2011-03-14T00:39:11 *** methril has quit IRC 2011-03-14T00:39:35 *** kiwichris has joined #rtems 2011-03-14T01:00:01 *** methril has joined #rtems 2011-03-14T01:48:30 *** antgreen has quit IRC 2011-03-14T03:13:11 *** sebhub has joined #rtems 2011-03-14T03:17:04 good morning 2011-03-14T03:17:20 sebhub, hi 2011-03-14T03:27:43 *** core-ix has quit IRC 2011-03-14T03:59:26 *** dr__house has joined #rtems 2011-03-14T04:06:21 *** core-ix has joined #rtems 2011-03-14T04:23:52 *** mmu_man has joined #rtems 2011-03-14T04:48:55 *** dr__house has quit IRC 2011-03-14T06:02:06 *** mmu_man has quit IRC 2011-03-14T06:45:09 *** antgreen has joined #rtems 2011-03-14T06:49:25 *** antgreen has quit IRC 2011-03-14T06:49:42 *** antgreen has joined #rtems 2011-03-14T07:18:56 *** methril has quit IRC 2011-03-14T07:31:12 *** methril has joined #rtems 2011-03-14T08:06:15 *** mmu_man has joined #rtems 2011-03-14T08:28:26 *** DrJoel has joined #rtems 2011-03-14T08:28:26 *** ChanServ sets mode: +o DrJoel 2011-03-14T08:40:33 *** madrazr has joined #rtems 2011-03-14T09:07:59 *** zwj has joined #rtems 2011-03-14T09:31:21 *** antgreen has quit IRC 2011-03-14T09:32:13 *** mmu_man has quit IRC 2011-03-14T09:40:18 *** arvind_khadri has joined #rtems 2011-03-14T09:49:06 *** madrazr has quit IRC 2011-03-14T10:15:23 *** mmu_man has joined #rtems 2011-03-14T10:32:26 *** arvind_khadri has quit IRC 2011-03-14T10:32:47 *** lcpfnyvc has quit IRC 2011-03-14T10:56:42 *** dr__house has joined #rtems 2011-03-14T10:58:07 *** dr__house has quit IRC 2011-03-14T11:02:26 *** madrazr has joined #rtems 2011-03-14T11:07:11 *** sebhub has quit IRC 2011-03-14T11:08:46 *** zwj has quit IRC 2011-03-14T11:09:52 *** dr__house has joined #rtems 2011-03-14T11:09:52 *** dr__house has joined #rtems 2011-03-14T11:11:02 quiet in here today 2011-03-14T11:24:21 *** arvind_khadri has joined #rtems 2011-03-14T11:24:21 *** arvind_khadri has joined #rtems 2011-03-14T11:50:05 *** madrazr has quit IRC 2011-03-14T12:11:23 *** core-ix has quit IRC 2011-03-14T12:13:29 *** arvind_khadri has quit IRC 2011-03-14T12:16:01 *** mmu_man has quit IRC 2011-03-14T12:22:19 *** dr__house has quit IRC 2011-03-14T12:27:20 *** arvind_khadri has joined #rtems 2011-03-14T12:27:21 *** arvind_khadri has joined #rtems 2011-03-14T12:56:34 *** mmu_man has joined #rtems 2011-03-14T13:01:31 *** arvind_khadri has quit IRC 2011-03-14T13:03:55 *** dr__house has joined #rtems 2011-03-14T13:04:03 *** dr__house has joined #rtems 2011-03-14T13:32:17 *** dr__house has quit IRC 2011-03-14T13:48:35 *** mmu_man_ has joined #rtems 2011-03-14T13:50:46 *** mmu_man has quit IRC 2011-03-14T14:07:43 DrJoel: hello. i managed to compile i386 toolchain, only had some wrong parameter to gcc and binutils 2011-03-14T14:28:45 *** madrazr has joined #rtems 2011-03-14T14:31:55 *** madrazr has quit IRC 2011-03-14T14:35:51 *** mmu_man_ is now known as mmu_man 2011-03-14T14:43:53 rokka: that's good news. Did you manage to run a test on qemu? 2011-03-14T15:26:57 kiwichris: you around? 2011-03-14T15:31:14 *** antgreen has joined #rtems 2011-03-14T15:42:15 *** core-ix has joined #rtems 2011-03-14T16:11:15 *** kiwichris has quit IRC 2011-03-14T16:11:21 *** DrJoel has quit IRC 2011-03-14T16:33:44 *** hex-code has joined #rtems 2011-03-14T16:37:24 *** demute has joined #rtems 2011-03-14T16:40:39 *** projectgus has joined #rtems 2011-03-14T17:15:44 *** kiwichris has joined #rtems 2011-03-14T17:27:37 *** dr__house has joined #rtems 2011-03-14T17:27:37 *** dr__house has joined #rtems 2011-03-14T17:47:48 *** demute has quit IRC 2011-03-14T18:08:16 *** methril has quit IRC 2011-03-14T18:11:20 *** methril has joined #rtems 2011-03-14T18:30:57 *** dr__house has quit IRC 2011-03-14T18:52:03 *** mmu_man has quit IRC 2011-03-14T20:02:17 *** DrJoel has joined #rtems 2011-03-14T20:02:17 *** DrJoel has joined #rtems 2011-03-14T20:02:17 *** ChanServ sets mode: +o DrJoel 2011-03-14T20:06:11 * kiwichris waves to Joel 2011-03-14T20:07:10 hey.. just got in from the hot tub.. the brackets on the lid snapped when I started to shut it. It too me and two of the boys to put it on after it fell. :( Time for the repair guys. :( 2011-03-14T20:08:09 The hot tub sounds nice. I dig my calf muscle on the weekend and could do with a hot tub 2011-03-14T20:08:25 dig -> did 2011-03-14T20:08:40 I have found an API wrinkle in the file system. 2011-03-14T20:08:55 it is nice to have one. We haven't used ours since the snow collapsed the canopy over it. I got a new one up yesterday and this was the first time using it. And this... 2011-03-14T20:09:08 seek or stat? 2011-03-14T20:09:42 In the handlers. If you look in libio.h the file specific handlers for fstat take a pathloc while all others take a rtems_libio_t. 2011-03-14T20:10:18 The stat call which is the file system handler one, ie from stat, takes a pathlog 2011-03-14T20:10:37 that's not consistent. Is it hard to fix? 2011-03-14T20:10:51 This means you can use the same handler in the file handlers and the file system handlers 2011-03-14T20:11:24 How-ever this assumes the node_access etc in use the file system has the same interface for fstat and stat 2011-03-14T20:11:53 This is not the case the RFS. The RFS uses an ino in node_access for file system calls and a file handler pointer for files. 2011-03-14T20:12:45 The fix is not difficult but is it worth it. Seb's change to remove path_conf from the iop broke the RFS and it did not show up until now. 2011-03-14T20:12:55 I wonder if we should have made stat just do an open, fstat close 2011-03-14T20:13:20 what about lstat? 2011-03-14T20:13:20 The call to the file system is correct. I would not do an open. Far too slow 2011-03-14T20:13:30 Just looking. 2011-03-14T20:14:06 Hmmm a define hack that uses the stat code. This is a file system call. 2011-03-14T20:14:27 It takes a path and find the pathloc so that is ok. 2011-03-14T20:14:44 Figure out what's right and do what you have to. If you need help, just yell. 2011-03-14T20:14:49 The issue is the overloading of the node_access variable in the iop 2011-03-14T20:14:58 FYI I have local changes for the coverity fatfs issues. And a couple of ones in the shell 2011-03-14T20:15:08 It add a small amount of code to the DOSFS 2011-03-14T20:15:29 * DrJoel can't believe the scale of tragedy in Japan.. I thought Christchurch looked bad .. but this looks horrible 2011-03-14T20:15:44 It is very sad. 2011-03-14T20:17:09 John E from the Elvis site has apparently managed to find a new place to move to. He posted photos of his place and it looked awful.. 2011-03-14T20:18:07 kiwichris: it is :( 2011-03-14T20:18:26 one of my clients distributed photos internally of some techs that flew by helicopter to one of our datacenters there 2011-03-14T20:18:55 something punched a hole in the wall and water flooded in, but none of the employees there can be found (yet they may be at a shelter with no way to get out) 2011-03-14T20:19:24 the building was built to widthstand worse than this but not several dozen tons of debris smashing into it at 30km/h 2011-03-14T20:19:39 Fear mongering here is worrying about San Andreas being next.. plots of earthquakes recently make you suspicious. We are kind of close to New Madrid fault here 2011-03-14T20:19:50 Nothing can withstand that. 2011-03-14T20:20:07 DrJoel: yeah :( 2011-03-14T20:20:45 I watched a report from a coastal town of 17K with about 10K missing. Nothing but destruction for miles inland 2011-03-14T20:20:51 how far inliand was the data center 2011-03-14T20:21:48 almost 1.5mi 2011-03-14T20:22:01 or is that km i forget 2011-03-14T20:22:04 km probably 2011-03-14T20:22:23 not far enough for this one. 2011-03-14T20:22:27 DrJoel: well there were several problems, the first one was that people usually