/srv/irclogs.ubuntu.com/2011/01/25/#ubuntu-kernel.txt

=== bjf is now known as bjf[afk]
=== zul_ is now known as zul
neur0nhello?00:56
dsfgcan anyone answer a computery question03:27
jjohansendsfg: I can try03:29
dsfgit's a dumb question and not about ubuntu03:29
dsfgbut how reliable is the info you get when you lookup an IP address to see which country it;s in?03:29
virtualdnot reliable at all since you can use a proxy or other form of tunnel, but i think it's usually right03:31
dsfgok im trying to find out where a person is, i got his IP03:32
dsfgand he is really dumb with computers03:32
dsfgso i dont think he even knows what  a proxy is03:32
dsfgi just want to see if he's lying about being in california on vacation, when i look up his IP it confirms he is lying but maybe the IP is lying03:32
dsfginstead03:32
jjohansendsfg: its fairly reliable, if there isn't a proxy03:32
dsfgeverywhere says the IP is on costa rica, so should i trust that?03:33
dsfgi know he isnt smart enouhg to use a proxy03:33
dsfgbut is there anything else that would cause a perosns IP to be in costa rica03:33
jjohansendsfg: well there is potentially bugs, or cache poisoning03:36
dsfgbut for the most part it's reliable?03:36
jjohansenerr make that stale cache03:36
jjohansenyes03:36
dsfgIP depends on your ISP not your computer right?03:37
dsfgi hope that question makes sense03:37
jjohansenyes, mostly depends on your ISP03:37
dsfgso, if someone had a laptop they bought in costa rica, but was in california using it03:38
dsfgwould they have a US or costa rican IP?03:38
jjohansenUS,03:38
jjohansenits not like cell phones03:39
jjohansenyour "unique" device id would be the ethernet mac03:39
dsfgwhat if he was on a cellphone03:39
dsfgin california03:39
dsfgusing 3g03:39
dsfgwould he have a US ip or costa rica?03:40
jjohansenhehe, that is an interesting problem03:40
jjohansenit would depend03:40
dsfgon what03:40
jjohansenit would depend on how the cellular provided routed cell data traffic, and contracts with other cell carriers03:41
dsfgi'm 99% sure he is on a laptop03:41
dsfgusing wifi03:41
dsfgand not a cellphone weith 3g03:41
dsfgit's not illegal to find someones IP and trace it to see what country theyre in, is it?03:42
jjohansengenerally you could accept that it will come from the cell carriers IP pool, so in US us carrier pool, in costa rica - that phone companies pool03:42
jjohansenuh no03:42
dsfgi feel like i did something wrong03:42
dsfgand went snooping!03:42
jjohansenwell it might be, depends where you are03:42
dsfgcanada03:42
jjohansenno03:43
dsfgbasically i think this guy is lying about being on vacation in california03:43
dsfgcause his IP says he;'s still in costa rica03:43
jjohansengeo location is done all the time, for targeting adds or giving you access to servers caching the data closer to you03:43
jjohansenuh I am not even going to go there03:44
dsfgbut i guess i cant be sure03:44
dsfgi can be like 99% sure03:44
jjohansenit is possible, especially if you are judging from email, as that will likely go out to his isp smtp server which is in costa rica03:45
dsfgnope03:45
dsfgwasnt email03:45
jjohansenI really can't tell you how likely it is, but geo location is right more often than not03:45
dsfgokay cool03:46
dsfgn ow ive been lied to :(03:46
dsfgwell ty vm guys03:50
diwicanyone still around for ubuntu-audio-dev meeting?11:17
diwicsorry wrong channel11:18
apwdiwic, heh :)12:58
diwicapw, yeah I was a little late and was in a hurry :-)13:03
diwicapw, I've done worse - the window was not scrolled down completely and so I kept writing and writing wondering why my text didn't show up ;-)13:04
apwdiwic, heh i've done that too, one does look a bit of a waz 13:13
=== smb` is now known as smb
ari-tczewhello13:41
ari-tczewmjg59: do we really need separate binary package for smartdimmer?13:42
ari-tczewcan smartdimmer be used with other drivers than nvidia?13:42
ari-tczewhttp://bazaar.launchpad.net/~ubuntu-branches/ubuntu/natty/nvclock/natty/changes?filter_file_id=smartdimmer.install-20090626033400-d3ab2kzykpm31he0-18413:43
=== Quintasan_ is now known as Quintasan
=== sconklin-away is now known as sconklin
