| --- | Log | opened Wed Apr 06 00:00:30 2005 |
| 02:15 | --- | ---> AaronWL [aaronwl@2002:44e4:80::dead:1337] has joined #xen |
| 02:39 | --- | <<-- rusty [~rusty@bh02i525f01.au.ibm.com] has quit (Quit: Client exiting) |
| 03:11 | mael | Hi |
| 03:14 | --- | ---> jimix [~jimix@80.187.147.89] has joined #xen |
| 03:38 | --- | ---> athomas [~athomas@ppp-0-96.lond-a-1.access.uk.tiscali.com] has joined #xen |
| 03:49 | --- | <<-- jimix [~jimix@80.187.147.89] has quit (Quit: Download Gaim: http://gaim.sourceforge.net/) |
| 03:50 | --- | ---> jimix [~jimix@80.187.147.89] has joined #xen |
| 06:05 | --- | <<-- jimix [~jimix@80.187.147.89] has quit (Ping timeout: 480 seconds) |
| 06:37 | --- | <--- Sir_Ahzz [~ahzz@c-24-0-215-3.hsd1.tx.comcast.net] has left #xen (Client exiting) |
| 06:37 | --- | ---> Sir_Ahzz [~ahzz@c-24-0-215-3.hsd1.tx.comcast.net] has joined #xen |
| 06:37 | --- | Channel: mode/#xen [+o Sir_Ahzz] by ChanServ |
| 07:34 | --- | ---> jimix [~jimix@212.44.26.183] has joined #xen |
| 07:51 | --- | <<-- jimix [~jimix@212.44.26.183] has quit (Read error: No route to host) |
| 08:49 | --- | ---> rharper [~rharper@pixpat.austin.ibm.com] has joined #xen |
| 09:03 | --- | ---> jimix [~jimix@212.44.26.183] has joined #xen |
| 09:04 | --- | <<-- jimix [~jimix@212.44.26.183] has quit (Quit: ) |
| 09:04 | --- | ---> jimix [~jimix@212.44.26.183] has joined #xen |
| 09:10 | --- | <<-- JViz [Anomaly@cpe-024-167-184-225.triad.res.rr.com] has quit (Ping timeout: 480 seconds) |
| 09:12 | --- | <<-- jimix [~jimix@212.44.26.183] has quit (Ping timeout: 480 seconds) |
| 09:21 | --- | ---> hollis [~hollis@user-0vvde2g.cable.mindspring.com] has joined #xen |
| 09:26 | --- | ---> JViz [Anomaly@cpe-024-167-184-225.triad.res.rr.com] has joined #xen |
| 09:29 | --- | ---> jimix [~jimix@212.44.26.183] has joined #xen |
| 09:35 | --- | <<-- DEac- [~deac@xdsl-81-173-138-85.netcologne.de] has quit (Read error: Operation timed out) |
| 09:37 | --- | <<-- jimix [~jimix@212.44.26.183] has quit (Ping timeout: 480 seconds) |
| 09:40 | --- | ---> hbaum [~hbaum@bi01p1.co.us.ibm.com] has joined #xen |
| 09:49 | --- | ---> DEac- [~deac@xdsl-195-14-216-81.netcologne.de] has joined #xen |
| 09:52 | --- | <<-- plars [~plars@pixpat.austin.ibm.com] has quit (Ping timeout: 480 seconds) |
| 09:53 | --- | ---> jeroney [~jeroney@adsl-68-91-216-6.dsl.austtx.swbell.net] has joined #xen |
| 09:58 | --- | ---> plars [~plars@pixpat.austin.ibm.com] has joined #xen |
| 10:05 | --- | ---> jimix [~jimix@212.44.26.183] has joined #xen |
| 10:08 | jeroney | jimix: how is the xen summit? |
| 10:08 | rharper | hey jimix |
| 10:09 | hollis | BK news: http://kerneltrap.org/node/4966 |
| 10:10 | jeroney | hollis: the news is out, I had Steve French in my office about 3 days ago talking about what should be used nowt that BK is going away |
| 10:11 | hollis | that's funny, because the PR is dated yesterday |
| 10:12 | mael | BK will still be used for a little while, no? |
| 10:14 | jimix | sorry guys.. the network has been unstable in the room |
| 10:14 | jimix | its ok.. still intro tho |
| 10:15 | jimix | discussing 4th generation tools |
| 10:20 | jimix | can't .. keep this .. network..up |
| 10:20 | hollis | :) |
| 10:20 | hollis | mael: dunno |
| 10:21 | mael | well it seems that they will provide a BK version that allow for a slick transition |
