TheGrebs.com:IRC Logs:#xen:2005:05:24

---	Log	opened Tue May 24 00:00:52 2005
00:17	---	<<-- rread [~robert@datarithm.com] has quit (Remote host closed the connection)
00:47	tessier_	I would ask him but he is not available at the moment. The only thing I have heard about these blades is second hand from someone else.
00:48	tessier_	I think good hardware is worth paying for if you have the money.
00:48	tessier_	Unfortunately like many employers mine does not.
00:48	tessier_	And always goes for the cheapest solutions.
00:48	tessier_	They just lost an hd on the accounting/inventory system. Fsck'd right in the arse they are.
01:15	---	<<-- liberie [~root@dsl027-160-029.atl1.dsl.speakeasy.net] has quit (Ping timeout: 480 seconds)
01:16	---	---> cartel_ [~cartel@shinobi.thoughtcrime.org.nz] has joined #xen
01:16	cartel_	hey all
01:16	cartel_	looking for some doc / help on pci device exclusion and pci = [ 'x,y,z' ]
01:32	cartel_	nm, seems to be working now
01:34	---	<<-- Tv [~Tv@hq.inoi.fi] has quit (Quit: Client exiting)
01:54	---	---> woody__ [~woody@bi01p1.co.us.ibm.com] has joined #xen
01:54	---	<<-- woody_ [~woody@bi01p1.co.us.ibm.com] has quit (Read error: Connection reset by peer)
01:58	---	---> tessier [~treed@203.210.214.65] has joined #xen
02:08	mael	hi
02:21	---	---> betonamu [~Administr@dhcp-243-187.mag.keio.ac.jp] has joined #xen
02:23	betonamu	hello, anybody here? i run "xm list", and it shows one VCPU column, but what does VCPU mean in xen?
02:32	mael	you must run "unstable", no?
02:32	mael	I think this mean "Virtual CPU"
02:33	betonamu	mael: yes, unstable
02:33	mael	xen 3 will support multiple cpu for domains
02:33	mael	so they added the VCPU stuff
02:33	mael	(if I understood correctly)
02:33	betonamu	multiple cpu for domain? that means one domain runs on multiple cpu?
02:34	mael	It mean you can emulate SMP system
02:34	mael	the domain sees 'n' VCPU
02:34	mael	and the VCPU can be real CPU or not
02:35	betonamu	ah that is great. how many cpu can i set for one domain? any limition?
02:35	betonamu	and how to set that?
02:35	mael	well, I don't know :)
02:35	mael	I haven't been playing with unstable yet :)
02:36	mael	but I guess some devs here might tell
02:36	betonamu	that is nice to emulate SMP. so developer can write program for smp without having the real hardware :)
02:37	mael	betonamu: domain/context switching overhead will give you a limit
02:37	mael	and I presume this will happen very fast
02:37	betonamu	how to calculate/estimate that limit?
02:37	mael	you should test and tell me!
02:38	betonamu	:) the problem is that i dont know how to test
02:39	betonamu	looks like there are smt like vcpu_to_cpu setting in sxp
02:39	betonamu	and "cpumap"
02:40	mael	I hope I was right anyway, because as I told you, this is just guessing
02:40	betonamu	yeah but that makes sense to me
02:40	betonamu	or it works this way: this is not to emulate SMP
02:41	betonamu	if you have N cpu, you can configure to give some to a domain
02:41	betonamu	like CPU1, CPU3 and CPU4 - not CPU2
02:41	betonamu	i guess this is more practical
02:42	betonamu	virtual CPUs are always count from 1 up
02:42	mael	betonamu: well I'm not sure, because I think cpu pinning already existed in xen 2
02:43	mael	I'm download unstable sources...
02:43	mael	+ing
02:43	betonamu	yes i am reading code rite now to figure this out
02:46	mael	there a #define MAX_VIRT_CPUS 32 in xen/include/public/xen.h
02:47	betonamu	hmm.. just found a small bug in xm/main.py :)
02:47	mael	and MAX_VIRT_CPUS seem to be used for the vcpu stuff
02:48	betonamu	to make (a patch) or not to make :)
02:53	betonamu	mael: if i want to compile a kernel module for domU, what i must do?
