Back to Home / #uml / 2007 / 12 / Prev Day | Next Day
#uml IRC Logs for 2007-12-14

---Logopened Fri Dec 14 00:00:50 2007
00:38|-|jdike [~jdike@pool-72-93-105-51.bstnma.fios.verizon.net] has quit [Quit: Leaving]
02:59|-|balbir [~balbir@122.167.182.219] has quit [Ping timeout: 480 seconds]
03:10|-|balbir [~balbir@122.167.179.183] has joined #uml
03:22|-|dstuxo [~dstuxo@86.34.180.179] has joined #uml
03:28|-|flips [~phillips@phunq.net] has joined #uml
03:29<flips-#uml->>question: is skas still in active development?
03:34<Magotari-#uml->>flips: Yes.
03:34<flips-#uml->>cool
03:51|-|ftumch [~James@james.1ec.aaisp.net.uk] has joined #uml
03:52<peterz-#uml->>Magotari: do you get xterm consoles with rc5-mm1?
04:06<Magotari-#uml->>peterz: rc5-mm1 does not work for me. I don't know.
04:07<peterz-#uml->>hmm, right. the thing just sits there
04:07<peterz-#uml->>when I look at it, it thinks its done booting
04:07<peterz-#uml->>just don't get any terms
04:10<peterz-#uml->>hmm, init looks stuck
04:11<Magotari-#uml->>You can try init=/bin/bash, it might give you enough functionality to work for a while.
04:12<Magotari-#uml->>Another thing that I do to dodge around the problems with xterms was putting the consoles onto a host port.
04:13<Magotari-#uml->>con=port:9000 might be the key, I think.
04:13<Magotari-#uml->>Then do 'telnet localhost 9000'
04:13<peterz-#uml->>yeah, lemme try that
04:20<peterz-#uml->>something is not right
04:47|-|dstuxo [~dstuxo@86.34.180.179] has quit [Ping timeout: 480 seconds]
05:07|-|dstuxo [~dstuxo@86.34.180.179] has joined #uml
05:24|-|dstuxo [~dstuxo@86.34.180.179] has quit [Ping timeout: 480 seconds]
05:25|-|dstuxo [~dstuxo@86.34.180.179] has joined #uml
05:30|-|dstoica [~dstuxo@86.34.180.179] has joined #uml
05:35|-|dstuxo [~dstuxo@86.34.180.179] has quit [Ping timeout: 480 seconds]
05:44|-|dstuxo [~dstuxo@86.34.180.179] has joined #uml
05:44|-|kokoko1 [~Slacker@203.148.65.19] has joined #uml
05:49|-|dstoica [~dstuxo@86.34.180.179] has quit [Ping timeout: 480 seconds]
06:01|-|krau [~cktakahas@200.184.118.132] has quit [Quit: Varei!!!]
06:09|-|karol [~karol@chello089076073248.chello.pl] has quit [Quit: Lost terminal]
06:37|-|Magotari [~karol@chello089076073248.chello.pl] has quit [Quit: Lost terminal]
07:03|-|Infinito [argos@200-103-110-89.gnace701.dsl.brasiltelecom.net.br] has joined #uml
07:47|-|namit [~test_acco@83-70-39-95.b-ras1.prp.dublin.eircom.net] has quit [Quit: Leaving.]
07:53|-|namit [~test_acco@83-70-39-95.b-ras1.prp.dublin.eircom.net] has joined #uml
07:58|-|dstoica [~dstuxo@86.34.180.179] has joined #uml
08:00|-|dstuxo [~dstuxo@86.34.180.179] has quit [Ping timeout: 480 seconds]
09:08|-|dang [~dang@aa-redwall.nexthop.com] has joined #uml
09:21|-|dstoica [~dstuxo@86.34.180.179] has quit [Quit: Leaving]
09:21|-|dstuxo [~dstuxo@86.34.180.179] has joined #uml
09:31|-|kokoko1 [~Slacker@203.148.65.19] has quit [Ping timeout: 480 seconds]
10:02<dang-#uml->>Woot.
10:02<dang-#uml->>Running my UMLs on my newly booted skas3 box killed the host.
10:03<dang-#uml->>completely: no ping or anything.
10:03<dang-#uml->>Back to the old kernel, I guess...