| 10:22 | mael | and I don't think that arch/darcs/svn tools are 100% operational for linux kernel dev |
| 10:22 | hollis | mael: so far I don't see anything that says you can't continue to use the free BK you already have |
| 10:22 | mael | yeah true |
| 10:22 | hollis | although that may just be a temporary oversight |
| 10:22 | mael | the only limit seems to be that 64000 changes stuff |
| 10:28 | --- | <<-- jimix [~jimix@212.44.26.183] has quit (Ping timeout: 480 seconds) |
| 10:38 | jonmason | this might be a good opportunity for CVS (or another) to step up |
| 10:39 | hollis | CVS?!? |
| 10:39 | rharper | I don't think anyone wants cvs to step up |
| 10:39 | hollis | you're a funny man :) |
| 10:39 | rharper | even if it could |
| 10:48 | jonmason | What is the best version control app? |
| 10:49 | hollis | what is the best OS? ;) |
| 10:50 | Tv | mael: I just had a darcs pull of linux take 1GB of RAM. |
| 10:50 | Tv | My best hope is bazaar-ng |
| 10:51 | jonmason | hollis: touche |
| 10:55 | mael | Tv: argh |
| 11:02 | --- | <<-- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit (Quit: leaving) |
| 11:02 | --- | ---> soffi [~soffi@Tolvudeild-SC-gw.vv.simnet.is] has joined #xen |
| 11:03 | --- | ---> jimix [~jimix@212.44.26.183] has joined #xen |
| 11:09 | --- | <--- jimix [~jimix@212.44.26.183] has left #xen () |
| 11:12 | @Sir_Ahzz | dang, lots of people in here now. :) |
| 11:12 | @Sir_Ahzz | mooo Robot101 |
| 11:13 | soffi | heh |
| 11:13 | | * soffi is trying eagerly to start a flamewar on xen-users |
| 11:14 | @Sir_Ahzz | 8-P |
| 11:18 | --- | ---> hollis [~hollis@pixpat.austin.ibm.com] has joined #xen |
| 11:25 | Robot101 | Sir_Ahzz: moo |
| 11:27 | @Sir_Ahzz | how's things going for you lately? |
| 11:27 | @Sir_Ahzz | haven't chatted in a while. :) |
| 11:29 | Robot101 | Sir_Ahzz: alright... I'm reading up on Xen and trying it out... trying to think of a cunning PhD proposal :) |
| 11:29 | @Sir_Ahzz | heh |
| 11:30 | @Sir_Ahzz | what area? |
| 11:30 | Robot101 | er, Xen. the more cluster/server oriented, the more likely to attract funding. ;) |
| 11:31 | @Sir_Ahzz | ahhh |
| 11:32 | Tv | Robot101: remote snapshots of machine state, ready to take over should be primary machine fail |
| 11:32 | Tv | Robot101: minimizing the memory requirements of guests, in order to be able to run more of them in one server |
| 11:33 | Robot101 | hot standby has the potential to be very cool, but also very crack |
| 11:34 | Robot101 | the live migration already has provision to copy dirtied pages across the local network whilst stuff is still running |
| 11:34 | Tv | better security to dom0 so you can give clients a shell, and all they can access is their own guest machines |
| 11:34 | Tv | some sort of framebuffer support for guests |
| 11:35 | Robot101 | I already suggested some kind of graphical device abstraction |
| 11:35 | Robot101 | framebuffer's a bit crappy though, something a little higher level like GL primitives might let you presend a 3d graphic card abstraction |
| 11:35 | Robot101 | *present |
| 11:39 | | * knewt would suggest svk for version control |
| 11:41 | Tv | knewt: svk wasn't really distributed, last I looked |
| 11:41 | knewt | Tv: not svn, svk |
| 11:41 | Tv | knewt: when moving stuff between repositories, it batched things together into a single commit |
| 11:42 | Tv | yes, svk |
| 11:42 | knewt | batching or not is a choice |
