Back to Home / #xen / 2005 / 04 / Prev Day | Next Day
#xen IRC Logs for 2005-04-28

---Logopened Thu Apr 28 00:00:23 2005
01:24--- <<-- DEac- [~deac@xdsl-81-173-162-125.netcologne.de] has quit (Quit: Verlassend)
01:25--- ---> DEac- [~deac@xdsl-81-173-162-125.netcologne.de] has joined #xen
01:29--- ---> sleon [test@e180056038.adsl.alicedsl.de] has joined #xen
01:29sleon hi
01:29sleon which python twisted plugins do i need ?
01:38sleon i mean i do: xend start
01:38sleon and it won't start!!
01:41sleon S.O.S
02:13--- <<-- sleon [test@e180056038.adsl.alicedsl.de] has quit (Quit: Leaving)
02:31--- <<-- rusty [~rusty@bh02i525f01.au.ibm.com] has quit (Quit: Client exiting)
03:12--- ---> sleon [~sleon@129.69.221.149] has joined #xen
03:12sleon hi
03:12sleon i have xen 2.0.5 built from source
03:12sleon xend starts but then xm list complains that connection was refused
03:12sleon what is the cause for it?
03:47--- ---> anticw [cw@adsl-67-124-119-21.dsl.snfc21.pacbell.net] has joined #xen
03:47--- Channel: mode/#xen [+o anticw] by ChanServ
03:47--- Channel: mode/#xen [+o cf0989b8] by anticw
03:50--- ---> athomas [~athomas@ppp-0-91.lond-a-3.access.uk.tiscali.com] has joined #xen
04:00sleon hallo
04:00--- ---> Beaky [~chatzilla@ncg-adsl.demon.co.uk] has joined #xen
04:00sleon i have a problem
04:00sleon when i start xend start
04:00sleon and then do xm list
04:00sleon i get : connection refused
04:01--- <<-- Beaky [~chatzilla@ncg-adsl.demon.co.uk] has quit (Remote host closed the connection)
04:02--- <<-- hebutterworth [~harry@blueice1n1.de.ibm.com] has quit (Quit: Leaving)
04:05--- ---> Beaky [~chatzilla@ncg-adsl.demon.co.uk] has joined #xen
04:15--- <<-- sleon [~sleon@129.69.221.149] has quit (Remote host closed the connection)
04:22--- ---> sleon [~sleon@129.69.221.149] has joined #xen
04:22sleon hi
04:22sleon i can not start xend
04:22sleon heeeeeeeeeeeeeeeeeelp
04:41sleon hallloo
05:01--- <<-- nextime [~nextime@213-140-6-96.fastres.net] has quit (Quit: nextime has no reason)
05:02sleon HELP HELP HELp
05:07--- ---> nextime [~nextime@213-140-6-96.fastres.net] has joined #xen
05:09sleon nextime, help me
05:09sleon nextime, i do: xend start
05:09sleon no error is reported
05:09sleon then i do : xm list
05:09sleon i get connection 111 refused
05:09sleon (is xend running)
05:09sleon so i start it, but it won't start
05:16--- <--- sleon [~sleon@129.69.221.149] has left #xen (Verlassend)
05:29--- <<-- drbyte [~byte@intern146.lnk.telstra.net] has quit (Quit: Leaving)
05:35--- <<-- DEac- [~deac@xdsl-81-173-162-125.netcologne.de] has quit (Ping timeout: 480 seconds)
05:48--- ---> DEac- [~deac@xdsl-84-44-145-3.netcologne.de] has joined #xen
06:59--- ---> visik7 [~ciao@host55-41.pool80182.interbusiness.it] has joined #xen
07:22--- ---> drbyte [~byte@intern146.lnk.telstra.net] has joined #xen
08:05--- ---> hebutterworth [~harry@blueice4n1.de.ibm.com] has joined #xen
08:19--- <<-- athomas [~athomas@ppp-0-91.lond-a-3.access.uk.tiscali.com] has quit (Quit: Leaving)
08:22--- <<-- hebutterworth [~harry@blueice4n1.de.ibm.com] has quit (Read error: Operation timed out)
08:22--- ---> hebutterworth [~harry@blueice1n1.uk.ibm.com] has joined #xen
08:23--- <<-- knewt [~jmb@p213.54.76.1.tisdip.tiscali.de] has quit (Ping timeout: 480 seconds)
08:25--- <<-- hebutterworth [~harry@blueice1n1.uk.ibm.com] has quit (Remote host closed the connection)
08:27--- ---> knewt [~jmb@p213.54.76.1.tisdip.tiscali.de] has joined #xen
08:39--- <<-- drbyte [~byte@intern146.lnk.telstra.net] has quit (Quit: Leaving)
08:40--- User: *** unriel is now known as riel
08:40riel ok, time to make memory ballooning work the way it was meant to ;)
08:55--- ---> athomas [~athomas@ppp-0-250.lond-a-1.access.uk.tiscali.com] has joined #xen
09:05--- User: *** rh_out is now known as rharper
