TheGrebs.com:IRC Logs:#xen:2007:02:12

---	Log	opened Mon Feb 12 00:00:59 2007
00:10	\|-\|	paelzer [~kvirc@blueice1n1.de.ibm.com] has joined #xen
00:14	\|-\|	DoberMann_ changed nick to DoberMann
00:15	\|-\|	sputhenp [~sputhenp@202.80.58.210] has joined #xen
00:29	<tessier>	icblenke: I'll do your benchmark and publish it. I've never even seen a vmware license or purchased a copy. :)
01:19	\|-\|	DoberMann changed nick to DoberMann[PullA]
01:23	\|-\|	mastermind [~mastermin@mastermind.kaltenbrunner.cc] has quit [Quit: Client exiting]
01:48	\|-\|	redii [~tyler@c-67-182-108-193.hsd1.ca.comcast.net] has joined #xen
01:49	<redii>	good evening. Anyone had any problems getting a 3ware 9500-sl4 sata raid card working under xen-3.0.4.1?
02:01	\|-\|	DoberMann[PullA] changed nick to DoberMann
02:52	\|-\|	mejlholm\|uni [~mejlholm@optiplex19.cs.aau.dk] has joined #xen
02:57	\|-\|	guillaume_t [~guillaume@129.184.84.11] has quit [Quit: Leaving]
03:03	\|-\|	movement [~moz@82.152.219.10] has quit [Ping timeout: 480 seconds]
03:15	\|-\|	movement [~moz@82.153.7.144] has joined #xen
03:21	\|-\|	djrise [~djrise2b@109-125.252-81.static-ip.oleane.fr] has joined #xen
03:21	<djrise>	hello everyone
03:22	<djrise>	can i compil a 2.6.18 kernel with 3.03 tarball ?
03:22	<djrise>	I don't uderstand very well
03:22	<djrise>	in debian repository i have a 2.6.18 kernel with xen patch
03:29	\|-\|	mulix [~mulix@87.69.40.180.cable.012.net.il] has quit [Remote host closed the connection]
03:53	\|-\|	Sangmin [~orfeo61@user-10cmc8l.cable.mindspring.com] has joined #xen
03:53	<Sangmin>	hi guys
03:54	<Sangmin>	Can anyone explain me the difference between xenheap and domheap??
03:55	<Sangmin>	or any reference for it?
03:59	<Sangmin>	hello?
04:08	\|-\|	riel [~riel@2002:4a5c:3b41:1:213:72ff:fe17:4a9c] has quit [Server closed connection]
04:08	\|-\|	surriel [~riel@74.92.59.66] has joined #xen
04:25	\|-\|	djrise [~djrise2b@109-125.252-81.static-ip.oleane.fr] has quit [Quit: Quitte]
05:06	\|-\|	bernarde [~bernarde@32.104.18.240] has joined #xen
05:27	<mejlholm\|uni>	Sangmin, domheap is the pool of all free pages. xenheap is a fix sized (roughly 10MB on x86) pool used for domain metadata
05:27	<mejlholm\|uni>	if I recall correctly
05:28	<mejlholm\|uni>	memory allocated from the xenheap is mapped in (through superpages)
05:28	<mejlholm\|uni>	at least in 3.0.2, don't know if it may have changed
05:48	\|-\|	jimix [~jimix@c-67-189-187-152.hsd1.ma.comcast.net] has joined #xen
05:51	\|-\|	Shoragan [~shoragan@datenfreihafen.org] has quit [Server closed connection]
05:51	\|-\|	Shoragan [~shoragan@datenfreihafen.org] has joined #xen
05:51	\|-\|	Shoragan [~shoragan@datenfreihafen.org] has quit [Max SendQ exceeded]
05:51	\|-\|	Shoragan [~shoragan@datenfreihafen.org] has joined #xen
05:57	\|-\|	tonyb [~tony@pythia.bakeyournoodle.com] has quit [Server closed connection]
05:57	\|-\|	tonyb [~tony@pythia.bakeyournoodle.com] has joined #xen
05:59	\|-\|	waldi [~waldi@bblank.thinkmo.de] has quit [Server closed connection]
05:59	\|-\|	waldi [~waldi@bblank.thinkmo.de] has joined #xen
06:02	\|-\|	mdday [~mdday@cpe-024-163-122-235.nc.res.rr.com] has joined #xen
06:12	\|-\|	brianw [~ahzz@pool-71-170-81-199.dllstx.fios.verizon.net] has quit [Ping timeout: 480 seconds]
06:15	\|-\|	visik7 [~visi@host50-38-dynamic.6-87-r.retail.telecomitalia.it] has joined #xen
06:53	\|-\|	DavidS [~david@vpn.uni-ak.ac.at] has joined #xen
06:53	\|-\|	jwb_ [~jwboyer@rchp4.rochester.ibm.com] has joined #xen
07:07	\|-\|	dhendrix [~dhendrix@c-24-7-121-175.hsd1.ca.comcast.net] has quit [Server closed connection]
07:07	\|-\|	dhendrix [~dhendrix@c-24-7-121-175.hsd1.ca.comcast.net] has joined #xen
07:09	<Sangmin>	hey mejlholm.. are you a guy who has been working on page sharing stuff?