| 11:43 | knewt | you just tell it to do an incremental merge if you don't want the batching |
| 11:46 | knewt | anyway, i must concentrate on some crouching and some hiding. bbl. |
| 11:49 | --- | ---> matta-lt [~matta@69.93.28.254] has joined #xen |
| 12:01 | --- | <<-- athomas [~athomas@ppp-0-96.lond-a-1.access.uk.tiscali.com] has quit (Quit: Leaving) |
| 12:33 | --- | <<-- matta-lt [~matta@69.93.28.254] has quit (Quit: Hey! Where'd my controlling terminal go?) |
| 12:38 | --- | <<-- plars [~plars@pixpat.austin.ibm.com] has quit (Ping timeout: 480 seconds) |
| 12:57 | --- | <<-- muli_ [~muli@nesher3.haifa.il.ibm.com] has quit (Server closed connection) |
| 12:58 | --- | ---> muli_ [~muli@nesher3.haifa.il.ibm.com] has joined #xen |
| 12:58 | aliguori_ | Robot101: this previrtualization stuff is really interesting.. if you're looking for a PhD topic, that's where I'd start |
| 13:13 | --- | <<-- hollis [~hollis@pixpat.austin.ibm.com] has quit (Quit: Lost terminal) |
| 13:21 | --- | ---> hollis [~hollis@pixpat.austin.ibm.com] has joined #xen |
| 13:31 | --- | <<-- hbaum [~hbaum@bi01p1.co.us.ibm.com] has quit (Quit: Client exiting) |
| 13:49 | --- | <<-- unriel [~riel@nat-pool-bos.redhat.com] has quit (Ping timeout: 480 seconds) |
| 13:51 | --- | ---> tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has joined #xen |
| 13:59 | --- | ---> plars [~plars@pixpat.austin.ibm.com] has joined #xen |
| 13:59 | soffi | my flamewar excercise was a success |
| 14:00 | aliguori_ | uh oh |
| 14:03 | soffi | :-p |
| 14:08 | --- | ---> Arnald [~Arnald@81-86-116-102.dsl.pipex.com] has joined #xen |
| 14:09 | --- | <<-- aliguori_ [~anthony@pixpat.austin.ibm.com] has quit (Quit: Leaving) |
| 14:15 | --- | ---> homebaum [~michael@wbar1.sea1-4-5-031-104.sea1.dsl-verizon.net] has joined #xen |
| 14:16 | hollis | anybody build today's xeno-unstable? I see a build break in audit.c |
| 14:17 | jonmason | hollis: I am running on Mondays w/o any probs |
| 14:17 | soffi | me too |
| 14:17 | hollis | gcc -v ? |
| 14:17 | soffi | from monday |
| 14:17 | jonmason | 3.3.5 |
| 14:18 | jonmason | hollis: gcc version 3.3.5 (Gentoo Linux 3.3.5-r1, ssp-3.3.2-3, pie-8.7.7.1) |
| 14:18 | soffi | gcc version 3.4.2 20041017 (Red Hat 3.4.2-6.fc3) |
| 14:18 | rharper | hollis: gcc version 3.3.5 (Debian 1:3.3.5-5) builds unstable |
| 14:19 | hollis | huh |
| 14:20 | hollis | ah, x86_64/32 mixup |
| 14:20 | hollis | thanks for the sanity check :) |
| 14:20 | jonmason | anytime |
| 14:41 | Robot101 | aliguori: previrtualization? |
| 14:49 | --- | <<-- soffi [~soffi@Tolvudeild-SC-gw.vv.simnet.is] has quit (Quit: Leaving) |
| 14:57 | aliguori | Robot101: the stuff about automated para-virtualization |
| 14:57 | aliguori | modifying the compiler to automatically identifying sensitive instructions and generating virtualization friendly instructions instead |
| 14:57 | aliguori | very cool stuff |
| 14:58 | aliguori | i'm curious to see how Ian et al will react.. if it performs as well as they claim it could eliminate a ton of code in xen |
| 14:58 | --- | <<-- jeroney [~jeroney@adsl-68-91-216-6.dsl.austtx.swbell.net] has quit (Ping timeout: 480 seconds) |
| 14:59 | Robot101 | most of the code in xen seems to be working around architectural limitations in x86 afaict :) |
| 15:01 | Robot101 | what kind of thing is quicker to insert instructions for than trap/hypercall into xen? |