02:53	mael	betonamu: there's a vcpu.c in ia64 arch only
02:54	mael	betonamu: recompile using arch=xen
02:54	betonamu	mael: ah... but finally what does that vcpu do? probably we must wait for some developers to popup here to ask
02:55	betonamu	mael: then copy to domU/lib/modules/<version> ?
02:56	mael	betonamu: probably
02:56	mael	I have only used monolithic kernel so far
02:57	betonamu	any chance to compile the module from inside domU?
02:57	mael	I think the domU must be able to load the module, so it must be on a FS it has access to
02:57	mael	betonamu: why not?
02:57	betonamu	yes of course the module must stay inside domU filesystem
02:58	mael	the domU/lib/... stuff was not very clear :)
02:58	betonamu	mael: what clear?
02:59	mael	I was not sure about the meaning
03:00	mael	clear=obvious
03:00	betonamu	i meant what do you mean by "clear"?
03:01	betonamu	that is for kernel module of domU. that is very clear
03:01	mael	I couldn't decide where you planned to put the module :)
03:01	betonamu	of course in the normal place
03:02	betonamu	the kernel code to load module doesnt change, rite?
03:02	mael	I don't think so
03:02	betonamu	(only machine-dependent code are changed for xen para-virtualization)
03:02	betonamu	i am sure
03:05	betonamu	lots of linux syscall remain unchanged in xen. and one of them (sys_create_module) is used to load kernel module
03:05	---	---> Shaun [ndci@ip68-111-70-41.oc.oc.cox.net] has joined #xen
03:07	---	---> enum [~Administr@ip-207-145-127-226.lax.megapath.net] has joined #xen
03:08	enum	yo, this seems to be a serious issue. I am using xen 2.0.6, and when using my normal debian kernel everything is fine. But with xen in dom0 I am getting an error when I do a lot of hd activity
03:08	enum	hda: Lost Interrupt
03:11	enum	I feel that I should mention this exact problem is happening on 2 seperate machines
03:18	betonamu	enum: you should post to devel list, so some gurus might help you
03:19	mael	ping rusty (strange you're still there at that time :))
03:37	rusty	mael: ack
03:38	betonamu	rusty: what is vcpu in unstable?
03:38	rusty	betonamu: virtual number of CPUs the domain can use.
03:39	mael	rusty: does it works on all arch?
03:39	rusty	mael: Don't know, actually.
03:39	mael	ok
03:40	mael	I think we should send a mail on xen-devel to ask for a new entry in the FAQ
03:42	mael	it seems to be an interesting feature in the next release but there's almost no info about it in the documentation/presentation nor on the devel ML
03:43	betonamu	rusty: does that mean we can emulate SMP on an UP?
03:45	rusty	betonamu: that's part of it, yes.
03:45	mael	betonamu: aliguori and unriel have been doing test with vcpu > cpu
03:45		* mael have been avidly digging in the devel ML
03:47	---	<<-- rusty [~rusty@bh02i525f01.au.ibm.com] has quit (Quit: Client exiting)
03:48	betonamu	sweet :)
03:50	tessier_	mael: The VCPU can be real CPU or not? So I can sell virtual hosting services and tell my customers they are on an 8-way box and charge them a lot of money? :)
03:50	mael	tessier: yeah it seems this would be possible
03:51	betonamu	tessier_ : great way to make money :)
03:51	mael	betonamu: http://lists.xensource.com/archives/cgi-bin/namazu.cgi?query=vcpu&submit=Search%21&idxname=xen-devel&max=20&result=normal&sort=score
03:51		* tessier_ thinks evil thoughts
03:51		* tessier_ rents time to the DOE on his 1024-way box
03:51	betonamu	mael: nice
03:52	betonamu	will take a closer look
04:08	---	---> athomas [~athomas@ppp-0-69.lond-a-3.access.uk.tiscali.com] has joined #xen
05:01	tessier_	Anyone using Xen on RHEL4?
05:01	tessier_	My xen kernel installed from binaries seems to be missing modules required to see my / on lvm
05:02	tessier_	Wondering if there is an easy fix or if I need to recompile my own xen kernel from the source
05:30	---	<<-- betonamu [~Administr@dhcp-243-187.mag.keio.ac.jp] has quit (Ping timeout: 480 seconds)
06:27	---	User: *** lilo_ is now known as lilo
06:28	---	---> hebutterworth [~harry@blueice3n1.uk.ibm.com] has joined #xen
07:47	---	---> rusty [~rusty@ppp60-148.lns1.cbr1.internode.on.net] has joined #xen
07:48	rusty	I suspect you are not really on IRC at 5am your time...