07:17	<mejlholm\|uni>	jup
07:24	\|-\|	AllenJB [~ajb42@raptor.ukc.ac.uk] has quit [Server closed connection]
07:25	\|-\|	AllenJB [~ajb42@raptor.ukc.ac.uk] has joined #xen
07:28	\|-\|	visik7 [~visi@host50-38-dynamic.6-87-r.retail.telecomitalia.it] has quit [Ping timeout: 480 seconds]
07:29	\|-\|	guillaume_t [~guillaume@129.184.84.11] has joined #xen
07:30	\|-\|	Cybaspace [Cybaspace@host-14.kbcclearing.demon.nl] has joined #xen
07:34	<Cybaspace>	Hi there! I have a small problem with a windows 2003 VM. When I create the VM (1 CPU/512MB/10GbHD) nothing happens. I expected to see that blue windows install screen, but I don't get a graphical console. The only thing I see is a CPUload of 99% of that VM. Anyone experienced the same yet?
07:51	\|-\|	ns [~nivedita@210.211.241.45] has joined #xen
07:52	\|-\|	Sangmin [~orfeo61@user-10cmc8l.cable.mindspring.com] has quit []
07:58	\|-\|	dhendrix_work [~dhendrix@216-239-45-4.google.com] has quit [Server closed connection]
07:58	\|-\|	dhendrix_work [~dhendrix@216-239-45-4.google.com] has joined #xen
08:06	\|-\|	yoavw [~Mack@80.178.86.105.adsl.012.net.il] has quit [Ping timeout: 480 seconds]
08:09	\|-\|	mejlholm\|uni [~mejlholm@optiplex19.cs.aau.dk] has quit [Quit: Leaving]
08:10	\|-\|	Hunger [Hunger.hu@213.163.11.138] has quit [Server closed connection]
08:12	\|-\|	Hunger [Hunger.hu@Hunger.hu] has joined #xen
08:16	\|-\|	mcr [~mcr@desk.marajade.sandelman.ca] has joined #xen
08:16	<mcr>	do I need to restart xend if I want to start another physical interface up with a bridge?
08:16	<mcr>	(the dom0 won't be part of this bridge)
08:20	\|-\|	KriP [kp@yancey.unixworld.dk] has joined #xen
08:20	\|-\|	djrise [~djrise2b@109-125.252-81.static-ip.oleane.fr] has joined #xen
08:20	<djrise>	hi
08:21	<djrise>	does i must compil kernel with CONFIG_XEN_BLKDEV_FRONTEND=y?
08:21	\|-\|	mcr [~mcr@desk.marajade.sandelman.ca] has left #xen []
08:24	<Cybaspace>	Anyone has a clue why Win2k3 can't boot and why the VM uses 99% CPU? (question posted a few lines above) The graphical console is dead too...
08:30	\|-\|	dansmith [~dan@gw0.danplanet.com] has quit [Quit: Leaving]
08:39	\|-\|	Netsplit osmotic.oftc.net <-> scorpio.oftc.net quits: guillaume_t, KriP, dhendrix, djrise
08:39	\|-\|	Netsplit over, joins: djrise, KriP, guillaume_t
08:39	\|-\|	Netsplit over, joins: dhendrix
08:41	\|-\|	agriffis changed nick to agrif_away
08:41	\|-\|	agrif_away changed nick to agriffis
08:58	\|-\|	mejlholm [~mejlholm@port79.ds1-abc.adsl.cybercity.dk] has joined #xen
08:58	\|-\|	surriel changed nick to riel
08:59	\|-\|	aliguori [~anthony@cpe-70-112-17-156.austin.res.rr.com] has joined #xen
09:05	\|-\|	dansmith [~dan@bi01p1.co.us.ibm.com] has joined #xen
09:08	\|-\|	ns [~nivedita@210.211.241.45] has quit [Ping timeout: 480 seconds]
09:09	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has joined #xen
09:19	\|-\|	sputhenp [~sputhenp@202.80.58.210] has quit [Quit: Leaving]
09:22	\|-\|	sputhenp [~sputhenp@202.80.58.210] has joined #xen
09:27	\|-\|	ivan [~ikelly@skynet.skynet.ie] has joined #xen
09:27	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
09:31	\|-\|	stekloff [~stekloff@pool-71-245-98-220.ptldor.fios.verizon.net] has joined #xen
09:37	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has joined #xen
09:38	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has joined #xen
09:39	\|-\|	hollisb [~hollisb@bi01p1.co.us.ibm.com] has joined #xen
09:41	\|-\|	brianw [~ahzz@pool-71-170-81-199.dllstx.fios.verizon.net] has joined #xen
10:03	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has quit [Ping timeout: 480 seconds]
10:09	\|-\|	sputhenp [~sputhenp@202.80.58.210] has quit [Quit: Leaving]
10:11	\|-\|	nou [Chaton@causse.larzac.fr.eu.org] has quit [Server closed connection]
10:11	\|-\|	nou [Chaton@causse.larzac.fr.eu.org] has joined #xen
10:13	\|-\|	dansmith_ [~dan@bi01p1.co.us.ibm.com] has joined #xen
10:13	\|-\|	dansmith_ [~dan@bi01p1.co.us.ibm.com] has quit []
10:14	\|-\|	dansmith [~dan@bi01p1.co.us.ibm.com] has quit [Quit: Leaving]
10:14	\|-\|	dansmith [~dan@bi01p1.co.us.ibm.com] has joined #xen
10:16	\|-\|	tom [~tom@nonserviam.net] has joined #xen
10:16	\|-\|	tom changed nick to PerlSaiyan
10:17	<PerlSaiyan>	I used xm mem-max domu1 3072, and then xm mem-set domu1 3072, but I didn't see an increase in memory in the domU, although xm list shows the memory at the 3 meg line now
10:17	<PerlSaiyan>	am I missing something?
