| --- | Log | opened Fri May 20 00:00:24 2005 |
| 00:52 | --- | ---> Shaun [ndci@ip68-111-70-41.oc.oc.cox.net] has joined #xen |
| 01:03 | --- | <<-- soffi [~soffi@proxy.du.vdsl.is] has quit (Quit: Leaving) |
| 01:16 | --- | ---> soffi [~soffi@Tolvudeild-SC-gw.vv.simnet.is] has joined #xen |
| 01:33 | --- | <<-- lilo [~lilo@lilo.usercloak.oftc.net] has quit (Quit: brb again 8) |
| 01:41 | --- | ---> lilo [~lilo@lilo.usercloak.oftc.net] has joined #xen |
| 01:51 | --- | ---> franR [~franr@inextensia.net] has joined #xen |
| 02:32 | --- | <<-- rusty [~rusty@bh02i525f01.au.ibm.com] has quit (Quit: Client exiting) |
| 02:48 | --- | ---> cfreak [cfreak@dsl-084-056-104-177.arcor-ip.net] has joined #xen |
| 03:50 | --- | ---> athomas [~athomas@ppp-0-55.lond-b-2.access.uk.tiscali.com] has joined #xen |
| 04:08 | --- | ---> Beaky [~chatzilla@ncg-adsl.demon.co.uk] has joined #xen |
| 06:41 | --- | ---> monrad [~monrad@213083190130.sonofon.dk] has joined #xen |
| 07:23 | --- | <<-- monrad [~monrad@213083190130.sonofon.dk] has quit (Read error: Operation timed out) |
| 07:34 | --- | <<-- Dougie_ [~Doug@shade.idmf.net] has quit (Remote host closed the connection) |
| 07:39 | --- | ---> monrad [~monrad@213083190130.sonofon.dk] has joined #xen |
| 07:40 | --- | <<-- DEac- [~deac@xdsl-213-196-197-104.netcologne.de] has quit (Ping timeout: 480 seconds) |
| 07:51 | --- | ---> Dougie_ [~Doug@shade.idmf.net] has joined #xen |
| 07:51 | --- | ---> DEac- [~deac@xdsl-213-196-200-152.netcologne.de] has joined #xen |
| 07:54 | --- | <<-- Beaky [~chatzilla@ncg-adsl.demon.co.uk] has quit (Remote host closed the connection) |
| 08:05 | --- | <<-- lilo [~lilo@lilo.usercloak.oftc.net] has quit (Remote host closed the connection) |
| 08:06 | --- | ---> lilo [~lilo@lilo.usercloak.oftc.net] has joined #xen |
| 08:58 | --- | ---> rharper [~rharper@pixpat.austin.ibm.com] has joined #xen |
| 09:05 | --- | <<-- plars [~plars@pixpat.austin.ibm.com] has quit (Ping timeout: 480 seconds) |
| 09:14 | --- | ---> plars [~plars@pixpat.austin.ibm.com] has joined #xen |
| 09:19 | --- | ---> hollis [~hollis@user-0vvde2g.cable.mindspring.com] has joined #xen |
| 09:30 | --- | <<-- homebaum [~michael@wbar1.sea1-4-5-031-104.sea1.dsl-verizon.net] has quit (Remote host closed the connection) |
| 09:35 | --- | ---> homebaum [~michael@wbar1.sea1-4-5-031-104.sea1.dsl-verizon.net] has joined #xen |
| 09:43 | --- | ---> aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen |
| 09:51 | --- | <<-- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit (Quit: leaving) |
| 09:57 | --- | ---> Tv [~Tv@hq.inoi.fi] has joined #xen |
| 10:31 | --- | <<-- franR [~franr@inextensia.net] has quit (Read error: Connection reset by peer) |
| 10:48 | --- | ---> hollis [~hollis@pixpat.austin.ibm.com] has joined #xen |
| 10:53 | --- | <<-- cfreak [cfreak@dsl-084-056-104-177.arcor-ip.net] has quit (Quit: .) |
| 10:55 | --- | ---> stekloff [~stekloff@bi01p1.co.us.ibm.com] has joined #xen |
| 10:57 | --- | <<-- plars [~plars@pixpat.austin.ibm.com] has quit (Ping timeout: 481 seconds) |
| 11:02 | --- | <<-- hebutterworth [~harry@blueice1n1.uk.ibm.com] has quit (Quit: Leaving) |
| 11:29 | --- | ---> plars [~plars@pixpat.austin.ibm.com] has joined #xen |