apwcommit 9bea7f23952d5948f8e5dfdff4de09bb9981fb5f14:48
apwAuthor: Rusty Russell <rusty@rustcorp.com.au>14:48
apwDate:   Sat Jun 5 11:17:37 2010 -060014:48
apw    module: fix bne2 "gave up waiting for init of module libcrc32c"14:48
sconklinapw:   https://wiki.ubuntu.com/Kernel/KernelBisection15:03
apwsconklin, overall makes sense15:29
sforsheeapw: bug 21580215:49
ubot2Launchpad bug 215802 in linux "rtl8187 link quality poor" [Undecided,Confirmed] https://launchpad.net/bugs/21580215:49
sforsheeapw: http://launchpadlibrarian.net/58772782/rtl8187.patch15:51
=== bjf[afk] is now known as bjf
apwbjf, looks like the CPU count change on maverick has popped the abi on amd64 ... pre-proposed build failed16:16
bjfapw, looking16:17
bjfapw, there's already an abi bump in that batch, i'll dig into it and work it out though16:20
bjfapw, right after the last "start new release" is an "bump abi"16:20
tgardnerbjf, apw: I already bumped the ABI on Maverick.16:21
bjftgardner, apw, i'll look into the problem16:21
tgardnerbjf, no, I'm saying that I already fixed the build failure and re-pushed with the ABI bump. I simply rebased it so taht the build is bisectable16:22
bjftgardner, AH!16:22
bjf##16:22
bjf## Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting16:22
bjf##      agenda: https://wiki.ubuntu.com/KernelTeam/Meeting16:22
bjf##16:22
tgardnerbjf, I'm surprised there isn't an ABI bump required for Lucid. Lemme do a quick build16:23
apwtgardner, ahh good16:48
tgardnerbjf, I'm gonna rebase maverick master-next once again in order t0 catchup with master16:48
bjftgardner, ack16:54
tgardnerbjf, just a quick check on gitweb to make sure I'm not gonna trash something....16:55
bjf##16:57
bjf## Kernel team meeting in 3 minutes16:57
bjf##16:57
JFo<-lunch17:22
diwicapw, this channel works for me17:23
apwdiwic, so i am not aware of there being a work-item gating someone for this docs17:23
apwso where can i find out who is waiting on it17:23
diwicapw, that's beyond my knowledge as I'm not really familiar with the work item infrastructure17:24
apwdiwic, well to be on the release team radar its gonna be in the work items system17:25
apwbut anyhow, who is waiting on it?17:25
diwicapw, well, I'd say alessio / TheMuso17:25
diwicapw, for making the lowlatency flavour17:25
diwicapw, which, as I understand it, they want to have in universe for Natty 17:26
apwwell i am pretty sure allessio has already made one, so doesn't really need any documentation overall, its unlikely they are blocked17:26
diwicapw, are we having information on how to best rebase with the official channels w r t SRU and security?17:27
apwi am not sure i was expecting that to be in the documentation anyhow17:28
apwrebases are all advertised on release in our repos17:28
diwicapw, and I'm not sure alessio's kernel follows all conventions we want in its current state17:28
apwdiwic, no but last time he dropped out when we tried to help him and it got lost in the noise17:29
diwicapw, but a document would help others to help out if one person "drops out"17:29
apwyep, and it is on my todo list, but i'd not heard any real consumers so its not been a priority17:30
=== ogra is now known as Guest34014
apwpoor Guest34014 17:30
diwicapw, ok so my expectations on this document would be instructions for how to get SRU and security updates for a derivative flavour, as well as "convention" instructions17:32
apwthat can be encapsulated in one line however, "follow updates to master"17:33
apwas all cves and sru's go there first17:33
diwicapw, so now you've made half the document ;-) 17:34
apwi guess i am not understadning why this is so hard for people 17:34
diwicapw, can one get notifications when updates to master occur?17:35
diwicapw, or do we have to poll once a day/hour/etc?17:35
tgardnerbjf, have we formally requested regression testing for kvm on the various releases?17:36
apwdiwic, there should be emails out to the kernel-team and installer when they occur in general17:36
bjftgardner, i believe it was discussed but not sure we formally requested such testing17:37
apwdiwic, also they should apper on the per-series changes list i guess17:37