07:52	mael	dunno
07:52	knewt	i was this morning :/
07:54	mael	knewt: so, how is your dm work going?
08:06	---	<<-- tessier [~treed@203.210.214.65] has quit (Ping timeout: 480 seconds)
08:17	---	---> tessier [~treed@222.253.72.54] has joined #xen
08:19	---	---> katzj [~katzj@wlanconf-nat-pool-bos.redhat.com] has joined #xen
08:31	---	Netsplit iridium.oftc.net <-> jupiter.oftc.net quits: @Sir_Ahzz, cartel, pwagland, hbaum
08:31	---	Netsplit iridium.oftc.net <-> jupiter.oftc.net quits: hebutterworth, grifferz, knewt, katzj, Shaun, JViz, viewbee, woody__, tim, cartel_, (+6 more, use /NETSPLIT to show all of them)
08:31	---	Netsplit iridium.oftc.net <-> jupiter.oftc.net quits: tessier_, demon, surriel, schweeb, lilo, muli_, cfreak
08:32	---	Netsplit over, joins: cartel, pwagland, hbaum, Sir_Ahzz
08:32	---	Netsplit over, joins: Rai
08:32	---	Channel: ServerMode/#xen [+o Sir_Ahzz] by jupiter.oftc.net
08:33	---	Netsplit over, joins: Shaun, JViz, Hunger, katzj, tessier, hebutterworth, woody__, cartel_, movement, knewt (+5 more)
08:33	---	Netsplit over, joins: lilo, schweeb, muli_, surriel, tessier_, demon, cfreak
08:48	---	<<-- knewt [~jmb@p213.54.100.97.tisdip.tiscali.de] has quit (Ping timeout: 480 seconds)
08:57	---	---> rharper [~rharper@pixpat.austin.ibm.com] has joined #xen
08:57	---	---> xai [~pasta@cpe-70-112-17-10.austin.res.rr.com] has joined #xen
09:04	---	---> knewt [~jmb@p213.54.94.151.tisdip.tiscali.de] has joined #xen
09:08	---	<<-- rusty [~rusty@ppp60-148.lns1.cbr1.internode.on.net] has quit (Quit: Client exiting)
09:08	---	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/
09:15	---	---> hollis [~hollis@user-0vvde2g.cable.mindspring.com] has joined #xen
09:21	---	---> liberie [~root@dsl027-160-029.atl1.dsl.speakeasy.net] has joined #xen
09:21	---	---> yarihm [~yarihm@vpn-global-015-dhcp.ethz.ch] has joined #xen
09:47	---	<<-- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit (Quit: leaving)
09:50	---	<<-- yarihm [~yarihm@vpn-global-015-dhcp.ethz.ch] has quit (Quit: Leaving)
10:07	---	---> jimix [~jimix@ip13.194.susc.suscom.net] has joined #xen
10:23	knewt	erm. i somehow don't think there's /any/ chance i'll be able to make it to Canada, when i couldn't even make it to the UK :(
10:38	---	---> aliguori- [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen
11:32	---	---> hollis [~hollis@pixpat.austin.ibm.com] has joined #xen
11:54	---	<<-- athomas [~athomas@ppp-0-69.lond-a-3.access.uk.tiscali.com] has quit (Quit: Leaving)
11:55	---	<<-- jimix [~jimix@ip13.194.susc.suscom.net] has quit (Quit: jimix)
11:57	---	---> DEac- [~deac@xdsl-213-196-200-210.netcologne.de] has joined #xen
12:08	---	<<-- hebutterworth [~harry@blueice3n1.uk.ibm.com] has quit (Quit: Leaving)
12:29	---	---> bunoc [~bun@YahooBB219206220072.bbtec.net] has joined #xen
12:45	---	Netsplit oxygen.oftc.net <-> jupiter.oftc.net quits: tessier_, aliguori-, cfreak
12:45	---	Netsplit oxygen.oftc.net <-> jupiter.oftc.net quits: DEac-, grifferz, rharper, knewt, katzj, demon, Shaun, JViz, surriel, schweeb, (+13 more, use /NETSPLIT to show all of them)
12:45	---	Netsplit oxygen.oftc.net <-> jupiter.oftc.net quits: Rai, pwagland, hbaum
12:46	---	Netsplit over, joins: schweeb, cartel, DEac-, aliguori-, knewt, rharper, surriel, lilo, cfreak, demon (+18 more)
12:46	---	Netsplit over, joins: Rai
12:46	---	Channel: ServerMode/#xen [+o Sir_Ahzz] by jupiter.oftc.net
13:06	---	---> niv [~nivedita@bi01p1.co.us.ibm.com] has joined #xen
13:33	---	<<-- aliguori- [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit (Ping timeout: 480 seconds)
13:40	---	---> aliguori- [~anthony@pixpat.austin.ibm.com] has joined #xen
13:56	---	<<-- pwagland [~paul@kungfucoder.org] has quit (Quit: BitchX-1.1-final -- just do it.)