| 11:35 | --- | ---> niv [~nivedita@bi01p1.co.us.ibm.com] has joined #xen |
| 11:37 | --- | <<-- athomas [~athomas@ppp-0-55.lond-b-2.access.uk.tiscali.com] has quit (Quit: Leaving) |
| 11:49 | cdub | pre-virtualiztion...hmm... |
| 11:52 | riel | pre? not para? ;) |
| 11:52 | tessier | What is that? |
| 12:04 | --- | ---> tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has joined #xen |
| 12:09 | aliguori | mmmm, pre-virtualization |
| 12:10 | aliguori | tessier: basically, it uses a modified binutils to essentially implement software VT-x |
| 12:10 | tessier | VT-x? |
| 12:10 | aliguori | automatically identify problem instructions and call-out to the hypervisor.. but they've optimized it by having it call a domain-space routine that can do smart things like batch hypercalls |
| 12:10 | aliguori | intel's virtualization extensions |
| 12:11 | aliguori | a.k.a VMX or Vanderpool |
| 12:11 | tessier | ah |
| 12:11 | tessier | I can't wait. Running Windows on Linux....sweet. |
| 12:12 | aliguori | heh, i think we're a bit away from that being all that practical |
| 12:12 | hollis | yes, as soon as you rebuild Windows with virtualization annotations |
| 12:12 | aliguori | there's a lot of difficult problems to solve although the recent VMX driver frontends are a good start |
| 12:13 | aliguori | hollis: just ask bill, i'm sure he'll give you the source code... |
| 12:14 | hollis | you mean http://l4ka.org/projects/virtualization/afterburn/ ? |
| 12:14 | riel | the problem is, what does that do to kernel modules? |
| 12:15 | riel | and debugging symbols ... |
| 12:15 | hollis | riel: what are the problems? |
| 12:15 | riel | if a user has a crash on a previrtualized system, is the crash dump going to be useful ? |
| 12:16 | hollis | why not? |
| 12:16 | riel | ie. are the symbols in the binary still where the debuginfo package says they were at build time ? |
| 12:16 | hollis | the debuginfo package would be built by the same previrtualizing toolchain, no? |
| 12:17 | aliguori | riel: I don't think symbol locations change at all |
| 12:17 | aliguori | riel: instructions are padded with nops |
| 12:17 | aliguori | so that when it's pre-virtualized, the only thing that changes is the nop-block |
| 12:18 | aliguori | riel: i reckon they'll have to introduce an "afterburn" hypercall and annotate linux such that it pre-virtualizes modules at load-time |
| 12:22 | riel | mmmm, interesting |
| 12:24 | cdub | toolchain reliance sounds problematic |
| 12:25 | hollis | cdub: we already rely on toolchains ;) |
| 12:25 | cdub | i guess malicious code is still not at ring level 0 |
| 12:25 | cdub | hollis: hehe, true |
| 12:25 | cdub | hollis: i meant to keep from emitting non-virtualizable insn's |
| 12:27 | hollis | why is that problematic? seems safer to do in the toolchain than to do by hand -- you know you won't miss anything |
| 12:30 | cdub | hollis: toolchain can't possibly stop malicious code |
| 12:30 | hollis | cdub: you were just talking about non-virtualizable instructions |
| 12:31 | hollis | the toolchain can certainly recognize those |
| 12:31 | hollis | and the compiled code is running inside a domain anyways |
| 12:32 | cdub | right, safety comes from ring > 0 |