09:49--- <<-- muli [~muli@alhambra.mulix.org] has quit (Ping timeout: 480 seconds)
09:59--- User: *** jon-out is now known as jonmason
10:00--- <<-- Beaky [~chatzilla@ncg-adsl.demon.co.uk] has quit (Remote host closed the connection)
10:19--- <<-- visik7 [~ciao@host55-41.pool80182.interbusiness.it] has quit (Remote host closed the connection)
10:21--- ---> knewt_ [~jmb@p213.54.96.138.tisdip.tiscali.de] has joined #xen
10:25--- ---> hbaum [~hbaum@bi01p1.co.us.ibm.com] has joined #xen
10:25--- <<-- knewt [~jmb@p213.54.76.1.tisdip.tiscali.de] has quit (Read error: Connection reset by peer)
10:38--- ---> cfreak [cfreak@dsl-084-056-100-138.arcor-ip.net] has joined #xen
10:41--- User: *** knewt_ is now known as knewt
10:41knewt anyone know if there's likely to be a difference in speed in C between "num >> shift << shift" or "num & ~(size-1)"?
10:46rharper not sure, but I recognize &~ more easily than >> shift << , then again, I dont do a lot of bitshifting
10:49jonmason knewt: I've had people argue with me that <</>> is easier to read
10:50jonmason I would think the perf would be roughly equilivant
10:52knewt well, quick test on a 50,000x50,000 loop gives 18.5s for the 1st and 17s for the 2nd. probably not worth worrying about then
10:53knewt so i think i'll go for the 1st, as it's easier for me to see what it's doing at a glance
11:13--- ---> hebutterworth [~harry@blueice4n1.uk.ibm.com] has joined #xen
11:14--- ---> matta-lt [~matta@69.93.28.254] has joined #xen
11:23--- ---> MarkWilliamson [~MarkW@maw48.kings.cam.ac.uk] has joined #xen
11:26--- ---> edsuom [~edsuom@207-118-65-127.dyn.centurytel.net] has joined #xen
11:26--- ---> strafbomber [~admin@gw-ha1.gw.hosteurope.de] has joined #xen
11:26strafbomber hello
11:27hebutterworth| hey, MarkWilliamson, am currently trying to hook my back-end to my front-end through my new xend usb code.
11:28hebutterworth| Can unload and load driver modules OK if I limit functionality to just driver status messages.
11:28strafbomber i\xB4m using xen on a suse 9.3 machine, in a xen console i get the following error message "Couldnt get a file descriptor referring to the console", what can i do? i think i have to adjust something in my inittab, right?
11:28strafbomber (sorry for bad english)
11:29hebutterworth| But if I send a be_create message for a USB device as a result of receiving a be driver status up message the console locks up for all domains.
11:30hebutterworth| Any idea what would cause everything to lock up like that. Looks like some kind of deadlock in xend to me.
11:30MarkWilliamso| hebutterworth: great news!
11:30MarkWilliamso| good timing as well, 2.4 dom0 doesn't have long to live ;-)
11:30MarkWilliamso| (in the unstable tree, that is)
11:31hebutterworth| Well, there is still much work to do and progress is very slow.
11:31MarkWilliamso| after the consoles lock up, does Xend respond to anything?
11:31hebutterworth| I can do an xm destroy on the domains
11:31MarkWilliamso| oooh, that's a bit weird.
11:32MarkWilliamso| anything interesting in xend's logs
11:32hebutterworth| Nothing in the log.
11:32hebutterworth| I have my driver doing printfs
11:32hebutterworth| If I disable the be_create message
11:33hebutterworth| then the back-end sends the driver status up message
11:33hebutterworth| which is received by xend which pretends to send the be_create
11:33hebutterworth| then I can unload the back end driver module
11:33hebutterworth| but if I try to enable the be create message
11:34hebutterworth| then the back end driver hangs halfway through sending the be driver status up
11:34hebutterworth| and xend never sees a response to the be_create
11:34MarkWilliamso| ok that's quite weird
11:35MarkWilliamso| what era Xend is this - when did you pull it?