10:17	<PerlSaiyan>	(xen 3.0.3-0)
10:19	<danpb>	you can't increase the max mem beyond what it was booted with - you can only decrease, and then increase it back to the original limit
10:19	\|-\|	hbaum [~hbaum@bi01p1.co.us.ibm.com] has joined #xen
10:19	<PerlSaiyan>	ah, not even with mem-max
10:25	\|-\|	rpg [~rpg@bi01p1.co.us.ibm.com] has joined #xen
10:27	\|-\|	PerlSaiyan [~tom@nonserviam.net] has quit [Quit: [BX] Leggo my Eggo!]
10:31	\|-\|	rpg [~rpg@bi01p1.co.us.ibm.com] has quit [Read error: Operation timed out]
10:35	\|-\|	weasel [weasel@weasel.noc.oftc.net] has quit [Quit: Changing server]
10:36	\|-\|	weasel [weasel@2001:858:10f:6::1] has joined #xen
10:36	\|-\|	cableman [~cableman@0x503fe19a.hrnxx3.adsl-dhcp.tele.dk] has joined #xen
10:37	<icblenke>	http://news.com.com/Novell+offers+faster+virtualized+Windows/2100-1012_3-6158196.html
10:38	<icblenke>	I take it this means that Novell is shipping Windows PV drivers for SuSE 10.1 for use with opensource Xen HVM.
10:42	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has joined #xen
10:47	\|-\|	rpg [~rpg@bi01p1.co.us.ibm.com] has joined #xen
10:56	\|-\|	jerone [~jerone@bi01p1.co.us.ibm.com] has joined #xen
10:58	\|-\|	aliguori [~anthony@cpe-70-112-17-156.austin.res.rr.com] has quit [Quit: Leaving]
11:00	\|-\|	visik7 [~visi@host177-108-dynamic.56-82-r.retail.telecomitalia.it] has joined #xen
11:04	\|-\|	Cybaspace [Cybaspace@host-14.kbcclearing.demon.nl] has quit [Ping timeout: 480 seconds]
11:04	\|-\|	tonyb [~tony@pythia.bakeyournoodle.com] has quit [Read error: Operation timed out]
11:09	\|-\|	Netsplit osmotic.oftc.net <-> scorpio.oftc.net quits: guillaume_t, cableman, KriP, dhendrix, djrise
11:09	\|-\|	tonyb [~tony@pythia.bakeyournoodle.com] has joined #xen
11:09	\|-\|	Netsplit over, joins: cableman
11:09	\|-\|	Netsplit over, joins: djrise, KriP, guillaume_t
11:09	\|-\|	Netsplit over, joins: dhendrix
11:10	\|-\|	Basic_py [~Basic@warden.real-time.com] has quit [Quit: Leaving]
11:12	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
11:14	\|-\|	DoberMann changed nick to DoberMann[PullA]
11:16	\|-\|	ivan [~ikelly@skynet.skynet.ie] has quit [Quit: leaving]
11:20	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has joined #xen
11:20	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has quit [Ping timeout: 480 seconds]
11:23	\|-\|	DavidS [~david@vpn.uni-ak.ac.at] has quit [Quit: Leaving.]
11:35	\|-\|	buggs [~noidentd@shadow.splashground.de] has quit [Server closed connection]
11:36	\|-\|	buggs [~noidentd@shadow.splashground.de] has joined #xen
11:40	\|-\|	djrise [~djrise2b@109-125.252-81.static-ip.oleane.fr] has quit [Remote host closed the connection]
11:41	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has joined #xen
11:41	\|-\|	msinhore [~msinhore@correio.samurai.com.br] has joined #xen
11:50	\|-\|	mulix [~mulix@87.69.40.180.cable.012.net.il] has joined #xen
11:51	\|-\|	DoberMann[PullA] changed nick to DoberMann
11:51	\|-\|	aw [~awilliam@c-24-9-84-32.hsd1.co.comcast.net] has quit [Quit: Leaving]
11:53	\|-\|	Basic_py [~Basic@gatekeeper.real-time.com] has joined #xen
11:57	\|-\|	msinhore [~msinhore@correio.samurai.com.br] has quit [Remote host closed the connection]
12:15	\|-\|	aw [~awilliam@ftcrel4.hp.com] has joined #xen
12:20	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has joined #xen
12:25	\|-\|	visik7 [~visi@host177-108-dynamic.56-82-r.retail.telecomitalia.it] has quit [Remote host closed the connection]
12:28	\|-\|	swingle [~swingle@238.Red-81-40-68.staticIP.rima-tde.net] has joined #xen
12:28	<swingle>	hi
12:29	<swingle>	im running winXP as an hvm guest with xen but im getting a terrible performance
12:29	<swingle>	ive done some benchmarks and its worse than VirtualPC
12:29	<swingle>	(which is worse than vmWare)
12:30	<swingle>	i read on the webpage that it should perform much better, how can i fix it?