| 12:32 | cdub | (for malicious code trying to use non-virt. insn) |
| 12:33 | hollis | cdub: you should read the afterburner page... |
| 12:33 | hollis | pure virtualization is slow |
| 12:33 | hollis | para virtualization can require extensive code modifications |
| 12:34 | hollis | pre virtualization solves both these problems |
| 12:34 | cdub | hollis: so does not using x86 |
| 12:34 | hollis | it is still virtualization |
| 12:34 | hollis | cdub: no, I don't think that's true |
| 12:35 | cdub | hollis: how would it help zseries? |
| 12:35 | hollis | cdub: I can speak for PowerPC... and if you look at the PPC Linux kernel, you will see extensive modifications to support various hypervisors |
| 12:35 | hollis | the pure" alternative would be |
| 12:35 | cdub | hollis: right, i assumed that was mostly for management |
| 12:36 | hollis | management? |
| 12:36 | cdub | (I hadn't looked at ppc) |
| 12:36 | tessier | aliguori: We are a ways away from windows running under xen? I thought the tlb fixups of the new cpu's would make it cake? |
| 12:37 | hollis | cdub: if you're interested, have a look at the "ppc_md" structure |
| 12:37 | cdub | hollis: as in asking for resources, finding where to direct i/o, etc. |
| 12:37 | hollis | cdub: you can see for example ppc_md.hpte_insert() being called in arch/ppc64/mm/init.c |
| 12:38 | cdub | hollis: ah, indeed |
| 12:38 | hollis | (hpte being "hash page table entry", i.e. creating an MMU translation) |
| 12:38 | hollis | and that function pointer is chosen depending on the environment (pSeries hypervisor, legacy iSeries hypervisor, bare metal) |
| 12:39 | hollis | the pure virtualization alternative of course would be to try to modify the hash table, trap into hypervisor, and figure it out from there |
| 12:39 | hollis | but that's slower than just asking the hypervisor politely in the first place |
| 12:40 | cdub | hollis: ok, i see |
| 12:41 | hollis | so with previrtualization, we would make a little note in the code when inserting an HPTE, and let the toolchain+init code fix it up for us |
| 12:41 | hollis | without having to explicitly call a function pointer |
| 12:41 | hollis | (and all the abstraction that entails) |
| 12:43 | cdub | hollis: interesting, just tag code, and let toolchain figure it out for you |
| 12:43 | hollis | yeah |
| 12:43 | cdub | nice |
| 12:43 | cdub | hollis: thanks for explanation |
| 12:43 | hollis | sure |
| 12:45 | --- | <<-- monrad [~monrad@213083190130.sonofon.dk] has quit (Quit: Leaving) |
| 13:10 | aliguori | tessier: well, once you get windows booting, you need to implement para-virtualized drivers for windows, get decent SVGA support, etc. |
| 13:11 | riel | *sigh* |
| 13:11 | riel | that ioemu code just won't build right in an RPM |
| 13:11 | riel | mkdir -p "" |
| 13:11 | riel | mkdir: cannot create directory `': No such file or directory |
| 13:13 | --- | ---> monrad [~monrad@213083190130.sonofon.dk] has joined #xen |
| 13:17 | --- | <<-- tessier [~treed@222.253.82.200] has quit (Ping timeout: 480 seconds) |
| 13:19 | Tv | riel: sound like "$UNSET_VAR" |
| 13:19 | Tv | riel: just wait till you hit the rm -rf "$UNSET_VAR/" ;) |
| 13:20 | riel | Tv: install: all |
| 13:20 | riel | mkdir -p "$(bindir)" |