11:35hebutterworth| Maybe a week before the xen summit
11:35MarkWilliamso| Well it shouldn't be strictly a deadlock because that Xend is lock free (that's going to change)
11:36hebutterworth| I thought that trying to send the be_create before responding to the driver status up might be an issue. What do you think?
11:37MarkWilliamso| hmmmm. could be upsetting the state machine in the backend I suppose. seems suprising if it is, tho.
11:38hebutterworth| No, the state machine in the backend can cope with it. I had to write new state machines to cope with shutdown for unloading the modules.
11:38hebutterworth| Having unloadable modules was supposed to save me debug time.
11:39MarkWilliamso| yeah
11:39MarkWilliamso| can you put some tracing in the backend drivers "control.c" to see what messages it thinks it's getting?
11:40hebutterworth| It already has printfs, the backend doesn't see any control messages: the be create never arrives.
11:40hebutterworth| At least, the console never shows that it arrives.
11:40hebutterworth| Also, it dosn't get into the kern.log.
11:40hebutterworth| So it probably really hasn't arrived.
11:41hebutterworth| But I also trace trying to send it in xend.
11:41MarkWilliamso| hmmmm. it's strange because the control rings AFAIK can cope with out of order messages.
11:41hebutterworth| So my best guess was that I try to send it but xend won't send it until after sending the response to the driver status up.
11:41hebutterworth| And the response for that won't get sent until after I send the be_create.
11:42hebutterworth| But that was just a wild guess.
11:42MarkWilliamso| maybe some sort of ordering is being enfornced...
11:42MarkWilliamso| it's a bit difficult to tell with divergent code bases
11:42hebutterworth| I've not changed anything much outside the usb code.
11:43MarkWilliamso| I can't think of any obvious reason why the generic code in Xend or the kernel would choke on this
11:43hebutterworth| What's the status of the xend rewrite in Cambridge?
11:43MarkWilliamso| Although it's possible your mods are tickling a latent "feature"
11:44--- ---> muli [~muli@alhambra.mulix.org] has joined #xen
11:45MarkWilliamso| Mike wray has rewritten the core Xend stuff to ditch Twisted and use language level threads
11:45hebutterworth| I can restrict the function in the back-end driver to eliminate the new code there from the equation. Maybe I can get the messages working that way.
11:45MarkWilliamso| The class hierarchy is flattened out so it's easier to follow the structure and control flow.
11:45MarkWilliamso| So basically, expect a "Xend++" in the unstable tree soonish, if it's not there already.
11:45hebutterworth| Still python?
11:46MarkWilliamso| This is a separate stream to the registry and modular tools.
11:46MarkWilliamso| Yes, it's the same codebase just spruced up and robustified. It'll form the basis of the 3.0 tools. The full rewrite will probably come later.
11:47hebutterworth| Is there any public documentation of these efforts that I can keep an eye on? Wiki pages for example?
11:48hebutterworth| There hasn't been any discussion on xen-devel for a while.
11:48MarkWilliamso| Errrm not much, no.
11:49MarkWilliamso| under the developer todo list there's quite a good overview of what's going on (you may like to add yourself although the page might be locked right nov)
11:49MarkWilliamso| but it's rather sparse on detail
11:49hebutterworth| OK.
11:56--- <<-- athomas [~athomas@ppp-0-250.lond-a-1.access.uk.tiscali.com] has quit (Quit: Leaving)
11:57riel python is fine, once twisted is gone it should all be readable again
11:59--- <<-- strafbomber [~admin@gw-ha1.gw.hosteurope.de] has quit (Ping timeout: 480 seconds)
11:59MarkWilliamso| riel: the new code seems a bit easier to get into and hack on
11:59MarkWilliamso| so hopefully it'll enable the knowledge to be spread a bit more
12:00hebutterworth| I didn't find twisted a big deal. I've had more difficulty trying to reverse engineer the intention of the control messages w.r.t driver module unload.
12:00hebutterworth| Came to the conclusion that be driver status down meant that the be had already freed up the fe resources
12:01hebutterworth| but fe driver status down was a request from the fe to get xend to free up the fe resources.