| 15:01 | aliguori | everything. |
| 15:01 | aliguori | that's what makes para-virtualization appealing |
| 15:02 | aliguori | inserting instructions can be done once |
| 15:02 | aliguori | at boot time |
| 15:02 | aliguori | and perhaps even cache |
| 15:02 | aliguori | cached* |
| 15:02 | aliguori | a lot of the nasty instructions (like popf, cpuid, etc.) can be modified once and no trap is necessary at all |
| 15:03 | aliguori | i think they have a mechanism to batch the mmio and page table writes.. but it's hard to tell until they release more info |
| 15:03 | aliguori | basically, it sounds lot like what xen does today but just all automated |
| 15:11 | Robot101 | do you have an URL or something? |
| 15:12 | Robot101 | Xen *is* autpmated - doing code rewriting at runtime is horrible - the hoops VMWare has to jump through... |
| 15:12 | knewt | hmm, not that i think it'd be a good idea, but i wonder what they'd do about a kernel module that rewrites itself at runtime? |
| 15:15 | Robot101 | the whole point of the xen stuff is to try and avoid making modifications beyond the OS, including crazy code rewriting |
| 15:15 | --- | ---> jeroney [~jeroney@pixpat.austin.ibm.com] has joined #xen |
| 15:19 | aliguori | Robot101: no url yet, information is still sketchy |
| 15:19 | aliguori | Robot101: i think the basic idea though is to automatically do what Xen does by hand in the compiler |
| 15:20 | aliguori | it's not rewriting stuff at runtime like vmware.. it's quite reasonable to do in the compiler |
| 15:20 | Robot101 | yeah, you said boot time/caching though, which scared me :) |
| 15:20 | aliguori | oh |
| 15:21 | aliguori | yeah, so, they've suggested identifying the spots that need to be changed, padding them with nops, and then patches those things at boot time |
| 15:21 | Robot101 | there is still stuff that you must have hypercalls for |
| 15:21 | aliguori | so that different hypervisors can patch in different ways |
| 15:21 | aliguori | sure |
| 15:21 | aliguori | and you can still have them :-) |
| 15:21 | aliguori | it just removes a lot of the tediousness of porting an OS to xen |
| 15:21 | Robot101 | so my question is what currently eg traps that can be dealt with without a trap/hypercall? or is this about replacing traps with hypercalls? |
| 15:22 | aliguori | ok, so, nothing is trapped in Xen today |
| 15:22 | aliguori | b/c Xen modifies that stuff by hand |
| 15:22 | aliguori | well, that's not true, but imagine it is |
| 15:22 | aliguori | but consider something like VMX |
| 15:22 | aliguori | in VMX, you have to register a trap for the CPUID instruction |
| 15:23 | aliguori | b/c you can use CPUID to figure out what ring your in |
| 15:23 | aliguori | so, you can either trap it and emulate it |
| 15:23 | aliguori | or, if you're really smart |
| 15:23 | aliguori | you can replace any instances of cpuid with cpuid that ands the results with the proper mask to hide the privilege levels |
| 15:24 | Robot101 | hrm |
| 15:24 | aliguori | a lot of the 17 sensitive instructions can be handled that way |
| 15:24 | aliguori | some of the stuff, like sgdt can be handled nicely if you make a pointer to the domain's gdt available in the shared xen-space |
| 15:25 | aliguori | the upper 64MB of memory |
| 15:25 | aliguori | so you can replace sgdt ops with movs from that well-known address |