| 13:20 | riel | in the ioemu Makefile |
| 13:28 | --- | ---> tessier [~treed@222.253.82.200] has joined #xen |
| 13:34 | aliguori | what's up with newlines in subject lines.. what client actually supports that? |
| 13:34 | aliguori | i've never seen it so much as i do on the xen-devel list |
| 13:35 | --- | <<-- Shaun [ndci@ip68-111-70-41.oc.oc.cox.net] has quit (Ping timeout: 480 seconds) |
| 13:35 | --- | <<-- DEac- [~deac@xdsl-213-196-200-152.netcologne.de] has quit (Quit: Verlassend) |
| 13:36 | --- | <<-- soffi [~soffi@Tolvudeild-SC-gw.vv.simnet.is] has quit (Quit: Leaving) |
| 14:00 | tessier | Newlines in email subject lines? That is a big no no. |
| 14:00 | tessier | It would screw up header parsing etc. I would think. |
| 14:01 | hollis | maybe that explains the xen-commit mails that look funny |
| 14:01 | hollis | I think their script does that |
| 14:01 | hollis | (takes the whole multi-line description and uses it as the subject) |
| 14:09 | --- | <<-- crisen [~crisen@spank.terdmonk.com] has quit (Ping timeout: 480 seconds) |
| 14:11 | --- | ---> DEac- [~deac@xdsl-84-44-213-205.netcologne.de] has joined #xen |
| 14:49 | --- | ---> pwagland [~paul@kungfucoder.org] has joined #xen |
| 14:49 | pwagland | Hi all! |
| 14:51 | pwagland | I was here last night, and just wanted to thank hollis and rharper for helping out. WIth your help I figured out what the problem is, and now have a working HD with DMA :-) |
| 14:51 | rharper | sweet |
| 14:51 | hollis | pwagland: what was the problem? |
| 14:51 | pwagland | Turns out that the generic driver was being loaded first, which meant that the via driver could not initialise the hardware. |
| 14:52 | pwagland | And it would appear that the generic driver cannot configure DMA on the via. |
| 14:52 | hollis | so you unconfigured the generic driver? |
| 14:53 | pwagland | Yeah. unconfigured it, turned the via from module into 'Y' and now everything is fine. |
| 14:53 | pwagland | I could probably have left the via as a module, but... :-) |
| 14:54 | hollis | cool |
| 14:54 | pwagland | The secret was knowing that hdparm should work from dom0. AFAIU under Xen1 xen set the DMA up? |
| 14:54 | pwagland | At least that is the only thing that I could surmise from my google searches. |
| 14:55 | pwagland | Knowing that it was the Dom0 kernel at fault then I could work out what to do to resolve it. |
| 14:55 | pwagland | It may be worth while adding that as a FAQ, or in the docs somewhere, I read both and couldn't find the information. |
| 14:57 | hollis | pwagland: maybe you could find a place to add a note in the Wiki? |
| 15:01 | pwagland | OK... |
| 15:05 | --- | ---> cfreak [cfreak@dsl-084-056-104-177.arcor-ip.net] has joined #xen |
| 15:06 | Tv | I hate moving iso images to dom0 to burn them. |
| 15:12 | pwagland | And I hate waking up at 6am... but sometimes you gotta do what you gotta do ;-) |
| 15:28 | --- | ---> mael_ [~Mael@kali.tirnamban.org] has joined #xen |
| 15:47 | --- | Netsplit unununium.oftc.net <-> uranium.oftc.net quits: grifferz, rharper, niv, stekloff, hollis, cartel, mael_, Rai, tab, Hunger, (+4 more, use /NETSPLIT to show all of them) |
| 15:49 | --- | Netsplit over, joins: Hunger, murble, tab, grifferz |