12:01hebutterworth| These assumptions seem to make it possible to construct state machines driven by the existing messages that allow driver module unload.
12:01* riel is trying to find the place where variables get replaced for %d in the config file ;)
12:01riel extra = "4 VMID=%d usr=/dev/sda6 mem=%dM" % vmid maxmem
12:02riel that breaks, it appears currently only one variable can be replaced ;)
12:03hebutterworth| Thanks for your help, Mark, will try stripping down the back end to eliminate it as a cause of the problem.
12:04knewt riel: ... % (vmid, maxmem)
12:04riel knewt: ohhhhh
12:04riel nope, still complaining
12:04riel File "/usr/lib/python2.4/site-packages/xen/xm/opts.py", line 418, in load
12:04riel execfile(defconfig, globals, locals)
12:04riel File "/etc/xen/two", line 116
12:04riel extra = "4 VMID=%d usr=/dev/sda6 mem=%dM" % (vmid maxmem)
12:04riel ^
12:04riel SyntaxError: invalid syntax
12:05knewt you missed the comma
12:05--- ---> strafbomber [~admin@gw-ha1.gw.hosteurope.de] has joined #xen
12:05riel File "/etc/xen/two", line 24, in vmid_check
12:05riel val = int(val)
12:05riel TypeError: int() argument must be a string or a number
12:05riel ;)
12:05matta-lt riel: did the fixed xen rpm happen to get into last nights build?
12:05--- ---> dash [~washort@adsl-159-172-100.bhm.bellsouth.net] has joined #xen
12:05riel matta-lt: yes, fixed kernel-xen*
12:05matta-lt ok
12:05--- <<-- knewt [~jmb@p213.54.96.138.tisdip.tiscali.de] has quit (Quit: frickin' reboot)
12:06matta-lt 27-Apr-2005 15:22 14M
12:06matta-lt got it
12:07MarkWilliamso| riel: the config files are Python scripts - could can do *anything* in them (muhahahaha)
12:07riel oh, so they're _run_ and not parsed ?
12:08riel that would explain why I can't find parsing code ;)
12:08dash yep
12:08riel ok thanks for that hint ;)
12:08dash it's usually a good strategy for config files, at least that way you don't have to guess about quoting rules :)
12:08riel (this is how you can tell I spend too much time in the kernel)
12:13riel doh!
12:14* riel was being stupid and forget to set the vmid variable ;)
12:14riel this was xmexample2
12:14MarkWilliamso| riel: I've done that too
12:15MarkWilliamso| yes, they're executed but it happens in a little sandbox to avoid variable name clashes. any syntax you can use in python (regexps, substitutions, etc.) should work, if you want to make your config files really nasty :-)
12:17Tv also, configvar = file(".../config/var").read() is ok, if you don't like the python :)
12:17Tv that's the _best_ part of programmable configuration
12:18--- ---> admin_ [~admin@gw-ha1.gw.hosteurope.de] has joined #xen
12:19--- <<-- strafbomber [~admin@gw-ha1.gw.hosteurope.de] has quit (Read error: No route to host)
12:24--- ---> hollis [~hollis@pixpat.austin.ibm.com] has joined #xen
12:27--- <<-- admin_ [~admin@gw-ha1.gw.hosteurope.de] has quit (Quit: Verlassend)
12:29--- ---> the_hydra [~a_mulyadi@202.147.200.25] has joined #xen
12:30edsuom Newbie here: Is it realistic to think I could move my existing Gentoo 2.installation to run as a guest OS on a new system?
12:31edsuom One potential problem is it's all Athlon XP optimized, and new system will be Opteron.
12:35edsuom No reply? Doesn't sound promising.
12:35hollis I think it's realistic, but I haven't done it so I don't know
12:36matta-lt edsuom: athlon and opteron use the same optimization flags for gcc so it should be fine
12:36the_hydra edsuom: newbie to newbie-- have u check the Xen wiki?
12:36* edsuom blushes
12:37--- ---> knewt [~jmb@p213.54.96.138.tisdip.tiscali.de] has joined #xen
12:37matta-lt the problem would be if you moved to a p4 arch system...
12:38edsuom Thanks.
