| --- | Log | opened Thu Jan 31 00:00:06 2008 |
| --- | Day | changed Thu Jan 31 2008 |
| 00:00 | -!- | jdike [~jdike@pool-72-93-84-53.bstnma.fios.verizon.net] has quit [Quit: Leaving] |
| 00:32 | -!- | Syivester [~root@59.172.73.92] has joined #uml |
| 00:33 | <Syivester:#uml> | hello pals |
| 00:33 | <Syivester:#uml> | anybody has tried to install UML from source? |
| 00:36 | <Syivester:#uml> | I had downloaded the 2.6.16 source tree, but no success on build it for UML by the instructions posted on the website |
| 00:38 | <Syivester:#uml> | my host machine runs FC7, even uml_utilities can not make successfully. Anyone can tell me what's wrong? |
| 00:53 | -!- | baroni [baroni@189-041-221-076.xd-dynamic.ctbcnetsuper.com.br] has quit [Ping timeout: 480 seconds] |
| 00:53 | -!- | Syivester [~root@59.172.73.92] has quit [Quit: Leaving.] |
| 00:58 | -!- | balbir [~balbir@59.145.136.1] has joined #uml |
| 02:55 | -!- | Urgleflogue [~plamen@83.228.65.158] has joined #uml |
| 05:05 | -!- | kokoko1 [~Slacker@203.148.64.252] has joined #uml |
| 05:29 | -!- | balbir [~balbir@59.145.136.1] has quit [Ping timeout: 480 seconds] |
| 06:19 | -!- | Blissex [~Blissex@host217-43-110-171.range217-43.btcentralplus.com] has joined #uml |
| 07:44 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has joined #uml |
| 08:00 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has quit [Quit: Quitte] |
| 08:38 | -!- | Magotari [~karol@chello089076077063.chello.pl] has quit [Ping timeout: 480 seconds] |
| 08:41 | -!- | dang [~dang@nemesis.fprintf.net] has quit [Quit: Leaving.] |
| 08:48 | -!- | SNy [6c36a0fdc1@bmx-chemnitz.de] has quit [Server closed connection] |
| 08:48 | -!- | SNy [d5f54ff054@bmx-chemnitz.de] has joined #uml |
| 09:20 | -!- | Magotari [~karol@chello089076077063.chello.pl] has joined #uml |
| 10:02 | -!- | dang [~dang@aa-redwall.ghs.com] has joined #uml |
| 10:05 | <kokoko1:#uml> | hi? |
| 10:23 | <dang:#uml> | Morning, kokoko1 |
| 10:25 | <kokoko1:#uml> | facing this problem ---> http://www.guardian.co.uk/technology/2008/jan/31/internet.blackout.asia?gusrc=rss&feed=networkfront |
| 10:26 | <kokoko1:#uml> | net is damn slow |
| 10:26 | <dang:#uml> | Wow, that sucks. |
| 10:27 | <kokoko1:#uml> | dang, you are at work ? |
| 10:27 | <dang:#uml> | Yes. |
| 10:27 | <kokoko1:#uml> | i am sure you got fat ass pipes in there ;) |
| 10:27 | <kokoko1:#uml> | oracle? |
| 10:28 | <dang:#uml> | No... |
| 10:28 | <dang:#uml> | I work for Green Hills. http://www.ghs.com |
| 10:29 | <dang:#uml> | Decent pipes, tho, as such things go. |
| 10:31 | <kokoko1:#uml> | oh thought you works for oracle, |
| 10:31 | <kokoko1:#uml> | i am working as sysadmins (telecommute) for canadian company. |
| 10:32 | <dang:#uml> | Must be hard to telecommute if your internet is slow... |
| 10:32 | <kokoko1:#uml> | true, seven my ssh sessions sucks atm :( |
| 10:32 | <kokoko1:#uml> | s/seven/even |
| 10:33 | <kokoko1:#uml> | cool greenhills big clients |
| 10:33 | <kokoko1:#uml> | have big* |
| 10:34 | <dang:#uml> | Yep. |
| 10:34 | -!- | Syivester [~root@58.49.28.44] has joined #uml |
| 10:34 | <dang:#uml> | It's a good company to work for. |
| 10:35 | <kokoko1:#uml> | yeah looks big .. |
| 10:35 | <kokoko1:#uml> | i work for https://affordableprogrammers.com/ |
| 10:36 | <Syivester:#uml> | Does anybody tried to build UML from 2.6.16 source tree successfully? |