| 15:49 | --- | Netsplit over, joins: mael_ |
| 15:49 | --- | Netsplit over, joins: niv |
| 15:49 | --- | Netsplit over, joins: stekloff, hollis, rharper |
| 15:49 | --- | Netsplit over, joins: Rai, hbaum, rread |
| 15:50 | --- | Netsplit over, joins: cartel, @Sir_Ahzz |
| 15:51 | --- | <<-- knewt [~jmb@zeus.pimb.org] has quit (Ping timeout: 480 seconds) |
| 16:02 | pwagland | Another interesting question for you all.... |
| 16:03 | pwagland | kernel: unregister_netdevice: waiting for vif7.0 to become free. Usage count = 1 |
| 16:03 | pwagland | I am figuring that this is bad... |
| 16:03 | pwagland | But, is it known? |
| 16:06 | pwagland | Or perhaps more interesting, is there anyone here? :-) |
| 16:23 | pwagland | http://wiki.xensource.com/xenwiki/HowToEnableDMA |
| 16:23 | pwagland | If anyone has any ideas about the unregister_netdevice I am all ears! :-) |
| 16:35 | jonmason | pwagland: I am here, and vif not shutting down is bad |
| 16:36 | jonmason | pwagland: what did you do to break it? |
| 16:46 | --- | ---> knewt [~jmb@p213.54.74.139.tisdip.tiscali.de] has joined #xen |
| 16:53 | --- | <<-- cilkay [~cilkay@CPE00d0b743a22f-CM0011ae01fcbe.cpe.net.cable.rogers.com] has quit (Quit: Leaving) |
| 16:55 | --- | <<-- monrad [~monrad@213083190130.sonofon.dk] has quit (Remote host closed the connection) |
| 16:56 | --- | <<-- knewt [~jmb@p213.54.74.139.tisdip.tiscali.de] has quit (Quit: leaving) |
| 16:57 | --- | ---> knewt [~jmb@p213.54.74.139.tisdip.tiscali.de] has joined #xen |
| 16:57 | --- | ---> monrad [~monrad@213083190130.sonofon.dk] has joined #xen |
| 16:57 | --- | <<-- monrad [~monrad@213083190130.sonofon.dk] has quit (Quit: ) |
| 17:00 | --- | <<-- knewt [~jmb@p213.54.74.139.tisdip.tiscali.de] has quit (Quit: ) |
| 17:01 | --- | ---> knewt [~jmb@zeus.pimb.org] has joined #xen |
| 17:01 | --- | ---> cilkay [~cilkay@CPE00d0b743a22f-CM0011ae01fcbe.cpe.net.cable.rogers.com] has joined #xen |
| 17:02 | --- | ---> soffi [~soffi@proxy.du.vdsl.is] has joined #xen |
| 17:15 | mael_ | hi soffi |
| 17:16 | soffi | hi mael |
| 17:16 | soffi | 'sup |
| 17:16 | mael_ | I found an open source project doing iscsi target |
| 17:16 | soffi | cool |
| 17:16 | mael_ | thought it might be of interest for you |
| 17:16 | mael_ | http://iscsitarget.sourceforge.net/ |
| 17:16 | soffi | tellmetellmetellme |
| 17:17 | soffi | old news :( |
| 17:17 | mael_ | as openiscsi is only the initiator stuff |
| 17:17 | soffi | thanks though :) |
| 17:17 | mael_ | you knew it? |
| 17:17 | mael_ | :\ too bad I thought you'll be happy! |
| 17:19 | soffi | :) |
| 17:19 | soffi | I'm always happy |
| 17:19 | mael_ | that's cool then :) |
| 17:20 | mael_ | but did you try the iscsi target stuff? |
| 17:24 | soffi | yup |
| 17:24 | soffi | works well |
| 17:32 | --- | ---> crisen [~crisen@spank.terdmonk.com] has joined #xen |
| 17:33 | --- | <<-- rharper [~rharper@pixpat.austin.ibm.com] has quit (Quit: Leaving) |
| 17:35 | mael_ | well g'night |
| 17:35 | --- | <<-- mael_ [~Mael@kali.tirnamban.org] has quit (Quit: Fermeture du client) |
| 17:59 | --- | <<-- knewt [~jmb@zeus.pimb.org] has quit (Ping timeout: 480 seconds) |
| 18:05 | --- | <<-- soffi [~soffi@proxy.du.vdsl.is] has quit (Quit: This computer has gone to sleep) |