10:04|-|hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has joined #uml
11:12|-|jdike [~jdike@pool-72-93-105-51.bstnma.fios.verizon.net] has joined #uml
11:12|-|dstuxo [~dstuxo@86.34.180.179] has quit [Read error: Connection reset by peer]
11:13<jdike-#uml->>Hi guys
11:21|-|karol [~karol@abgs186.neoplus.adsl.tpnet.pl] has joined #uml
11:21|-|karol changed nick to Magotari
11:26|-|dstuxo [~dstuxo@86.34.180.179] has joined #uml
11:29|-|dstuxo [~dstuxo@86.34.180.179] has quit [Remote host closed the connection]
11:35|-|mgross [~mgross@pool-71-245-99-169.ptldor.fios.verizon.net] has joined #uml
12:01<dang-#uml->>jdike: Today, I booted my 2.6.23-skas3 kernel, started my UML, and my host went bye bye...
12:01<dang-#uml->>No panic (at least the LEDs weren't flashing) but it was completely hung, and not responding to ping.
12:02|-|newbie [~jeff@ita4fw1.itasoftware.com] has joined #uml
12:02<dang-#uml->>Guest kernel is stock 2.6.24-rc3 (with NOHZ on)
12:03<Magotari-#uml->>Ouch.
12:03<Magotari-#uml->>Does it happen always or just sometimes?
12:03<jdike-#uml->>wow
12:03<dang-#uml->>I only tried it once.
12:04<dang-#uml->>I had to have it working, so I booted back into the old kernel.
12:04<dang-#uml->>I can try again monday, I guess, but probably not before then, as this is a work machine.
12:04<jdike-#uml->>well, I'll need some sort of debug info
12:04<dang-#uml->>Yeah, I figured. :)
12:04<jdike-#uml->>sysrq T would be a good start
12:04<dang-#uml->>On the host?
12:04<jdike-#uml->>yup
12:05<jdike-#uml->>and you would have seen any console messages?
12:05<dang-#uml->>Nope. I was in X.
12:05<jdike-#uml->>OK, switch back to tty1 next time and see what shows up
12:05<dang-#uml->>'k
12:08|-|tyler29 [~tyler@ARennes-257-1-12-166.w81-250.abo.wanadoo.fr] has joined #uml
12:15<Magotari-#uml->>Stupid. Forgot about putting my work machine on the network. No ssh. Now I am stuck here on a tiny laptop. Argh.
12:15<Magotari-#uml->>Which means little to no work can be done.
12:15<Magotari-#uml->>At least testwise.
12:18<Magotari-#uml->>I really wish I could ssh to my school computer and test uml there. Sadly, the whole place is mounted noexec, and I'm not surprised.
12:18<Magotari-#uml->>Bad people in my school. Bad bad bad.
12:38[~]jdike #uml looks at the O_NONBLOCK mconsole thing#uml-> looks at the O_NONBLOCK mconsole thing
12:39<Magotari-#uml->>Erm, I have a patch for it already.
12:39<Magotari-#uml->>If you mean the hang at bootup. I even tested it a bit, seemed fine.
12:39<jdike-#uml->>OK, send it over
12:40<jdike-#uml->>it's just setting O_NONBLOCK right after the socket is created, right?
12:40<Magotari-#uml->>I wish it were this easy. The patch is 50km away, in a now not connected to a network computer. It was a oneliner...
12:40<Magotari-#uml->>Yeah, sorta. os_set_fd_blocking(fd, 0);
12:40<Magotari-#uml->>Pretty much that.
12:40<jdike-#uml->>yup
12:40<Magotari-#uml->>Just after the unix socket was created.
12:40<jdike-#uml->>I'll just do it here then
12:40<Magotari-#uml->>Right.
12:41|-|namit [~test_acco@83-70-39-95.b-ras1.prp.dublin.eircom.net] has quit [Read error: Connection reset by peer]
12:41<jdike-#uml->>hmm
12:41<jdike-#uml->>that function needs more work anyway
12:42<jdike-#uml->>nothing gets cleaned up if anything fails
12:45|-|namit [~test_acco@83-70-39-95.b-ras1.prp.dublin.eircom.net] has joined #uml
12:48<jdike-#uml->>winch_interrupt also dies
12:49<jdike-#uml->>but there's no mconsole hang
12:49<Magotari-#uml->>Progress.
12:50<jdike-#uml->>except I haven't fixed anything
12:50|-|namit [~test_acco@83-70-39-95.b-ras1.prp.dublin.eircom.net] has quit [Read error: Connection reset by peer]
12:54|-|namit [~test_acco@83-70-39-95.b-ras1.prp.dublin.eircom.net] has joined #uml
13:00<jdike-#uml->>sleep has started hanging on shutdown
13:03<Magotari-#uml->>Is the whole lot of recent problems caused by everyone switching to skas3?
13:07<jdike-#uml->>which recent problems?