| 15:25 | aliguori | i'm not sure if that's what they do, but i think that's the idea |
| 15:35 | --- | ---> niv [~nivedita@bi01p1.co.us.ibm.com] has joined #xen |
| 15:59 | --- | <<-- joink [~joink@186.80-202-132.nextgentel.com] has quit (Server closed connection) |
| 15:59 | --- | ---> joink [~joink@186.80-202-132.nextgentel.com] has joined #xen |
| 16:07 | --- | <<-- JViz [Anomaly@cpe-024-167-184-225.triad.res.rr.com] has quit (Read error: Connection reset by peer) |
| 16:07 | --- | ---> JViz` [Anomaly@cpe-024-167-184-225.triad.res.rr.com] has joined #xen |
| 16:12 | --- | <<-- matta [~matta@69.93.28.254] has quit (Ping timeout: 480 seconds) |
| 16:13 | --- | ---> matta [~matta@69.93.28.254] has joined #xen |
| 16:18 | --- | <<-- jeroney [~jeroney@pixpat.austin.ibm.com] has quit (Quit: Leaving) |
| 16:21 | --- | <<-- matta [~matta@69.93.28.254] has quit (Ping timeout: 480 seconds) |
| 16:23 | --- | ---> matta [~matta@69.93.28.254] has joined #xen |
| 16:29 | --- | ---> matta-lt [~matta@69.93.28.254] has joined #xen |
| 16:30 | --- | <<-- cf0989b8 [~cf0989b8@ns.372broadway.com] has quit (Ping timeout: 480 seconds) |
| 16:41 | --- | <<-- matta-lt [~matta@69.93.28.254] has quit (Read error: Operation timed out) |
| 16:42 | --- | <<-- matta [~matta@69.93.28.254] has quit (Ping timeout: 480 seconds) |
| 17:05 | knewt | heh, i see Ted got himself into a huff |
| 17:08 | --- | ---> cf0989b8 [~cf0989b8@ns.372broadway.com] has joined #xen |
| 17:24 | --- | ---> jeroney [~jeroney@adsl-68-91-216-6.dsl.austtx.swbell.net] has joined #xen |
| 17:45 | --- | <<-- hollis [~hollis@pixpat.austin.ibm.com] has quit (Quit: leaving) |
| 17:47 | --- | ---> Bluefox [~bluefox@pcp484971pcs.whtmrs01.md.comcast.net] has joined #xen |
| 17:51 | --- | <<-- Nigelenki [~bluefox@pcp484971pcs.whtmrs01.md.comcast.net] has quit (Ping timeout: 480 seconds) |
| 18:12 | | * niv thought it was just spam composers who talked like that... |
| 18:20 | --- | ---> visik7 [~ciao@host155-3.pool80183.interbusiness.it] has joined #xen |
| 18:20 | visik7 | wonderful :) |
| 18:21 | visik7 | the users of this channel are triples from the last time I visit it |
| 18:24 | jonmason | visik7: that's a good thing, yes? |
| 18:24 | visik7 | yes |
| 18:27 | visik7 | I'm trying to compile the unstable tree but I've some problems :/ |
| 18:28 | jonmason | visik7: what's happening? |
| 18:28 | visik7 | compiling errors |
| 18:28 | jonmason | specifically? |
| 18:28 | visik7 | oh it's some time ago |
| 18:28 | visik7 | after the merge of freebsd domU tree |
| 18:28 | visik7 | -spare |
| 18:29 | jonmason | I am running unstable, and it seems to be okay |
| 18:30 | visik7 | mmm |
| 18:30 | visik7 | maybe in the w-end I'll try to recompile it |
| 18:39 | --- | <<-- rharper [~rharper@pixpat.austin.ibm.com] has quit (Quit: Leaving) |
| 18:48 | --- | <<-- jeroney [~jeroney@adsl-68-91-216-6.dsl.austtx.swbell.net] has quit (Quit: Leaving) |
| 18:49 | visik7 | jonmason how much Unstable is nstable ? |
| 18:49 | visik7 | s/nstable/unstable |
| 18:50 | jonmason | visik7: I'm not pushing it too hard (aside from networking tests) buit I haven't had any problems (aside from a build problem that I think was my fault) |
| 18:51 | visik7 | mmm |
| 18:51 | visik7 | Btw I notice that using the crashme program with a xen kernel is simplier to crash a system than with a non xen kernel |
| 18:52 | jonmason | not too familuar with crashme |
