TheGrebs.com:IRC Logs:#xen:2005:03:16

---	Log	opened Wed Mar 16 00:00:25 2005
01:22	---	---> Fidget007 [~rpage@adsl-214-114-38.gnv.bellsouth.net] has joined #xen
02:15	---	---> xoxoxo [~xoxoxo@YahooBB219206220072.bbtec.net] has joined #xen
03:04	---	---> atari [~atari@213.144.146.89] has joined #xen
03:10	---	---> muli_ [~muli@nesher3.haifa.il.ibm.com] has joined #xen
03:12	---	<<-- DEac- [~deac@xdsl-213-196-200-108.netcologne.de] has quit (Read error: Operation timed out)
03:27	---	---> DEac- [~deac@xdsl-213-168-123-61.netcologne.de] has joined #xen
03:46	---	---> murb [~murble@debian.home.boughton.org.uk] has joined #xen
03:47	murb	gutenmorgen!
03:51	muli_	hello murb
04:39	atari	quiiitsch!
05:39	---	<<-- atari [~atari@213.144.146.89] has quit (Ping timeout: 480 seconds)
05:50	---	---> mael [~mael@194.214.199.130] has joined #xen
05:50	mael	hi
05:58	mael	is there any xen dev on this chan?
05:58	mael	or is it more a end-user oriented chan?
05:59	visik7	the second u says
05:59	muli_	mael, it's probably more of a "quiet oriented" channel
05:59	muli_	there are some developers, but I don't think any of the Cambridge guys are here
06:00	mael	hum ok
06:01	mael	then I should ask the xen-dev ML I suppose
06:01	muli_	you can ask here as well
06:01	muli_	but you're more likely to get an answer there
06:01	mael	I have a question on pacifica/vanderpool impact on xen
06:01	muli_	what's the question?
06:02	mael	well I tried to decipher intel's specs to see if vanderpool will have more performance impact on a full-virtualization or on a para-virtualization model
06:03	muli_	this was mentioned on the list, IIRC
06:03	muli_	Ian said that they definitely plan to continue with para-virtualization
06:03	mael	yes of course :)
06:03	muli_	but take advantage of VT if possible, and for "legacy" OS
06:04	mael	k then I'll try arder on the ML archives :)
06:05	mael	ah yeah I was a bad boy
06:05	mael	it was not so hard to find
06:05	mael	sorry to bother :)
06:05	muli_	that's ok
06:06	mael	if I'm correct though, pacifica's specs are not public yet
06:15	Tv	para-virt will probably always be faster, just because there's less layering
06:15	Tv	no need to pretend the network really is a pcnet32, like vmware used to do
06:16	Tv	(the real question is, will the difference be big enough to justify the extra work)
06:16	mael	yep this is the point
06:17	mael	running an OS off the shelf is stille an advantage
06:17	mael	-e
06:23	Tv	although with full virt you can't provide the nice console interfaces, like uml and xen do
06:23	Tv	vmware "lets display vga in a window" sucks
06:24	mael	for the workstation version you mean?
06:24	Tv	yup
06:24	Tv	I don't know how gsx/esx do that
06:24	mael	I don't think esx work this way
06:25	Tv	but, no matter what they do, they are unable to provide the same higher-lever console interface
06:26	Tv	they _need_ to emulate a vga
06:26	mael	yeah probably
06:26	Tv	just because they pretend that the guest is running on real hardware
06:26	mael	I should test esx to see
06:26	muli_	Tv, there's no technical reason why you couldn't do a console interface over vga text mode
06:26	Tv	muli_: won't be the same thing
06:26	muli_	graphical mode is an entirely different beast though
06:27	Tv	muli_: as in, guest writes character to screen location 2,3 -> you need to talk vt100 or something to make that effect happen
06:27	Tv	totally different experience
06:27	Tv	also, uml can export your tty1..8 as xterms in parallel
06:27	Tv	can't do that by emulating hardware
06:28	Tv	and stuff like trying to provide a HZ tick to guests..
