TheGrebs.com:IRC Logs:#uml:2004:04:07

---	Log	opened Wed Apr 07 17:14:06 2004
17:14	---	---> mikegrb [michael@mikegrb.netop.oftc.net] has joined #uml
17:14	-!-	Irssi: #uml: Total of 57 nicks 0 ops, 0 halfops, 0 voices, 57 normal
17:14	-!-	Irssi: Channel: Join to #uml was synced in 10 secs
17:14	---	---> sighup [~sighup@thegrebs.com] has joined #uml
17:14	jdike	Darkman: do you have the 2.6.4 -> 2.6.5 patch?
17:14	Darkman	jdike: uhm? no?
17:14	caker	mikegrb: thanks
17:14	Darkman	blaisors dir?
17:14	mikegrb	caker: no prob bob
17:14	mikegrb	er s/bob/chris/
17:14	jdike	Darkman: well, you could look at include/asm-i386/system.h and see if it changed
17:15	sikkh	:)
17:15	sikkh	Darkman: I know a little c, I'm on it, wih me luck ;)
17:15	sikkh	wish
17:17	sikkh	uml-patch-2.6.4-1.bz2 applies pretty well
17:18	Darkman	yep
17:19	Darkman	seems to be a 2.6.5 issue
17:19	Darkman	compiling 2.6.4 now
17:20		* jdike writes a diary entry
17:20	jdike	could be a long one
17:20	caker	It's been a while
17:20	jdike	almost two months
17:20	jdike	just checked that
17:21		* jdike needs to be less slack
17:22	jdike	btw, what does sighup do?
17:22	mikegrb	what doesn't he do is the better question ;)
17:22	mikegrb	jdike++
17:22	mikegrb	sighup: karma jdike
17:23	sighup	jdike has karma of 1
17:23	mikegrb	he has been a little lagged lately
17:23	mikegrb	does math
17:23	Darkman	name -a
17:23	Darkman	Linux (none) 2.6.4-1um #1 Wed Apr 7 23:22:35 CEST 2004 i686 unknown
17:23	Darkman	works
17:23	mikegrb	can ask him when someone was last around
17:23	caker	sighup: seen jdike
17:23	sighup	jdike was last seen on #uml 1 minutes and 30 seconds ago, saying: btw, what does sighup do? [1081372928]
17:23	mikegrb	can give you headlines from rss feeds
17:24	mikegrb	can tell you how many and what type of linodes are available
17:24	sikkh	Darkman: compiling 2.6.5-um1 now ;)
17:24	caker	sighup: uml?
17:24	sighup	caker: bugger all, i dunno
17:24	caker	sighup: uml is User-Mode Linux is a safe, secure way of running Linux versions and Linux processes. Run buggy software, experiment with new Linux kernels or distributions, and poke around in the internals of Linux, all without risking your main Linux setup.
17:24	sighup	OK, caker.
17:24	Darkman	sikkh: good look;)
17:24	sikkh	Darkman: wanna patch?
17:25	Darkman	sikkh: what was it?
17:25	---	<<-- frediz [~frediz@82.226.101.28] has quit (Read error: Connection reset by peer)
17:25	sikkh	Darkman: it's very simple issue
17:25	Darkman	hmkay
17:25	Darkman	patch to me
17:25	Darkman	.oO( damn, only two days left till the workshop.. ;)
17:26	caker	Darkman: What is this workshop you keep mentioning?
17:26	jdike	caker: ccc
17:26	jdike	caker: he's pimping UML for me
17:26	Darkman	caker: a workshop about uml.. first it was a 'can someone talk about uml security' and jeff asked me if i could.. then it was 'could you do a complete uml workshop? providing some kernels etc.etc.'
17:26	mikegrb	heh
17:27	Darkman	and i said yes...
17:27	Darkman	stupid me ;)
17:27	Darkman	so i started to work it out.. but it becomes more and more...
17:27	caker	ccc?
17:28	Darkman	caker: chaos computer club
17:29	mikegrb	oh spiffy
17:31	sikkh	jdike: http://kem.p.lodz.pl/~peter/2.6.5-uml1.fix is it ok?
17:31	Darkman	sikkh: compiling, thanks
17:32	caker	sikkh++
17:32	sikkh	:)
17:32	sikkh	wait till jdike says if it's not fscked up ;)
17:32	sikkh	hope not
17:32	Darkman	sikkh: for me at least its ok if its run now
17:33	jdike	sikkh: no, that sucks
17:33	Darkman	g
17:33	sikkh	see ;)
17:33	jdike	sikkh: but it probably works
17:33	sikkh	;)
17:33	sikkh	I love quick hacks :)
17:33	Darkman	sighup: jdike just rules, hmkay? ;)
17:33	sighup	Darkman: huh?
