3 new commits but no start release since the last UBUNTU: Ubuntu-xxx closure10:12
ppisatishall i start a new release or do i keep committing on top of these?10:13
apwppisati, generally we expect whoever notices the startnewrelease is missing to add it, thoguh it doesn't matter when it occurs, so feel free to add it10:14
ppisatiapw: do i pop off the 3 commits and add a new release there or do i add it now?10:14
apwppisati, it totally doesn't matter the order, so just stick it next10:14
ppisatiok10:14
=== jjohansen is now known as jj-afk
* ogasawara back in 2014:31
sconklinmoin15:15
loolI've started an ubuntu-devel@ thread to discuss removal of the armel versatile flavor; I didn't Cc: ubuntu-kernel@, but will pass the result of the discussion in case it's favorable (nobody objects to the removal)15:18
tgardnerlool, isn't the versatile flavor required for qemu ?15:19
tgardnerlool, nm, just saw your email15:23
ppisatiwho decides which patch-es close a CVE? because i'm skeptical about CVE-2010-425115:31
ubot2ppisati: The socket implementation in net/core/sock.c in the Linux kernel before 2.6.34 does not properly manage a backlog of received packets, which allows remote attackers to cause a denial of service (memory consumption) by sending a large amount of network traffic, as demonstrated by netperf UDP tests. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-4251)15:31
ppisatisince the mentioned patches are not enought to close the DDOS scenario15:31
ppisatiand in fact, later, there's another patch related to it but the CVE doesn't mention it as necessary15:31
apwppisati, i would say in large part the security team who read the background, but should you find an apparent discrepancy do feel free to refer it back to them, better to waste a bit of their time checking than to miss it15:39
ppisatiapw: ok15:39
statAnyone here able to let me know a bit about the PAE kernel?16:03
stat(I'm just going to talk anyways. Ignore me if I'm irrelevant)16:06
statIt seems that the "big memory" flag for the PAE kernel is disabled by default now. That means you only get a maximum of 3.5GB. If you only get an extra 512MB, why even bother with PAE?16:07
apwstat, big memory flage?  which one are you referring to ?16:09
statI'm not entirely sure to be honest. From what I understand though, is that there is a flag that you must enable to get up to 64GB, and you need to recompile the current PAE kernel to get that flag enabled. I'll see if I can find the link where I was reading this.16:12
statOk, my mistake. It's called "High Memory Support (64GB)" not "Big Memory"16:13
apwstat, the PAE kernel doubled the number of bits available to store the location of physical page, which allows you to have more physical memory16:13
statI know what PAE is for. I've just installed it, but now it seems 3.5GB instead of the 4GB that it should, which apparently is due to this flag.16:13
stathttp://www.larmeir.com/2009/07/enabling-pae-on-a-32-bit-ubuntu-desktop-supporting-up-to-64-gb/16:14
apwstat, not necessarilly, physical layout of your ram may preclude seeing all 4GB16:14
apwcan you pastebin yout e820 information from your dmesg16:14
smbI think the issue is that you get 3.5G per process because kernel has to be mapped there too. but with PAE you can address all mem so the virtual 4G could be spread over different physical memory16:15
stathttp://pastebin.com/evE63C7P16:15
=== wendar is now known as allison
=== allison is now known as wendar
apwstat, ok and how do you know you only have 3.5gb, ie. where is that being expressed16:21
stat/proc/meminfo16:22
apwcan you paste the overall number pls16:23
statMemTotal:        3600576 kB16:23
apwstat, none of this makes any sense.  the point of the generic-pae kernel is that that it does set HIGHMEM64G16:29
apwcan you paste the whole dmesg please from this boot16:29
apwthat blog post is simply not applicable, as the kernle you have should have the flag you are aski16:30
apwasking to be on16:30
stathttp://pastebin.com/zLLZZLim16:31
=== wendar is now known as allison
apwstat, how much memory did you have on the -generic kernel ?16:43
apwi am suspectin16:43
apwi am suspecting from the physical layout that it as much less than 3GB16:44
statUmmm, I can't remember. I'm pretty sure it was 3GB. It was definitely more than 2.5 if that's what your suspecting.16:44
smbor probably asked the other way round: how much ram is there for real16:44
stat4GB. 2 2GB sticks.16:45
apwi am suspecting its 2.6 as you have a full G over 4 GB16:45
statThat... might be.16:45
apwor do you ... argle stupid hex numbers16:46
statHang on, I'll reboot it without PAE and double-check.16:46
apwstat, ok16:46
apwcking, this is 512MB right?16:46
apw[    0.000000]  modified: 0000000100000000 - 0000000120000000 (usable)16:47
* apw needs a sanity check16:47
smbapw, But for that (4G) it sounds reasonable to have 3.6G left after taking away kernel, page tables and other reservations, doesn't it?16:47