have hours of sirens and warnings 2011-03-14T20:22:35 this tsunami hit in 30mins 2011-03-14T20:22:52 the second is that it was incredibly massive 2011-03-14T20:23:05 that was mentioned for this coastal town.. plus it only had one road in/out and it got clogged.. then slammed 2011-03-14T20:23:26 yeah well all those communities are buffeted by mountains on either side so people climed up (there are trails) 2011-03-14T20:23:39 also schools in japan generally built on hills (for this very reason) 2011-03-14T20:24:04 but not enough for this one unfortunately.. :( 2011-03-14T20:24:18 no :( a few towns were dessimated 2011-03-14T20:24:58 If the reactors don't blow up, it will be amazing.. they are well beyond their design parameters 2011-03-14T20:25:19 http://news.bbcimg.co.uk/media/images/51662000/jpg/_51662576_51662575.jpg incredible photo 2011-03-14T20:25:23 DrJoel: they're light water recators 2011-03-14T20:25:43 the second explosion should not have happened 2011-03-14T20:25:54 they're using fire trucks to pump water over the fuel rods 2011-03-14T20:25:58 they ran out of gas :( 2011-03-14T20:26:34 and uncovered rods == hydrogen, plus compressed steam as soon as that hits the air == boom 2011-03-14T20:26:35 that's a dreadful picture 2011-03-14T20:26:44 so the explosions (both) probably happened as soon as they tried to vent 2011-03-14T20:27:15 it is but it is not as bad as it sounds, remember the reactors were turned off in the first seconds of the quake 2011-03-14T20:27:23 chernoyble was running at full blast 2011-03-14T20:27:42 also with a light water reactor the only worry you have is a steam explosion which is pretty bad 2011-03-14T20:27:51 but at this point all they need to do is keep the fuel rods cool if they can do that they'll be fine 2011-03-14T20:28:06 the disaster plans couldn't account for the mutltiple events.. running out of gas.. they probably never conceived of that occurring 2011-03-14T20:28:09 but who knows there was a 3rd explosion now, with no word on the cause 2011-03-14T20:28:28 DrJoel: it's worse than that 2011-03-14T20:28:45 the generators were put on the ground beacuse they thought their tsunami wall was enough 2011-03-14T20:28:57 the offsite power didn't have multiple redundant connections so the quake knocked that off 2011-03-14T20:29:05 and they only had enough batteries to run for hours not days 2011-03-14T20:29:24 I guess they never read Lucifer's Hammer. 2011-03-14T20:30:07 nuclear plant ends up behind dikes after a comet strike 2011-03-14T20:30:35 oh well.. time to head to bed. Michele was watchign something I didn't want to so I headed for the computer 2011-03-14T20:30:37 night all 2011-03-14T20:30:40 *** DrJoel has quit IRC 2011-03-14T21:13:49 *** dr__house has joined #rtems 2011-03-14T21:13:49 *** dr__house has joined #rtems 2011-03-14T21:36:11 *** dr__house has quit IRC 2011-03-14T21:53:25 *** kiwichris has quit IRC 2011-03-14T22:24:39 *** kiwichris has joined #rtems 2011-03-14T22:31:40 *** arvind_khadri has joined #rtems 2011-03-14T22:31:40 *** arvind_khadri has joined #rtems 2011-03-14T22:45:29 *** arvind_khadri has quit IRC 2011-03-14T22:51:00 *** arvind_khadri has joined #rtems 2011-03-14T22:51:00 *** arvind_khadri has joined #rtems 2011-03-14T23:24:17 *** arvind_khadri has quit IRC 2011-03-14T23:37:25 *** arvind_khadri has joined #rtems 2011-03-14T23:37:25 *** arvind_khadri has joined #rtems 2011-03-14T23:59:44 *** arvind_khadri has quit IRC 2011-03-15T00:35:06 *** mmu_man has joined #rtems 2011-03-15T00:38:21 *** agrier has quit IRC 2011-03-15T00:40:41 *** methril has quit IRC 2011-03-15T00:44:19 *** methril has joined #rtems 2011-03-15T01:13:00 *** projectgus has quit IRC 2011-03-15T01:15:37 *** lcpfnyvc has joined #rtems 2011-03-15T01:21:22 *** core-ix has quit IRC 2011-03-15T01:40:27 *** methril has quit IRC 2011-03-15T01:43:36 *** arvind_khadri has joined #rtems 2011-03-15T01:43:36 *** arvind_khadri has joined #rtems 2011-03-15T01:44:40 *** methril has joined #rtems 2011-03-15T01:48:42 *** mmu_man has quit IRC 2011-03-15T02:10:41 *** core-ix has joined #rtems 2011-03-15T02:31:19 *** sebhub has joined #rtems 2011-03-15T02:31:25 good morning 2011-03-15T02:31:40 sebhub, hi 2011-03-15T02:35:12 *** agrier has joined #rtems 2011-03-15T02:54:54 sebhub, hi 2011-03-15T03:37:55 *** madrazr has joined #rtems 2011-03-15T03:42:27 *** madrazr has quit IRC 2011-03-15T05:05:51 *** arvind_khadri has quit IRC 2011-03-15T05:17:40 *** mmu_man has joined #rtems 2011-03-15T05:18:06 *** arvind_khadri has joined #rtems 2011-03-15T05:18:07 *** arvind_khadri has joined #rtems 2011-03-15T06:00:26 *** mmu_man has quit IRC 2011-03-15T06:36:00 *** mmu_man has joined #rtems 2011-03-15T07:27:13 *** peerst has joined #rtems 2011-03-15T08:27:41 *** mmu_man has quit IRC 2011-03-15T08:38:29 *** core-ix has quit IRC 2011-03-15T08:38:44 *** zwj has joined #rtems 2011-03-15T08:52:18 *** mmu_man has joined #rtems 2011-03-15T09:06:14 *** dr__house has joined #rtems 2011-03-15T09:06:14 *** dr__house has joined #rtems 2011-03-15T09:10:37 *** madrazr has joined #rtems 2011-03-15T09:19:44 *** dr__house has quit IRC 2011-03-15T09:19:49 *** dr__house` has joined #rtems 2011-03-15T09:20:04 *** dr__house` is now known as dr__house 2011-03-15T10:25:10 *** madrazr has quit IRC 2011-03-15T11:04:42 *** sebhub has quit IRC 2011-03-15T11:12:46 *** zwj has quit IRC 2011-03-15T11:13:53 *** mmu_man has quit IRC 2011-03-15T11:18:09 *** arvind_khadri has quit IRC 2011-03-15T11:33:21 *** arvind_khadri has joined #rtems 2011-03-15T11:44:53 *** arvind_khadri has quit IRC 2011-03-15T12:32:51 *** mmu_man has joined #rtems 2011-03-15T12:35:00 *** arvind_khadri has joined #rtems 2011-03-15T14:02:42 *** antgreen has quit IRC 2011-03-15T14:38:01 *** DrJoel has joined #rtems 2011-03-15T14:38:02 *** ChanServ sets mode: +o DrJoel 2011-03-15T14:58:33 *** gedare has joined #rtems 2011-03-15T14:58:37 *** core-ix has joined #rtems 2011-03-15T15:53:13 *** arvind_khadri has quit IRC 2011-03-15T16:04:55 *** gedare has quit IRC 2011-03-15T16:20:31 DrJoel: sure.. i tested hello.exe 2011-03-15T16:21:11 does it really run on real pc and ethernet works with ne2000 eth adapter? 2011-03-15T16:32:49 I have seen it on real pc's..laptops, embedded PCs, desktops, etc but a real ne2000 NIC is hard to come by these days LOL 2011-03-15T16:33:10 qemu has been broken on and off for ne2k. Try the Intel fxp driver with like an i82551 2011-03-15T16:35:31 i'll try now to boot hello.exe with grub on my laptop 2011-03-15T16:41:24 *** dr__house has quit IRC 2011-03-15T16:44:22 * DrJoel heads home 2011-03-15T16:44:25 *** DrJoel has quit IRC 2011-03-15T16:50:31 worked :) 2011-03-15T17:39:26 *** projectgus has joined #rtems 2011-03-15T17:50:22 error: segment doesn't fit in memory reserved for the OS (0x0 < 0x100000) 2011-03-15T17:50:23 :< 2011-03-15T17:59:52 try modifying the linker script 2011-03-15T18:00:06 to increase the size of the region that hosts the segment 2011-03-15T18:00:39 I had this kind of error today with my linker for a PIC18F ^^ 2011-03-15T18:24:43 got it working.. 2011-03-15T18:24:49 i had wrong compiler parameters :) 2011-03-15T18:37:27 *** dr__house has joined #rtems 2011-03-15T18:37:27 *** dr__house has joined #rtems 2011-03-15T18:40:30 Fallenou, for PIC18F? 2011-03-15T18:40:38 Fallenou, what are you doing with PIC18F? 2011-03-15T18:40:45 some robotics stuff 2011-03-15T18:40:49 motor controlling 2011-03-15T18:40:50 PID 2011-03-15T18:40:55 srevoing 2011-03-15T18:40:57 Fallenou, nice!! 2011-03-15T18:40:57 servoing* 2011-03-15T18:41:12 did you use sdcc? or other compiler? 