14:01	---	<<-- bunoc [~bun@YahooBB219206220072.bbtec.net] has quit (Remote host closed the connection)
14:06	---	<<-- soffi [~soffi@Tolvudeild-SC-gw.vv.simnet.is] has quit (Quit: Leaving)
14:21	---	<<-- aliguori- [~anthony@pixpat.austin.ibm.com] has quit (Ping timeout: 480 seconds)
14:27	---	---> mikenc [~mikenc@199.108.226.254] has joined #xen
14:29	---	---> aliguori- [~anthony@pixpat.austin.ibm.com] has joined #xen
14:57	---	<<-- aliguori [~anthony@pixpat.austin.ibm.com] has quit (Quit: Leaving)
15:08	---	User: *** aliguori- is now known as aliguori
15:18	---	<<-- aliguori [~anthony@pixpat.austin.ibm.com] has quit (Read error: Connection reset by peer)
15:19	---	---> aliguori [~anthony@pixpat.austin.ibm.com] has joined #xen
15:31	---	---> soffi [~soffi@proxy.du.vdsl.is] has joined #xen
15:55	---	<<-- schweeb [~chris@schweeb.org] has quit (Ping timeout: 480 seconds)
16:16	xai	network? why is everyone leaving?
16:16	---	<<-- katzj [~katzj@wlanconf-nat-pool-bos.redhat.com] has quit (Read error: Connection reset by peer)
16:16	xai	or joining.
16:50	---	<<-- soffi [~soffi@proxy.du.vdsl.is] has quit (Quit: This computer has gone to sleep)
16:58	knewt	mael: been doing some other stuff just lately, but i'm going to get back to it shortly :)
16:59	---	---> soffi [~soffi@proxy.du.vdsl.is] has joined #xen
17:17	---	<<-- hollis [~hollis@pixpat.austin.ibm.com] has quit (Quit: leaving)
17:22	---	<<-- aliguori [~anthony@pixpat.austin.ibm.com] has quit (Quit: Leaving)
17:24	---	---> monski [~monrad@213083190130.sonofon.dk] has joined #xen
17:38	---	<<-- rharper [~rharper@pixpat.austin.ibm.com] has quit (Quit: Leaving)
17:39	---	<<-- soffi [~soffi@proxy.du.vdsl.is] has quit (Quit: This computer has gone to sleep)
17:46	---	<<-- knewt [~jmb@p213.54.94.151.tisdip.tiscali.de] has quit (Ping timeout: 480 seconds)
17:49	---	---> knewt [~jmb@zeus.pimb.org] has joined #xen
17:55	---	<<-- knewt [~jmb@zeus.pimb.org] has quit (Remote host closed the connection)
17:55	---	---> knewt [~jmb@zeus.pimb.org] has joined #xen
17:56	---	User: *** monski is now known as monrad
18:00	---	<<-- knewt [~jmb@zeus.pimb.org] has quit (Remote host closed the connection)
18:05	---	---> knewt [~jmb@p213.54.74.1.tisdip.tiscali.de] has joined #xen
18:18	---	---> aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen
18:29	---	---> knewt_ [~jmb@zeus.pimb.org] has joined #xen
18:29	---	<<-- tessier [~treed@222.253.72.54] has quit (Ping timeout: 480 seconds)
18:30	---	---> soffi [~soffi@proxy.du.vdsl.is] has joined #xen
18:32	---	<<-- knewt [~jmb@p213.54.74.1.tisdip.tiscali.de] has quit (Ping timeout: 480 seconds)
18:34	---	<<-- knewt_ [~jmb@zeus.pimb.org] has quit (Remote host closed the connection)
18:39	---	---> knewt [~jmb@p213.54.69.164.tisdip.tiscali.de] has joined #xen
18:42	---	---> katzj [~katzj@c-66-30-206-55.hsd1.ma.comcast.net] has joined #xen
18:52	---	<<-- knewt [~jmb@p213.54.69.164.tisdip.tiscali.de] has quit (Quit: leaving)
18:52	---	---> knewt [~jmb@p213.54.69.164.tisdip.tiscali.de] has joined #xen
19:11	---	<<-- niv [~nivedita@bi01p1.co.us.ibm.com] has quit (Quit: Quitting)
19:14	---	---> niv [~nivedita@bi01p1.co.us.ibm.com] has joined #xen
19:20	---	---> rusty [~rusty@bh02i525f01.au.ibm.com] has joined #xen
19:23	rusty	niv: ping?