| 10:36 | <dang:#uml> | 2.6.16? Not in years... |
| 10:36 | <Syivester:#uml> | is there new source tree? |
| 10:36 | <dang:#uml> | I think I may have used 2.6.16, a while ago, but it would have been based on -bb, i think. |
| 10:37 | <dang:#uml> | Heck yeah. 2.6.24 just came out. |
| 10:37 | <Syivester:#uml> | I just failed when run the command "make ARCH=um" |
| 10:37 | <kokoko1:#uml> | Syivester, why not to try sometime new? |
| 10:37 | <Syivester:#uml> | 2.6.16 |
| 10:38 | <kokoko1:#uml> | something* |
| 10:38 | <Syivester:#uml> | ok |
| 10:38 | <dang:#uml> | Syivester: Stock 2.6.16? Or with one of the UML patchsets applied? |
| 10:38 | <dang:#uml> | (ie, where did you get it?) |
| 10:38 | <Syivester:#uml> | from the kernel.org? |
| 10:38 | <dang:#uml> | Okay, so stock. I'd try a recent kernel. |
| 10:39 | <Syivester:#uml> | while I got 2.6.16 kernel as it is instructed by the UML homepage |
| 10:39 | <dang:#uml> | The older ones had problems, and you generally needed to patch them. |
| 10:39 | <dang:#uml> | That's a bit out of date. :) |
| 10:39 | <Syivester:#uml> | really?! as my host machine runs FC7, i also tried FC7's kernel source tree |
| 10:39 | <Syivester:#uml> | 2.6.21-1.3194.fc7 , also failed |
| 10:40 | <Syivester:#uml> | :( |
| 10:40 | <dang:#uml> | The most recent patchset for 2.6.16 was from Sept 2006 |
| 10:40 | <Syivester:#uml> | where can i get the patch set for UML in 2.6.16? |
| 10:40 | <dang:#uml> | Try 2.6.24 from kernel.org |
| 10:40 | <dang:#uml> | I'm running that fine. |
| 10:40 | <Syivester:#uml> | ok |
| 10:40 | <dang:#uml> | http://www.user-mode-linux.org/~blaisorblade/patches/guest/ |
| 10:41 | <dang:#uml> | That has patchsets for older kernels. |
| 10:41 | <Syivester:#uml> | ok, thanks! |
| 10:42 | <Syivester:#uml> | what does bb1 and bb2 means? |
| 10:42 | <kokoko1:#uml> | dang, 2.6.24 on guest right? |
| 10:42 | <dang:#uml> | Yes. |
| 10:42 | <kokoko1:#uml> | how about host did you patch it with skas+ ? |
| 10:43 | <dang:#uml> | I currently have 2.6.24 on one of my hosts, and 2.6.23.something on the other. |
| 10:43 | <dang:#uml> | No. |
| 10:43 | -!- | Blissex [~Blissex@host217-43-110-171.range217-43.btcentralplus.com] has quit [Remote host closed the connection] |
| 10:43 | <dang:#uml> | I tried that, but I'm on 64-bit hosts, and need 32-bit guests, and that's not working yet. |
| 10:43 | <dang:#uml> | (or wasn't when I tried it, at least...) |
| 10:43 | <dang:#uml> | (I've been offline for 2 days, maybe Jeff fixed it...) |
| 10:43 | <kokoko1:#uml> | right |
| 10:46 | -!- | jdike [~jdike@pool-72-93-84-53.bstnma.fios.verizon.net] has joined #uml |
| 10:46 | <jdike:#uml> | Hi guys |
| 10:47 | <kokoko1:#uml> | hi jdike |
| 10:49 | -!- | the_hydra [~driver_bu@61.247.7.140] has joined #uml |
| 10:49 | <the_hydra:#uml> | hi jeff |
| 10:49 | -!- | the_hydra [~driver_bu@61.247.7.140] has quit [] |
| 10:52 | <Syivester:#uml> | i patched uml-2.6.16-bs2.patch.bz2 and uml-2.6.16-bs2-broken-out.tar.bz2 by run "bunzip2 *.bz2 | patch -p0" in the 2.6.16 kernel source tree, but still can make ARCH=um successfully! |
| 10:52 | -!- | balbir [~balbir@122.167.176.218] has joined #uml |
| 10:52 | <Syivester:#uml> | why? :( |
| 10:52 | <Syivester:#uml> | may i need apply the bb1 patches to the source tree also? |