diwicapw, ok, so one can listen to them either manually or automatic (the latter preferred) - I'm not saying it's hard, it's just a lot of things to be aware of17:37
bjftgardner, part of the discussion was that QA was using that for testing rather than real HW and we had an problem with that17:37
apwbjf, i presume you mean 'just kvm' in that context17:38
bjfapw, correct17:38
apwdiwic, i am not against writing it, its just not at the top of my list.  i have a2 to worry about17:38
apwdiwic, i don't think there is much blocing people without it, so its not obviously a priority17:38
apwdiwic, i suspect anyone i would want making kernels ought to be up on how the packaging works well enough17:39
apwdiwic, to figure it out, or we should worry more about other things going wrong.  like linux-libc-dev getting wacked17:39
tgardnerbjf, in this case I'd like to verify that we've not regressed KVM gues support on the host under test.17:39
bjftgardner, sconklin and I will bring it up with QA17:40
tgardnerbjf, reference CVE-2010-3698 as a basis for your discussion17:41
ubot2tgardner: The KVM implementation in the Linux kernel before 2.6.36 does not properly reload the FS and GS segment registers, which allows host OS users to cause a denial of service (host OS crash) via a KVM_RUN ioctl call in conjunction with a modified Local Descriptor Table (LDT). (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-3698)17:41
diwicapw, ok. so I'll try to communicate that this document won't be created within the coming weeks and given that, #ubuntu-kernel and the ML would be helpful in general for any specific questions.17:41
apwdiwic, as you know we are short of bodies right now even for the committed effort17:41
apwdiwic, seems about right yes17:41
diwicapw, ok, thanks.17:43
diwicapw, I'll send something out on the ubuntu-studio-devel ML about it17:43
=== sforshee is now known as sforshee-lunch
=== Guest34014 is now known as ogra_
=== bjf changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Maverick Kernel Version: 2.6.35 || Ubuntu Kernel Team Meeting - February-01 - 17:00 UTC || If you have a question just ask, and do wait around for an answer!
diwichow many bits is a pointer in a 32-bit-pae kernel?18:30
tgardnerdiwic, 3218:32
diwicok, thanks18:32
diwictgardner, so sizeof(buf2 - 8) would evaluate to "4"?18:34
jjohansendiwic: no18:34
jjohansensizeof(buf2) - 8 might depending on what buf2 is18:35
tgardnersizeof(void *) will evaulate to 418:35
diwicjjohansen, so I have discovered a bug where it says sizeof(buf2 - 8), it should say "sizeof(buf2) - 8"18:36
jjohansendiwic: but you need to be careful with sizeof and structs, as C can pad them and return sizes different than you might expect18:36
diwicjjohansen, so the question is what "sizeof(buf2 - 8)" would evaluate to in a 32-bit-pae kernel18:36
tgardnerdiwic, sizeof(buf2 - 8) is definitely a bug18:36
jjohansendiwic: ugh, I am surprised that even compiles18:37
diwicjjohansen, I'm curious if it could be a security thing or if I should just fix it and send upstream18:37
diwicfwiw, buf2 is a char[24]18:37
jjohansendiwic: can you point me at the code?18:38
diwicjjohansen, sound/pci/hda/hda_eld.c:384 18:39
jjohansendiwic: natty?18:39
tgardnerapw, I pushed CONFIG_NR_CPUS=256 for natty amd64 -generic. Am boot testing on hoover.18:39
diwicjjohansen, that line has been in there since 200818:39
diwicjjohansen, in function hdmi_show_short_audio_desc18:40
jjohansendiwic, okay looking18:40
tgardnerdiwic, it doesn't even make sense.18:40
tgardnerit really should be sizeof(buf2)-818:41
diwictgardner, so far we agree :-)18:41
tgardnerdiwic, have you confirmed that it fixes the problem?18:41
diwictgardner, the question is if it could evaluate to something larger than buf218:41
diwictgardner, nope, don't have hw18:42
jjohansendiwic: I don't see how it has any security concerns18:42
tgardnerdiwic, you could write a short test program to see what it does.18:42
diwicjjohansen, good. I'll then just send a patch upstream.18:42
diwictgardner, I did so and it seems to always evaluate to "8" on my amd6418:43
diwictgardner, so I guess it evaluates to a pointer18:43
jjohansendiwic: yep, buf2 - 8 works out as a pointer18:43
