Back to Home / #uml / 2007 / 07 / Prev Day | Next Day
#uml IRC Logs for 2007-07-09

---Logopened Mon Jul 09 00:00:42 2007
01:03|-|hfb [~hfb@75.80.37.175] has quit [Quit: Leaving]
01:42|-|_horst [~horst@a89-182-68-181.net-htp.de] has joined #uml
01:48|-|_horst [~horst@a89-182-68-181.net-htp.de] has quit [Remote host closed the connection]
---Logclosed Mon Jul 09 03:36:19 2007
---Logopened Mon Jul 09 03:36:24 2007
03:36|-|mikegrb [~michael@mail.thegrebs.com] has joined #uml
03:36|-|Ekipa kanalu #uml: Wszystkich: 28 |-| +op [0] |-| +voice [0] |-| normalnych [28]
03:37|-|Kanal #uml zsynchronizowany w 78 sekundy
04:01|-|Ancalagon [~PtitKosmi@215.208-78-194.adsl-fix.skynet.be] has joined #uml
05:26|-|kokoko1 [~Slacker@203.148.65.8] has joined #uml
07:56|-|jdike [~jdike@pool-71-174-247-82.bstnma.fios.verizon.net] has joined #uml
07:56<jdike>Hi guys
07:59[~]jdike checks out 2.6.22
08:01<peterz>jdike: seen that linux-m post on multiple address-spaces per process?
08:01<peterz>linux-mm even
08:04<jdike>no
08:04[~]jdike isn't on linux-mm
08:04<peterz>right, suspected as much
08:04<jdike>Peter Chubb?
08:04<peterz>yep
08:04<jdike>yeah, he has a very special purpose hack
08:05<jdike>he has indexes, which aren't usable outside the process
08:05<peterz>right, that hurts a little
08:05<peterz>but it made me think of the skas patches
08:05<jdike>yup
08:05<jdike>indeeed
08:05<jdike>s/e//
08:06<jdike>is he claiming they should go in?
08:06<jdike>or is it "look what I've done!"?
08:06<peterz>yeah
08:06<peterz>asking if something similar could be used by others
08:06<jdike>well, he know I can use it
08:07<jdike>and he knows what would need changing
08:07<peterz>right
08:07<peterz>well this might be a good moment to propose the skas stuff again :-)
08:08<jdike>the interface is a non-starter
08:08<peterz>obviously
08:08<jdike>that's always been the problem
08:08<peterz>that was most fugly
08:08<peterz>oh, for skas
08:08<peterz>right
08:08<jdike>yeah, for skas
08:09<peterz>you did propose syscalls once, right?
08:09<jdike>Linus dinged that idea
08:09<peterz>perhaps now that we're adding syscalls left right and center
08:09<jdike>yeah
08:09<jdike>the *at syscalls do exactly what Linux said he didn't like
08:09<peterz>address space as file descriptor
08:09[~]peterz runds
08:09<peterz>s/d//
08:10<jdike>that was always the idea, and there's nothing wrong with that
08:10<peterz>how does one select?
08:10<peterz>not using the regular ops I take it
08:10<jdike>just where do you get it (open(/proc/mm), new_mm(), ...
08:10<jdike>)
08:10<peterz>/dev/mm
08:10<jdike>and how do you operate on it?
08:11<peterz>read/write/select don't seem like useful operations, but then I haven't really pondered it
08:11<jdike>a new_mm() syscall, I think would be better
08:11<peterz>yeah
08:11<jdike>fmmap/fmprotect,fmunmap
08:12<peterz>ah, map in another mm
08:12<jdike>yeah, that's essential for me
08:12<peterz>what would happen if people mistake that for mmap this fd
08:12<jdike>peter's (chubb) thing doesn't need that
08:12<peterz>he does LoL, right?