12:35	\|-\|	msinhore [~msinhore@correio.samurai.com.br] has joined #xen
12:38	<icblenke>	find the SuSE 10.1 beta Windows PV drivers and try those.
12:38	<swingle>	drivers for what?
12:38	<icblenke>	http://news.com.com/Novell+offers+faster+virtualized+Windows/2100-1012_3-6158196.html
12:40	<swingle>	but my problem is not networking, the CPU is terribly slow
12:41	<icblenke>	are you _sure_ it's not disk or networking slowdown?
12:41	<icblenke>	try giving SANDRA or something similar a try.
12:41	<swingle>	those might be slow too but i dont care
12:41	<icblenke>	aye, but you will care. particularly if you're paging off of slow disk.
12:42	<swingle>	i read xen's overhead on the CPU is about 10%
12:42	<swingle>	in my case its over 60%
12:42	<icblenke>	and I'll argue that most of it is in IDE emulation.
12:43	<swingle>	you need ide emulation to add floating point numbers toguether?
12:43	<icblenke>	what is your benchmark?
12:43	<swingle>	hmm ive tried 3 or 4 but i cant remember the names
12:43	<icblenke>	when I ran Sandra, I was getting pretty darned good CPU and memory access numbers.
12:44	<swingle>	anyway you can notice it slow just by looking at it
12:44	<icblenke>	are you using VNC to access it?
12:44	<swingle>	yep
12:44	<icblenke>	what encoding?
12:44	<swingle>	raw
12:44	<icblenke>	locally?
12:44	<swingle>	yes
12:44	<icblenke>	still, raw is very expensive. try hextile.
12:45	<swingle>	hmm im not sure about this but shouldnt it be the cheapest?
12:45	<icblenke>	it's very resource intensive.
12:45	\|-\|	Netsplit synthon.oftc.net <-> scorpio.oftc.net quits: guillaume_t, KriP, cableman, tonyb, dhendrix
12:45	\|-\|	Netsplit over, joins: tonyb, KriP, guillaume_t
12:45	\|-\|	Netsplit over, joins: dhendrix
12:45	<icblenke>	ie, you ship around large bitmap chunks.
12:45	<swingle>	hmm ok
12:45	<icblenke>	I'm guessing you've tried SDL, and the speed is no different.
12:46	<swingle>	you are right :)
12:46	<icblenke>	leading back to my assertion that you should look at IO performance bottlenecks.
12:46	<icblenke>	try Sandra.
12:46	<swingle>	ok i will
12:46	<icblenke>	I'm sure there are others. I'm not a windows guy though.
12:47	<swingle>	hmm do you know the url for sandra?
12:47	<icblenke>	I'd have to google it again myself.
12:47	<swingle>	google gives me tons of matches
12:47	<swingle>	92 millon to be precise :)
12:47	<icblenke>	try sandra benchmark.
12:48	<swingle>	ok
12:48	<icblenke>	keep whittling down your search.
12:49	\|-\|	hbaum [~hbaum@bi01p1.co.us.ibm.com] has quit [Read error: Connection reset by peer]
12:49	<swingle>	btw did you do anything special in your config file?
12:49	<swingle>	like assigning all the cpus to the domain or disabling acpi ...
12:51	<icblenke>	I converted my image from vwmare directly to xen. If you're going to try installing windows under Xen, I suggest using 3.0.4
12:51	<swingle>	im using unstable
12:51	<swingle>	directly from mercurial
12:52	<icblenke>	you'll have to get someone else here to help you out then. I'm not insane enough to run off of unstable.
12:52	<swingle>	lol
12:53	<swingle>	the tarball they have on the webpage doesnt even boot with my computers
12:53	<swingle>	it gives a kernel panic quite early
12:53	<icblenke>	a xen panic, or a kernel panic?
12:53	<swingle>	dom0 kernel panic
12:53	<swingle>	xen starts fine
12:53	<icblenke>	any specific panic?
12:54	<swingle>	something with iommu on its name
12:54	<swingle>	i cant remember the exact message
12:54	<swingle>	hmm im thinking of something
12:55	<swingle>	ive got a 64 bit machine but im running a 32 bit windows as a guest
12:55	<swingle>	can that be the reason for the poor performace?
12:55	\|-\|	hbaum [~hbaum@bi01p1.co.us.ibm.com] has joined #xen
12:55	\|-\|	Netsplit osmotic.oftc.net <-> scorpio.oftc.net quits: guillaume_t, KriP, tonyb, dhendrix
12:55	\|-\|	Netsplit over, joins: tonyb, KriP, guillaume_t
12:56	\|-\|	Netsplit over, joins: dhendrix
12:56	<danpb>	swingle: no 64/32 bits is irrelevant - fullvirt is just generally rubbish performance unless you have paravirt drivers
12:56	<swingle>	danpb, but there are no drivers for the cpu
12:56	<danpb>	sadly no one has seen fit to release their paravirt drivers for Windows as open source :-(
12:57	<icblenke>	I'm actively digging through SuSE's stuff now looking for their PV drivers, btw. It looks like they're keeping it out of the opensource realm.