13:08<jdike-#uml->>OK, sleep doesn't hang on rc5
13:09<jdike-#uml->>bisection time
13:10<Magotari-#uml->>Caker, me, dang, you... At least four people had a bunch of problems in the last two or three days. Many related to the new skas3 patch.
13:10<caker-#uml->>I'm not running the new skas3 patch
13:10<Magotari-#uml->>Oh. Well, related to skas3 then.
13:11<jdike-#uml->>not me
13:11<jdike-#uml->>what problems have you and caker had that are associated with skas3?
13:11<jdike-#uml->>the only one I know of besides dang's hang is userspace returning -EINVAL under gdb
13:12<Magotari-#uml->>caker has a spin of some kind in init.
13:12<jdike-#uml->>yeah, but that's not obviously skas3
13:12<caker-#uml->>which we suspect was compiler or config related, but I haven't had time to test that yet
13:12<jdike-#uml->>he's been a skas3 customer since the beginning
13:12[~]caker #uml flexes#uml-> flexes
13:13<Magotari-#uml->>No surprise there. I would have been too, if it was supported for 23 from the start.
13:13<caker-#uml->>I still think you should have called it SPAZ :>
13:13<Magotari-#uml->>Maybe for skas4... :)
13:14<jdike-#uml->>hehe
13:14<Magotari-#uml->>I had a host of problems in the last few days, all reported by now. None really skas3 oriented.
13:14<jdike-#uml->>BTW, I'm working on a skas3 replacement
13:14<Magotari-#uml->>I just wonder why so many reports in the last few days...
13:14<Magotari-#uml->>OOOHHHHHH!
13:14<Magotari-#uml->>Yum.
13:15<Magotari-#uml->>Cannot wait to test THAT. :)
13:15<jdike-#uml->>I think I know how to do all the skas3 pieces in Linus-acceptable ways
13:16<Magotari-#uml->>Hmm... So it will be skas3 mainlined, eh? I thought a new concept of doing things...
13:16<Magotari-#uml->>Still looking forward to the testing phase though.
13:19<jdike-#uml->>instead of /proc/mm, I have system calls - new_mm and switch_mm
13:20<jdike-#uml->>for now, switch_mm will hop onto the stub in the process and run mmap et al
13:22<jdike-#uml->>I have it working inside UML with a little test program
13:22<jdike-#uml->>next step is to port it to the host and get UML using it
13:23<jdike-#uml->>then I'll RFC it to LKML
13:23|-|balbir [~balbir@122.167.179.183] has quit [Ping timeout: 480 seconds]
13:23<Magotari-#uml->>Then I will test it. :) Hanging around kernel developers has benefits, like an advance warning to look for an RFC. Yum.
13:24<Magotari-#uml->>Wish I were one of you guys, and hey, maybe one day...
13:24<jdike-#uml->>everyone starts somewhere
13:26<Magotari-#uml->>You can expect me to be more help in a year or so.
13:27<Magotari-#uml->>I'm at the 'add tiny feature / fix trivial bug' phase.
13:34|-|balbir [~balbir@122.167.179.170] has joined #uml
13:39<Magotari-#uml->>Ah, yes. Speaking of trivial bugs, there was one which I was supposed to take care of... log printing OK twice. I'm on it.
13:44<jdike-#uml->>it was my risk-free cleanup segfault handling patch apparently
13:47|-|mgross [~mgross@pool-71-245-99-169.ptldor.fios.verizon.net] has quit [Quit: Ex-Chat]
13:47|-|mgross [~mgross@pool-71-245-99-169.ptldor.fios.verizon.net] has joined #uml
13:47|-|mgross [~mgross@pool-71-245-99-169.ptldor.fios.verizon.net] has quit [Remote host closed the connection]
13:54<fo0bar-#uml->>Magotari: what new skas3?
13:55[~]fo0bar #uml lives in his own little fantasy world, stuck in a fork of 2.6.18 :)#uml-> lives in his own little fantasy world, stuck in a fork of 2.6.18 :)
14:00<fo0bar-#uml->>well, not "stuck" of course, but after a year of experimentation, I had a nice stable patchlist against debian stable's 2.6.18 kernel
14:01|-|Blissex [~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk] has joined #uml
14:06<Magotari-#uml->>fo0bar: The patch for 2.6.23.
14:07<Magotari-#uml->>Skas used to go only up to 2.6.22, now it got into 2.6.23.
14:07<Magotari-#uml->>Talking about skas3 here, of course.
