2011-02-14T00:37:28 *** projectgus has quit IRC 2011-02-14T01:11:07 *** lcpfnyvc has joined #rtems 2011-02-14T01:42:34 morning 2011-02-14T01:47:55 *** lcpfnyvc has quit IRC 2011-02-14T01:48:21 *** lcpfnyvc has joined #rtems 2011-02-14T02:32:29 *** methril_ has quit IRC 2011-02-14T02:45:04 *** methril_ has joined #rtems 2011-02-14T05:22:31 *** madrazr has joined #rtems 2011-02-14T05:32:50 *** madrazr has quit IRC 2011-02-14T05:38:25 *** madrazr has joined #rtems 2011-02-14T07:39:38 *** madrazr has quit IRC 2011-02-14T07:53:43 *** madrazr has joined #rtems 2011-02-14T07:53:52 *** madrazr has joined #rtems 2011-02-14T08:32:39 *** antgreen has joined #rtems 2011-02-14T09:11:22 *** madrazr has quit IRC 2011-02-14T09:37:50 *** bswimley has joined #rtems 2011-02-14T10:58:46 *** DrJoel has joined #rtems 2011-02-14T10:58:46 *** DrJoel has joined #rtems 2011-02-14T10:58:46 *** ChanServ sets mode: +o DrJoel 2011-02-14T11:22:15 hi DrJoel :) 2011-02-14T11:23:02 hey! 2011-02-14T11:23:17 I filled a PR for the patch I was talking about yesterday 2011-02-14T11:23:43 let it be clear, I am not asking for a commit in RTEMS CVS tree, I am just asking for your opinion on the patch :) 2011-02-14T11:23:53 but you can commit it if you think it's correct 2011-02-14T11:25:08 and btw I have a question about RTEMS, is it for now capable of handling multi core/multi processor systems ? 2011-02-14T11:25:24 like running threads in parallel on several cores/processors ? 2011-02-14T11:25:35 I saw the description and if you are copying incoming packets into a buffer, the buffer had better be >= size of MTU 2011-02-14T11:26:25 The MP support now is for systems with separate CPUs connected by a bus -- communicating with message passing. Each CPU/board runs its own entire copy of the OS. 2011-02-14T11:26:47 18:03 <@DrJoel> I saw the description and if you are copying incoming packets into a buffer, the buffer had better be >= size of MTU 2011-02-14T11:26:51 that's what I think too 2011-02-14T11:27:03 oh ok 2011-02-14T11:27:10 We are working (slowly) on SMP support. The Score has been refactored to have a true scheduler handler class. This allows for SMP aware algorithm 2011-02-14T11:27:23 excellent :) 2011-02-14T11:27:28 I redid the core data structures so they are now instanced once per core. 2011-02-14T11:27:32 that would be awesome to have the same kind of support linux has 2011-02-14T11:27:48 (for SMP) 2011-02-14T11:27:53 There is a pr now with the next step .. treats a single core system as a single core SMP one from an OS object view. 2011-02-14T11:28:19 Next patches add BSP required support for pc386 and leon3. We still need SMP aware scheduler and locking addressed. 2011-02-14T11:28:33 So it is close and yet feels far away.. 2011-02-14T11:28:42 hehe yes I understand 2011-02-14T11:29:08 so I have the confirmation I needed 2011-02-14T11:29:12 Jennifer is working on the first alternate scheduler .. very lightweight simple scheduler for small systems. Helps TIny/RTEMS and verified pluggable scheduler infrastructure works as expected 2011-02-14T11:29:20 I will ask lekernel to test the driver again 2011-02-14T11:29:33 it is getting there. If someone needed it and pays some $$$, then it will get finished quicker LOL 2011-02-14T11:29:44 ahah I suppose yes 2011-02-14T11:29:54 a good GSoC subject so ? :) 2011-02-14T11:30:05 or maybe too big :x 2011-02-14T11:31:19 I think the pieces left are