08:12<jdike>yeah
08:13<jdike>OK, maybe there are better names
08:13<jdike>the one technical problem is the number of args that fmmap would have
08:14<jdike>7
08:14<peterz>-ETOOMANY
08:14<jdike>I fear that
08:14<jdike>although the arg block works there
08:15<peterz>you could use this wrapper syscall linus has been on about for this
08:15<jdike>that's the other idea
08:15<jdike>mm_indirect
08:15<jdike>I've cooled on that idea, although it seems to be popular for other purposes
08:16<peterz>indeed, and if that is what is going to happen, might as well use it
08:16<jdike>yep
09:57|-|linbot [~supybot@ns.theshore.net] has quit [Ping timeout: 480 seconds]
10:11|-|hfb [~hfb@pool-72-87-254-188.lsanca.dsl-w.verizon.net] has joined #uml
10:38|-|jdike [~jdike@pool-71-174-247-82.bstnma.fios.verizon.net] has quit [Quit: Leaving]
10:38|-|jdike [~jdike@pool-71-174-247-82.bstnma.fios.verizon.net] has joined #uml
10:54|-|hfb [~hfb@pool-72-87-254-188.lsanca.dsl-w.verizon.net] has left #uml [Leaving]
11:07|-|Netsplit synthon.oftc.net <-> charon.oftc.net quits: kybe_, Ancalagon, Blissex, albertito, peterz, SNy, flatronf700B, apic, weasel, Hunger
11:07|-|Netsplit over, joins: Hunger, albertito
11:09|-|Blissex [~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk] has joined #uml
11:09|-|peterz [~peterz@i55087.upc-i.chello.nl] has joined #uml
11:09|-|SNy [54b006a92b@bmx-chemnitz.de] has joined #uml
11:09|-|weasel [weasel@weasel.noc.oftc.net] has joined #uml
11:09|-|kybe_ [~kyrre@legolas.iu.hio.no] has joined #uml
11:10|-|Ancalagon [~PtitKosmi@215.208-78-194.adsl-fix.skynet.be] has joined #uml
11:10|-|apic [~andi@paranerd-geekophren.name] has joined #uml
11:10|-|flatronf700B [~flatronf7@ns1.clipsalportal.com] has joined #uml
12:01|-|ram [~ram@pool-71-117-236-168.ptldor.fios.verizon.net] has joined #uml
12:12|-|Ancalagon [~PtitKosmi@215.208-78-194.adsl-fix.skynet.be] has quit [Max SendQ exceeded]
12:17|-|jjkola [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has joined #uml
12:28<jjkola>hello
12:29<jjkola>I tried compiling 2.6.22 both as uml and skas3, and both failed to compile
12:32<jjkola>uml choked on hostfs and skas3 choked with mm
12:32[~]peterz builds
12:33<peterz>builds just fine it seems
12:35<jjkola>errors with uml: http://pastebin.ca/610838
12:36<jjkola>erros with skas3: http://pastebin.ca/610840
12:36<jjkola>erros -> errors
12:37|-|_horst [~horst@a89-182-68-181.net-htp.de] has joined #uml
12:39<peterz># grep HOSTFS uml-build/.config
12:39<peterz>CONFIG_HOSTFS=y
12:39<peterz>still builds
12:39<jjkola>with uml version I have only applied mconsole exec patch to vanilla kernel
12:40<jdike>I know about the hostfs module errors
12:41<jdike>a patch is forthcoming
12:41<jdike>looks like the skas patch needs an update
12:42<jjkola>with skas3 version I have applied patch for 2.6.20-v9-pre9 and applied all rejected chunks by hand to vanilla kernel
13:02|-|jjkola1 [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has joined #uml
13:08|-|jjkola [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has quit [Ping timeout: 480 seconds]
13:19|-|jjkola1 changed nick to jjkola
14:05<jjkola>when I built 2.6.22 uml version and run that I got segmentation fault
14:05<jjkola>http://pastebin.ca/611014
14:06<jjkola>the host is 2.6.21.5 with skas3 patch
14:07<jjkola>if you need more information the kernel is still running under gdb
14:09<jdike>get a backtrace?
14:12<jjkola>http://pastebin.ca/611028
14:13<jdike>this is stock a 2.6.22 UML?