12:39--- <<-- edsuom [~edsuom@207-118-65-127.dyn.centurytel.net] has quit (Remote host closed the connection)
12:42--- <<-- hebutterworth [~harry@blueice4n1.uk.ibm.com] has quit (Quit: Leaving)
12:44--- <<-- the_hydra [~a_mulyadi@202.147.200.25] has quit (Quit: )
13:12--- <<-- mael [~mael@nat.inha.fr] has quit (Ping timeout: 480 seconds)
13:20--- <<-- MarkWilliamson [~MarkW@maw48.kings.cam.ac.uk] has quit (Ping timeout: 480 seconds)
13:24--- ---> niv [~nivedita@bi01p1.co.us.ibm.com] has joined #xen
13:41--- ---> mael [~mael@nat.inha.fr] has joined #xen
13:43--- ---> aliguori [~anthony@pixpat.austin.ibm.com] has joined #xen
13:43--- <<-- nextime [~nextime@213-140-6-96.fastres.net] has quit (Ping timeout: 480 seconds)
13:57--- <<-- DEac- [~deac@xdsl-84-44-145-3.netcologne.de] has quit (Quit: Verlassend)
14:04--- ---> DEac- [~deac@xdsl-81-173-163-107.netcologne.de] has joined #xen
14:12--- ---> tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has joined #xen
14:15--- <<-- jonmason [~jonmason@pixpat.austin.ibm.com] has quit (Remote host closed the connection)
14:27--- ---> jonmason [~jonmason@pixpat.austin.ibm.com] has joined #xen
14:46matta-lt riel: new fedora en build doesn't work too well... gets to booting, and then
14:46matta-lt slab sgpool-8: redzone mismatch in slabp c7f08000, objp c7f086e0, bufctl 0xfffe
14:46matta-lt Redzone: 0x0/0x170fc2a5.
14:46matta-lt Last user: [<c014cdb5>](mempool_create+0xe5/0x110)
14:46matta-lt 000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
14:46matta-lt 010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
14:47matta-lt ------------[ cut here ]------------
14:47matta-lt kernel BUG at drivers/xen/netback/netback.c:101!
14:47matta-lt invalid operand: 0000 [#1]
14:47matta-lt a few pages of OOPS data
14:50riel matta-lt: which kernel and xen version ?
14:50riel you need xen-2-20050423 for kernel 1275
14:51matta-lt yep
14:51matta-lt http://download.fedora.redhat.com/pub/fedora/linux/core/development/i386/Fedora/RPMS/xen-2-20050424.i386.rpm
14:51matta-lt that's 0424...
14:51riel oh right, 23 was my test build here
14:51riel 24 is the exact same sources, built in the build system
14:51matta-lt do you want me to e-mail you what I have?
14:52riel I'll try it here ;)
14:53--- <<-- cfreak [cfreak@dsl-084-056-100-138.arcor-ip.net] has quit (Quit: .)
14:54riel on what kind of system are you trying it ?
14:54riel it's starting up fine here
14:55riel if you open a bug, please open it on bugzilla.xensource.com
14:55riel if it turns out to be fedora specific, I'll just assign it to myself ;)
15:04matta-lt riel: opteron
15:05riel running it on a xeon here - no problem
15:07--- ---> hebutterworth [~harry@blueice2n1.uk.ibm.com] has joined #xen
15:09* riel wonders about paravirtualised xen on PPC ...
15:10* aliguori thinks riel should poke hollis on that one
15:10aliguori :-)
15:10riel yeah, I guess I should ;)
15:11hollis yeah, that's the idea...
15:12hollis although at least for now we would like to fit into the pSeries hypervisor ABI; i.e. no PPC Linux patches needed
15:12--- <<-- hebutterworth [~harry@blueice2n1.uk.ibm.com] has quit (Quit: Leaving)
15:18riel *nod*
15:24aliguori hollis: are you guys gonna implement xen-style drivers for ppc after the initial papr drivers?
15:25aliguori hollis: is it possible to get xen running on the g5s (after modifying the kernel of course)?
15:47hollis the current Apple G5s have hypervisor mode disabled, which is unfortunate
15:47aliguori hollis: but you could still paravirtualize them right?
15:47hollis mostrows is working on a hypervisor model for non-hypervisor-enabled PPC, but that will have a radically different interface from the pSeries model
15:47knewt an individual work_struct only ever runs once at any particular moment in time, right?
15:48aliguori hollis: ahh
15:48hollis aliguori: so for that mode, I assume there wouldn't be any problem using the Xen drivers (assuming they don't need much work)
15:48hollis aliguori: since the kernel needs to be all hacked up anyways
15:48aliguori hollis: how much work do you think it would be?