| 10:54 | <dang:#uml> | Syivester: Did you make ARCH=um menuconfig on a completely clean tree? |
| 10:54 | <dang:#uml> | (or config...) |
| 10:54 | <Syivester:#uml> | yes |
| 10:54 | <Syivester:#uml> | definitely |
| 10:54 | <jdike:#uml> | do you absolutely need 2.6.16? |
| 10:55 | <dang:#uml> | That's what I suggested... |
| 10:55 | <Syivester:#uml> | no, i just follow the instructions posted on the UML homepage |
| 10:55 | <Syivester:#uml> | only two commands are executed. 1) make defconfig ARCH=um 2) make ARCH=um |
| 10:56 | <jdike:#uml> | it does say |
| 10:56 | <jdike:#uml> | You need to start by getting a kernel tree (generally, the more recent, the better) |
| 10:56 | <Syivester:#uml> | on clean 2.6.16 source tree downloaded according to the instructions on the homepage |
| 10:56 | <jdike:#uml> | and linux-2.6.16 in this case |
| 10:56 | <jdike:#uml> | i.e. that was just an example |
| 10:56 | <Syivester:#uml> | .....oh, i c |
| 10:56 | <jdike:#uml> | "generally, the more recent, the better" |
| 10:57 | <jdike:#uml> | 2.6.16 isn't that recent |
| 10:57 | <Syivester:#uml> | it still needs patches for UML, right? |
| 10:57 | <jdike:#uml> | no |
| 10:57 | <Syivester:#uml> | ok, i will try to download 2.6.24 and try it later |
| 10:58 | <jdike:#uml> | 2.6.24 is good |
| 10:58 | <Syivester:#uml> | thanks! |
| 10:59 | <Syivester:#uml> | the two commands are right? i mean 1) make defconfig ARCH=um 2) make ARCH=um |
| 10:59 | <dang:#uml> | If you want default settings, yes. |
| 11:00 | <dang:#uml> | Do make distclean first... |
| 11:00 | <dang:#uml> | (no ARCH=um) |
| 11:00 | <Syivester:#uml> | ok |
| 11:05 | -!- | Syivester [~root@58.49.28.44] has left #uml [] |
| 11:31 | -!- | dang [~dang@aa-redwall.ghs.com] has quit [Ping timeout: 480 seconds] |
| 11:31 | <kokoko1:#uml> | any idea what could be reason if uml do not execute /etc/rc.local |
| 11:32 | <kokoko1:#uml> | we have the problem on one of uml and it do not execuate our iptables script when uml reboot |
| 11:33 | <kokoko1:#uml> | lrwxrwxrwx 1 root root 13 2008-01-29 16:54 /etc/rc.local -> rc.d/rc.local |
| 11:33 | <kokoko1:#uml> | iptables script is xecutable |
| 11:34 | <kokoko1:#uml> | fedora behave differently on some hosts |
| 11:34 | <jdike:#uml> | whatever's supposed to run it, doesn't |
| 11:37 | -!- | dang [~dang@aa-redwall.ghs.com] has joined #uml |
| 11:38 | <Magotari:#uml> | jdike: Thought I'd say, I'm finally getting back into working mood. I'll start helping on the channel in a few days, and working on that multiple console too. Finally getting useful again, I hope. |
| 11:53 | <jdike:#uml> | cool |
| 12:05 | <Magotari:#uml> | Mostly I had to pick myself up from the ground after developing a nice idea for a few days, then reading up on what a VDSO is and quickly becoming very very sad. |
| 12:07 | <jdike:#uml> | hehe |
| 12:07 | <jdike:#uml> | what was the idea? |
| 12:09 | <Magotari:#uml> | Well... |
| 12:09 | <Magotari:#uml> | Syscalls are expensive right? |
| 12:09 | <Magotari:#uml> | But what if we put a bit of code in memory... |
| 12:09 | <Magotari:#uml> | The bit of code would jump to the kernel. |
| 12:09 | <Magotari:#uml> | Then the kernel would do a syscall. |
| 12:09 | <Magotari:#uml> | But we are avoding doing an interrupt. |
| 12:10 | <Magotari:#uml> | Now, this is important for uml. |