| 18:52 | jonmason | maybe that is a fun thing to investigate....for you |
| 18:52 | visik7 | it's a program that generate a rondom code and execute it |
| 18:53 | jonmason | with the puropse of thing to crash the kernel? |
| 18:53 | jonmason | s/thing/trying/g |
| 18:53 | visik7 | yes and mostly it succed |
| 18:53 | jonmason | then it is a good program :-) |
| 18:54 | visik7 | :) |
| 18:54 | visik7 | crashme generates strings of random bytes and then attempts to execute |
| 18:54 | visik7 | them. Used to test kernel stability. |
| 18:57 | jonmason | this channel gets quiet at night |
| 18:58 | --- | ---> hollis [~hollis@user-0vvde2g.cable.mindspring.com] has joined #xen |
| 18:58 | jonmason | hollis: did you ever sort out that unstable prob? |
| 18:59 | hollis | yes, I guess it was just an x86-64 problem |
| 18:59 | hollis | which I don't care about, so... |
| 19:02 | jonmason | lol |
| 19:02 | jonmason | make jerone work it |
| 19:03 | hollis | shrug, or the Cambridge guys. there are more people doing x86-64 than PPC |
| 19:03 | jonmason | naa, there is what...5 of yoyu |
| 19:03 | jonmason | err...you |
| 19:04 | hollis | depends on how you count I guess |
| 19:04 | jonmason | with my fingers for anything less than 10 |
| 19:04 | jonmason | after that, the shoes come off |
| 19:14 | aliguori | jerone's said that there's files missing from the x86-64 port... |
| 19:15 | aliguori | i haven't tried compiling it myself.. knowing hollis though, i'd expect the ppc port to work before the x86-64 port :-) |
| 19:15 | hollis | it wasn't a missing file, it was an int/long mismatch in audit.c |
| 19:15 | aliguori | so x86-64 compiles with trivial change? |
| 19:16 | hollis | shrug |
| 19:18 | aliguori | hm |
| 19:19 | aliguori | perhaps i'll give it a try tomorrow |
| 19:41 | --- | <--- niv [~nivedita@bi01p1.co.us.ibm.com] has left #xen (Leaving) |
| 20:07 | --- | <<-- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit (Quit: leaving) |
| 20:18 | --- | <<-- drbyte [~byte@c210-49-121-44.eburwd3.vic.optusnet.com.au] has quit (Ping timeout: 480 seconds) |
| 20:18 | --- | ---> nextime [~nextime@213-140-22-64.fastres.net] has joined #xen |
| 20:29 | --- | ---> drbyte [~byte@203-206-88-160.dyn.iinet.net.au] has joined #xen |
| 20:30 | --- | ---> rusty [~rusty@bh02i525f01.au.ibm.com] has joined #xen |
| 20:56 | --- | <<-- tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has quit (Quit: Everybody wants to go to heaven, but nobody wants to die.) |
| 21:27 | --- | <<-- dst [~dst@p4b23e3d4.np.schlund.de] has quit (Server closed connection) |
| 21:27 | --- | ---> dst [~dst@p4b23e3d4.np.schlund.de] has joined #xen |
| 21:43 | --- | <<-- DEac- [~deac@xdsl-195-14-216-81.netcologne.de] has quit (Ping timeout: 480 seconds) |
| 21:50 | --- | <<-- homebaum [~michael@wbar1.sea1-4-5-031-104.sea1.dsl-verizon.net] has quit (Quit: Client exiting) |
| 21:55 | --- | ---> DEac- [~deac@xdsl-84-44-148-71.netcologne.de] has joined #xen |
| 22:26 | --- | <<-- buggs [~noidentd@n80-237-228-135.cnet.hosteurope.de] has quit (Server closed connection) |
| 22:26 | --- | ---> buggs [~noidentd@n80-237-228-135.cnet.hosteurope.de] has joined #xen |
| 23:21 | --- | <<-- AaronWL [aaronwl@2002:44e4:80::dead:1337] has quit (Quit: I love you.) |
| --- | Log | closed Thu Apr 07 00:00:09 2005 |
![[FSF Associate Member]](/assets/fsf.png){kind=small}