too scary to be student projects for most students. Would rather offer projects that we expect most can do a good job on. 2011-02-14T11:31:59 *** antgreen has quit IRC 2011-02-14T11:32:41 ok fair enough :) 2011-02-14T11:33:05 gedare bloom seems to think the fixe makes sense too 2011-02-14T11:33:10 thanks for your opinion 2011-02-14T11:33:30 FYI there is an issue in the Milkymist BSP network driver 2011-02-14T11:33:33 That's good. Do you read the configured mtu size and malloc the buffer 2011-02-14T11:33:50 that makes sometime the FIFO overflow 2011-02-14T11:33:54 I am trying to fix that 2011-02-14T11:34:10 I am not sure wether this memory corruption problem was related to it or not 2011-02-14T11:34:14 we'll see with more testing 2011-02-14T11:34:36 DrJoel: I am allocating 4 packet structures, each is given the max Ethernet MTU 2011-02-14T11:35:00 so 1500 + some other bytes for preamble and FCS 2011-02-14T11:36:02 4 packets structure because the ethernet controller has a FIFO with 4 slots for 4 packets 2011-02-14T11:36:20 (the device uses DMA to stream the packets into memory) 2011-02-14T11:36:50 ooohh... you definitely had better make them larger than they were!!! 2011-02-14T11:40:46 DrJoel: do you think my way of managing packet arrival is efficient ? 2011-02-14T11:40:55 is there a way to receive them faster ? 2011-02-14T11:41:03 in order to free slots quicker in the FIFO 2011-02-14T11:41:24 for example implementing like a fifo with like 10 packets in software 2011-02-14T11:41:32 and then change the DMA address in the op half 2011-02-14T11:41:33 of the driver 2011-02-14T11:43:19 top half* 2011-02-14T11:43:27 instead of waiting to be in the bottom half 2011-02-14T11:45:05 have you run ttcp? That will give you the throughput measure. You may not need to optimize 2011-02-14T12:31:51 *** antgreen has joined #rtems 2011-02-14T13:13:10 *** ssachdeva has joined #rtems 2011-02-14T13:13:41 *** ssachdeva has left #rtems 2011-02-14T13:29:22 *** antgreen has quit IRC 2011-02-14T13:37:38 *** antgreen has joined #rtems 2011-02-14T13:59:49 DrJoel: no I didn't try it, I will :) 2011-02-14T14:00:11 DrJoel: the reason why I would like to optimize is the RX FIFO gets full sometimes 2011-02-14T14:00:15 and it freezes the board 2011-02-14T14:00:18 I don't know why 2011-02-14T14:18:03 Ahhh.. is there any chance of an interrupt logic error? .. what does the linux driver do? 2011-02-14T14:23:56 humm dunno 2011-02-14T14:24:11 and I don't know if the linux driver is a good example 2011-02-14T14:24:25 I will have a look too 2011-02-14T14:35:07 if it doesn't lock up, then it isn't a bad one. LOL 2011-02-14T14:38:51 well yes :) 2011-02-14T14:40:58 *** madrazr has joined #rtems 2011-02-14T14:41:09 well DrJoel as you can see line 250 https://github.com/tmatsuya/linux-2.6/blob/master/drivers/net/milkymist_minimac.c 2011-02-14T14:41:18 it does not seem to handle anything in the top half :o 2011-02-14T14:41:21 hum hum 2011-02-14T14:41:28 will have to look over this 2011-02-14T14:41:39 Usually something like that is a hint 2011-02-14T14:41:52 sure 2011-02-14T14:51:39 OK i've just been told that the linux driver may not be working at all 2011-02-14T14:51:50 ^^" 2011-02-14T14:52:17 anyway I think I have enough pointers to start investigating the code now :) 2011-02-14T14:52:22 thanks ! 