19:30	rusty	aliguori: ping?
19:30	niv	heya rusty
19:30	niv	just sent you an email
19:31	rusty	niv: yes, so I jumped on IRC.
19:31	niv	how goes it?
19:31	rusty	Just writing my OLS paper.
19:32	niv	What's your topic?
19:32	rusty	Anyway, Mike has switched xend to store its persistent data in the store, and I'm trying to get the xenbus driver working.
19:32	aliguori	hey rusty
19:32	rusty	niv: nfsim
19:32	niv	Ah, I remember now
19:32	rusty	aliguori: When I ^] the console I can't reconnect.
19:33	aliguori	in vm-tools? yeah, in the latest version it connects the controlling terminal.. i have to add an option to connect a pty
19:33	aliguori	rusty: is the store changing in the near future?
19:33	rusty	aliguori: and I wasn't sure how to use the -B option: I want /dev/hdb to appear as /dev/sda inside.
19:33	aliguori	rusty: what version are you using?
19:33	rusty	aliguori: 0.0.4
19:33	aliguori	cvs is quite different.. there's a vm-vbd command
19:33	niv	rusty: just what I was about to ask, and is it settled?
19:33	aliguori	ahh, i see
19:34	---	<<-- tessier_ [~treed@203.210.216.1] has quit (Read error: Connection reset by peer)
19:34	niv	aliguori: when are you releasing another version of vm-tools?
19:34	aliguori	niv: the plan is tomorrow
19:34	aliguori	i just got save/restore working
19:35	niv	cool!
19:35	rusty	niv: well, it's kinda settled. We need to get a prototype so we can properly critique the layout.
19:36	aliguori	rusty: are the tools responsible for seeting up the data layout?
19:36	rusty	aliguori: your code is also not really designed to drop into the xen tree AFAICT.
19:36	aliguori	no, it's not
19:36	niv	rusty: ok, also, what else will be impacted?
19:36	knewt	rusty: heh, you're going to canada then i guess?
19:37	rusty	aliguori: Yes, but the layout is very much in flux.
19:37	rusty	niv: well, the devices will be impacted, and the tools. Not much else.
19:37	rusty	knewt: You mean OLS? Yes.
19:38	rusty	aliguori: looks like we're going to go for a UUID for each domain. I'll probably add that as an arg to xs_introduce, and then the domain can ask the store "who am I?"
19:39	aliguori	rusty: how's the uuid going to be generated?
19:39	niv	xenbus being neither in xen nor a bus ;)
19:39	knewt	rusty: yeah. if i couldn't afford to get to the uk, i don't think i'm going to be able to make canada unfortunately :(
19:39	rusty	aliguori: the tools make it up. Then they set up /domain/<uuid>/....
19:40	rusty	aliguori: the killer is that people want names in Kanji, so using names was going to be nasty.
19:40	aliguori	rusty: of course
19:41	rusty	aliguori: and domain ids are local, and people were talking about clusters, so I gave in 8)
19:41	aliguori	names are not going to be enough too.. i imagine people are going to want groups too
19:41	aliguori	rusty: yeah, this was inevitable :-)
19:41	rusty	aliguori: Yeah, tools problem 8)
19:42	aliguori	as it should be :-)
19:43	aliguori	rusty: does this mean the store is not going to just deal in strings or are you going to standardize on utf8 or something?