2011-03-15T19:05:23 *** dr__house has quit IRC 2011-03-15T19:09:55 *** mmu_man has quit IRC 2011-03-15T19:20:27 *** dr__house has joined #rtems 2011-03-15T21:29:44 *** dr__house has quit IRC 2011-03-16T00:01:04 *** verm__ has quit IRC 2011-03-16T02:34:57 *** sebhub has joined #rtems 2011-03-16T02:35:10 good morning 2011-03-16T02:58:33 *** madrazr has joined #rtems 2011-03-16T03:22:09 *** sebhub has quit IRC 2011-03-16T03:22:17 *** sebhub has joined #rtems 2011-03-16T03:35:08 *** core-ix has quit IRC 2011-03-16T04:28:57 *** core-ix has joined #rtems 2011-03-16T04:40:44 *** madrazr has quit IRC 2011-03-16T05:57:48 *** cloud_ has joined #rtems 2011-03-16T05:58:02 *** hex-code has quit IRC 2011-03-16T06:54:25 *** mmu_man has joined #rtems 2011-03-16T07:01:15 *** cloud_ has quit IRC 2011-03-16T07:16:00 *** methril has quit IRC 2011-03-16T07:29:39 *** mwalle has quit IRC 2011-03-16T07:29:40 *** Fallenou has quit IRC 2011-03-16T07:29:40 *** Fallenou has joined #rtems 2011-03-16T07:31:23 *** methril has joined #rtems 2011-03-16T07:33:35 *** mwalle has joined #rtems 2011-03-16T07:38:37 *** peerst has left #rtems 2011-03-16T08:15:16 *** antgreen has joined #rtems 2011-03-16T08:22:15 *** cloud_ has joined #rtems 2011-03-16T08:22:42 Hello 2011-03-16T08:23:04 hi 2011-03-16T08:24:39 is there a way to replace function, that already loaded to device, without reloading whole application? 2011-03-16T08:26:39 you need dynamic linking for this 2011-03-16T08:28:16 i've extracted the function from the elf file and not sured that addresses in this function is not absolute 2011-03-16T08:28:35 how i can using the dinamical linking? 2011-03-16T08:32:57 in function that i've extracted i don't have a _rodata and i don't know how to extract this data from elf file and link together with my function 2011-03-16T08:46:25 sorry, what do you want to do and why can't you just update your application as a whole? 2011-03-16T08:47:18 well let's imagine that my object is in open space :) 2011-03-16T08:48:00 it should be controlled every second and i want, for example, update part of navigation system 2011-03-16T08:48:23 so it would be perfect if i could just replace existing function with new one 2011-03-16T08:49:27 ok, you can do this with dynamic linking, but this will not work out of the box with rtems, you may ask on the mailing list 2011-03-16T08:50:34 so 2011-03-16T08:50:57 did i understand right, that if i use rtems, i need to update whole application every time when i want to do even minor changes? 2011-03-16T08:51:14 this is the standard way, yes 2011-03-16T08:51:24 ok, thank you very much 2011-03-16T08:52:08 *** cloud_ has quit IRC 2011-03-16T09:05:07 *** zwj has joined #rtems 2011-03-16T10:21:21 *** DrJoel has joined #rtems 2011-03-16T10:21:21 *** ChanServ sets mode: +o DrJoel 2011-03-16T10:24:16 sebhub: got a second for a lpc2478 setup question? 2011-03-16T10:26:01 yes 2011-03-16T10:28:04 have you hooked the USB/serial up on Linux? 2011-03-16T10:28:40 we see data from it but can't TX to it to stop u-boot autobooting. I am wondering if there are hw handshaking settings we need to change in minicom 2011-03-16T10:29:17 which board do you use? 2011-03-16T10:29:31 lpc2478.. the embedded artists board you guys gave me 2011-03-16T10:29:50 Richard is trying to integrate it into the lab rack and I don't remember much 2011-03-16T10:29:51 hm, it worked out of the box here 2011-03-16T10:30:10 not hardware handshaking 2011-03-16T10:30:14 then must be a minicom setting 2011-03-16T10:30:23 disable hardware handshaking? 2011-03-16T10:31:06 maybe your u-boot image is not right 2011-03-16T10:31:15 please have a look at the README of the bsp 2011-03-16T10:31:36 i can send you the docs and hex files 2011-03-16T10:32:02 we see u-boot so it isn't that bad. When we hit space, the baord doesn't respond 2011-03-16T10:32:29 what is the u-boot version? 2011-03-16T10:32:52 hold on... 2011-03-16T10:40:13 UBoot 1.1.6.. 2011-03-16T10:40:29 this should be ok 2011-03-16T10:40:35 was hardware flow control. Has to be disabled. Now accessible on Linux and Windows.. stupid problem.. 2011-03-16T10:41:14 Can you email me a known good executable? I assume there is a mkimage command I don't know to get from elf -> uboot image 2011-03-16T10:43:48 do you have the mkimage tool? 2011-03-16T10:46:35 yes. 2011-03-16T10:51:33 mkimage -A arm -O rtems -T kernel -C gzip -a a0000000 -e a0000040 -n "RTEMS Application" -d app.bin.gz app.img 2011-03-16T10:53:17 Got it. Thanks.. 2011-03-16T10:54:14 Richard is trying to make progress on the rack of boards in the RTEMS Lab... the goal is to provide remote access to them. 2011-03-16T10:55:22 this is cool 2011-03-16T10:58:44 has taken a while.. started setup at old office.. boxed it up.. started again here.. then had construction.. then boxed it up again.. now getitng closer 2011-03-16T11:04:54 have to go, bye 2011-03-16T11:05:04 *** zwj has quit IRC 2011-03-16T11:05:42 i booted my laptop with i386 version of 4.11 :) 2011-03-16T11:06:05 and got my qt-ide environment building apps for it 2011-03-16T11:07:44 i tried to port nbench to work on rtems (http://www.tux.org/~mayer/linux/bmark.html) 2011-03-16T11:08:02 but it blows stack when it runs string sort test 2011-03-16T11:08:17 it doesn't matter how big stack i reserve 2011-03-16T11:08:28 *** sebhub has quit IRC 2011-03-16T11:08:29 must debug this today:) 2011-03-16T11:08:59 qemu + i386 gdb should be easy? 2011-03-16T11:09:54 and the nbench works well with the string test disabled :) it shows about same performance on qemu than with booting laptop directly 2011-03-16T11:10:14 qemu must be using kvm 2011-03-16T11:10:52 of course native linux nbench show better results on my laptop.. maybe because it has better gcc optimizations (-march=native vs -march=i386) 2011-03-16T11:16:47 yes qemu+i386 gdb is very easy.. just add -s to the qemu invocation 2011-03-16T11:16:50 tar remote :1234 2011-03-16T11:16:58 and set break points.. away you go 2011-03-16T11:17:42 you can use another set of gcc options.. check c/src/lib/libbsp/i386/pc386/make/custom/*cfg.. if there is enough difference, then we can consider another variant 2011-03-16T11:17:46 hold on .. brb 2011-03-16T11:59:31 ok:) 2011-03-16T12:00:29 *** madrazr has joined #rtems 2011-03-16T12:02:14 * DrJoel is back 2011-03-16T12:06:29 :) 2011-03-16T13:12:58 ok.. ddd is working 2011-03-16T13:13:11 and of course direct gdb 2011-03-16T13:13:27 qt ide + i386-rtems411-gdb does not work 2011-03-16T13:23:09 LOL.. it is always the one combination you want to work that does not. LOL 2011-03-16T13:24:01 yep 2011-03-16T13:24:19 im not sure how can i debug blowing stack 2011-03-16T13:25:06 You could turn on RTEMS Stack Checker and periodically check on its usage 2011-03-16T13:25:07 can i ask how much stack does certain task use 2011-03-16T13:25:17 from api 2011-03-16T13:25:21 *** dr__house has joined #rtems 2011-03-16T13:25:21 *** dr__house has joined #rtems 2011-03-16T13:25:31 You can print it with routines in stack checker 2011-03-16T13:25:57 cpukit/libmisc/stackchk ... easy api.. one configuration option to enable putting pattern in stack 2011-03-16T13:27:34 *** madrazr has quit IRC 2011-03-16T13:31:46 yes i have stackchecker o n 2011-03-16T13:33:48 then you can call rtems_stack_checker_report_usage(void) and get a report on stack usage. 2011-03-16T13:34:02 yep.. i think it will flood the screen but lets try 2011-03-16T13:34:13 it crashes either way. :D 2011-03-16T13:34:29 Is there a large string buffer declared on the stack. 2011-03-16T13:34:50 it does some malloc free stuff 2011-03-16T13:35:11 but that should not do anything with stack? 2011-03-16T13:35:13 My alternate trick is to make the stack huge -- 4MB or more if possible -- and see if it runs 2011-03-16T13:35:24 i tested 16MB stack 2011-03-16T13:35:27 nope.. malloc/free are separate 2011-03-16T13:35:27 still blows 2011-03-16T13:35:33 WOW! does it use alloca? 