2011-02-14T14:52:53 wow! RTEMS ahead of linux for a change... go! 2011-02-14T14:52:59 ;) 2011-02-14T14:53:07 lm32 port of Linux is a shame 2011-02-14T14:53:13 it's dirty and does not work 2011-02-14T14:53:19 RTEMS is way better for this arch :) 2011-02-14T14:54:22 Takeshi MATSUYA is trying to get it cleaner 2011-02-14T14:54:26 but it's hell of a task :) 2011-02-14T15:06:52 Can you get an interrupt when the RX queue is half full? 2011-02-14T15:09:34 with the actual ip core design no 2011-02-14T15:09:42 but I think it could be done 2011-02-14T15:10:21 I think it would be pretty simple to do this in the verilog code 2011-02-14T15:10:43 but what I don't know is if it would make meeting timing any hardere for sebastien 2011-02-14T15:11:34 harder* 2011-02-14T15:11:47 sometimes a small change breaks everything 2011-02-14T15:14:12 sounds like sebastien needs a poke. :D 2011-02-14T15:14:33 sorry i didn't get the joke :p 2011-02-14T15:16:52 I think you need his help 2011-02-14T15:17:48 I will try without bothering him first :p 2011-02-14T15:17:59 Maybe there is no need for an extra interrupt 2011-02-14T15:18:21 I want to be sure before asking him to do any fpga design chirurgical operation 2011-02-14T15:23:39 I understand. I just thuoght he might have a hint on the best way to handle things in the current implemetnation 2011-02-14T15:26:38 he told be to do more job in the top half 2011-02-14T15:26:46 and having a pool of packet buffers 2011-02-14T15:27:58 and to remove the crc32 function from the driver and use the one provided by rtems core :p 2011-02-14T15:28:31 which I tried when developping the driver, without any luck 2011-02-14T15:28:40 (for the crc32) 2011-02-14T15:30:35 writing the in_cksum code in optimized assembly is tough. :-D 2011-02-14T15:31:00 yes I guess :p 2011-02-14T15:31:13 I just tried using the ethernet_crc32_be function from rtem 2011-02-14T15:31:20 and it generated wrong crc 2011-02-14T15:31:27 *** methril_ is now known as methril 2011-02-14T15:31:28 but maybe I did it wrong :) 2011-02-14T15:31:30 I will try again 2011-02-14T15:31:45 endianness can drive me crazy sometimes 2011-02-14T15:54:18 *** madrazr has quit IRC 2011-02-14T15:56:27 *** DrJoel has quit IRC 2011-02-14T15:56:59 *** antgreen has quit IRC 2011-02-14T16:11:05 *** projectgus has joined #rtems 2011-02-14T16:55:29 can anyone give me a hint about the fate of ticks_executed in the TCB (struct Thread_Control) in RTEMS? 2011-02-14T16:55:37 looks like it once existed, but now (pre-4.10) does not? 2011-02-14T17:12:48 aha, cpu_time_used is what I should have been using :) 2011-02-14T18:04:26 *** bswimley has quit IRC 2011-02-15T00:15:30 *** projectgus has quit IRC 2011-02-15T00:34:26 *** methril has quit IRC 2011-02-15T00:34:56 *** methril has joined #rtems 2011-02-15T01:09:30 morning 2011-02-15T02:16:10 * agrier tests spanning tree, just for the heck of it 2011-02-15T02:35:03 *** mini-HOWTO has joined #rtems 2011-02-15T03:04:03 *** bazinski has joined #rtems 2011-02-15T04:19:19 *** madrazr has joined #rtems 2011-02-15T05:35:10 *** ssachdeva has joined #rtems 2011-02-15T06:39:17 *** madrazr has left #rtems 2011-02-15T06:43:01 *** lcpfnyvc has quit IRC 2011-02-15T06:43:29 *** lcpfnyvc has joined #rtems 2011-02-15T06:47:55 *** madrazr has joined #rtems 2011-02-15T07:59:14 *** dr__house has joined #rtems 2011-02-15T07:59:14 *** dr__house has joined #rtems 2011-02-15T08:34:23 *** dr__house has quit IRC 2011-02-15T08:34:52 *** dr__house has joined #rtems 2011-02-15T09:51:23 *** bazinski has