15:49aliguori hollis: wrt to the linux port
15:49hollis I haven't heard anything from mostrows recently, but I think it will be significant
15:49hollis pushing the kernel into user mode is no small thing
15:50aliguori hm, so i shouldn't buy an apple machine anytime soon then :-)
15:50hollis x86 has the whole ring distinction, which gives you a little wiggle room for a hypervisor
15:50riel on PPC you have the segments
15:50hollis yeah. we hope that one day Apple will stop disabling hypervisor mode in their systems, but...
15:50riel those might be useful
15:51hollis riel: yes, they're better than nothing ;) but that doesn't address privileged instructions, only memory protection
15:51knewt hollis: they probably will at some point, but charge a lot more for them *g*
15:54riel knewt: didn't apple start the "charge a lot more" thing already ? ;)
15:54knewt riel: any idea about my work_struct question above?
15:54riel knewt: sorry, nope
15:58--- ---> hebutterworth [~harry@blueice4n1.uk.ibm.com] has joined #xen
15:59hebutterworth| I just spent a whole day debugging a problem that turned out to be an error in whitespace in the usb xend code I wrote.
15:59hollis ouch. been there... :)
15:59aliguori hehe.. mmm, python
15:59knewt ouch. that's where whitespace-sensitive languages get annoying
15:59hollis actually I've been there with C and braces
16:00hebutterworth| So, I was calling packMsg, which takes a lot of parameters
16:00hollis the whitespace was correct, but the braces were imaginary ;)
16:00hebutterworth| and I was putting the parameters below...
16:00hebutterworth| (
16:00hebutterworth| ...
16:00hebutterworth| )
16:01hebutterworth| Seemd ok to me with only 3 days python experience
16:01hebutterworth| Executed
16:01hebutterworth| failed silently
16:01hebutterworth| by hanging messages
16:01hebutterworth| I want to take the person that made whitespace significant
16:02hebutterworth| and apply a hot poker
16:03Robot101 I know so many people who've suggested fixing python by adding braces back
16:03rharper mm packMsg , I know that was my favorite part of xend
16:04hollis hebutterworth: well like I said above, you can make whitespace mistakes in C too, just in the other direction :)
16:04hebutterworth| So, I format parameters like that in C specifically because you can align matching parentheses either horizontally or vertically which means you never get them mixed up
16:06hebutterworth| seems I'll need to use ugly toothbrush layout or an explicit line continuation in python
16:12yosh how were you trying to format things exactly?
16:13yosh () should do implict line continutation
16:13--- ---> visik7 [~ciao@host55-41.pool80182.interbusiness.it] has joined #xen
16:13hebutterworth| as above, with the parentheses on their own line
16:13hebutterworth| the problem was that I put the open parenthesis at the start of the next line
16:13yosh oh
16:13yosh that'd do it, yeah
16:14yosh if you put the open parenthesis on the same line, it'd have been ok
16:14hebutterworth| yes, that's what fixed it.
16:14yosh ok, I didn't know what you meant by "toothbrush layout"
16:15hebutterworth| toothbrush is what you get when all the parameters are at the end of a line
16:15hebutterworth| under each other
16:15hebutterworth| looks like a toothbrush. difficult to maintain
16:15hebutterworth| like this...
16:15hebutterworth| function_name(
16:16hebutterworth|
16:16hebutterworth| param,
16:16hebutterworth| param,
16:16hebutterworth| oops
16:16hebutterworth| try again
16:16hebutterworth| function_name(
16:16hebutterworth| param,
16:16hebutterworth| param,
16:16hebutterworth| param,
16:16hebutterworth| param)
16:17hebutterworth| sort of. Anyway, very ugly.
16:17--- ---> nextime [~nextime@213-140-6-96.fastres.net] has joined #xen
16:19hebutterworth| I've calmed down enough to drive home now :-) Calling it a day.
16:19--- <<-- hebutterworth [~harry@blueice4n1.uk.ibm.com] has quit (Quit: Leaving)
16:20knewt hmm. if there's only ever one work_struct running at a time in a workqueue, i wonder if there's any point /not/ creating it as single-threaded
16:24--- User: *** anticw is now known as cw
16:24@cw did anyone decide when/if the xensource were going to drop bk?
16:26Robot101 darcs is the one true way
16:26demon does anyone have any suggestions on where I'd need to add code to in xend to handle an extra argument for a different network setup script? I want to have the ability to auto-bridge a Xen domain's network device to a specific VLAN, but I've been foiled previously trying to dig through the xend code
16:28hollis cw: it makes sense to me to stay with bk until git becomes more usuable/better understood...