| 12:10 | <jdike:#uml> | you need a mode change, from usermode to kernel mode |
| 12:10 | <Magotari:#uml> | To do a 'hypocall'. |
| 12:10 | <Magotari:#uml> | Thus avoiding the host kernel for some tasks. |
| 12:10 | <jdike:#uml> | otherwise the process can execute system calls on the host itself |
| 12:11 | <Magotari:#uml> | Yes I had a solution for it, don't remember it anymore thought. |
| 12:12 | <Magotari:#uml> | The other idea was to place a shared bit of memory, the kernel would check it. If it were set, the kernel would check a few more shared bits of code and do a syscall with them. |
| 12:12 | <Magotari:#uml> | Apalling Polling. |
| 12:13 | <Magotari:#uml> | Then I read about VDSOs. Pretty much my idea. |
| 12:15 | <jdike:#uml> | better to have a good idea that's already been done than to have an idea that no one has done because it sucks |
| 12:15 | <Magotari:#uml> | Ah yes, I remember. The bit of memory would be execute only. After the execution would be placed at the area, it would jump to a different area. Because we are not readable, we can't read it and jump straight to the second area. The second area escalates to the kernel. |
| 12:16 | <Magotari:#uml> | Of course we have an area with leads right to the kernel in the memory, even if we can't find it easily... Security problem. |
| 12:16 | <jdike:#uml> | yup |
| 12:16 | <jdike:#uml> | someone will find it |
| 12:16 | <Magotari:#uml> | Jeff, I have lots of such ideas. That is my problem. First fat32, then slirp, then VDSOs. |
| 12:16 | <Magotari:#uml> | I also wrote my own pidof, because I did not know it exists... |
| 12:16 | <jdike:#uml> | also execute permission implies read permission on lots of CPUs |
| 12:17 | <Magotari:#uml> | I keep doing it to myself, hardly motivating that. |
| 12:17 | <Magotari:#uml> | Whoever does not know UNIX is destinied to reimplement it, poorly. |
| 12:19 | <Magotari:#uml> | An yes. An idea for the security hole. The first area could also place a special value somewhere into memory, the second area would not execute without that value set. Again, this relies on execute only. |
| 12:19 | <Magotari:#uml> | By now it's probably cheaper to just int 80. |
| 12:22 | <jdike:#uml> | this is still security through obscurity |
| 12:22 | <Magotari:#uml> | Yeah, that is true. Good point. Useless security. |
| 12:23 | <jdike:#uml> | nothing prevents an attacker from implementing his own first area |
| 12:23 | <jdike:#uml> | and the second area can't tell the difference |
| 12:24 | <Magotari:#uml> | Yeah, here I yield. (for now) |
| 12:24 | <jdike:#uml> | hehe |
| 12:32 | <Magotari:#uml> | Bad idea: One could make a program execute a series of nop instructions. For example cmov (eax, eax); nop; cmov (ebx, ebx);. The chance that this will occur in nature is close to zero. An occurence of such a string of instructions would instruct the kernel to read out the syscall data from a shared place in memory. I think it is obivous why this is a baaaad idea. |
| 12:39 | -!- | Urgleflogue [~plamen@83.228.65.158] has quit [Ping timeout: 480 seconds] |
| 12:42 | <Magotari:#uml> | I give up for good. I'm just getting closer and closer to VDSOs. (A bit of read-only code in memory which jumps to the kernel, then kernel escalates, reads data from memory and makes a syscall out of that.) |
| 12:46 | -!- | kos_tom [~thomas@76.Red-80-33-152.staticIP.rima-tde.net] has joined #uml |
