2009-05-05T00:07:45 *** syslogd_ has joined #glendix 2009-05-05T00:21:05 *** syslogd has quit IRC 2009-05-05T00:41:01 *** TheAppleMan has joined #glendix 2009-05-05T01:22:40 *** syslogd has joined #glendix 2009-05-05T01:35:23 *** syslogd_ has quit IRC 2009-05-05T02:12:08 *** jrm8005 has quit IRC 2009-05-05T04:24:46 *** TheAppleMan has quit IRC 2009-05-05T06:35:58 *** jrm8005 has joined #glendix 2009-05-05T06:38:44 *** Murmuria has joined #glendix 2009-05-05T06:39:21 jrm8005, this netdev ML is noisy! 75 emails since mine!! 2009-05-05T06:39:40 I have absolutely no hope of a response here 2009-05-05T06:40:03 Murmuria: dude, it's been 1 day. 2009-05-05T06:40:41 Murmuria: if you had posted to LKML, there would have been about 500 emails since yours. 2009-05-05T06:40:50 oh! 2009-05-05T06:40:58 netdev is *not* high traffic. 2009-05-05T06:43:51 * Murmuria is bamboozled 2009-05-05T06:45:12 ha, dude I said it would probably be a couple of days, if anything. 2009-05-05T08:03:30 *** jrm8005 has quit IRC 2009-05-05T08:16:30 *** Murmuria has quit IRC 2009-05-05T09:35:44 *** kristoffer has joined #glendix 2009-05-05T12:46:35 *** kristoffer_ has joined #glendix 2009-05-05T12:47:38 *** kristoffer has quit IRC 2009-05-05T15:07:04 *** less1 has quit IRC 2009-05-05T15:25:31 *** less1 has joined #glendix 2009-05-05T15:45:19 *** less1 has quit IRC 2009-05-05T15:59:02 *** less1 has joined #glendix 2009-05-05T16:18:37 *** less1 has quit IRC 2009-05-05T16:20:28 *** olejorgenb has joined #glendix 2009-05-05T16:37:28 *** less1 has joined #glendix 2009-05-05T16:39:08 *** TheAppleMan has joined #glendix 2009-05-05T16:54:29 *** less1 has quit IRC 2009-05-05T16:56:16 *** less1 has joined #glendix 2009-05-05T17:22:14 *** less1 has quit IRC 2009-05-05T17:38:34 *** jrm8005 has joined #glendix 2009-05-05T17:41:52 *** TheAppleMan has quit IRC 2009-05-05T17:47:09 *** TheAppleMan has joined #glendix 2009-05-05T18:30:25 *** KillerX has joined #glendix 2009-05-05T18:30:30 hey KillerX 2009-05-05T18:31:16 hey jrm8005 2009-05-05T18:32:15 been a while since we've been online contemporaneously... do we have anything on the horizon? 2009-05-05T18:32:36 yep, we have a gsoc student working on per-process namespaces 2009-05-05T18:32:49 and another student who didn't get through, but is still working on completing all system calls 2009-05-05T18:33:08 are they working now, or not as of yet? 2009-05-05T18:33:14 i'll be spending some time to write a few of the filesystems, #c and #e specifically 2009-05-05T18:33:23 all I've seen in hg is netfs commits. 2009-05-05T18:33:33 yep, no commits as of yet 2009-05-05T18:33:34 KillerX: I assume those will be 9P 2009-05-05T18:33:55 I don't expect any code from the students before june 2009-05-05T18:34:05 oh, really? 2009-05-05T18:38:42 *** Murmuria has joined #glendix 2009-05-05T18:39:01 Murmuria: I noticed your commit with the broken code 2009-05-05T18:39:20 in the future, I recommend surrounding that kind of stuff with #ifdef BROKEN or some such. 2009-05-05T18:40:37 I added a comment with a warning in the commit 2009-05-05T18:40:49 I dont know about ifdef broken... 2009-05-05T18:41:11 I'm saying if you commit broken code, make it not compile by default. 2009-05-05T18:41:12 jrm8005, I never knew there is a formal way of sending in not working code 2009-05-05T18:41:16 can you tell me mroe? 2009-05-05T18:41:28 oh, hmm 2009-05-05T18:41:34 just make it so that broken things do not compile by default with #ifdef 2009-05-05T18:41:35 ok, I get it 2009-05-05T18:41:50 but the broken part is just one line 2009-05-05T18:42:01 sock->ops->connect 2009-05-05T18:42:10 comment that line, and the code will compile 2009-05-05T18:42:43 so I should pack that line with the ifdef? 2009-05-05T18:42:49 in the future if the parts are bigger (they will be), do #ifdef so the code atleast compiles if not behaves correctly 2009-05-05T18:42:51 Murmuria: it doesn't matter, if it's broken and known to be broken, it needs to be compiled out. 2009-05-05T18:42:57 exactl 2009-05-05T18:42:58 y 2009-05-05T18:43:50 KillerX: where in the man pages are the kernel services documented? 2009-05-05T18:45:22 jrm8005: if you mean in plan9, the man pages don't differentiate between services offered by the kernel and userspace programs 2009-05-05T18:45:38 though every man page will have a sources section which usually indicates what is where 2009-05-05T18:45:46 KillerX: yeah, I mean all the #c #p stuff. 2009-05-05T18:47:04 jrm8005: #c is documented in http://www.cs.bell-labs.com/magic/man2html/3/cons 2009-05-05T18:47:05 Title: Plan 9 /sys/man/3/cons (at www.cs.bell-labs.com) 2009-05-05T18:47:14 ah, I just saw it's section 3. 