16:29@cw yeah, i just wanna rm bk here :)
16:29hollis there are nightly snapshots, no? :)
16:30* knewt was glad when the sfioball/etc stuff was given out by lm, as i /definitely/ shouldn't have been using bitkeeper
16:30aliguori demon: the bridge code is all a script--not actually in xend
16:30aliguori demon: look in /etc/xen/scripts/
16:31demon aliguori: I know where it is
16:31demon but I need to be able to add handling in xend for an argument
16:31demon the shell script isn't the problem
16:32demon it's that I'm rather unfamiliar with python
16:32rharper demon: if you are modifying xm interface, you will want to see tools/python/xen/xm/main.py
16:32rharper thats the top
16:33rharper it gets a bit ugly after that, creating new "ops" that the xend server handles by calling other handlers
16:34@cw hollis: are there? of unstable?
16:36rharper cw: http://www.cl.cam.ac.uk/Research/SRG/netos/xen/downloads/xen-unstable-src.tgz
16:36rharper http://www.cl.cam.ac.uk/Research/SRG/netos/xen/downloads.html
16:36rharper for testing as well
16:37knewt demon: you need to start with "preprocess_vifs" and "configure_vifs" in xen/xm/create.py, then you need to edit "configure", "reconfigure", "sxpr", and "vifctl_params" in xen/xend/server/netif.py, along with creating your own method def for your extra argument. and finally you need to edit "vifctl" in xen/xend/Vifctl.py
16:38demon knewt: ahh... nice
16:38demon that's what I needed
16:39demon I'm actually writing my own code (in PHP... yeah, I know) to act as a client to xend
16:39demon because I'm trying to get everything integrated so I can have a pretty management frontend
16:39demon but that's just the info I was looking for :)
17:05* caker tests vm-tools-0.0.9a
17:06hollis aliguori: don't forget to include a brief description of vm-tools in your announce mails, for people who haven't been following along...
17:07--- User: *** riel is now known as unriel
17:09caker aliguori: vm-tools won't build against -testing unless you remove the last arg from tools/vm-build-linux.c:
17:09caker ret = xc_linux_build(xc_handle, domid, kernel,
17:09caker ramdisk, cmdline, evtchn, flags, 1);
17:10visik7 how can I assign 2 mac address to a single interface so I can obtain 2 dhcp ip with 1 ethernet ?
17:11aliguori caker: oops, yeah, that's a mistake
17:11aliguori hollis: ok, thanks
17:12knewt visik7: err, you can only have 1 mac address per network interface
17:13visik7 knewt I know but I'm able to have multiple dhcp request with a single interface using bridging module one for the bridged iface and one for a xen machine
17:14visik7 do it's not possible to have a multiple mac address card without run a xen domu ?
17:16--- <<-- hbaum [~hbaum@bi01p1.co.us.ibm.com] has quit (Quit: Client exiting)
17:20knewt visik7: surely the easiest way if you need more than one ip assigned over dhcp is to tell the dhcp client to request more than one, and configure the server appropriately?
17:21visik7 the server is not under my control and if see that I'm requesting 2 ip from the same mac address it lock me
17:21visik7 locks
17:21visik7 someone tell me to use vconfig
17:21visik7 but I dunno how to
17:21demon er... vconfig is for setting up VLANs
17:22demon which I don't think is going to help you
17:22--- <<-- aliguori [~anthony@pixpat.austin.ibm.com] has quit (Read error: Operation timed out)
17:22demon I don't think most NICs support multiple MAC addresses assigned to one card... you might be able to use bridging to do something
17:22visik7 ok so I've solved creating a vpn in loopback so tap0 and tap1 bridging tap0 with eth0 and assign 2 different mac addresses one to br0 and the other to tap1
17:23visik7 but it sucks
17:23visik7 and really doesn't work in loopback but only if tap1 stay on another machine (dunno why)
17:24visik7 do u have other solution ?