19:44	rusty	aliguori: it can handle arbitrary data, although everything so far is a string.
19:47	aliguori	rusty: so does that mean paths must be ascii?
19:47	aliguori	yeah, i guess that makes sense
19:47	rusty	Paths must be a-zA-Z0-9-
19:47	niv	rusty: is the layout still going to resemble the unreadable mess that some other implementations generate?
19:47	aliguori	oh, ok
19:47	rusty	And I'm thinking about removing A-Z.
19:47	rusty	niv: I am fighting against that one now.
19:48	rusty	niv: I have the OF tree as an example of how not to do it.
19:48	niv	rusty: yep, that's the one I saw, and really, seemed ghastly
19:48	rusty	niv: I'm trying to scare them straight 8)
19:49	niv	rusty: good :) more power to ya!
19:50	rusty	aliguori: basically xs_linux_build takes two new args: a evtchn and a memory_t *: the evtchn is the evtchn set up by the tool, and the memory_t is the mfn of the shared page set up by the xc_linux_build() function
19:50	---	---> tessier_ [~treed@222.253.77.38] has joined #xen
19:51	rusty	aliguori: ie (from my hacked vm-tools-0.0.4):
19:51	rusty	+ /* FIXME: handle errors. */
19:51	rusty	+ xsh = xs_daemon_open();
19:51	rusty	+ fprintf(stderr, "Opened connection to daemon: %p %s\n",
19:51	rusty	+ xsh, xsh ? "" : strerror(errno));
19:51	rusty	+ if (xsh) {
19:51	rusty	+ if (xc_evtchn_bind_interdomain(xc_handle, DOMID_SELF, domid,
19:51	rusty	+ &store_local_evtchn,
19:51	rusty	+ &store_evtchn) != 0)
19:51	rusty	+ perror("xc_evtchn_bind_interdomain failed");
19:52	rusty	+
19:52	rusty	+ if (build_domain(xc_handle, domid, store_evtchn, &store_mfn) == -1) {
19:52	---	<<-- monrad [~monrad@213083190130.sonofon.dk] has quit (Quit: Leaving)
19:52	aliguori	yeah, i've got a store locally and have a version of vm-tools that builds against it :-)
19:52	aliguori	i haven't committed it to our cvs because we're still going to ship against xen-unstable (which doesn't include the store yet)
19:53	aliguori	rusty: do you know what's going on in mike wray's tree btw? it looks like christian and steve are doing a ton of xend work in xen-unstable
19:53	rusty	aliguori: the store will go in RSN. Other than the change to the xs_introduce function it's unchanged fotr weeks
19:53	rusty	aliguori: yes, Mike mentioned that. He's going to have some merge fun soon.
19:55	aliguori	rusty: as soon as it goes in, we'll do a store-friendly release :-) i'll be very happy to get rid of the block device setup code once the store can do that
19:55	niv	rusty: so we are going to have something fairly soon that will incorporate all of these various efforts?
19:56	rusty	niv: from the call last night, some of the stuff will just go into the main tree. Mike and I need to actually get the block devices using this so we can nail down the correct layout of the store.
19:58	rusty	niv: The various parts are: (1) the xenstore code itself, (2) the xc_linux_build and start_info_t changes for the store shared page, (3) the tools changes to use the new xc_linux_build, (4) the tools changes to actually populate the store, (5) the driver changes to actually use the store.
19:59	rusty	(1) can go in today. (2) needs (3) so compile doesn't break, but hopefully late this week. (4) and (5) are an iterative process as the layout changes.
20:01	niv	rusty: I don't have access to Mike Wray's tree - is this publicly available? I wanted the API that the drivers will use
20:02	rusty	niv: ah, that kinda doesn't exist yet. I'm writing the xenbus driver, but it as the comment at the top says:
20:02	rusty	"When this code grows up, it's going to be a bus driver!"
20:03	aliguori	:-)
20:03	niv	heh ;)
20:04	rusty	niv: they may also query the store directly. THe interface is the same as xs.h in userspace.
20:04	niv	ah, good to know..
20:07	rusty	niv: I can send you my current kernel code.
20:07	niv	rusty: yes, please! many thanks..
20:13	rusty	niv: sent.
20:14	niv	I accidentally booted to an old version of Xen (April 12th tarball) and was running tests without noticing and found that all udp datagrams get dropped if less than 1470bytes thereabouts
20:14	niv	going from dom1 to remote host
20:15	niv	rusty: thanks! rcvd..