2011-03-16T13:35:55 hmm? 2011-03-16T13:36:12 if i remember right, alloca is like malloc but from stack 2011-03-16T13:36:39 you can always bundle up the source before you leave in a kit and email it to me. I can try it on sparc or powerpc and see if it fails in an easier to spot way 2011-03-16T13:37:35 nope 2011-03-16T13:37:51 oh ok 2011-03-16T13:37:52 ok.. let's try this brutal method.. 2011-03-16T13:37:59 qemu has a trace/debug log 2011-03-16T13:39:33 add the argument -d in_asm,cpu and let it run. It will do an instruction and cpu state trace and put it in /tmp/qemu.log. From there you should be able to search back from the end to see what went wrong. Unfortunately this is a way to debug I am familiar with 2011-03-16T13:41:03 ok i'll try it 2011-03-16T13:50:37 hmm 2011-03-16T13:50:41 empty file 2011-03-16T13:57:56 *** mmu_man has quit IRC 2011-03-16T13:59:59 that's not right. :( 2011-03-16T14:01:01 /usr/bin/qemu -m 128 -cpu pentium -boot a -fda /home/joel/qemu/pc386_fda -hdc fat:/home/joel/qemu/hd --serial msmouse --no-reboot -d in_asm,cpu 2011-03-16T14:01:10 that worked for me 2011-03-16T14:10:02 does it do log real time? 2011-03-16T14:10:13 or dump log at end of program 2011-03-16T14:11:13 i now have: nbench_rtems_test.tar.gz 2011-03-16T14:11:36 it has simple main.c and files from nbench 2011-03-16T14:12:14 it dumps as it goes 2011-03-16T14:12:19 no makefile.. you have to add emfloat.c hardware.c misc.c nbench0.c nbench1.c sysspec.c from nbench directory to project 2011-03-16T14:12:42 ok.. I think I can handle that.. 2011-03-16T14:12:45 compiles with i386-rtems4.11 2011-03-16T14:13:10 where should i send it? 2011-03-16T14:13:43 I wonder if it uses FPU and it is disabled... hmmmmm 2011-03-16T14:14:08 my oar account joel DOT sherrill AT oarcorp DOT com 2011-03-16T14:14:11 those tests run fine but string tests crash 2011-03-16T14:14:15 :D 2011-03-16T14:14:25 ok 2011-03-16T14:17:54 email it and I will take a quick look 2011-03-16T14:20:24 there it is 2011-03-16T14:20:38 maybe :) 2011-03-16T14:21:33 got it.. give me 5 minutes before you run off 2011-03-16T14:21:37 it runs fine on my laptop and ubuntu :) 2011-03-16T14:21:54 resources are bigger but still :) 2011-03-16T14:24:48 maybe my i386 toolchain or bsp is bad 2011-03-16T14:25:55 I doubt it. :-D 2011-03-16T14:26:43 if you have different tickpersecond than 1000 maybe have to change constant in sysspec.c TicksToSecs and TicksToFracSecs 2011-03-16T14:26:57 but main.c configures 1000 2011-03-16T14:27:23 I have a shell prompt.. now what 2011-03-16T14:27:27 nbench 2011-03-16T14:27:54 [/] # nbench 2011-03-16T14:27:55 [/] # 2011-03-16T14:27:55 [/] # nbench 2011-03-16T14:27:55 [/] # nbench 2011-03-16T14:27:55 [/] # nbench 2011-03-16T14:27:55 [/] # 2011-03-16T14:28:01 :D 2011-03-16T14:28:05 It isn't printing anything on sis 2011-03-16T14:28:47 strange.. can you copy 2011-03-16T14:28:51 char **args[1] = {"./nbench"}; nbench_main(1, args); 2011-03-16T14:29:02 just before rtems_shell_main_loop 2011-03-16T14:29:46 Unexpected trap ( 4) at address 0x02002D60 2011-03-16T14:29:46 fp disabled 2011-03-16T14:29:50 hold on... 2011-03-16T14:32:22 running now..sis has a limit on how long a run it can simulate .. may have to move to tsim 2011-03-16T14:34:14 ok 2011-03-16T14:35:06 if you run straight from Init you may have to make stack bigger for it 2011-03-16T14:36:15 that makes sense.. trying to build other BSPs now on another machine to get other stabs at it 2011-03-16T14:36:29 :l.. if i run from straight Init 2011-03-16T14:36:33 it doesn't print anything 2011-03-16T14:36:42 hmm.. 2011-03-16T14:37:09 it is stuck in some big loop doing math 2011-03-16T14:38:29 oh 2011-03-16T14:38:48 running pc386 on qemu on another machine .. how long does it take to print more than header 2011-03-16T14:39:17 1s? 2011-03-16T14:39:50 it works well only if i run it from another task 2011-03-16T14:39:57 don't know why 2011-03-16T14:40:15 did not test from Init until now:l 2011-03-16T14:41:02 I added floating point attribute to init task here 2011-03-16T14:41:22 sis is in the middle of the string test 2011-03-16T14:41:30 NUMERIC SORT : 3.9736 : 0.10 : 0.03 2011-03-16T14:41:41 nbench1.c: DoStringSort AllocateMemory done 2011-03-16T14:41:41 0.044722 : 0.02 : 0.00 2011-03-16T14:41:46 Remember this is a 14Mhz V7 2011-03-16T14:42:08 :-o 2011-03-16T14:42:21 should i add floating point too? 2011-03-16T14:42:38 #define CONFIGURE_INIT_TASK_ATTRIBUTES (RTEMS_FLOATING_POINT) 2011-03-16T14:42:38 how? 2011-03-16T14:42:45 ok 2011-03-16T14:44:15 did you change #define CONFIGURE_EXECUTIVE_RAM_SIZE (32*1024*1024) 2011-03-16T14:46:32 can't get it run from Init :l 2011-03-16T14:46:42 it just doesn't do anything else than print header 2011-03-16T14:46:48 I removed it 2011-03-16T14:47:14 If you have unified memory, it uses all that is available so that is redundant 2011-03-16T14:47:20 ok 2011-03-16T14:48:45 BITFIELD : 9.071e+05 : 0.16 : 0.03 2011-03-16T14:49:00 seems to work for you? 2011-03-16T14:49:41 on sis.. qemu hasn't printed anything past the header yet 2011-03-16T14:50:07 o_O 2011-03-16T14:50:23 that is same behaviour than here.. 2011-03-16T14:50:29 12 minutes of CPU time so far 2011-03-16T14:50:29 if you call it from Init 2011-03-16T14:50:48 but strange that it did not work from another task? 2011-03-16T14:51:12 you can skip the while(1) stuff in task_nbench 2011-03-16T14:51:27 RTEMS_DEFAULT_ATTRIBUTES on other task.. needs to be RTEMS_FLOATING_POINT I think 2011-03-16T14:51:28 to make it run without flag from shell cmd 2011-03-16T14:51:41 oh.. 2011-03-16T14:52:18 shell task likely has FPU on 2011-03-16T14:52:33 *** core-ix has quit IRC 2011-03-16T14:53:10 yeh.. but i only signal another task with shell cmd 2011-03-16T14:53:35 via bad global variable:) 2011-03-16T14:53:48 :) 2011-03-16T14:53:50 still blows stack here 2011-03-16T14:54:00 with RTEMS_FLOATING_POINT 2011-03-16T14:54:30 i suspect my toolchain.. i have left out floating point support? 2011-03-16T14:54:32 sis just printed FP EMULATION : 0.24756 : 0.12 : 0.03 2011-03-16T14:54:32 Nothing like a 14Mhz SPARC Station 1+ .. ahhh those were the days 2011-03-16T14:54:35 or something 2011-03-16T14:54:59 ;) 2011-03-16T14:55:18 no.. you have to specifically enable it on a per task basis. 2011-03-16T14:55:35 Took 4m15 to run on the development server natively.. still crunching in qemu 2011-03-16T14:55:56 hmm 2011-03-16T14:56:05 18 minutes and nothing past header printed 2011-03-16T14:56:09 or is it stuck? 2011-03-16T14:56:17 qemu on what hardware? 2011-03-16T14:56:29 i have intel i5 laptop 2011-03-16T14:56:29 qemu x86 running on a Linux box 2011-03-16T14:56:42 Dual GenuineIntel Intel(R) Core(TM)2 Duo CPU E8500 @ 3.16GHz 3158MHz 2011-03-16T14:56:45 ok 2011-03-16T14:56:55 probably faster than this :) 2011-03-16T14:56:55 fairly fast box :) 2011-03-16T14:58:00 ok, i did not enable floating point support for tasks 2011-03-16T14:58:13 but it did not help to enable it :l 2011-03-16T14:58:55 is there good benchmark app for rtems? 2011-03-16T14:59:14 similar than nbench 2011-03-16T15:00:58 we probably have different compiler settings 2011-03-16T15:02:09 I would ask on the list to see what people have used. I focus on the tests with RTEMS and don't worry too much about the 3rd party ones. 2011-03-16T15:02:25 I am running this again on qemu with debugger attached so I can see what's happening 2011-03-16T15:03:03 sis is slow but just printed the fourier number 2011-03-16T15:03:52 better arch :) 2011-03-16T15:05:19 i'll run it tomorrow on arm-rtems4.10 2011-03-16T15:06:25 lpc2468 2011-03-16T15:07:54 Someone here is hooking up boards we hadn't gotten to into the remote access cage. lpc2478 is in the queue. 