tgardnerdiwic, more likely it evaluates to the last constant in the sizeof() expression.18:43
diwictgardner, nope, tried sizeof(buf2-35) which also evaluates to 818:44
tgardnerwell, regardless, that code is busticated18:44
jjohansenright, its an array - const which is identified as pointer arithmetic, so it returns a pointer to -8 from the start of the array and then takes the sizeof the pointer18:45
bjfapw, you reported a bug against the wiki theme, what was the package name that you filed that against ?18:49
tgardnerdiwic, don't forget to Cc: stable@kernel.org on tat patch.18:49
tgardnerin the commit message, not in email to LKML18:50
diwictgardner, agreed18:50
=== sforshee-lunch is now known as sforshee
=== sconklin is now known as sconklin-afk
tjaaltonsmb: hey, re bug 415353, do you mean that the proposed lucid kernel should have the patch you mentioned?19:08
ubot2Launchpad bug 415353 in linux "karmic/lucid installation slow on "detecting network hardware" with bnx2x" [Medium,Fix released] https://launchpad.net/bugs/41535319:08
tjaaltonbecause it's not any different to the earlier ones19:09
smbtjaalton, Hi, well not more differently what you tested last (which is the 90s delay)19:10
smbBut initially there were like hours delay reported19:10
tjaaltonthe installer probes them twice19:10
tjaaltonand the timeout seems to be 100s and not 30 like on an installed system19:11
tjaaltonwhich is weird..19:11
* tgardner --> lunch19:11
smbwell the  100s (is near 90)19:11
smbwhich is trying tree times a 30s timeout19:11
tjaaltonok19:11
tjaaltonI'll try the kernel package19:11
smbtjaalton, much appreciated19:12
=== diwic is now known as diwic_afk
tormodfyi I get warnings about CONFIG_ACPI_PROCFS_POWER when booting the latest .38 mainline snapshot19:18
=== cmagina is now known as cmagina-shovelin
tjaaltonsmb: success! though, there are ugly errors too19:22
tjaaltonsmb: http://users.tkk.fi/~tjaalton/foo/dmesg.txt19:24
smbtjaalton, If you place the dmesg into the bug report, I will have a look tomorrow. Possibly the backport needs some tweaking.19:24
tjaaltonalright, will do19:24
* smb has moved to the beer part of the evening and should not be trusted anymore. :)19:24
* tjaalton will follow shortly19:25
smbtjaalton, but quickly glancing this may tell me that I need to pull in a few more changes from later. I went around some parts which seemed to add a new list of dependency relationships19:28
tjaaltoncool19:29
=== yofel_ is now known as yofel
=== diwic_afk is now known as diwic
=== diwic is now known as diwic_afk
tgardnerapw, my emerald ran 3600 seconds of stress on 2.6.38-1 master-next19:54
=== cmagina-shovelin is now known as cmagina
* jjohansen -> lunch20:10
tgardnerzul, do you remember what xen version we used for Hardy?20:33
zultgardner: barely...it might have been 3.420:34
zultgardner: gimme a sec and i can give you a better answer20:35
zultgardner: 3.320:37
tgardnerzul, maybe 'UBUNTU: build/custom: Hello Xen 3.0.5'20:37
zultgardner: sounds right20:37
=== charlie-tca__ is now known as charlie-tca
tgardnerzul, yep, I don't see any major updates after that. mostly small fixes.20:38
zultgardner: yeah it was when i was volunteering20:38
tgardnerman, that was awhile ago.20:39
zultgardner: it was...i still had hair :)20:46
JFoI just love it when the Launchpad server disappears from under a script 21:24
lifelesswhat happened?21:24
bjfJFo, can you accept nominations ?21:26
lifelessJFo: what happened?21:26
=== diwic_afk is now known as diwic
JFobjf, I can21:28
JFolifeless, I had a script running along nicely and then a huge stack trace basically meaning that the server disappeared21:28
bjfjfo, please accept all my nominations https://bugs.launchpad.net/ubuntu/+source/linux/+bug/70764921:28
ubot2Launchpad bug 707649 in linux "CVE-2010-4079" [Undecided,New]21:28
JFoWe have gotten a lot of help from various people in the #ubuntu-bugs channel. I'd like to thank charlie-tca specifically and several others in there whose nicks escape me at the moment for directing questions about kernel bugs to me. There have also been a ton of requests through irc to look at specific bugs. In those cases, and many like them, I am directing folks to the bug triage pages of our wiki. I hope that some of these folks will become re21:29