| 12:51 | -!- | ram [~ram@bi01p1.co.us.ibm.com] has joined #uml |
| 12:57 | <Magotari:#uml> | jdike: Can you recommend any good resources on x86 or computer architerctures in general? I'm looking for a possible solution to this without using VDSOs. Even thought it most surely will be a much worse concept, I can't quite let it go. |
| 12:58 | <jdike:#uml> | Like a spec? |
| 12:58 | <jdike:#uml> | Intel has lots of them |
| 12:58 | <jdike:#uml> | dunno what to recommend wrt a book |
| 12:58 | <jdike:#uml> | ask on #kernelnewbies |
| 13:05 | <Magotari:#uml> | I think Intel specs will be good enough. |
| 13:05 | * | Magotari:#uml goes back to reading on rings. |
| 13:07 | -!- | ram_ [~ram@bi01p1.co.us.ibm.com] has joined #uml |
| 13:09 | <jdike:#uml> | http://www.intel.com/products/processor/manuals/ |
| 13:09 | <jdike:#uml> | Software Developer's Manual |
| 13:09 | -!- | ram [~ram@bi01p1.co.us.ibm.com] has quit [Read error: Connection reset by peer] |
| 13:10 | <Magotari:#uml> | Thanks a lot, Jeff. |
| 13:16 | -!- | kos_tom [~thomas@76.Red-80-33-152.staticIP.rima-tde.net] has quit [Ping timeout: 480 seconds] |
| 13:22 | -!- | dang [~dang@aa-redwall.ghs.com] has quit [Ping timeout: 480 seconds] |
| 13:51 | -!- | Urgleflogue [~plamen@83.228.65.158] has joined #uml |
| 14:53 | -!- | dang [~dang@aa-redwall.ghs.com] has joined #uml |
| 15:14 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has joined #uml |
| 16:00 | -!- | hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has joined #uml |
| 16:03 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has quit [Quit: Quitte] |
| 16:26 | -!- | dang [~dang@aa-redwall.ghs.com] has quit [Quit: Leaving.] |
| 17:10 | -!- | diatoid [~diatoid@c-67-181-137-73.hsd1.ca.comcast.net] has joined #uml |
| 18:15 | -!- | ElectricElf [~dbharris@bas1-toronto48-1279276148.dsl.bell.ca] has quit [Remote host closed the connection] |
| 18:15 | -!- | ElectricElf [~dbharris@bas1-toronto48-1279276148.dsl.bell.ca] has joined #uml |
| 18:49 | -!- | hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has quit [Quit: Leaving] |
| 19:00 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has joined #uml |
| 19:09 | -!- | diatoid [~diatoid@c-67-181-137-73.hsd1.ca.comcast.net] has left #uml [] |
| 19:15 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has quit [Quit: Quitte] |
| 20:10 | -!- | kokoko1 [~Slacker@203.148.64.252] has quit [Ping timeout: 480 seconds] |
| 20:44 | -!- | jdike [~jdike@pool-72-93-84-53.bstnma.fios.verizon.net] has quit [Quit: Leaving] |
| 20:56 | -!- | ram_ [~ram@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds] |
| 21:27 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has joined #uml |
| 21:58 | -!- | acklen [~kno@sdith.zotikos.com] has quit [Ping timeout: 480 seconds] |
| 22:02 | -!- | Infinito [argos@201-2-46-200.gnace701.dsl.brasiltelecom.net.br] has quit [Quit: Quitte] |
| 22:45 | -!- | dang [~dang@nemesis.fprintf.net] has joined #uml |
| 23:13 | -!- | balbir [~balbir@122.167.176.218] has quit [Ping timeout: 480 seconds] |
| 23:59 | -!- | VS_ChanLog [~stats@ns.theshore.net] has left #uml [Rotating Logs] |
| 23:59 | -!- | VS_ChanLog [~stats@ns.theshore.net] has joined #uml |
| --- | Log | closed Fri Feb 01 00:00:01 2008 |
![[FSF Associate Member]](/assets/fsf.png){kind=small}