2011-03-16T15:08:12 Our question is how to use usb serial port be able to power it off 2011-03-16T15:08:29 Even with external power attached, if you remove that power, it stays on. 2011-03-16T15:09:33 qemu is up to nbench1.c: DoStringSort AllocateMemory done 2011-03-16T15:09:54 thats my debug texst.. after that it crashes 2011-03-16T15:10:48 your EA board powers it self from usb? 2011-03-16T15:10:59 mine does that too 2011-03-16T15:11:13 yep 2011-03-16T15:11:59 of course you can modify usb-cable to be able to cut 5V off when you like 2011-03-16T15:12:02 :D 2011-03-16T15:12:21 or do not use usb-cable at all? 2011-03-16T15:12:34 it should have other COM ports? 2011-03-16T15:15:02 do you have electronics personel that can do mods like that? 2011-03-16T15:15:35 Hmmm... we do have someone who can a USB cable to do that 2011-03-16T15:16:28 qemu is at field 2011-03-16T15:16:37 your qemu is better than mine :l 2011-03-16T15:17:16 this confirm i have done something wrong 2011-03-16T15:18:04 do you want me to email you the main.c I have? 2011-03-16T15:18:09 sure 2011-03-16T15:19:47 shot back at you 2011-03-16T15:20:05 BITFIELD : 6.2821e+07 : 10.78 : 2.25 2011-03-16T15:20:05 FP EMULATION : 22.318 : 10.71 : 2.47 2011-03-16T15:20:07 from qemu 2011-03-16T15:21:15 my i386 plan: modern pc + pci card with some IO ADC/DAC and counters + rtems & pc386 & app = machine control / robotics with great computing power and realtime features :) 2011-03-16T15:21:49 that's getting to be a common profile. :-D I have a Winsystems board in the lab with their multi-io board on it 2011-03-16T15:22:30 http://rtemsramblings.blogspot.com/2011_02_01_archive.html is a blog about shell commands I wrote for it. 2011-03-16T15:22:49 I want to define an interface for those commands so I can use them again on another multiIO board 2011-03-16T15:24:34 hmm.. i would like to duplicate that hardware 2011-03-16T15:24:42 we have some pc104 2011-03-16T15:24:50 expensive board? 2011-03-16T15:25:56 not so long ago i read your entire blog.. but i must have missed that made that for i386 2011-03-16T15:26:06 :o 2011-03-16T15:29:26 *** core-ix has joined #rtems 2011-03-16T15:37:29 O_O, qemu prints results.. it just takes long time! 2011-03-16T15:37:37 with your main.c 2011-03-16T15:39:03 yep... just finished here while I was out.. running it on Winsystems 1Ghz embedded PC now as comparison 2011-03-16T15:39:25 I will send you those results via email 2011-03-16T15:39:30 ok 2011-03-16T15:39:50 thanks 2011-03-16T15:39:55 will be interesting to see if compiler settings help a lot on a modern CPU 2011-03-16T15:40:10 no problem. don't want rtems to get a bad reputation 2011-03-16T15:40:15 now i must check why this does run faster on thread and crash.. 2011-03-16T15:40:52 change attribute to floating point as a starting point. 2011-03-16T15:41:06 yes 2011-03-16T15:44:34 *** mmu_man has joined #rtems 2011-03-16T15:49:27 strange.. if i change the task priority from 51 to 49 it does not crash but act similarly than when it is run from Init.. 2011-03-16T15:49:48 I wonder what all the tasks are in the application. 2011-03-16T15:50:31 should be only Init and task for calculation.. 2011-03-16T15:51:58 Does Init fall into the shell or return 0? 2011-03-16T15:52:21 while(1){sleep(1)} now 2011-03-16T15:52:45 hard to say .. 2011-03-16T15:53:30 is the pcmmio for rtems available somewhere? 2011-03-16T15:55:53 not yet.. any suggestions on how to make it available? It is in a private repository at the moment.. maybe I should move it to a "multi_io" cvs module and reorg it a bit so I have space to insert an adapter layer to arbitrary hardware 2011-03-16T15:59:35 maybe if some other task interrupts (preemt)?calc task.. it crashes 2011-03-16T16:00:15 no idea.. shoot me your main and we can try again :-D 2011-03-16T16:01:13 *** lcpfnyvc has quit IRC 2011-03-16T16:01:39 it works form another task as long as i don't let init task interrupt it.. 2011-03-16T16:01:46 seems like it.. 2011-03-16T16:01:54 because init task is priority 50? 2011-03-16T16:02:09 and 49 for calc task works and 51 doesn't 2011-03-16T16:02:36 Even when calc task is floating point? 2011-03-16T16:02:54 yes 2011-03-16T16:03:57 calc task still has 16MB stack? 2011-03-16T16:04:01 i put 51 but disabled preemt.. init can't interrupt.. no crash.. 2011-03-16T16:04:18 1MB 2011-03-16T16:05:16 16MB crashes too 2011-03-16T16:05:23 if i interrupt it 2011-03-16T16:06:40 Weird.. there aren't really any shared resources between them unless their stacks collide. Init is 80K which is huge 2011-03-16T16:09:49 i sent my crasher 2011-03-16T16:11:06 maybe i should send my binary? 2011-03-16T16:12:08 the crashing main 2011-03-16T16:12:19 that i sent 2011-03-16T16:17:22 works well? 2011-03-16T16:17:53 getting there.. 2011-03-16T16:17:57 rtied sis first 2011-03-16T16:18:04 ok 2011-03-16T16:18:08 running on qemu 2011-03-16T16:18:23 my qemu params:qemu -m 128 -boot a -fda ./rtems-boot.img -hda ./hda.img -hdb fat:. l; 2011-03-16T16:19:41 hmm different cpu model than me.. running now.. but still made it to herald and when I break in gdb, it is runnign test code 2011-03-16T16:20:01 what if you hit keyboard..? 2011-03-16T16:20:10 can you check stackuse while running test 2011-03-16T16:21:02 Stack usage by thread 2011-03-16T16:21:02 ID NAME LOW HIGH CURRENT AVAILABLE USED 2011-03-16T16:21:02 0x09010001 IDLE 0x0017bde4 - 0x0017cdf3 0x0017cde8 4096 12 2011-03-16T16:21:02 0x0a010001 UI1 0x0017e1a8 - 0x001921b7 0x00191abc 81920 4588 2011-03-16T16:21:02 0x0a010002 NBM2 0x00192ad4 - 0x00292ae3 0x0029266c 1048576 2568 2011-03-16T16:21:03 0xffffffff INTR 0x00177a78 - 0x00178a5f 0x00000000 4056 528 2011-03-16T16:21:05 [/] # 2011-03-16T16:21:20 [/] # cpuuse 2011-03-16T16:21:20 ------------------------------------------------------------------------------- 2011-03-16T16:21:20 CPU USAGE BY THREAD 2011-03-16T16:21:21 ------------+----------------------------------------+---------------+--------- 2011-03-16T16:21:21 ID | NAME | SECONDS | PERCENT 2011-03-16T16:21:21 ------------+----------------------------------------+---------------+--------- 2011-03-16T16:21:25 0x09010001 | IDLE | 0.000000 | 0.000 2011-03-16T16:21:27 0x0a010001 | UI1 | 0.042320 | 0.034 2011-03-16T16:21:29 0x0a010002 | NBM2 | 120.967122 | 99.964 2011-03-16T16:21:31 ------------+----------------------------------------+---------------+--------- 2011-03-16T16:21:33 TIME SINCE LAST CPU USAGE RESET IN SECONDS: 121.010667 2011-03-16T16:21:35 ------------------------------------------------------------------------------- 2011-03-16T16:21:38 are you running on 4.10 or head? 2011-03-16T16:21:43 4.11 2011-03-16T16:21:49 vga console or com1? 2011-03-16T16:22:09 hmm. don't know? 2011-03-16T16:22:16 graphics or stdio 2011-03-16T16:22:34 in qemu? 2011-03-16T16:22:42 or? 2011-03-16T16:23:14 is the console in the graphics box qemu pops up 2011-03-16T16:23:34 I switched to vga console and it crashed.. 2011-03-16T16:23:39 :D 2011-03-16T16:23:45 it pops up 2011-03-16T16:23:51 window with QEMU 2011-03-16T16:23:54 title 2011-03-16T16:23:55 I have no idea what makes that different 2011-03-16T16:24:08 You can change the grub menu and make it use com1 2011-03-16T16:24:27 parma? 2011-03-16T16:24:29 param? 2011-03-16T16:25:13 timeout=0 2011-03-16T16:25:44 default=0 2011-03-16T16:25:44 serial --unit=0 --speed=9600 2011-03-16T16:25:44 terminal serial 2011-03-16T16:25:44 title= Automatically Run Test 2011-03-16T16:25:44 #kernel= (hd0,0)/test.exe --console=com1 2011-03-16T16:25:45 kernel= (hd0,0)/test.exe --console=vgacons 2011-03-16T16:29:57 mm.. no text.-. 