| 18:05 | --- | ---> knewt [~jmb@p213.54.74.139.tisdip.tiscali.de] has joined #xen |
| 18:15 | --- | <<-- knewt [~jmb@p213.54.74.139.tisdip.tiscali.de] has quit (Quit: leaving) |
| 18:23 | --- | ---> knewt [~jmb@zeus.pimb.org] has joined #xen |
| 18:29 | --- | <<-- hollis [~hollis@pixpat.austin.ibm.com] has quit (Quit: leaving) |
| 18:35 | @Sir_Ahzz | mael_ there's a channel for that I started as well. #ietd |
| 18:36 | @Sir_Ahzz | I use it on 2 machines with 1.4TB of disk space to export to a third machine, that then mounts the targets, manages the space with evms and re-exports as both iscsi targets and nfs mounts for the xen cluster here. |
| 18:36 | | * Sir_Ahzz runs off to dinner. |
| 18:50 | --- | <<-- cfreak [cfreak@dsl-084-056-104-177.arcor-ip.net] has quit (Quit: .) |
| 19:03 | --- | <<-- niv [~nivedita@bi01p1.co.us.ibm.com] has quit (Quit: Quitting) |
| 19:09 | --- | <<-- stekloff [~stekloff@bi01p1.co.us.ibm.com] has quit (Quit: Leaving) |
| 19:12 | --- | ---> soffi [~soffi@proxy.du.vdsl.is] has joined #xen |
| 19:24 | --- | <<-- tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has quit (Quit: bug, n: A son of a glitch.) |
| 19:38 | --- | ---> niv [~nivedita@bi01p1.co.us.ibm.com] has joined #xen |
| 19:38 | --- | <<-- niv [~nivedita@bi01p1.co.us.ibm.com] has quit (Quit: ) |
| 19:52 | --- | <<-- hbaum [~hbaum@bi01p1.co.us.ibm.com] has quit (Quit: Client exiting) |
| 20:00 | --- | <<-- aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit (Quit: Leaving) |
| 20:09 | --- | ---> badbunny [~badbunny@ppp-66-143-176-237.dsl.austtx.swbell.net] has joined #xen |
| 20:10 | --- | <<-- cilkay [~cilkay@CPE00d0b743a22f-CM0011ae01fcbe.cpe.net.cable.rogers.com] has quit (Quit: Leaving) |
| 20:14 | --- | <<-- badbunny [~badbunny@ppp-66-143-176-237.dsl.austtx.swbell.net] has quit (Quit: ) |
| 20:44 | --- | ---> badbunny [~badbunny@ppp-66-143-176-237.dsl.austtx.swbell.net] has joined #xen |
| 21:15 | --- | <<-- badbunny [~badbunny@ppp-66-143-176-237.dsl.austtx.swbell.net] has quit (Quit: Leaving) |
| 21:36 | --- | ---> Shaun [ndci@ip68-111-70-41.oc.oc.cox.net] has joined #xen |
| 21:42 | --- | ---> hollis [~hollis@user-0vvde2g.cable.mindspring.com] has joined #xen |
| 22:27 | --- | <<-- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit (Quit: leaving) |
| 22:59 | --- | <--- VS_ChanLog [~stats@ns.theshore.net] has left #xen (Rotating Logs) |
| 22:59 | --- | ---> VS_ChanLog [~stats@ns.theshore.net] has joined #xen |
| 23:04 | --- | <<-- lilo [~lilo@lilo.usercloak.oftc.net] has quit (Quit: brb) |
| 23:10 | --- | <<-- JViz [Anomaly@cpe-065-190-034-249.triad.res.rr.com] has quit (Quit: YES THEY DESERVE TO DIE, AND I HOPE THEY BURN IN HELL!) |
| 23:31 | --- | ---> JViz [Anomaly@cpe-065-190-034-249.triad.res.rr.com] has joined #xen |
| 23:46 | --- | <<-- tessier [~treed@222.253.82.200] has quit (Read error: Operation timed out) |
| 23:58 | --- | ---> tim [~tim@cpe-66-67-139-238.rochester.res.rr.com] has joined #xen |
| 23:59 | --- | <<-- tim [~tim@cpe-66-67-139-238.rochester.res.rr.com] has quit (Quit: ) |
| 23:59 | --- | ---> tim [~tim@cpe-66-67-139-238.rochester.res.rr.com] has joined #xen |
| --- | Log | closed Sat May 21 00:00:05 2005 |