12:57	<swingle>	so you are telling me i can improve networking performance but the cpu will suck anyway?
12:57	\|-\|	paelzer [~kvirc@blueice1n1.de.ibm.com] has quit [Quit: KVIrc 3.2.0 'Realia']
12:57	<danpb>	swingle: CPU performance is generally reasonable with full virt
12:57	<icblenke>	swingle: we're telling you that any perceived slowness is likely due to the emulated IDE chipset.
12:57	<danpb>	its primarily I/O (both disk & network) which sucks
12:58	<icblenke>	Xen HVM's emulated rtl8139 performance is abysmal.
12:58	<swingle>	cpu scores about 40%-50% of what it does without virtualization
12:58	<icblenke>	swingly: give me a benchmark that says this.
12:58	<swingle>	ill send it to the mailing list tomorrow
12:58	<swingle>	im not at work now
12:58	<icblenke>	good.
12:59	<swingle>	icblenke, now if its really true, what could be the reason?
13:00	<hollisb>	swingle: a slow user interface can make a system feel very very slow
13:00	<icblenke>	I'm not going to speculate about something I've never seen before and have no metrics to consider.
13:00	<hollisb>	swingle: so maybe if you fixed your UI issue, it would feel much faster
13:00	\|-\|	aw [~awilliam@ftcrel4.hp.com] has quit [Remote host closed the connection]
13:00	<swingle>	hollisb, it generates a list of prime numbers 1.5 times slower than the non virtualized OS
13:00	<swingle>	that has nothing to do with the UI
13:01	<icblenke>	while it is doing this, what is the RAM footprint like?
13:01	<icblenke>	is it paging at all?
13:01	<swingle>	no its not paging
13:02	<icblenke>	if you disconnect from the console, does it run faster?
13:02	<swingle>	no
13:02	\|-\|	aw [~awilliam@156.153.254.69] has joined #xen
13:02	<icblenke>	(remember, the framebuffer console is emulated in dom0, and thus causes CPU overhead in dom0 which would affect performance under a domU)
13:03	<swingle>	thats reasonable and ill check it out tomorrow
13:06	<swingle>	icblenke, ---> "Raw
13:06	<swingle>	The raw encoding simply sends width*height pixel values. All clients are required to support this encoding type. Raw is also the fastest when the server and viewer are on the same machine, as the connection speed is essentially infinite and raw encoding minimizes processing time."
13:08	<swingle>	are you sure changing the encoding would improve something?
13:10	\|-\|	tonyb_ [~tony@pythia.bakeyournoodle.com] has joined #xen
13:10	\|-\|	KriiP [kp@yancey.unixworld.dk] has joined #xen
13:10	\|-\|	dwh [~dhendrix@c-24-7-121-175.hsd1.ca.comcast.net] has joined #xen
13:10	\|-\|	Netsplit osmotic.oftc.net <-> scorpio.oftc.net quits: guillaume_t, KriP, tonyb, dhendrix
13:11	\|-\|	Netsplit over, joins: guillaume_t
13:16	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has quit [Read error: No route to host]
13:20	\|-\|	agriffis changed nick to agrif_away
13:22	<aliguori>	swingle: what encoding are you talking about changing?
13:22	<swingle>	the vncviewer encoding
13:22	<swingle>	im using it on the same machine with raw encoding
13:23	<aliguori>	swingle: if you're concerned about performance, don't use vnc
13:23	<swingle>	whan can i use then?
13:23	<aliguori>	this is hvm right?
13:23	<swingle>	yep
13:23	<aliguori>	sdl
13:23	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has joined #xen
13:23	<aliguori>	on the local machine
13:23	<swingle>	sdl forces me to keep the window open all the time
13:23	<aliguori>	swingle: well, if you do a test with the vnc window open, it chews a fair bit of cpu minimizing the rectangles
13:23	<aliguori>	to minimize traffic
13:24	<brendan>	hg qpush
13:24	<brendan>	oops, wrong win
13:24	<swingle>	i can remove those checks if it can fix my problems
13:25	\|-\|	matti [matti@acrux.romke.net] has quit [Server closed connection]
13:25	\|-\|	matti [matti@acrux.romke.net] has joined #xen
13:26	\|-\|	matti changed nick to Guest2363
13:29	<aliguori>	swingle: what exactly is your problem?
13:29	<swingle>	the cpu is scoring very low in all the benchmarks
13:29	<swingle>	40-50% of the real machine
13:29	<swingle>	its even worse than virtualPC
13:30	<aliguori>	swingle: i'm not terribly surprised about that
13:30	<aliguori>	what did you expect?
13:31	<swingle>	well, there are some fake benchmarks on the webpage then
13:31	<swingle>	i expected 10% overhead
13:31	<aliguori>	swingle: well, that's certainly not realistic
13:31	<swingle>	or at least something better than virtualPC which doesnt use hw virtualization at all
13:31	<aliguori>	where did you get that from?
13:31	<aliguori>	hardware virtualization != performance improvement
13:31	<swingle>	i found some benchmarks somewhere on the page
13:32	<murb>	aliguori: well not yet.