2011-03-16T16:30:14 this is strange.. if I am reading this the stack protection area is unmodified 2011-03-16T16:30:22 add --serial stdio to your qemu invocation 2011-03-16T16:36:40 ok.. no crash on stdio 2011-03-16T16:36:46 strange 2011-03-16T16:37:05 yes.. I am in gdb when it is reported now.. rtying to see what's up 2011-03-16T16:37:28 i'll test quickly on the laptop directly.. brb.. 2011-03-16T16:40:25 crash 2011-03-16T16:40:44 but i can't test the com1 with laptop, probably it doesn't crash 2011-03-16T16:41:11 theres is no hardware serialport .. :) 2011-03-16T16:41:53 ok.. we will trust it. It is crashing coming out or an ISR dispatch. I wonder if it isn't the thread's stack but the interrupt stack and it is being misinterpreted or misreoprted for a blown interrupt stack.. still looking .. 2011-03-16T16:42:33 thread stack is not corrupted as best I can tell and is in range 2011-03-16T16:47:03 ok 2011-03-16T16:48:00 and it should be because it works well on linux and sparc? 2011-03-16T16:48:12 I set hardware watchpoints and the "marker" at the end of the stack was never written.. so it is getting confused somehow... 2011-03-16T16:48:44 it works well in the other configuration. This may be i386 specific.. since it is traces back to an interrupt preemption 2011-03-16T16:49:04 interesting 2011-03-16T16:49:29 yes.. hints so far are confusing.. 2011-03-16T16:50:54 which arch has most active users today? 2011-03-16T16:51:06 is i386 rare? 2011-03-16T16:52:33 arm, i386, powerpc, sparc, and coldfire if I had to guess. SPARC is high number of users due to space cpus. ARM is very popular. PC gets a lot of use. I have helped on a robotic project using embedded 386 .. very common setup 2011-03-16T16:52:48 okay 2011-03-16T16:53:36 it is also possible Jennifer and I broke something with the SMP changes since 4.10. I probably should try this with 4.10 and see if it breaks there 2011-03-16T16:53:45 just thinking how somebody could mis this crashing thing.. if it isn't bug in the app 2011-03-16T16:54:10 that's what bothers me.. I take any hint of a bug pretty seriously. 2011-03-16T16:55:25 :l 2011-03-16T16:55:59 *** dr__house has quit IRC 2011-03-16T17:04:06 I added some more prints in the stack check code.. 2011-03-16T17:04:09 printk( "sp_ok=%d pattern_ok=%d\n", sp_ok, pattern_ok ); 2011-03-16T17:04:09 printk( "pattern_p=%p size=%d\n", pattern, PATTERN_SIZE_BYTES ); 2011-03-16T17:04:28 sp is OK and pattern_ok =0 so it thinks pattern area is not ok 2011-03-16T17:04:41 pattern is at 0x194a84 for 16 bytes 2011-03-16T17:04:47 gdb) x/x 0x194a84 2011-03-16T17:04:47 0x194a84: 0xfeedf00d 2011-03-16T17:04:47 (gdb) 2011-03-16T17:04:47 0x194a88: 0x0bad0d06 2011-03-16T17:04:47 (gdb) 2011-03-16T17:04:48 0x194a8c: 0xdeadf00d 2011-03-16T17:04:50 (gdb) 2011-03-16T17:04:52 0x194a90: 0x600d0d06 2011-03-16T17:04:54 (gdb) 2011-03-16T17:04:58 0x194a94: 0xa5a5a5a5 2011-03-16T17:05:00 (gdb) p/x Stack_check_Pattern 2011-03-16T17:05:02 $1 = {pattern = {0xfeedf00d, 0xbad0d06, 0xdeadf00d, 0x600d0d06}} 2011-03-16T17:05:04 Doesn't the stack memory match the master pattern 2011-03-16T17:06:57 *** antgreen has quit IRC 2011-03-16T17:07:08 same in plain simple english?:> 2011-03-16T17:08:05 false alarm for unknown reason? 2011-03-16T17:08:14 the memcmp is actually a repz cmpi.. I wonder what could go wrong there 2011-03-16T17:08:24 That's where I am leaning 2011-03-16T17:09:25 at least it is easily reproducible 2011-03-16T17:10:25 yes... that's a good thing I suppose :( 2011-03-16T17:10:49 can we make the test simpler 2011-03-16T17:11:33 wait a second.. maybe a light bulb.. this happens during a context switch during the string operations? 2011-03-16T17:11:50 seems like so 2011-03-16T17:12:22 x86 has those repeating string instructions.. this is the memcmp in the stack checker (generated by gcc) 2011-03-16T17:12:24 0x0010f8e8 <+116>: repz cmpsb %es:(%edi),%ds:(%esi) 2011-03-16T17:13:00 The string test is likely full of those... I wonder if an interrupt and context switch confuses it 2011-03-16T17:14:44 hmm.. 2011-03-16T17:16:59 slen=strncmp((char *)(strarray+*(optrarray+a)), (char *)(strarray+*(optrarray+b)),slen); 2011-03-16T17:17:22 clip from code 2011-03-16T17:20:39 can you tell in the objdump what code gcc generates? 2011-03-16T17:22:06 sounds difficult..:D, i was making test app with str operations.. 2011-03-16T17:23:23 ok.. I am moving to 4.10 to see if this is new or existing 2011-03-16T17:28:29 I don't know if this is good or bad.. 4.10 has the same behaviour 2011-03-16T17:33:26 :l 2011-03-16T17:33:37 how about 4.6 2011-03-16T17:33:46 * kiwichris wave to Joel 2011-03-16T17:33:57 no.. not that LOL 2011-03-16T17:34:05 * DrJoel waves to kiwichris .. weird problem 2011-03-16T17:34:28 Been knocked out by a flue for the last couple of days 2011-03-16T17:34:48 we have what appears to be a bogus stack check failure report on pc386 on Byte nbench which works when using com1 as console.. but not vga 2011-03-16T17:35:17 VGA is non-blocking 2011-03-16T17:36:40 Hmmm... hint.. 2011-03-16T17:47:06 crashes without strncmp 2011-03-16T17:47:38 i did only stringtest and disables strncmp from it 2011-03-16T17:47:43 disabled.. 2011-03-16T17:55:04 hmmm... 2011-03-16T17:56:09 I tried not freeing any memory for real and it also crashed the same way 2011-03-16T17:57:36 I commented out the StrHeapSort loop and it doesn't break now 2011-03-16T18:02:10 memmove is at the bottom of swapping the elements around and does have an x86 rep movsb in the middle 2011-03-16T18:02:45 This all seems low level stuff 2011-03-16T18:03:12 i replaced malloc with alloca 2011-03-16T18:03:20 different crash :) 2011-03-16T18:03:23 kiwichris: yes. I am suspicious that there is some conflict of some odd sort in the direction or zero flag when using the string instructions 2011-03-16T18:03:32 rokka: that I would have expected.. alloca is from the stack 2011-03-16T18:03:51 Has something changed in the 386 compiler ? 2011-03-16T18:04:03 the pattern area for the stack is OK. The master pattern is OK. the sp is OK 2011-03-16T18:04:18 I have trie 4.10 and 4.11. Could easily go back to 4.9 I think 2011-03-16T18:04:24 FAULTY THREAD WILL BE SUSPENDED 2011-03-16T18:04:49 it is also possible the code is doing something it shouldn't.. it is old code and this string/sorting stuff is nasty. But I don't see what it is diong that would break like this 2011-03-16T18:06:01 bad pointers :) 2011-03-16T18:06:36 we pop eflags on the last thing out of an isr... the failure appears to be when we do a _Thread_Dispatch from the end of an ISR... hmmm... I wonder if the direction flag is wrong somehow 2011-03-16T18:06:45 The direction flag code changed in 4.3 2011-03-16T18:07:23 The 4.3 release notes state "GCC no longer places the cld instruction before string operations. Both i386 and x86-64 ABI documents mandate the direction flag to be clear at the entry of a function. It is now invalid to set the flag in asm statement without reseting it afterward." 2011-03-16T18:09:28 memmove in newlib-cvs doesn't appear to honor that 2011-03-16T18:09:45 There is an option called '-mcld' 2011-03-16T18:11:14 last change to memmove.S in newlib was when the copyright was changed from Cygnus to RedHat LOL 2011-03-16T18:11:51 Ha 2011-03-16T18:12:11 How to test this? Recompile RTEMS with -mcld? 2011-03-16T18:12:30 I suppose so. I do not know. Does newlib need to be rebuilt ? 2011-03-16T18:12:41 eventually but we might get lucky with this test case 2011-03-16T18:13:59 it is too late for me to think more about this. I think I will try to write this up in the morning and email newlib's list. Sounds like something subtle and the right people are there. 