14:07|-|namit [~test_acco@83-70-39-95.b-ras1.prp.dublin.eircom.net] has quit [Ping timeout: 480 seconds]
14:07<fo0bar-#uml->>Magotari: where's that? I don't see anything since 2.6.20 on blaisorblade
14:07<Magotari-#uml->>I will probably make my stand as to the kernel on the 23 version, very stable and supports my hardware.
14:07|-|dgraves [~agraves@inet-netcache2-o.oracle.com] has quit [Remote host closed the connection]
14:07<Magotari-#uml->>fo0bar: Sorta unofficial patch thingy.
14:08<Magotari-#uml->>Look at uml mailing lists.
14:08<Magotari-#uml->>I think uml-devel
14:08<fo0bar-#uml->>ahh, cool
14:08<Magotari-#uml->>That is for the host kernel though.
14:10<Magotari-#uml->>And blaisorblade has been inactive for a few months now, as far as I can tell.
14:14<fo0bar-#uml->>ok, found it
14:17<fo0bar-#uml->>oh hey, x86 merge in 2.6.24
14:17<fo0bar-#uml->>that'll be fun
14:18<Magotari-#uml->>:)
14:19<Magotari-#uml->>Yeah, it was fun too. Problems all over the place, at least so far.
14:19<Magotari-#uml->>The fallout has been dimisishing, but I'd be afraid to put a .24 kernel on a production system.
14:20|-|tyler29 [~tyler@ARennes-257-1-12-166.w81-250.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
14:24|-|newbie [~jeff@ita4fw1.itasoftware.com] has left #uml []
14:29|-|Blissex [~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk] has quit [Remote host closed the connection]
14:35|-|tyler29 [~tyler@ARennes-257-1-11-194.w81-250.abo.wanadoo.fr] has joined #uml
14:44|-|ram [~ram@pool-96-225-204-220.ptldor.fios.verizon.net] has joined #uml
14:56<Magotari-#uml->>jdike: Not sure if you know but that Reiser person just posted a bugfix for some /dev/urandom thing, discovered with valgrind.
14:57<Magotari-#uml->>Seems like he may be making some progress.
15:00|-|tyler29 [~tyler@ARennes-257-1-11-194.w81-250.abo.wanadoo.fr] has quit [Quit: ++]
15:23|-|Magotari [~karol@abgs186.neoplus.adsl.tpnet.pl] has quit [Read error: Connection reset by peer]
15:23|-|Magotari [~karol@abgs186.neoplus.adsl.tpnet.pl] has joined #uml
15:42|-|Magotari [~karol@abgs186.neoplus.adsl.tpnet.pl] has quit [Read error: Connection reset by peer]
15:51|-|aroscha [~aroscha@193.238.156.96] has joined #uml
16:23<fo0bar-#uml->>Warning: attempt to assign a globally valid ethernet address to a device
16:23<fo0bar-#uml->>:(
16:24<fo0bar-#uml->>what if I have my own company_id? huh? huh? :)
16:24<fo0bar-#uml->>err OUI
16:26[~]jdike #uml doubts fo0bar has his own ethernet range#uml-> doubts fo0bar has his own ethernet range
16:27<fo0bar-#uml->>ok, so I was stupid in my early days and assumed "reserved" in the OUI database meant "like RFC1918"
16:28<fo0bar-#uml->>turns out it means "the owner of this block doesn't want to reveal its identity yet"
16:28<fo0bar-#uml->>as such, I currently have about 20 production guests that technically use "LE INFORMATION COMMUNICATION INC" of Korea ethernet cards :)
16:29<fo0bar-#uml->>but, my "what if" scenario could be valid.. a kconfig or command line override would be useful
16:32<jdike-#uml->>hehe
16:33<jdike-#uml->>is that an actual error and you get a random MAC instead?
16:33<jdike-#uml->>or just a warning
16:33<fo0bar-#uml->>it sends to KERN_WARNING, then changes the mac to random
16:34<jdike-#uml->>hmm
16:34<jdike-#uml->>I suppose we could emit the warning and let you keep the MAC
16:34<fo0bar-#uml->>arch/um/drivers/net_kern.c, around 319
16:36<fo0bar-#uml->>all other cases in that block seem to be genuinely valid (multicast, out of range, etc) and reason for random
16:37<jdike-#uml->>yeah
16:51|-|aroscha [~aroscha@193.238.156.96] has quit [Quit: aroscha]
17:07|-|Micksa [~mslade@203.55.194.182] has quit [Ping timeout: 480 seconds]
17:09<jdike-#uml->>yeah, Reiser looks like he's getting somewhere
17:17|-|Micksa [~mslade@203.55.194.182] has joined #uml
17:43<jdike-#uml->>get_mm and switch_mm build in a host kernel
17:43<jdike-#uml->>so I have the world's first skas4 host kernel
17:45|-|dang [~dang@aa-redwall.nexthop.com] has quit [Quit: Leaving.]