13:32	<aliguori>	swingle: if you can find them again, i'd like to see them
13:32	<swingle>	ok, ill try
13:32	<swingle>	executing code directly on the cpu should be faster than emulating it shouldnt it?
13:33	<aliguori>	swingle: virtualpc does not emulate
13:33	<aliguori>	it does dynamic translation
13:33	<aliguori>	dynamic translation does run code directly on the cpu
13:33	<swingle>	i didnt know that
13:33	<swingle>	but anyway hw virt should score much lower
13:34	<aliguori>	swingle: why?
13:34	<swingle>	it runs code on the real cpu too
13:34	<swingle>	oh sorry
13:34	<swingle>	should -> shouldnt
13:36	<swingle>	brb
13:36	\|-\|	swingle [~swingle@238.Red-81-40-68.staticIP.rima-tde.net] has quit [Quit: Leaving]
13:42	\|-\|	pvanhoof [~pvanhoof@d54C0EE14.access.telenet.be] has joined #xen
13:42	\|-\|	rpg [~rpg@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
13:53	\|-\|	redii [~tyler@c-67-182-108-193.hsd1.ca.comcast.net] has quit [Quit: redii]
13:53	\|-\|	swingle [~swingle@221.Red-81-40-154.staticIP.rima-tde.net] has joined #xen
13:53	<swingle>	http://www.intel.com/technology/itj/2006/v10i3/3-xen/6-benchmark-performance.htm <--- the fake benchmark
13:54	<icblenke>	Interesting change control note in SuSE's logs: "Backport xen-unstable changesets 12040 to address spurious
13:54	<icblenke>	interrupts with PV drivers in HVM guests."
13:54	<icblenke>	now I wonder if 12040 made it to 3.0.4
13:55	<aliguori>	swingle: and you ran those same benchmarks?
13:55	<swingle>	no but i got some completely different results
13:55	<swingle>	they are not even similar
13:55	<aliguori>	swingle: with different benchmarks
13:56	<aliguori>	swingle: are you really surprised that they choose benchmarks that looked relatively good?
13:56	<aliguori>	that's a pretty common occurance
13:56	<swingle>	the problem is that they are wrong by a 60%
13:56	<aliguori>	it's not wrong if you aren't using the exact same benchmark
13:56	<icblenke>	vs the benchmarks you used, probably.
13:56	<swingle>	thats not chosing the right benchmark, thats lieing
13:56	<icblenke>	benchmarks must be compared apples for apples.
13:57	<swingle>	hehe icblenke thx for the explanation ;)
13:57	<aliguori>	swingle: fwiw, most benchmarks aren't related to reality anyway.
13:57	<icblenke>	agreed. benchmarks are like statistics 90% of them are made up on the spot.
13:59	\|-\|	bernarde [~bernarde@32.104.18.240] has quit [Remote host closed the connection]
14:03	<aliguori>	movement: does the opensolaris xen kernel support the supervisor_kernel_mode feature?
14:04	[~]	aliguori is curious about how widely supported it is
14:06	\|-\|	agrif_away changed nick to agriffis
14:06	\|-\|	rpg [~rpg@bi01p1.co.us.ibm.com] has joined #xen
14:09	\|-\|	infernix [~nix@spirit.infernix.net] has quit [Ping timeout: 480 seconds]
14:14	<movement>	aliguori: no
14:14	<movement>	aliguori: tbh I don't know what it's for.
14:14	<aliguori>	movement: it's so that the xen guest can run in ring 0
14:15	<movement>	right, but why?
14:15	<aliguori>	the original idea was that it could be your "bare metal" version of xen
14:15	<aliguori>	this was before paravirt_ops
14:15	<aliguori>	but i think it's also being used to accelerate xen guests on other hypervisors
14:15	<aliguori>	at least, that's my guess
14:16	<waldi>	it seems that vmx does not work with supervisor-mode-kernel
14:16	<movement>	ok, I must be being slow here, presuming that you can boot the same kernel on both metal and xen (which we can't and most likely won't), what use is supervisor_kernel_mode?
14:16	<movement>	since xen won't be there anyway
14:17	<icblenke>	walkdi: eh?
14:17	<aliguori>	movement: i'm trying to determine myself how useful it actually is
14:17	<icblenke>	so, can you run VMWare, virtual-box, kqemu (etc etc) under a supervisor mode kernel?
14:18	<aliguori>	icblenke: a supervisor-mode-kernel is not capable of running other xen guests
14:18	<aliguori>	just the dom0 kernel
14:18	<waldi>	incorrect
14:18	<waldi>	paravitual guests works
14:18	<aliguori>	waldi: it can run other guests?
14:18	<aliguori>	really?
14:18	<icblenke>	so you can run a Xen paravirt setup and vmware/virtual-box/kqemu instead of HVM?
14:18	<aliguori>	waldi: do you actually use supervisor-mode-kernel?
14:20	<icblenke>	I keep hearing people talk about it, but I've found no proof that such a thing is possible.
14:21	<movement>	is the idea some other hypervisor sets a XENFEAT flag? htf would that work?