20:16	niv	Wonder if there is other weirdness that is getting by our testing...
20:18	---	<--- niv [~nivedita@bi01p1.co.us.ibm.com] has left #xen (Leaving)
20:20	---	---> Vip- [~weasel@S0106000c41cd470e.cg.shawcable.net] has joined #xen
20:20	Vip-	Why am I getting an "unable to mount root fs via NFS" when booting the xen.gz?
20:21	aliguori	Vip-: you don't have a root= parameter in the kernel module line probably
20:21	Vip-	aligouri: just a sec, this is what I have:
20:22	---	---> bunoc [~bun@YahooBB219206220072.bbtec.net] has joined #xen
20:22	Vip-	title Fedora Core (2.6.11-initrd-xen0)
20:22	Vip-	root (hd0,1)
20:22	Vip-	kernel /boot/xen.gz dom0_mem=262144 noreboot
20:22	Vip-	module /boot/vmlinuz-2.6.11.10-xen0 ro root=/dev/sda2
20:22	Vip-	module /boot/initrd-2.6.11.10-xen0.img
20:23	Vip-	I've noticed on boot that sda2 doesn't show up as found. I can see hdc , the cdrom drive, but nothing else
20:23	aliguori	Vip-: oh
20:24	Vip-	And I never get to a line saying "booting domain 0" or something similar
20:24	aliguori	i bet you didn't enable your scsi controller in the xen kernel config
20:24	Vip-	ali: :-) I got that too...just a sec :-)
20:24	aliguori	that error your getting is a linux kernel one
20:24	Vip-	CONFIG_SCSI_SATA=y
20:25	aliguori	it means that it cannot find the root device.. usually, it occurs when root= isn't set at all
20:25	Vip-	CONFIG_SCSI_SATA_PROMISE=m
20:25	Vip-	Yeah, hrmmm....doh!
20:25	aliguori	does your initrd have that module in it?
20:25	aliguori	how did you make that initrd?
20:25	Vip-	Uh...does this need to be enabled? "# CONFIG_BLK_DEV_IDE_SATA is not set"
20:26	aliguori	i dunno all that much about scsi configuration
20:26	Vip-	mkinitrd /boot/initrdsomething <kernel-rev-to-use>
20:26	aliguori	but i'd recommend compilng your scsi drivers in statically
20:26	aliguori	i've had mixed results with fedora's mkinitrd script and xen
20:27	Vip-	ok, I'll try a static compile.
20:27	aliguori	still, that's not quite the error i'd expect if that was the problem
20:27	aliguori	it usually says something like invalid block device
20:27	Vip-	Yeah, in this case it tries NFS boot.
20:28	Vip-	Like I said, I don't see sda2 recognized during the boot
20:33	Vip-	Should I remove the initrd line from the boot menu?
20:34	aliguori	the initrd is necessary in fedora because of the way they have their udev setup
20:34	Vip-	ah , ok
20:37	---	<<-- soffi [~soffi@proxy.du.vdsl.is] has quit (Quit: This computer has gone to sleep)
20:40	---	---> tessier [~treed@222.253.85.43] has joined #xen
20:41	Vip-	ali: so make no changes to the menu.lst for grub?
20:41	---	---> soffi [~soffi@proxy.du.vdsl.is] has joined #xen
20:41	aliguori	Vip-: yeah, you should be ok with what you have.. have you had any additional luck?
20:41	Vip-	I just finished the static driver compile...doing mkinitrd
20:42	Vip-	No module sata_via found for kernel 2.6.11.10-xen0, aborting
20:43	knewt	it's easy enough to work around the udev issues requiring an initrd. just create the minimal necessary device nodes
20:44	knewt	i admit i have to use an initrd, but that's because my boot device doesn't even exist until it's been "discovered" by userland
20:45	Vip-	knewt: what do you mean "minimal necessary device nodes"?
20:49	knewt	i believe that /dev/console and /dev/null are the minimal required. might be mis-remembering though
20:51	knewt	oh, you might need /dev/zero actually i think
20:52	Vip-	ok
20:53	Vip-	knewt: I have a sata drive, any ideas on how to get it working?
20:54	knewt	you've built the necessary drivers into the kernel?