06:29	Tv	trying to fake hardware is a lot harder than paravirt
06:29	muli_	tv, for the "no seek" case, translating vga console to serial is fairly simple
06:30	muli_	I agree that for seek semantics you need a multiline protocol of some sort
06:30	muli_	as for exporting different virtual consoles as xterms in parallel, if it works with virtual consoles on hardware, there's no reason you can't virtualize it
06:31	muli_	whether it makes sense or is it worth it - that depends, as always
06:31	Tv	it's just continuous video memory to hardware
06:31	Tv	and it's displaying at different offsets
06:31	Tv	it has no clue which ones are different ttys and which ones are just scrolled ttys
06:32	Tv	(that's pretty much my point; paravirt allows higher level abstraction than emulating hardware)
06:32	Tv	of course, the best of both worlds is full hardware emu + paravirt for optimization, when guest OS supports it
06:48	---	---> VS_StatsLog [~stats@ns.theshore.net] has joined #xen
06:49	---	<<-- VS_StatsLog [~stats@ns.theshore.net] has quit (Remote host closed the connection)
06:49	---	---> VS_ChanLog [~stats@ns.theshore.net] has joined #xen
07:00	---	<<-- VS_ChanLog [~stats@ns.theshore.net] has quit (Remote host closed the connection)
07:00	---	---> VS_ChanLog [~stats@ns.theshore.net] has joined #xen
07:17	mael	Tv: will xen v3+ support an hybrid approach?
07:19	Tv	mael: I have no clue.
07:19	Tv	I guess there'll be something VT/AMD-similar-thingie support
07:29	---	---> atari [~atari@213.144.146.89] has joined #xen
08:31	---	Netsplit oxygen.oftc.net <-> xenon.oftc.net quits: atari, Fidget007, plars, xoxoxo, muli_
08:34	---	Netsplit over, joins: atari, muli_, xoxoxo, Fidget007, plars
08:35	---	Channel: services.oftc.net changed the topic of #xen to: Xen support central. big iron tool for white box PCs. INFO -> http://www.cl.cam.ac.uk/Research/SRG/netos/xen/index.html
08:46	---	<<-- Tv [~Tv@hq.inoi.fi] has quit (Quit: Client exiting)
08:59	---	<<-- muli_ [~muli@nesher3.haifa.il.ibm.com] has quit (Quit: My damn controlling terminal disappeared!)
09:00	---	---> Tv [~Tv@hq.inoi.fi] has joined #xen
09:38	---	---> Beirdo__ [~gjhurlbu@li11-45.members.linode.com] has joined #xen
09:38	---	<<-- Beirdo__ [~gjhurlbu@li11-45.members.linode.com] has quit (Quit: )
09:50	---	<<-- Beirdo [~gjhurlbu@beirdo.usercloak.oftc.net] has quit (Quit: leaving)
09:50	---	---> Beirdo_ [~gjhurlbu@li11-45.members.linode.com] has joined #xen
09:52	---	User: *** Beirdo_ is now known as Beirdo
10:04	---	---> hollis [~hollis@user-0vvde2g.cable.mindspring.com] has joined #xen
11:44	---	<<-- xoxoxo [~xoxoxo@YahooBB219206220072.bbtec.net] has quit (Read error: Connection reset by peer)
11:45	---	---> hbaum [~hbaum@bi01p1.co.us.ibm.com] has joined #xen
11:54	---	<<-- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit (Quit: leaving)
14:14	---	<<-- smurfix [~smurf@run.smurf.noris.de] has quit (Read error: Connection reset by peer)
15:38	---	<<-- DEac- [~deac@xdsl-213-168-123-61.netcologne.de] has quit (Read error: Operation timed out)
15:53	---	---> DEac- [~deac@xdsl-213-168-107-115.netcologne.de] has joined #xen
15:55	visik7	xen 2.0.4 debian package is crappy
15:55	Tv	2.0.5 deb exist
15:55	visik7	where ?
15:55	Tv	people.debian.org/~doogie
15:56	Tv	there's one bug in postinst, you need to fix that manually
15:57	visik7	are u doogie ?
15:57	Tv	dpkg -i xen_*.deb, edit
15:57	Tv	/var/lib/dpkg/info/xen.postinst
15:57	Tv	to say
15:57	Tv	update-rc.d xend start 20 2 3 4 5 . stop 21 0 1 6 .
15:57	Tv	update-rc.d xendomains start 21 2 3 4 5 . stop 20 0 1 6 .
15:57	Tv	then dpkg --configure --pending
15:57	Tv	no, doogie is doogie
15:57	visik7	ah ok
15:57	visik7	:)
15:57	visik7	'couse in the 2.0.4 package there was a problem removing xen package
15:58	visik7	the script xendomains stop return a status of error
15:58	Tv	if it still happens with the 2.0.5 debs, please email details to {doogie,tv}@debian.org
15:59	visik7	'couse xend was stopped before xendomains and so xendomains can't find xend
15:59	Tv	2.0.5 debs fixed that I think
15:59	Tv	see the changelog
15:59	visik7	ok thank you
16:00	visik7	anyway I need to use Xen-unstable so for now no 2.0.5 upgrade :)
16:05	---	---> lanzm [~lanzm@adsl-212-101-16-178.solnet.ch] has joined #xen
16:08	visik7	got err compiling unstable :(((
16:18	visik7	how can I compile unstable branch ??