17:35	Darkman	# uname -a
17:35	Darkman	Linux (none) 2.6.5-1um #1 Wed Apr 7 23:34:34 CEST 2004 i686 unknown
17:35	Darkman	well :)
17:35	Darkman	thanks sikkh
17:35	sikkh	np, I'm the quick hack masta ;)
17:36	jdike	sikkh: you might post that to uml-devel for the impatient
17:36	sikkh	jdike: I'm on it
17:36	jdike	sikkh: remember to mention that I highly disapprove of it though :-)
17:36	Darkman	yeah, do that
17:36	Darkman	g
17:36	sikkh	jdike: of course ;)
17:37	sikkh	Darkman: what's the email ? ;)
17:37	Darkman	ok, i got 2.4.25-um1 running, 2.6.5-um1, 2.4.25-skas for the host, should be enough
17:37	Darkman	sikkh: which one?
17:37	---	<<-- umluser [~d0c9fe10@red.onlinemedicalbilling.com] has quit (Quit: CGI:IRC)
17:38	sikkh	Darkman: uml list
17:38	Darkman	user-mode-linux-devel@lists.sourceforge.net
17:39	sikkh	thx
17:39	Darkman	np
17:43	sikkh	jdike: could you point me 'in the right direction' as it should be done?
17:44	jdike	sikkh: find out where my cmpxchg is coming from and get rid of it
17:44	sikkh	jdike: ok
17:45	Darkman	anyone can remember about the suse root_fs posted on the lists lately?
17:49	Darkman	no matter, got it ;)
17:51		* caker gives sf a tune-up
17:52	sikkh	http://prdownloads.sourceforge.net/user-mode-linux/uml-patch-2.4.24-2.bz2
17:52	mikegrb	it needs a lot more then a tune-up :/
17:54	Darkman	anyone knows about the gentoo rootfs? did it hit the ground?
17:57	caker	http://uml.linode.com/uml-patch-2.4.24-2.bz2 (no sf download page crap)
17:58	knewt	is there a difference between not specifying a sub pointer in a fs-driver operations block and specifying NULL?
17:58	sikkh	knewt: I guess NULL is more proper
17:59	sikkh	knewt: cause vfs can check for it
17:59	jdike	knewt: depends
18:00	knewt	jdike: i'm looking at the hppfs super_operations block
18:00	jdike	knewt: if it's bss, then it all starts as zero
18:00	knewt	you specify NULL for delete_inode
18:01	knewt	there are then two entries with specific sub pointers, and the rest is left unspecified
18:01	jdike	knewt: yeah, those are NULL
18:01	jdike	knewt: if you were to kmalloc the block, then you'd have garbage in the uninitialized fields
18:02	Darkman	jdike: got a few rejects with 24-um2 on 2.4.25.. yfi
18:02	Darkman	fyi even
18:02	jdike	Darkman: anything serious?
18:02	sikkh	jdike: uml and i386 definitions of __cmpxchg are the same
18:02	Darkman	include/linux/mm.h.rej
18:03	Darkman	include/linux/proc_mm.h.rej
18:03	Darkman	mm/Makefile.rej
18:03	Darkman	mm/mmap.c.rej
18:03	Darkman	mm/mprotect.c.rej
18:03	Darkman	mm/proc_mm.c.rej
18:03	Darkman	looks a bit related to the local root exploit
18:04	jdike	Darkman: those are all skas-related
18:04	Darkman	aww
18:04	jdike	Darkman: that wouldn't have happened if I didn't have the skas patch in my tree
18:04	Darkman	hold on
18:05	Newsome	Darkman: No rejects for me, but a few offset Hunks
18:05	knewt	ah, seems that some operations have defaults, and supplying NULL lets you stop that happening
18:05	Darkman	Newsome: same here, just used the wrong tree
18:05	Darkman	jdike: nevermind, sorry
18:05	sikkh	jdike: uml and i386 definitions of __cmpxchg are the same, should I purge it from uml (the whole header file basically)?
18:06	jdike	sikkh: yeah
18:06	sikkh	ok
18:06	jdike	sikkh: given that I probably just copied it in the first place
18:07	---	<<-- orospakr [~orospakr@216.221.36.160] has quit (Quit: later!)
18:07		* caker LXR's uml-patch-2.4.24-2
18:12	sikkh	jdike: I removed "include/asm/system-i386.h" but I have no ide how to fix build system now ;(
18:13	sikkh	jdike: where is the symbolic link made?
18:14	jdike	sikkh: that was all that was in system-i386?
18:14	---	<<-- G2 [~ghenry@212.159.59.85] has quit (Quit: oops)
18:15	jdike	sikkh: OK, just delete the cmpxchg stuff
18:15	sikkh	ok
18:15	jdike	sikkh: and leave the #include system-generic there
18:19	Newsome	on both 2.4.24 and 2.4.25, I get - aio.c:12:27: linux/aio_abi.h: No such file or directory
18:20	Newsome	Am I doing it wrong?
18:21	Newsome	aha
18:22	Darkman	Unknown HZ value! (94) Assume 100. <-- reported when executing 'uptime' with 2.6.5
18:22	Newsome	The makefile checks to see if it can find aio_abi.h, but the #include in the file isn't protected inside HAVE_AIO_ABI #include <linux/aio_abi.h>
18:22	Darkman	(debilian)
18:23	caker	Darkman: I get that with 2.6.4-um, too .. I think it needs a newer procps .. vmstat also segfaults
18:23	jdike	Newsome: oops, send patches
---	Log	closed Wed Apr 07 18:23:35 2004