20:54	Vip-	knewt: yes
20:54	Vip-	CONFIG_SCSI_SATA=y
20:55	Vip-	CONFIG_SCSI_SATA_PROMISE=y
20:55	knewt	verified that the same config works on a non-xen kernel?
20:55	Vip-	oh yeah...heh...not yet :-)
20:55	Vip-	Probably a good idea, huh?
20:55	aliguori	Vip-: if you've got the time, i'd recommend taking the fedora kernel config and porting it to xen
20:56	aliguori	that's what i did for my gentoo system and then i just saved that config...
20:56	aliguori	there's only a handful of extra xen specific parameters and then some normal ones that xen doesn't like (mostly isa related it seems)
20:57	Vip-	ali: I've tried that. When I do that it wants to take this x86_64 and compile i386 for it. ???
20:57	aliguori	??
20:57	aliguori	i'm not sure i parse that statement
20:57	Vip-	ali: I copied the /boot/config-<ver> to .config and ran "make ARCH=xen oldconfig"
20:57	aliguori	ohhh, i did it by hand
20:58	rusty	aliguori: you going to OLS?
20:58	Vip-	ali: then when I do a make ARCH=xen world , it goes off and sets all the architecture stuffs to i386 instead of x86_64
20:58	aliguori	rusty: yup
20:58	aliguori	Vip-: oh! you're on x86_64?
20:58	aliguori	heheh, ok, all bets are off then :-)
20:58	Vip-	ali: yeah
20:59	Vip-	ali: hahaha
20:59	aliguori	oh, that explains it :_0
20:59	aliguori	it's entirely possible that the scsi drivers are broken somehow
20:59	rusty	aliguori: cool. I'll be in Austin a couple of weeks before, too.
20:59	knewt	i have a pretty minimal configuration on my server. my laptop (standard ubuntu-built kernel right now) is 493 y, 1386 m. my server is 228 y, 0 m. g
20:59	Vip-	I got emails from some developers that it should work.
20:59	aliguori	rusty: cool!
21:00	Vip-	ali: I know...x86_64 is still in dev cycle :-)
21:00	aliguori	Vip-: I was hacking on x86_64 till the wee hours of the morning last night.. i don't think it stayed up without oops for more than 15 minutes at a time
21:00	knewt	Vip-: for now just run as if it's a bog standard i386 box
21:01	aliguori	Vip-: what kind of machine is this? my x86_64 machine has SATA drives and that's not a problem
21:02	Vip-	ali: AMD 64 3000+ w/ Asus K8V motherboard
21:02	Vip-	ali: really? heh
21:03	aliguori	yeah, i was hoping you'd say it was an intellistation :-)
21:03	Vip-	what's intellistation?
21:04	aliguori	it's an IBM machine
21:07	aliguori	ok, i'm gonna head in for the night..
21:07	---	<<-- aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit (Quit: Leaving)
21:07	Vip-	too many changes...I'm going to revert back
21:07	Vip-	to original settings and start over
21:07	---	<<-- mikenc [~mikenc@199.108.226.254] has quit (Read error: Connection reset by peer)
21:12	---	---> tim_ [~tim@cpe-66-67-139-238.rochester.res.rr.com] has joined #xen
21:12	---	<<-- tim [~tim@cpe-66-67-139-238.rochester.res.rr.com] has quit (Read error: Connection reset by peer)
21:19	---	<<-- Vip- [~weasel@S0106000c41cd470e.cg.shawcable.net] has quit (Quit: Leaving)
21:26	bunoc	cool. any presentations on Xen at OLS this year?
21:33	---	<<-- soffi [~soffi@proxy.du.vdsl.is] has quit (Quit: Leaving)
22:15	---	<<-- dwh [~sc@c-24-21-82-55.hsd1.or.comcast.net] has quit (Ping timeout: 480 seconds)
22:56	---	<<-- knewt [~jmb@p213.54.69.164.tisdip.tiscali.de] has quit (Ping timeout: 480 seconds)
22:59	---	<--- VS_ChanLog [~stats@ns.theshore.net] has left #xen (Rotating Logs)
22:59	---	---> VS_ChanLog [~stats@ns.theshore.net] has joined #xen
23:20	---	<<-- enum [~Administr@ip-207-145-127-226.lax.megapath.net] has quit (Ping timeout: 480 seconds)
---	Log	closed Wed May 25 00:00:58 2005