JFogular triagers, but that remains to be seen. :)21:29
lifelessJFo: I treat such problems as high severity issues21:29
lifelessJFo: did you get a OOPS in one of the headers? or could you pastebin the backtrace?21:29
ekoorehi to all21:30
JFoactually lifeless hang on a sec21:30
JFoI may be experiencing connectivity issues21:30
=== alex_jon1 is now known as alex_joni
ekoorei have a small problem with the linux  kernel in my device21:31
JFobjf, all accepted21:31
bjfJFo, your too kind21:31
JFolifeless, I have an install running and I told it to download updates as well. I don't normally21:31
JFo:)21:31
JFobjf, it is my pleasure :)21:32
JFolifeless, so I may have accused LP unnecessarily in this case21:32
ekooreJFo can you help me?21:33
lifelessJFo: when LP goes wrong, please do file a bug (even if its a dupe - that guides heat after all)21:33
JFolifeless, I will indeed21:33
lifelessJFo: however there isn't all that much we can do if you fill your net connection :P21:33
JFolet me conduct a test in any case and I will let you know the results via bug if necessary21:33
JFolifeless, :-P 21:33
JFoI wish21:33
JFoekoore, what is the nature of your problem?21:34
lifelessJFo: as  background, we have 2 ssl front ends, 2 ha proxy load balances, and ~ 20 frontend appservers behind those21:34
ekooreis a problemwith an acpi21:34
lifelessJFo: so for that stack to just go awol is a Big Fricking Deal21:34
JFolifeless, I can definitely imagine21:34
lifeless:)21:34
* JFo is about to rip his service provider a new one anyway21:34
JFoekoore, is there a bug open I can look at?21:35
ekooreJFo: not, let me explain21:36
JFolifeless, thanks for inquiring. I can imagine you are quite busy constantly21:36
JFo:)21:36
JFoekoore, ok21:36
ekoorei work in a company: www.ekoore.com21:36
lifelessJFo: I'm kept on my toes, yes :)21:36
JFoheh21:36
ekoorenow i have to prepare a new device21:36
ekooremultitouch with ubuntu21:36
JFook21:37
ekoorethis device have a G-Sensor21:37
JFoG-sensor?21:37
* JFo is unfamiliar21:37
ekooreyes, accelerometr21:37
bjfJFo, G-spot sensor21:37
ekooredo you know?21:37
JFobjf :)21:38
JFoekoore, I see21:38
ekooresensor of gravitation21:38
JFoI am vaguely familiar with those21:38
ekoorethis sensor cause a problem with acpi21:38
JFoI see21:39
ekoorewhen i rotate the tablet, the screen of the tablet power off21:39
ekoorei have a solution:21:39
JForotate as in from landscape to portrait?21:39
ekoorethis comand:21:39
ekooreecho disable > /sys/firmware/acpi/interrupts/gpe1121:40
ekoorei can write this comand in rc.locale file21:40
ekoorebut it is do only to the end of the boot sequence21:41
ekooreis possible disabling this interrupt directly in the device21:41
ekoore?21:41
ekooredirectly in the kernel?21:41
ekooreor to the start of boot sequence?21:42
ekooreJFo, do you think that is possible?21:43
JFothose are very good questions, but I don't know much about acpi events21:43
JFoand the person I would normally ask is away for the day21:44
JFoekoore, do you think you would mind sending the above information to the team mailing list?21:45
JFothat way the appropriate parties will see it and provide advice?21:45
ekooresure, this is not a problem21:45
ekooreinformation about the problem?21:46
ekoorewhere i have to write?21:48
JFoekoore, kernel-team@lists.ubuntu.com21:49
ekooreand they answare me in email? 21:50
JFoyes21:51
ekooreok JFo, thank you21:56
JFomy pleasure ekoore 21:57
JFo:)21:57
ekooreJFo, do you know other person that can help me?21:59
ekooreAfter that i send the email to the maling list21:59
JFoI don't know of anyone who is available in this time zone22:00
ekooreif they don't answare me22:00
JFobut the people I have in mind will see it on the mailing list22:00
ekoorewhere you come frome?22:00
JFowhat do you mean?22:01
ekoorewhere do you come from?22:01
JFoI am in the US22:01
ekooreoh, i am in italy22:02
JFocool22:02
ekoorecalifornia?22:02
JFono, I am in North Carolina22:03
ekooreoh beautifoul22:04
JFoyep, it is nice here22:04
=== cmagina-lunch is now known as cmagina
=== bjf is now known as bjf[afk]

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