2011-03-16T18:14:10 I need to get some coffee creamer and get home. 2011-03-16T18:14:11 night 2011-03-16T18:14:15 *** DrJoel has quit IRC 2011-03-16T18:15:19 i don't understand yet but maybe tomorrow.. :) 2011-03-16T18:18:58 night 2011-03-16T18:19:04 night 2011-03-16T18:40:37 *** mmu_man has quit IRC 2011-03-16T18:58:01 *** kiwichris has quit IRC 2011-03-16T19:10:30 *** soh_cah_toa has joined #rtems 2011-03-16T19:14:27 hey guys, this is my first gsoc and i need a little advice 2011-03-16T19:15:19 once the list of organizations is posted should i apply for one project or is it kinda like college where it's a good idea to apply for several? 2011-03-16T19:24:41 *** dr__house has joined #rtems 2011-03-16T19:24:41 *** dr__house has joined #rtems 2011-03-16T19:33:46 *** soh_cah_toa has quit IRC 2011-03-16T19:54:28 *** rtemsLogger has joined #rtems 2011-03-16T19:59:38 *** antgreen has joined #rtems 2011-03-16T20:49:47 *** dr__house has quit IRC 2011-03-16T23:33:35 *** antgreen has quit IRC 2011-03-17T00:16:31 *** madrazr has joined #rtems 2011-03-17T00:51:10 *** xiangfu has joined #rtems 2011-03-17T01:11:25 *** lcpfnyvc has joined #rtems 2011-03-17T01:33:12 *** lcpfnyvc has quit IRC 2011-03-17T01:33:23 *** lcpfnyvc has joined #rtems 2011-03-17T01:47:29 *** projectgus has quit IRC 2011-03-17T02:21:29 *** madrazr has quit IRC 2011-03-17T03:10:31 *** mmu_man has joined #rtems 2011-03-17T03:10:44 *** sebhub has joined #rtems 2011-03-17T03:10:52 good morning 2011-03-17T03:53:37 *** core-ix has quit IRC 2011-03-17T04:02:07 *** xiangfu has quit IRC 2011-03-17T04:05:38 *** mmu_man has quit IRC 2011-03-17T04:13:33 *** xiangfu has joined #rtems 2011-03-17T04:24:11 *** sebhub has quit IRC 2011-03-17T04:24:59 *** sebhub has joined #rtems 2011-03-17T04:34:43 *** mmu_man has joined #rtems 2011-03-17T04:35:46 *** core-ix has joined #rtems 2011-03-17T04:56:39 *** xiangfu has quit IRC 2011-03-17T05:24:41 *** madrazr has joined #rtems 2011-03-17T05:59:39 *** madrazr has left #rtems 2011-03-17T06:05:19 *** mmu_man has quit IRC 2011-03-17T06:47:39 sebhub: do you think 200us is ok for lpc2468 with 16bit bus for task waiting msg to recv queued message :)? 2011-03-17T06:48:02 57.6MHz 2011-03-17T06:48:06 cpu 2011-03-17T06:48:50 msg sent from ISR and recv from highest priority task 2011-03-17T07:44:28 *** DrJoel has joined #rtems 2011-03-17T07:44:28 *** ChanServ sets mode: +o DrJoel 2011-03-17T08:01:02 *** mmu_man has joined #rtems 2011-03-17T08:48:16 *** antgreen has joined #rtems 2011-03-17T09:19:19 *** sebhub has quit IRC 2011-03-17T09:19:37 *** sebhub has joined #rtems 2011-03-17T09:20:22 rokka: at this frequency this are a maximum of 11520 instuctions 2011-03-17T09:20:30 i think this is ok 2011-03-17T09:21:09 code execution from external 16-bit ram is not at 57.6MHz 2011-03-17T09:21:31 you achieve this only from the interal flash for sequential code 2011-03-17T09:22:10 you should check how fast your code execution works (instructions per second) 2011-03-17T09:22:19 is event much faster than queued message? 2011-03-17T09:23:11 *** dr__house has joined #rtems 2011-03-17T09:23:11 *** dr__house has joined #rtems 2011-03-17T09:25:09 rokka: on erc32 event readying a task is about 1/3 of message queue readying 2011-03-17T09:26:15 oh! 2011-03-17T09:26:19 much change to that the n 2011-03-17T09:26:23 must.. 2011-03-17T09:27:59 events have the lowest overhead 2011-03-17T09:28:33 you need only one object get (for the task) and the processing is very simple 2011-03-17T09:29:44 rokka: did you see the PR? That appears to solve it for me. nbench only uses about 22K stack so lowered it to 64K 2011-03-17T09:30:14 oh! great! url? 2011-03-17T09:30:16 :) 2011-03-17T09:30:41 so did we find bug again?:D 2011-03-17T09:31:31 https://www.rtems.org/bugzilla/show_bug.cgi?id=1768 2011-03-17T09:31:41 Changes in GCC make for hard bugs.. 2011-03-17T09:31:50 Impacts 4.9 and newer for x86 2011-03-17T09:32:03 One stinking instruction added fixes it 2011-03-17T09:35:32 did the ABI change in the x86? 2011-03-17T09:38:47 gcc stopped clearing the direction flag before string operations in gcc 4.3. 2011-03-17T09:39:06 not really an abi change.. just have to be more careful about it now 2011-03-17T09:51:58 checking for RTEMS Version... 4.10.0 2011-03-17T09:52:12 in rtems configure does this mean 4.10 ? 2011-03-17T09:52:19 have i checkouted wrong version.. 2011-03-17T09:53:20 --target=XXX-rtems4.XX 2011-03-17T09:53:41 i meant to check out 4.11 from cvs 2011-03-17T09:55:44 by the way what does PR mean? 2011-03-17T09:59:13 Problem Report 2011-03-17T09:59:20 okay:) 2011-03-17T09:59:37 ok.. i had malloc replaced with alloca -> crash 2011-03-17T09:59:49 but when i reverted to malloc test runs pretty fast without crash 2011-03-17T09:59:58 did you see speed up? 2011-03-17T10:00:01 The alloca() function allocates size bytes of space in the stack frame of the caller. This temporary space is automatically 2011-03-17T10:00:01 freed when the function that called alloca() returns to its caller. 2011-03-17T10:00:11 oh 2011-03-17T10:00:17 so it is supposed to crash.. 2011-03-17T10:00:25 yes.. appears to be faster on console.. 2011-03-17T10:01:16 with my qemu parameters it takes about 10s to complete whole program 2011-03-17T10:02:09 I suspect with the direction flag wrong, it may count up to 0. :( 2011-03-17T10:05:23 i can finish porting the nbench now 2011-03-17T10:05:32 because it works :) 2011-03-17T10:05:32 :-D 2011-03-17T10:05:42 I added it to examples-v2/benchmarks 2011-03-17T10:06:05 oh.. but its not ready.. last test needs input file from filesystem 2011-03-17T10:06:26 :) 2011-03-17T10:06:37 Look in filesystems in examples and it will show you how to make a filesystem image and untar it. 2011-03-17T10:06:43 *** sebhub has quit IRC 2011-03-17T10:06:44 Just a matter of getting a patch later 2011-03-17T10:06:48 yep 2011-03-17T10:09:49 does rtems have api to report cpu, l2 cache, libc version? 2011-03-17T10:15:52 *** zwj has joined #rtems 2011-03-17T10:16:10 hi,DrJoel 2011-03-17T10:19:19 zwj: hey .. 2011-03-17T10:19:26 * DrJoel is walking in and out of office so might be slow 2011-03-17T10:36:59 *** sebhub has joined #rtems 2011-03-17T10:37:05 im now running rtems from external 16bit ram and get 125us for task recv event from isr.. does it speed up when i start running rtems from internal flash? 2011-03-17T10:38:07 it should be faster 2011-03-17T10:38:28 where are your task stacks? 2011-03-17T10:38:43 you can use the internal ram for them 2011-03-17T10:40:12 see 2011-03-17T11:05:09 *** zwj has quit IRC 2011-03-17T11:48:29 *** mmu_man has quit IRC 2011-03-17T11:57:04 sebhub: how can i use it:D 2011-03-17T11:58:58 look in this file , watch for CONFIG* and use it for confdefs.h 2011-03-17T11:59:09 okey 2011-03-17T12:08:51 ok.. rtems_event_receive seems to be 10us faster now 2011-03-17T12:09:16 125us vs 117us 2011-03-17T12:09:24 with code execution from internal flash? 2011-03-17T12:09:36 no 2011-03-17T12:09:55 with bsp/stackalloc.h 2011-03-17T12:10:36 6.4% 2011-03-17T12:10:53 the external 16-bit ram is slow 2011-03-17T12:11:18 what is your goal? 2011-03-17T12:11:59 in case you use nested interrupt the task activation time may be considerable more 2011-03-17T12:12:18 nested interrupt? 2011-03-17T12:12:39 the lpc24xx supports 16 interrupt priority levels 2011-03-17T12:13:40 i have ISR with EXTINT0 and it uses rtems_event_send to signal task with highest priority 2011-03-17T12:13:41 you have no maximal task activation time only a minimal 2011-03-17T12:14:40 seems to take about 6500 clock cycles 2011-03-17T12:15:12 you have a maximal interrupt latency time for the highest priority interrupt if it exists 2011-03-17T12:15:55 the default is equal priority for all interrupts, so no highest priority interrupt exists 2011-03-17T12:16:19 how do i set interrupt priority for EXTINT00 2011-03-17T12:16:58 look in 2011-03-17T12:17:41 so if i don't use it, it should be at default? 2011-03-17T12:18:17 15 = lowest 2011-03-17T12:18:38 which is fastest:D 2011-03-17T12:18:48 if all irqs are same priority = fastest? 