quit IRC 2011-02-15T11:56:36 *** ssachdeva has quit IRC 2011-02-15T11:56:39 *** peerst has quit IRC 2011-02-15T11:56:44 *** ChanServ has quit IRC 2011-02-15T11:56:52 *** methril has quit IRC 2011-02-15T11:56:57 *** dr__house has quit IRC 2011-02-15T12:05:20 *** peerst has joined #rtems 2011-02-15T12:06:53 *** methril has joined #rtems 2011-02-15T12:10:07 *** ssachdeva has joined #rtems 2011-02-15T12:10:07 *** ChanServ has joined #rtems 2011-02-15T12:10:07 *** jordan.freenode.net sets mode: +o ChanServ 2011-02-15T12:10:13 *** ChanServ has quit IRC 2011-02-15T12:23:01 *** ChanServ has joined #rtems 2011-02-15T12:23:01 *** jordan.freenode.net sets mode: +o ChanServ 2011-02-15T12:26:56 *** methril is now known as Guest64317 2011-02-15T12:38:24 *** madrazr has left #rtems 2011-02-15T13:52:52 *** antgreen has joined #rtems 2011-02-15T14:21:24 *** ssachdeva has left #rtems 2011-02-15T14:27:19 *** agrier has quit IRC 2011-02-15T15:04:42 *** agrier has joined #rtems 2011-02-15T15:21:27 *** dr__house has joined #rtems 2011-02-15T15:21:28 *** dr__house has joined #rtems 2011-02-15T15:49:31 *** antgreen has quit IRC 2011-02-15T15:57:29 *** projectgus has joined #rtems 2011-02-15T16:21:20 *** dr__house has quit IRC 2011-02-15T16:53:40 *** DrJoel has joined #rtems 2011-02-15T16:53:54 *** ChanServ sets mode: +o DrJoel 2011-02-15T17:20:22 *** bswimley has joined #rtems 2011-02-15T18:22:52 *** DrJoel has quit IRC 2011-02-15T21:47:44 *** dr__house has joined #rtems 2011-02-15T21:47:44 *** dr__house has joined #rtems 2011-02-15T22:13:29 *** dr__house` has joined #rtems 2011-02-15T22:13:29 *** dr__house` has quit IRC 2011-02-15T22:13:29 *** dr__house` has joined #rtems 2011-02-15T22:13:32 *** dr__house has quit IRC 2011-02-16T00:47:10 *** dr__house` has quit IRC 2011-02-16T02:32:29 *** Guest64317 has quit IRC 2011-02-16T02:44:49 *** Guest64317 has joined #rtems 2011-02-16T02:45:04 *** bazinski has joined #rtems 2011-02-16T03:53:20 *** madrazr has joined #rtems 2011-02-16T04:16:04 *** madrazr has quit IRC 2011-02-16T04:18:17 *** madrazr has joined #rtems 2011-02-16T06:50:32 *** projectgus has quit IRC 2011-02-16T07:51:17 *** madrazr has quit IRC 2011-02-16T08:06:01 *** madrazr has joined #rtems 2011-02-16T08:47:50 *** projectgus has joined #rtems 2011-02-16T08:49:27 *** antgreen has joined #rtems 2011-02-16T10:14:58 *** bazinski has quit IRC 2011-02-16T12:00:32 *** peerst has left #rtems 2011-02-16T12:26:34 *** peerst has joined #rtems 2011-02-16T12:57:12 *** dr__house has joined #rtems 2011-02-16T12:57:12 *** dr__house has joined #rtems 2011-02-16T13:29:26 *** dr__house has quit IRC 2011-02-16T13:29:32 *** dr__house has joined #rtems 2011-02-16T13:29:32 *** dr__house has joined #rtems 2011-02-16T13:37:39 *** madrazr has quit IRC 2011-02-16T14:45:01 *** bswimley has quit IRC 2011-02-16T14:46:30 *** dr__house has quit IRC 2011-02-16T16:44:20 *** lcpfnyvc has quit IRC 2011-02-16T16:46:59 *** lcpfnyvc has joined #rtems 2011-02-16T18:01:14 *** dr__house has joined #rtems 2011-02-16T18:01:14 *** dr__house has joined #rtems 2011-02-16T18:02:26 *** dr__house has quit IRC 2011-02-16T23:44:48 *** antgreen has quit IRC 2011-02-17T00:04:13 *** lcpfnyvc has quit IRC 2011-02-17T00:04:32 *** lcpfnyvc has joined #rtems 2011-02-17T04:20:29 *** madrazr has joined #rtems 2011-02-17T06:01:47 *** lcpfnyvc has quit IRC 2011-02-17T06:16:41 *** lcpfnyvc has joined #rtems 2011-02-17T06:38:17 *** antgreen has joined #rtems 2011-02-17T08:46:38 *** bazinski has joined #rtems 2011-02-17T09:31:02 *** core-ix has joined #rtems 2011-02-17T09:46:20 *** bazinski has quit IRC 2011-02-17T09:46:46 *** dr__house has joined #rtems 2011-02-17T09:46:47 *** dr__house has joined #rtems 2011-02-17T09:50:40 *** antgreen has quit IRC 2011-02-17T09:57:50 *** bswimley has joined #rtems 2011-02-17T11:17:33 *** gedare has joined #rtems 2011-02-17T12:24:52 *** gedare has quit IRC 2011-02-17T12:25:38 *** madrazr has quit IRC 2011-02-17T12:37:30 *** core-ix has quit IRC 2011-02-17T12:52:23 *** DrJoel has joined #rtems 2011-02-17T12:52:23 *** ChanServ sets mode: +o DrJoel 2011-02-17T13:18:01 *** gedare has joined #rtems 2011-02-17T13:27:57 *** kristianpaul has quit IRC 2011-02-17T13:44:55 *** core-ix has joined #rtems 2011-02-17T15:26:54 *** gedare has quit IRC 2011-02-17T16:11:09 *** dr__house has quit IRC 2011-02-17T16:35:29 *** DrJoel has quit IRC 2011-02-17T17:04:25 *** Guest64317 has quit IRC 2011-02-17T17:04:50 *** bswimley has quit IRC 2011-02-17T17:08:04 *** methril has joined #rtems 2011-02-17T17:19:07 *** antgreen has joined #rtems 2011-02-17T17:48:59 *** kiwichris has joined #rtems 2011-02-17T17:49:22 hello all 2011-02-17T18:13:51 hi 2011-02-17T18:13:59 hi 2011-02-17T18:14:26 i don't quite well understand Graham problem 2011-02-17T18:14:29 about lm32_evr 2011-02-17T18:14:33 he seems to type the good commands 2011-02-17T18:21:07 Yes he does but if a tool set up is not 100% a check can fail and then the make breaks down 2011-02-17T18:21:27 I do not use the cygwin tools. Do you ? 2011-02-17T18:23:49 * kiwichris is out for a while 2011-02-17T20:20:46 *** kiwichris has quit IRC 2011-02-17T20:51:59 *** kiwichris has joined #rtems 2011-02-17T21:23:55 *** kristianpaul has joined #rtems 2011-02-17T21:23:55 *** kristianpaul has joined #rtems 2011-02-17T21:33:28 *** kiwichris has quit IRC 2011-02-17T22:48:19 *** kiwichris has joined #rtems 2011-02-17T23:36:25 *** projectgus has quit IRC 2011-02-18T02:32:29 *** methril has quit IRC 2011-02-18T02:45:04 *** methril has joined #rtems 2011-02-18T03:25:50 *** core-ix has quit IRC 2011-02-18T04:05:25 *** core-ix has joined #rtems 2011-02-18T04:06:58 *** madrazr has joined #rtems 2011-02-18T05:49:05 *** madrazr has left #rtems 2011-02-18T06:12:55 *** antgreen has quit IRC 2011-02-18T07:11:59 *** DrJoel has joined #rtems 2011-02-18T07:11:59 *** ChanServ sets mode: +o DrJoel 2011-02-18T07:47:26 *** tuxmania1 is now known as tuxmaniac 2011-02-18T07:47:36 *** tuxmaniac has joined #rtems 2011-02-18T08:17:47 *** kristianpaul has quit IRC 2011-02-18T08:18:05 *** kristianpaul has joined #rtems 2011-02-18T09:21:39 *** madrazr has joined #rtems 2011-02-18T09:23:10 *** bswimley has joined #rtems 2011-02-18T11:39:25 *** core-ix has quit IRC 2011-02-18T11:40:45 *** core-ix has joined #rtems 2011-02-18T11:47:22 *** madrazr has left #rtems 2011-02-18T12:11:05 *** gedare has joined #rtems 2011-02-18T12:11:47 DrJoel: I see the commits for the scheduler, overall I'm pleased with the changes made. 2011-02-18T12:12:11 I'm not sure on the name of the "information" field of Scheduler_Control, but I understand the change to void*, and it does simplify the interface a bit. 2011-02-18T12:44:20 I didn't have a great name.. suggestions appreciated. 