16:18	Tv	if you can't fix it, you probably shouldn't :)
16:18	Tv	try rolling back a changeset or two
16:19	visik7	I need a feature that only -unstable has
16:19	visik7	unfortunatly
16:22	visik7	is there a moment of the -unstable development that it compiles ok ?
16:24	---	---> lilo [~lilo@lilo.usercloak.oftc.net] has joined #xen
16:25	visik7	how can I change the bitkeeper pager ? is there a way? or am I forced to use 'more' ?
16:25	Tv	bk foo \| less
16:26	visik7	mmm
16:27	visik7	I thought that less interfere with it's own pager
16:30	Tv	you think wrong, then
16:30	visik7	:)
16:30		* Tv notes "... \| less \| less \| less" works just nicely
16:31	visik7	yes but for exemple when I'm grepping something in / and I reach the end of output less hangs
17:44	---	---> ijuz__ [~ijuz@p54ABB124.dip.t-dialin.net] has joined #xen
17:44	---	User: *** ijuz__ is now known as ijuz
17:47	ijuz	(xen 2.0.5) i took a clean 2.6.10 runned the mkbuildtree script, applied the patches in xen-2.0/patches, used a config with everything what i need and did a make ARCH=xen oldconfig and make bzImage modules
17:47	ijuz	(the xen options all default, just return)
17:47	ijuz	and i get this:
17:47	ijuz	drivers/xen/balloon/balloon.c: In function `balloon_process':
17:47	ijuz	drivers/xen/balloon/balloon.c:210: error: `phys_to_machine_mapping' undeclared (first use in this function)
17:47	ijuz	drivers/xen/balloon/balloon.c:210: error: (Each undeclared identifier is reported only once
17:47	ijuz	drivers/xen/balloon/balloon.c:210: error: for each function it appears in.)
17:47	ijuz	drivers/xen/balloon/balloon.c:210: error: `INVALID_P2M_ENTRY' undeclared (first use in this function)
17:47	ijuz	drivers/xen/balloon/balloon.c:253: warning: implicit declaration of function `scrub_pages'
17:47	ijuz	drivers/xen/balloon/balloon.c:275: warning: implicit declaration of function `mfn_to_pfn'
17:47	ijuz	wrong config? bug?
17:47	visik7	with which make ? the second one ?
17:48	ijuz	yes
17:48	visik7	anyway no bzImage
17:48	ijuz	oh
17:48	visik7	just make ARCH=xen
17:48	visik7	and
17:48	visik7	make ARCH=xen modules
17:49	visik7	nothing else
17:49	ijuz	ok, thanks
17:50	visik7	Tv I tried to rolling back a changeset or two but I got always a bunch of SCCS/s.ChangeSet: unmerged leaf <some version>
17:51	Tv	visik7: then you don't know how to use bk
17:51	Tv	bk undo -aCSET
17:52	visik7	I use the -r :P anyway u are right I dunno how to use bk :)
18:06	---	Netsplit uranium.oftc.net <-> olduuu.oftc.net quits: Nigelenki, muli
18:06	---	Netsplit over, joins: Nigelenki, muli
18:25	---	Netsplit jupiter.oftc.net <-> olduuu.oftc.net quits: Nigelenki, muli
18:27	---	Netsplit over, joins: muli
18:27	---	Netsplit over, joins: Nigelenki
18:45	---	Netsplit uranium.oftc.net <-> olduuu.oftc.net quits: Nigelenki
18:50	---	Netsplit over, joins: Nigelenki
18:54	---	Netsplit jupiter.oftc.net <-> olduuu.oftc.net quits: Nigelenki
18:56	---	Netsplit over, joins: Nigelenki
19:17	---	Netsplit uranium.oftc.net <-> olduuu.oftc.net quits: Nigelenki
19:28	---	<<-- hbaum [~hbaum@bi01p1.co.us.ibm.com] has quit (Quit: Client exiting)
19:30	---	---> Nigelenki [~bluefox@pcp484971pcs.whtmrs01.md.comcast.net] has joined #xen
22:36	---	<<-- Nigelenki [~bluefox@pcp484971pcs.whtmrs01.md.comcast.net] has quit (Quit: Linux sucks, it doesn't run all my viruses.)
23:35	---	---> Bluefox [~bluefox@pcp484971pcs.whtmrs01.md.comcast.net] has joined #xen
23:59	---	<--- VS_ChanLog [~stats@ns.theshore.net] has left #xen (Rotating Logs)
23:59	---	---> VS_ChanLog [~stats@ns.theshore.net] has joined #xen
---	Log	closed Thu Mar 17 00:00:51 2005