apwcking, and this is a smidge less than 3 right ?16:47
apw[    0.000000]  modified: 0000000000100000 - 00000000bffc6b00 (usable)16:47
apwso if i am reading this e820 right its is only advertising 3.5GB16:48
ckingapw, you got access to the machine?16:48
apwcking, nope,but i do have the whole e82016:49
apwhttp://pastebin.com/evE63C7P16:49
apwbut i think thats saying the bios is not advertising all the ram16:50
stat(Sorry that took so long. I suck at Grub2)16:53
apwstat, ok it looks like your bios is simply lying16:53
statMemTotal:        3096612 kB <-- Non PAE kernel.16:53
statMmk. So my motherboard is just crazy then?16:54
apw[    0.000000]  BIOS-e820: 0000000000000000 - 000000000009f800 (usable)16:54
apw[    0.000000]  BIOS-e820: 0000000000100000 - 00000000bffc6b00 (usable)16:54
apw[    0.000000]  BIOS-e820: 0000000100000000 - 0000000120000000 (usable)16:54
apware the 'usable' memory16:54
apwthe first two are basically the whole first 3GB, then you have .5GB from 4GB physical16:54
apwso 3.5 in total16:54
apwsometimes there is a layout option in your bios to say where the ram will be placed16:55
cking3583M useable IMHO16:55
apwbut given the information the machine is offering the kernel is doing the right things16:55
smb /me thinks there is also acpi tables placed 16:55
apwsmb, but those would be shown cut from the ram ... but its still not even there16:55
apwi suspect the ram is actually being placed under the io space and masked16:56
apwwhich is stupid16:56
ckingif e820 tables are wrong, that's all you get16:56
apw[    0.000000] e820 update range: 00000000d0000000 - 0000000100000000 (usable) ==> (reserved)16:56
apwi think that is the other bit being lost, i think that means its under the AGP apature16:56
smbhmm, maybe stolen graphics memory (on cheap cards)16:57
apwstat, what machine is this, we might find some help on if there is a bios wiggle16:57
apwcause you want different layouts for windows 32 bit and windows 64 bit ... so often there is a way16:58
hertonapw: yes, a bios option like "Remap memory above 4GB"16:58
apwherton, right16:58
statUhh an hp machine of sorts. 16:58
statDo you jsut want a device list?16:59
apwworth rooting about in the bios then, or searching for "remap memory above 4GB" and "machine name"16:59
apwon google16:59
apwheh ... really its the bios infor thats intreesting16:59
statSo do you want biosdecode info?17:00
ckingI have code to dump out the e820 from user space directly using the e820 BIOS call17:01
ckinghttp://kernel.ubuntu.com/git?p=cking/debug-code/.git;a=blob;f=e820/e820.c;h=fc4721eb80e50a7491dac4a059f746019f36d49e;hb=6cb227ece5ef420af9b341025b6a5e49e2d0e23217:01
statWill that be more useful if I reboot back into PAE?17:01
ckingbut I expect it will give you identical results as the kernel gives17:01
statcking: I am confused by your code. I tried that makefile that came with it, but it doesn't seem o have an actual target for building it.17:04
stathttp://pastebin.com/X0fVehMJ <-- biosdecode if you're interested. I'm going to fiddle with my BIOS settings as you've suggested.17:05
statThanks everyone for your help so far. :)17:05
ppisatilp#79980517:06
ppisatiplease accept my nomination17:06
tgardnerbug #79980517:07
ubot2Launchpad bug 799805 in linux-ti-omap4 "CVE-2010-4256" [Undecided,New] https://launchpad.net/bugs/79980517:07
tgardnerppisati, done17:07
ppisati10x17:07
statWell, I couldn't find anything in my BIOS about memory settings or remapping. I guess my BIOS is just lame. :P17:13
sforsheetgardner, one of the testers that had reported that the new rt2800 firmware made his rt3072 work is now saying it's causing problems (poor performance, machine won't shutdown/reboot). Is that justification to revert back to the old firmware, given that rt3072 didn't work at all previously?17:15
apwstat, did you see an option for the size of the AGP apature ?  and if so do you know how much video ram you have ?17:17
tgardnersforshee, yeah, I think so. I should just get the archive admon to reject it.17:17
tgardneradmin*17:17
statUhh, I did see a setting for video memory. It was set to auto, but could also be between 32-256.17:18
=== jj-afk is now known as jjohansen
tgardnersforshee, perhaps you should send an email to the SRU mailing list requesting rejection of the Lucid/Maverick packages that are currently in the unapproved queue17:19
tgardnersforshee, after that I'll just reset HEAD~2 on those 2 repos17:20
statapw: It looks like it's 256MB.17:22
sforsheetgardner, what's the address for the SRU mailing list? I'm not familiar with it.17:24
tgardnersforshee, uh, I forget. bjf, sconklin ?17:24
sconklinubuntu-kernel-sru@lists.ubuntu.com17:26
statOk, definitely 256MB according the memory range from lshw.17:27
=== allison is now known as wendar
* tgardner --> lunch19:00
=== AlanChicken is now known as AlanBell

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!