2011-02-18T12:44:54 The goal now is to have the plugin interface support a "simple priority" scheduler as an alternative and to be able to document it so another human being can implement one. 2011-02-18T13:03:23 Sounds good. Not sure on the name myself. information seems misleading though, since it is used for scheduler-private data 2011-02-18T13:22:28 We can easily change the name once we get a better one. 2011-02-18T13:22:47 FWIW hello and ticker now run with "Simple Priority Scheduler" which is a linear insertion in priority order. 2011-02-18T14:46:53 *** core-ix has quit IRC 2011-02-18T15:02:38 *** DrJoel has quit IRC 2011-02-18T15:40:01 *** core-ix has joined #rtems 2011-02-18T16:08:01 *** gedare has quit IRC 2011-02-18T16:11:36 *** bswimley has quit IRC 2011-02-18T16:11:40 *** dr__house has joined #rtems 2011-02-18T16:11:40 *** dr__house has joined #rtems 2011-02-18T16:33:24 *** kiwichris has quit IRC 2011-02-18T16:37:11 *** antgreen has joined #rtems 2011-02-18T16:58:27 *** dr__house has quit IRC 2011-02-18T20:25:19 *** dr__house has joined #rtems 2011-02-18T20:25:19 *** dr__house has joined #rtems 2011-02-18T21:13:10 *** dr__house` has joined #rtems 2011-02-18T21:13:10 *** dr__house` has joined #rtems 2011-02-18T21:13:13 *** dr__house has quit IRC 2011-02-18T22:00:40 *** peerst has quit IRC 2011-02-18T22:00:41 *** core-ix has quit IRC 2011-02-18T22:00:42 *** core-ix has joined #rtems 2011-02-18T22:08:51 *** dr__house` has quit IRC 2011-02-19T07:08:30 *** rokka__ has joined #rtems 2011-02-19T07:13:51 *** rokka_ has quit IRC 2011-02-19T07:43:37 *** peerst has joined #rtems 2011-02-19T08:05:07 *** kristianpaul has quit IRC 2011-02-19T08:12:08 *** kristianpaul has joined #rtems 2011-02-19T08:12:09 *** kristianpaul has joined #rtems 2011-02-19T09:07:27 *** dr__house has joined #rtems 2011-02-19T09:07:27 *** dr__house has joined #rtems 2011-02-19T11:13:50 *** ssachdeva has joined #rtems 2011-02-19T11:13:56 *** ssachdeva has left #rtems 2011-02-19T14:15:54 *** dr__house has quit IRC 2011-02-19T18:07:40 *** lcpfnyvc has quit IRC 2011-02-19T18:08:23 *** lcpfnyvc has joined #rtems 2011-02-19T19:38:07 *** dr__house has joined #rtems 2011-02-19T19:38:07 *** dr__house has joined #rtems 2011-02-20T01:37:54 *** dr__house has quit IRC 2011-02-20T02:32:53 *** methril has quit IRC 2011-02-20T02:45:02 *** methril has joined #rtems 2011-02-20T07:17:27 *** dr__house has joined #rtems 2011-02-20T07:17:27 *** dr__house has joined #rtems 2011-02-20T09:24:05 *** arvind_khadri has joined #rtems 2011-02-20T10:27:27 *** dr__house has quit IRC 2011-02-20T10:27:42 *** dr__house has joined #rtems 2011-02-20T10:27:42 *** dr__house has joined #rtems 2011-02-20T13:00:24 Why so silent :) ? 2011-02-20T13:11:15 boooo 2011-02-20T13:11:18 :) 2011-02-20T13:27:38 *** dr__house has quit IRC 2011-02-20T15:41:43 *** Fallenou has quit IRC 2011-02-20T15:44:29 *** Fallenou has joined #rtems 2011-02-20T16:05:09 *** peerst has left #rtems 2011-02-20T16:07:33 *** projectgus has joined #rtems 2011-02-20T16:41:44 *** peerst has joined #rtems 2011-02-20T18:22:54 *** kristianpaul has quit IRC 2011-02-20T18:57:07 *** kristianpaul has joined #rtems 2011-02-20T19:13:35 *** arvind_khadri has quit IRC 2011-02-20T19:14:04 *** arvind_khadri has joined #rtems 2011-02-20T19:14:04 *** arvind_khadri has joined #rtems 2011-02-20T20:23:14 *** dr__house has joined #rtems 2011-02-20T20:23:14 *** dr__house has joined #rtems 2011-02-20T21:11:45 *** dr__house has quit IRC 2011-02-20T23:39:52 *** arvind_khadri has quit IRC