2009-05-05T18:47:19 #e in http://www.cs.bell-labs.com/magic/man2html/3/env and so on 2009-05-05T18:47:20 Title: Plan 9 /sys/man/3/env (at www.cs.bell-labs.com) 2009-05-05T18:47:20 :) 2009-05-05T18:48:10 BTW, KillerX you read http://www.spinics.net/lists/netdev/msg96395.html 2009-05-05T18:48:19 I sent my first email to a Linux ML 2009-05-05T18:48:24 no replies though 2009-05-05T18:48:30 you got any clues? 2009-05-05T18:49:55 jrm8005, KillerX , BTW, by broken code, you mean, it does not compile? 2009-05-05T18:50:04 I dont think I nothiced that 2009-05-05T18:50:12 kernel crashes in runtime 2009-05-05T18:50:16 not compile time 2009-05-05T18:50:18 Murmuria: no, though that should ifdef'd out. 2009-05-05T18:50:35 ya, but I hope you got no compile time errors 2009-05-05T18:50:42 Murmuria: even though it's runtime, I think it should be ifdef'd out, since it faults the kernel 2009-05-05T18:51:08 ya 2009-05-05T18:51:08 I totally agree 2009-05-05T18:51:08 I just never knew how to do it 2009-05-05T18:51:10 Murmuria: We need more details on what exactly the crash is :0 2009-05-05T18:51:10 #ifdef BROKEN is fine 2009-05-05T18:51:23 stack trace maybe? 2009-05-05T18:51:40 the one that we get when the kernel crashes? 2009-05-05T18:51:49 I dont think that is human redable 2009-05-05T18:52:01 how else can I generate a stack trace in kernel space? 2009-05-05T18:52:13 KillerX: how do you want to handle kernel devices that are already provided by Linux? 2009-05-05T18:52:34 Murmuria: no logs are not human readable, they are meant to be read :p 2009-05-05T18:52:52 e.g., #r can already be handled by probing the rtc module (and whatever provides the nvram file) 2009-05-05T18:52:57 kernel crashes are logged somewhere? 2009-05-05T18:53:01 jrm8005: I imagine we could use union mounts, though I'm not quite clear on the specific details yet 2009-05-05T18:53:05 ok, I have to go now 2009-05-05T18:53:07 BBL 2009-05-05T18:53:39 Murmuria: is there some way to make your VM not automatically reboot when the kernel crashes? 2009-05-05T18:53:47 because you're probably missing a stackdump. 2009-05-05T18:56:13 well, when the kernel crashes, it dumps a trace... and waits for me to reset the VM 2009-05-05T18:56:26 but the trace only gives me a function name 2009-05-05T18:56:44 um, yeah, looking at the trace would help. 2009-05-05T18:56:44 which I already know 2009-05-05T18:56:44 it should give you register values and such, no? 2009-05-05T18:56:47 the other things in the trace are hex numbers 2009-05-05T18:56:53 hmm 2009-05-05T18:57:12 so I am supposed to understand these hexnumbers? they are register values? 2009-05-05T18:57:17 Murmuria: you can also use the BUG_ON macro. 2009-05-05T18:57:45 so before you make the call, you can do something like BUG_ON(!strcmp(somestring, "hostname")); 2009-05-05T18:58:24 for testing purposes, then if somestring does not contain "hostname", the kernel will crash. 2009-05-05T18:58:24 and give you diagnostics. 2009-05-05T19:02:31 ok, let me see 2009-05-05T19:03:16 I'll work something out on Friday... I am at University right now... I'll add the #ifdef now though 2009-05-05T19:03:23 ok 2009-05-05T19:03:28 thanks 2009-05-05T19:03:32 yep 2009-05-05T19:29:34 *** less1 has joined #glendix 2009-05-05T20:30:32 *** TheAppleMan has quit IRC 2009-05-05T21:20:12 *** synchris_ has joined #glendix 2009-05-05T21:36:58 *** synchris has quit IRC 2009-05-05T21:41:49 *** phirox has joined #glendix 2009-05-05T21:51:49 http://www.reddit.com/r/linux/comments/8hzqj/glendix_bringing_the_beauty_of_plan_9_to_linux/ 2009-05-05T21:51:50 Title: Glendix Bringing the beauty of Plan 9 to Linux : linux (at www.reddit.com) 2009-05-05T22:13:29 uriel: thanks 2009-05-05T22:13:35 it's looking pretty popular. 2009-05-05T22:25:09 yea, I was surprised myself 2009-05-05T22:25:13 hey KillerX! 2009-05-05T22:25:20 yeah, the plan9 articles usually don't get too popular. 2009-05-05T22:25:29 KillerX: how/where are you? :) 2009-05-05T22:59:21 He is on holiday spree 2009-05-05T22:59:30 laters... 2009-05-05T22:59:32 *** Murmuria has quit IRC 2009-05-05T23:11:48 ah... 2009-05-05T23:11:57 Just thought of something very nice to do with Glendix. 2009-05-05T23:12:27 If 8c (or whatever - but right now only x86) can run on Linux, then you could use it with distcc. 2009-05-05T23:12:36 I see no reason why that wouldn't work. 2009-05-05T23:12:58 stu8ball: that should work, yes. 2009-05-05T23:13:16 you just would do 'distcc 8c foo; 2009-05-05T23:13:19 ' 2009-05-05T23:13:33 (or in my case, 6c) 2009-05-05T23:13:38 *** less1 has quit IRC 2009-05-05T23:26:11 *** KillerX has quit IRC 2009-05-05T23:30:58 *** less1 has joined #glendix 2009-05-05T23:35:27 *** less1 has quit IRC 2009-05-05T23:57:35 *** jrm8005 has quit IRC