14:13<jdike>s/stock a/a stock/
14:13<jjkola>only mconsole exec patch is applied
14:14<jjkola>otherwise it's a stock kernel
14:14<jdike>hmm
14:15<jdike>that makes no sense though
14:15<jjkola>why?
14:16<jdike>a page fault shouldn't result in a >32 block I/O operation
14:16<jjkola>oh
14:17<jdike>whoa
14:17<jdike>it happens here too
14:17<jjkola>well, at least it seems to be reproducible : )
14:19<jdike>I'm getting a 48-block read
14:24<jdike>there's probably a one-line fix for this
14:41<jdike>http://pastebin.ca/611115
14:49[~]kokoko1 hates kernel-xen from fedora folks
14:50<kokoko1>ah i wish we could be using uml on all our infra, keep xen kernel sync with fedora update repos is become pita
14:50<kokoko1>i love rolling our own kernel :)
14:50<kokoko1>hi hi :)
15:01<jjkola>jdike: thanks for the patch, it helped
15:01<jdike>np, thanks for letting me know
15:01<jdike>its next stop is -stable
15:29|-|da-x [karrde@bzq-88-155-87-213.red.bezeqint.net] has quit [Remote host closed the connection]
15:36|-|jjkola1 [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has joined #uml
15:41|-|jjkola changed nick to Guest802
15:41|-|jjkola1 changed nick to jjkola
15:41|-|Guest802 [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has quit [Ping timeout: 480 seconds]
15:52|-|da-x [karrde@bzq-88-153-36-163.red.bezeqint.net] has joined #uml
15:54|-|da-x [karrde@bzq-88-153-36-163.red.bezeqint.net] has quit [Remote host closed the connection]
16:04|-|da-x [karrde@bzq-88-153-0-84.red.bezeqint.net] has joined #uml
16:10|-|da-x [karrde@bzq-88-153-0-84.red.bezeqint.net] has quit [Remote host closed the connection]
16:15|-|da-x [karrde@bzq-88-153-0-84.red.bezeqint.net] has joined #uml
16:38|-|jjkola [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has quit [Quit: *pop*]
16:38|-|jjkola [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has joined #uml
16:45|-|jjkola [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has quit [Quit: *pop*]
16:53|-|jjkola [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has joined #uml
17:05|-|_horst [~horst@a89-182-68-181.net-htp.de] has quit [Remote host closed the connection]
18:39|-|jjkola [~jjkola@dsl-olubrasgw1-fe56fb00-241.dhcp.inet.fi] has quit [Quit: *pop*]
19:35|-|mjf [~mjf@r5bb59.net.upc.cz] has joined #uml
19:35|-|mjf [~mjf@r5bb59.net.upc.cz] has quit []
20:04|-|jdike [~jdike@pool-71-174-247-82.bstnma.fios.verizon.net] has quit [Quit: Leaving]
20:28|-|ram [~ram@pool-71-117-236-168.ptldor.fios.verizon.net] has quit [Read error: Connection reset by peer]
20:28|-|ram [~ram@pool-71-117-236-168.ptldor.fios.verizon.net] has joined #uml
20:57|-|ram_ [~ram@pool-71-117-236-168.ptldor.fios.verizon.net] has joined #uml
20:58|-|ram [~ram@pool-71-117-236-168.ptldor.fios.verizon.net] has quit [Read error: Connection reset by peer]
21:07|-|Nem^ [~Nem@dslb-084-056-247-006.pools.arcor-ip.net] has quit [Ping timeout: 480 seconds]
21:08|-|ram [~ram@pool-71-117-236-168.ptldor.fios.verizon.net] has joined #uml
21:09|-|ram_ [~ram@pool-71-117-236-168.ptldor.fios.verizon.net] has quit [Read error: Connection reset by peer]
21:17|-|Nem^ [~Nem@dslb-084-056-228-102.pools.arcor-ip.net] has joined #uml
22:59|-|VS_ChanLog [~stats@ns.theshore.net] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [~stats@ns.theshore.net] has joined #uml
23:44|-|linbot [~supybot@ns.theshore.net] has joined #uml
---Logclosed Tue Jul 10 00:00:58 2007