14:21	[~]	aliguori scratches his head
14:21	<aliguori>	waldi: i'm looking at domctl, and sure enough, i see the check you're talking about
14:22	<aliguori>	movement: i assumed that they used it to run a guest, but i'm starting to doubt that
14:22	<aliguori>	so now, i really don't understand the value of it
14:22	<waldi>	movement: the feature flags are set in the elf header of the binary
14:23	<movement>	waldi: and other hypervisors are supposed to grok that or something?
14:23	<waldi>	icblenke: i don't think that kvm will work in a xen kernel with supervisor-mode
14:23	<movement>	I always assumed it was some kind of debug thing that the xen linux people found useful
14:23	<waldi>	icblenke: the memory management is even then done by xen
14:23	<aliguori>	waldi: so, do you know what it is useful for then?
14:23	<icblenke>	waldi: The question is, will any of the other virtualization systems work?
14:24	<icblenke>	virtual-box has its own ring-0 hypervisor it inserts. I'm guessing this is what vmware does as well.
14:25	<icblenke>	kqemu doesn't take advantage of any vmx/svm, so in theory should work all the same, maybe?
14:25	<waldi>	icblenke: no, vmware just does memory table switch and jump into ring3
14:26	<waldi>	and it change the interrupt handlers
14:26	<icblenke>	vmware workstation does use vmx if available.
14:26	<icblenke>	as of 5.5.x
14:26	<waldi>	not really. the use it for things where they know it is faster
14:26	\|-\|	infernix [~nix@spirit.infernix.net] has joined #xen
14:27	<icblenke>	no clue as to wether vmware-server or player do.
14:27	<icblenke>	they hybridize it, right. only using the vmx when it is advantageous to do so.
14:27	<waldi>	more or less the same
14:28	<icblenke>	waldi: so, I suppose my question remains: is it possble to run VMWare, virtual-box, or kqemu under a "supervisor mode" dom0?
14:29	<waldi>	i don't think so
14:29	<icblenke>	there were folks in the other #xen channel a month back that were claiming otherwise.
14:29	<waldi>	try it
14:30	\|-\|	mastermind [~mastermin@mastermind.kaltenbrunner.cc] has joined #xen
14:31	<icblenke>	http://www.linode.com/xen/irc/logs/xen.log-2006-12-06 <-- the last time I asked this question, 2 months ago.
14:35	<icblenke>	waldi: searching for the kernel .config flag for that now.
14:35	<waldi>	the linux kernel always claims to support this feature
14:36	<icblenke>	it's how the hypervisor treats dom0? looking for the xen compiletime magic to make that happen then.
14:38	<icblenke>	XENFEAT_supervisor_mode_kernel
14:44	<icblenke>	I'm not understanding what ELF has to do with supervisor_mode_kernel.. more googling...
14:44	<aliguori>	icblenke: the kernel marks that it supports that feature by having that string in a particular elf section
14:45	<aliguori>	so that when xen loads the kernel, it can check to see what the kernel supports
14:48	<icblenke>	Domain 0 kernel supports features = { 0000001f }, which is the lower 5 bits set to 1...
14:50	<icblenke>	apparently CONFIG_XEN_COMPAT_030002 makes that available
14:54	<icblenke>	this is beyond me at the moment. I'm just not seeing it. will have to look into it again sometime when I find the time.
14:57	\|-\|	KriiP [kp@yancey.unixworld.dk] has quit [Quit: Leaving]
15:08	\|-\|	agriffis changed nick to agrif_away
15:08	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
15:08	\|-\|	visik7 [~visi@host13-167-dynamic.54-82-r.retail.telecomitalia.it] has joined #xen
15:19	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has joined #xen
15:23	\|-\|	redii [~redii@63.80.44.254] has joined #xen
15:27	\|-\|	swingle [~swingle@221.Red-81-40-154.staticIP.rima-tde.net] has quit [Quit: Leaving]
15:33	\|-\|	mejlholm [~mejlholm@port79.ds1-abc.adsl.cybercity.dk] has quit [Quit: Leaving]
16:03	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has quit [Read error: Connection reset by peer]
16:03	\|-\|	Basic_py [~Basic@gatekeeper.real-time.com] has quit [Quit: Leaving]
16:07	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has joined #xen
16:19	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has quit [Quit: Leaving]
16:28	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has joined #xen
16:30	\|-\|	msinhore [~msinhore@correio.samurai.com.br] has quit [Quit: scp -C msinhore home:msinhore]
16:41	\|-\|	msinhore [~msinhore@correio.samurai.com.br] has joined #xen
17:05	\|-\|	msinhore [~msinhore@correio.samurai.com.br] has quit [Quit: scp -C msinhore home:msinhore]
17:05	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has quit [Quit: Leaving]
17:08	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has joined #xen
17:10	\|-\|	DoberMann changed nick to DoberMann[ZZZzzz]
17:12	\|-\|	nada [~nada@dhcp2.gdrpmibl.acd.net] has joined #xen
17:27	\|-\|	Guest2363 changed nick to matti
17:41	\|-\|	schultmc [~schultmc@zealot.progeny.com] has quit [Quit: Client exiting]
17:41	\|-\|	jerone [~jerone@bi01p1.co.us.ibm.com] has quit [Quit: jerone]
17:42	\|-\|	Basic_py [~Basic@warden.real-time.com] has joined #xen
17:47	\|-\|	schultmc [~schultmc@zealot.progeny.com] has joined #xen
17:49	\|-\|	dansmith [~dan@bi01p1.co.us.ibm.com] has quit [Quit: Leaving]
17:52	\|-\|	stekloff [~stekloff@pool-71-245-98-220.ptldor.fios.verizon.net] has quit [Quit: Ex-Chat]
17:52	\|-\|	aliguori [~anthony@bi01p1.co.us.ibm.com] has quit [Quit: Leaving]
18:06	\|-\|	MackAck [~Mack@bzq-88-155-109-230.red.bezeqint.net] has quit [Ping timeout: 480 seconds]
18:12	\|-\|	nada [~nada@dhcp2.gdrpmibl.acd.net] has quit [Read error: Connection reset by peer]
18:13	\|-\|	kurios [~nada@dhcp2.gdrpmibl.acd.net] has joined #xen
18:16	\|-\|	aw [~awilliam@156.153.254.69] has quit [Remote host closed the connection]
18:25	<visik7>	hi
18:26	<visik7>	is paravirtualized network and block device driver for windows already available ?