17:51--- ---> aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen
17:52--- <<-- rharper [~rharper@pixpat.austin.ibm.com] has quit (Quit: Leaving)
17:53--- <<-- hollis [~hollis@pixpat.austin.ibm.com] has quit (Quit: leaving)
18:06--- Netsplit orion.oftc.net <-> iridium.oftc.net quits: tierra, monrad, Robot101, @cw, Shaun, JViz, muli, @surriel, schweeb, lilo, (+11 more, use /NETSPLIT to show all of them)
18:06--- Netsplit oxygen.oftc.net <-> nova.oftc.net quits: DEac-, knewt, jonmason, niv, mael, dash, matta-lt
18:06--- Netsplit oxygen.oftc.net <-> nova.oftc.net quits: HoraPe, sunny
18:06--- Netsplit oxygen.oftc.net <-> xenon.oftc.net quits: yosh, demon, plars, caker, Dougie, unriel
18:06--- Netsplit oxygen.oftc.net <-> jupiter.oftc.net quits: xai, paavon, eigood, @Sir_Ahzz
18:06--- Netsplit over, joins: xai, paavon, @Sir_Ahzz, eigood
18:06--- Netsplit over, joins: @surriel
18:06--- Channel: services.oftc.net changed the topic of #xen to: Xen Homepage-> http://www.cl.cam.ac.uk/Research/SRG/netos/xen/index.html || Xen Wiki -> http://wiki.xensource.com || vm-tools : http://www.cs.utexas.edu/users/aliguori/vm-tools/
18:06--- Channel: mode/#xen [-o cf0989b8] by ChanServ
18:06--- Netsplit over, joins: caker, demon, yosh, Dougie, plars, unriel
18:06--- Channel: orion.oftc.net changed the topic of #xen to: Xen Homepage-> http://www.cl.cam.ac.uk/Research/SRG/netos/xen/index.html || Xen Wiki -> http://wiki.xensource.com || vm-tools : http://www.cs.utexas.edu/users/aliguori/vm-tools/
18:07--- Netsplit over, joins: tierra, Shaun, Tv, Hunger, lilo, aliguori, visik7, @cw, monrad, wenchien
18:07--- Channel: mode/#xen [-o surriel] by ChanServ
18:07--- Netsplit over, joins: jonmason, DEac-, mael, niv, knewt, dash, matta-lt, sunny, HoraPe
18:07--- Netsplit over, joins: muli, Robot101, tab, dst, schweeb, nextime, VS_ChanLog, JViz, matta, Nigelenki
18:14--- ---> soffi [~soffi@proxy.du.vdsl.is] has joined #xen
18:15--- <<-- matta-lt [~matta@69.93.28.254] has quit (Quit: Remember the... the... uhh.....)
18:40--- ---> drbyte [~byte@intern146.lnk.telstra.net] has joined #xen
19:13--- ---> jeroney [~jeroney@rrcs-24-173-101-48.sw.biz.rr.com] has joined #xen
19:16--- <<-- jeroney [~jeroney@rrcs-24-173-101-48.sw.biz.rr.com] has quit (Quit: )
19:21--- <<-- niv [~nivedita@bi01p1.co.us.ibm.com] has quit (Quit: Quitting)
19:24--- ---> rusty [~rusty@bh02i525f01.au.ibm.com] has joined #xen
19:57--- User: *** surriel is now known as riel
20:06--- Netsplit orion.oftc.net <-> nova.oftc.net quits: mael, HoraPe, dash, knewt, jonmason
20:06--- Netsplit orion.oftc.net <-> nova.oftc.net quits: sunny, DEac-
20:06--- Netsplit over, joins: jonmason, DEac-, mael, knewt, dash, sunny, HoraPe
20:07riel HoraPe !
20:07riel nice to see you here
20:10HoraPe hi rik
20:10HoraPe thanks
20:31knewt hmm. methinks that at 3:30am it might be a good idea to get some sleep and check over the spinlocks tomorrow, instead of just trying it out tonight
20:52--- ---> aliguori- [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen
20:52--- ---> aliguori-- [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen
20:53--- <<-- aliguori-- [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit (Quit: )
20:53--- <<-- aliguori- [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit (Quit: )
20:56--- <<-- aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit (Remote host closed the connection)
21:08--- ---> John_K [m4levolent@blk-222-164-135.eastlink.ca] has joined #xen
21:09John_K hello
21:34--- ---> katzj [~katzj@c-66-30-206-55.hsd1.ma.comcast.net] has joined #xen
22:44--- <<-- drbyte [~byte@intern146.lnk.telstra.net] 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
23:24--- ---> drbyte [~byte@intern146.lnk.telstra.net] has joined #xen
---Logclosed Fri Apr 29 00:00:27 2005