17:56|-|Magotari [~karol@abgs186.neoplus.adsl.tpnet.pl] has joined #uml
17:57<fo0bar-#uml->>nifty!
17:57[~]caker #uml drools#uml-> drools
17:58<fo0bar-#uml->>jdike: do you have a corresponding guest, or is it just a test suite at this point?
17:58<Magotari-#uml->>Sounds interesting. Log digging time.
17:58<fo0bar-#uml->>14:43 < jdike> get_mm and switch_mm build in a host kernel
17:58<fo0bar-#uml->>14:43 < jdike> so I have the world's first skas4 host kernel
17:58<Magotari-#uml->>Aha. Yeah. Nice.
17:59<Magotari-#uml->>This is great news.
18:08|-|cl4sh [~cl4sh@qik.ds.pg.gda.pl] has joined #uml
18:08<Magotari-#uml->>So SKAS3 for the masses out of box. Wonderful.
18:08<cl4sh-#uml->>hi all
18:08<Magotari-#uml->>Hey, cl4sh.
18:17<jdike-#uml->>fo0bar, just a trivial test program right now
18:18<jdike-#uml->>make a new mm, switch to it, map a page to demonstrate that it is in a new mm, then switch back
18:19<fo0bar-#uml->>nice
18:24<jdike-#uml->>http://rafb.net/p/Uo9i0633.html
18:24<jdike-#uml->>it dumps out /proc/self/maps at each step
18:30<jdike-#uml->>http://rafb.net/p/4nj1u596.html
18:30<jdike-#uml->>the little test program
18:41<Magotari-#uml->>Did anyone try applying the new skas3 thing onto a -stable kernel?
18:41|-|hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has quit [Quit: Leaving]
18:45<cl4sh-#uml->>Is it posiible to investigate vlan scuirity under UML?
18:47<Magotari-#uml->>Applies cleanly.
18:47<jdike-#uml->>it shoud apply
18:48<jdike-#uml->>should
18:48<jdike-#uml->>-stable patches are supposed to be minimal bug fixes
18:48<Magotari-#uml->>It worked, yeah.
18:48<Magotari-#uml->>Not even a single offset or fuzz comment.
18:53<fo0bar-#uml->>Magotari: I applied it against debian's 2.6.23 tree, but I've only compiled a guest so far. no hosts
18:53<Magotari-#uml->>I am preparing for a 23.10 host.
18:53<Magotari-#uml->>Then guest shortly after.
18:54<Magotari-#uml->>And then I'm going to curse my amount of ram. Loudly.
19:00<Magotari-#uml->>Hmm... "Sparse Memory". What does this button do? Hmm... ? Wonder what it can break...
19:00<jdike-#uml->>Linux tp.user-mode-linux.org 2.6.24-rc4-skas4 #4 SMP Fri Dec 14 18:35:55 EST 2007 i686 i686 i386 GNU/Linux
19:00<Magotari-#uml->>Wow.
19:03<jdike-#uml->>and my little test crashes
19:03<jdike-#uml->>hmm
19:30|-|ferret_0567 [~travis@72.191.26.86] has quit [Quit: leaving]
19:30|-|ferret_0567 [~travis@cpe-72-191-26-86.satx.res.rr.com] has joined #uml
19:57|-|ferret_0667 [~travis@72.191.26.86] has joined #uml
19:57|-|ferret_0667 [~travis@72.191.26.86] has quit []
20:50|-|remus [~remus@76.231.178.131] has quit [Remote host closed the connection]
22:17|-|cl4sh [~cl4sh@qik.ds.pg.gda.pl] has quit [Quit: leaving]
23:01|-|Infinito [argos@200-103-110-89.gnace701.dsl.brasiltelecom.net.br] has quit [Quit: Quitte]
23:10|-|Electric1lf [~dbharris@bas1-toronto48-1242461703.dsl.bell.ca] has joined #uml
23:12|-|ElectricElf [~dbharris@bas1-toronto48-1242461494.dsl.bell.ca] has quit [Read error: Operation timed out]
23:59|-|VS_ChanLog [~stats@ns.theshore.net] has left #uml [Rotating Logs]
23:59|-|VS_ChanLog [~stats@ns.theshore.net] has joined #uml
---Logclosed Sat Dec 15 00:00:06 2007