18:28	\|-\|	hbaum [~hbaum@bi01p1.co.us.ibm.com] has quit [Quit: Client exiting]
18:31	\|-\|	aliguori [~anthony@cpe-70-112-17-156.austin.res.rr.com] has joined #xen
18:53	\|-\|	rharper [~rharper@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
18:53	\|-\|	aw [~awilliam@c-24-9-84-32.hsd1.co.comcast.net] has joined #xen
19:02	\|-\|	agrif_away changed nick to agriffis
19:14	\|-\|	hollisb [~hollisb@bi01p1.co.us.ibm.com] has quit [Quit: leaving]
19:14	\|-\|	rpg [~rpg@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
19:16	\|-\|	dansmith [~dan@gw0.danplanet.com] has joined #xen
19:30	\|-\|	jimix [~jimix@c-67-189-187-152.hsd1.ma.comcast.net] has quit [Quit: jimix]
19:39	\|-\|	redii [~redii@63.80.44.254] has quit [Ping timeout: 480 seconds]
19:46	\|-\|	hollisb [~hollisb@user-0vvdf2d.cable.mindspring.com] has joined #xen
20:00	\|-\|	kurios [~nada@dhcp2.gdrpmibl.acd.net] has quit [Ping timeout: 480 seconds]
20:12	\|-\|	rharper [~rharper@r74-193-70-77.pfvlcmta01.grtntx.tl.dh.suddenlink.net] has joined #xen
20:12	<icblenke>	no.
20:13	\|-\|	pvanhoof [~pvanhoof@d54C0EE14.access.telenet.be] has quit [Remote host closed the connection]
20:18	\|-\|	MackAck [~Mack@80.178.86.105.adsl.012.net.il] has joined #xen
20:24	\|-\|	redii [~redii@c-67-182-108-193.hsd1.ca.comcast.net] has joined #xen
20:31	\|-\|	agriffis changed nick to agrif_away
20:46	\|-\|	agrif_away changed nick to agriffis
20:47	\|-\|	redii [~redii@c-67-182-108-193.hsd1.ca.comcast.net] has quit [Quit: redii]
21:01	\|-\|	visik7 [~visi@host13-167-dynamic.54-82-r.retail.telecomitalia.it] has quit [Read error: Connection reset by peer]
21:02	\|-\|	visik7 [~visi@host13-167-dynamic.54-82-r.retail.telecomitalia.it] has joined #xen
21:39	\|-\|	kalden [~kalden@fac34-3-82-231-26-11.fbx.proxad.net] has quit [Quit: Parti]
21:40	\|-\|	infernix [~nix@spirit.infernix.net] has quit [Remote host closed the connection]
21:40	\|-\|	infernix [~nix@spirit.infernix.net] has joined #xen
22:11	\|-\|	hivemind [~slaggsync@static-72-91-2-66.tampfl.fios.verizon.net] has joined #xen
22:12	<hivemind>	i'm trying to get a hvm to load, and i get "error 12: unable to allocate memory" i've got 2 other PV'ed DOMU's running.
22:13	<hivemind>	over a gig of ram free
22:14	\|-\|	hollisb [~hollisb@user-0vvdf2d.cable.mindspring.com] has quit [Quit: Leaving]
22:20	<hivemind>	anyone?
22:39	\|-\|	redii [~redii@c-67-182-108-193.hsd1.ca.comcast.net] has joined #xen
22:59	\|-\|	VS_ChanLog [~stats@ns.theshore.net] has left #xen [Rotating Logs]
22:59	\|-\|	VS_ChanLog [~stats@ns.theshore.net] has joined #xen
23:21	\|-\|	hivemind [~slaggsync@static-72-91-2-66.tampfl.fios.verizon.net] has quit [Quit: leaving]
23:26	\|-\|	DoberMann_ [~james@AToulouse-156-1-65-253.w90-16.abo.wanadoo.fr] has joined #xen
23:28	\|-\|	DoberMann[ZZZzzz] [~james@AToulouse-156-1-56-199.w90-16.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
23:49	\|-\|	sputhenp [~sputhenp@202.80.58.210] has joined #xen
---	Log	closed Tue Feb 13 00:00:31 2007