2011-03-17T12:19:20 what means fastest? 2011-03-17T12:20:06 since rtems is a real time system, there is an upper bound for the interrupt latency of the highest priority interrupt 2011-03-17T12:20:14 lowest amount of time from io triggering to task handler DIGITAL INPUT -> ISR -> TASK 2011-03-17T12:21:43 the lowest amount is useless, you have to consider the highest amount, and in general for task activation this is unlimited 2011-03-17T12:22:35 ok.. i want lowest number to the max latency ;) 2011-03-17T12:22:58 ok, but the max latency exists only for the highest priority interrupt 2011-03-17T12:23:08 not for the task activation 2011-03-17T12:23:46 actually in my application its not fatal it sometimes it is bigger value, if 90% is very little latency 2011-03-17T12:24:13 if you need something really fast on the lpc24xx use a fast interrupt and a software interrupt 2011-03-17T12:24:39 ok, then the default values will do 2011-03-17T12:24:47 try to use the internal flash for code 2011-03-17T12:25:00 i would if i could :D 2011-03-17T12:25:08 i would like to have uboot 2011-03-17T12:25:16 for lpc2468 2011-03-17T12:25:31 which starts rtems from internal flash 2011-03-17T12:26:07 wipe out u-boot, you only need your sdram config 2011-03-17T12:26:34 wipe out means? 2011-03-17T12:26:46 replace it with your application 2011-03-17T12:27:10 but i like to have bootloader 2011-03-17T12:27:17 or keep it and life with 120us 2011-03-17T12:27:19 for updating over ethernet 2011-03-17T12:27:42 so later i will compile uboot and then i see how fast it is.. :) 2011-03-17T12:27:59 it should not be worse than this i hope..:) 2011-03-17T12:28:39 120 us is ok for our app .. 450us is deadline 2011-03-17T12:31:06 no i get it... my rtems app will not fit the internal flash if i use uboot:-) it is supposed to be on external flash 2011-03-17T12:31:22 uboot loads it from external flash to ram and start it 2011-03-17T12:31:36 how much internal flash do you have? 2011-03-17T12:32:49 lpc2468 has 512kB? 2011-03-17T12:33:34 how does it differ if from running directly from ISR? 2011-03-17T12:33:47 if i use software interrupt? 2011-03-17T12:34:08 with fast interrupt 2011-03-17T12:34:17 you cannot use the operating system in a FIQ 2011-03-17T12:35:07 you have to service the fast things in the FIQ and then trigger a software interrupt and handle the high level stuff 2011-03-17T12:35:15 btw. what is it exactly what happens when i send event to task from isr? does it really take 6500 cycles to process? 2011-03-17T12:36:04 you don't execute that much instructions if you run from external 16-bit ram 2011-03-17T12:36:12 half of that? 2011-03-17T12:36:22 you should measure it 2011-03-17T12:36:52 by running from internal flash?:) 2011-03-17T12:37:51 no in this case you have to run from external ram 2011-03-17T12:38:13 ok how do i measure 2011-03-17T12:39:52 i have not template 2011-03-17T12:40:15 but what do you mean by measuring? 2011-03-17T12:41:21 measure the instructions per second for some benchmark program 2011-03-17T12:42:11 like nbench?-) 2011-03-17T12:43:55 interesting, does it run on rtems? 2011-03-17T12:44:21 oh, i have to leave, bye 2011-03-17T12:44:25 since yesterday, yes 2011-03-17T12:44:32 bye! 2011-03-17T12:44:49 and thanks for help 2011-03-17T12:45:46 cool, can you send me a link? sebastian.huber@embedded-brains.de 2011-03-17T12:46:25 *** sebhub has quit IRC 2011-03-17T12:46:29 i'll test it with lpc2468 to see if it works fine with it too 2011-03-17T12:46:38 oh there he goes already.. 2011-03-17T12:48:13 *** mmu_man has joined #rtems 2011-03-17T13:13:43 *** dr__house has quit IRC 2011-03-17T13:13:44 *** dr__house` has joined #rtems 2011-03-17T13:13:44 *** dr__house` has joined #rtems 2011-03-17T14:33:20 *** dr__house has joined #rtems 2011-03-17T14:33:20 *** dr__house has joined #rtems 2011-03-17T15:08:14 *** dr__house has quit IRC 2011-03-17T15:16:49 *** core-ix has quit IRC 2011-03-17T15:37:57 *** DrJoel has quit IRC 2011-03-17T16:50:56 *** projectgus has joined #rtems 2011-03-17T16:54:01 *** core-ix has joined #rtems 2011-03-17T18:26:26 *** kiwichris has joined #rtems 2011-03-17T18:37:05 *** mmu_man has quit IRC 2011-03-17T18:37:50 *** dr__house has joined #rtems 2011-03-17T20:40:39 *** dr__house has quit IRC 2011-03-18T00:00:03 *** mmu_man has joined #rtems 2011-03-18T00:09:17 *** kiwichris has quit IRC 2011-03-18T00:58:33 *** projectgus has left #rtems 2011-03-18T01:00:21 *** mmu_man has quit IRC 2011-03-18T01:50:43 *** xiangfu has joined #rtems 2011-03-18T01:55:05 *** xiangfu has quit IRC 2011-03-18T02:08:08 *** xiangfu has joined #rtems 2011-03-18T02:14:35 *** xiangfu has quit IRC 2011-03-18T02:14:59 *** xiangfu has joined #rtems 2011-03-18T02:30:16 *** xiangfu has quit IRC 2011-03-18T02:34:25 *** xiangfu has joined #rtems 2011-03-18T02:49:37 morning 2011-03-18T03:02:53 *** xiangfu has quit IRC 2011-03-18T03:03:17 *** xiangfu has joined #rtems 2011-03-18T03:09:14 *** sebhub has joined #rtems 2011-03-18T03:09:19 good morning 2011-03-18T03:23:35 *** xiangfu has quit IRC 2011-03-18T03:24:40 *** xiangfu has joined #rtems 2011-03-18T03:30:09 *** xiangfu has quit IRC 2011-03-18T03:30:34 *** xiangfu has joined #rtems 2011-03-18T03:35:07 *** xiangfu has quit IRC 2011-03-18T03:37:13 *** xiangfu has joined #rtems 2011-03-18T03:42:41 *** xiangfu has quit IRC 2011-03-18T03:44:56 *** xiangfu has joined #rtems 2011-03-18T03:46:59 *** core-ix has quit IRC 2011-03-18T03:50:35 *** xiangfu has quit IRC 2011-03-18T03:51:00 *** xiangfu has joined #rtems 2011-03-18T04:00:43 *** xiangfu has quit IRC 2011-03-18T04:01:19 *** xiangfu has joined #rtems 2011-03-18T04:06:57 *** xiangfu has quit IRC 2011-03-18T04:15:15 *** xiangfu has joined #rtems 2011-03-18T04:20:18 *** xiangfu has quit IRC 2011-03-18T04:22:32 *** xiangfu has joined #rtems 2011-03-18T04:28:31 *** core-ix has joined #rtems 2011-03-18T04:50:09 *** xiangfu has quit IRC 2011-03-18T04:50:28 *** xiangfu has joined #rtems 2011-03-18T05:00:48 *** xiangfu has quit IRC 2011-03-18T06:19:56 *** zwj has joined #rtems 2011-03-18T07:06:40 sebhub: nbench seems to work ok with i386-rtems4.11 but it didn't work with arm-rtems4.10 :l i must debug it later 2011-03-18T07:10:31 something screw the calculation tests 2011-03-18T07:13:57 *** mmu_man has joined #rtems 2011-03-18T07:19:23 *** methril has quit IRC 2011-03-18T07:31:16 *** methril has joined #rtems 2011-03-18T07:42:45 *** bazinski has joined #rtems 2011-03-18T07:50:06 *** mmu_man has quit IRC 2011-03-18T08:29:39 *** bazinski has quit IRC 2011-03-18T08:43:08 *** zwj has quit IRC 2011-03-18T08:48:50 *** mmu_man has joined #rtems 2011-03-18T08:54:15 *** dr__house has joined #rtems 2011-03-18T08:54:15 *** dr__house has joined #rtems 2011-03-18T08:58:20 *** mmu_man has quit IRC 2011-03-18T09:43:28 *** DrJoel has joined #rtems 2011-03-18T09:43:28 *** ChanServ sets mode: +o DrJoel 2011-03-18T09:44:15 sebhub: which BSP name is best for lpc2478? 2011-03-18T09:44:51 and would mind emailing me a known good executable? 2011-03-18T09:57:33 *** antgreen has quit IRC 2011-03-18T10:02:01 for the embedded artists board? 2011-03-18T10:02:33 the bsp name is lpc24xx_ea 2011-03-18T10:03:20 you can use the executables of the test suite 2011-03-18T10:09:22 ok.. just checking 2011-03-18T10:09:25 thank you 2011-03-18T10:10:08 do you know if the uboot supports the second (not usb) serial port? We want to be able to control the power on/off remotely and the usb power makes that tough 2011-03-18T10:12:49 *** lcpfnyvc has quit IRC 2011-03-18T10:15:30 maybe you can configure this via the environment variables 2011-03-18T10:48:19 *** mmu_man has joined #rtems 2011-03-18T10:58:53 *** dr__house has quit IRC 2011-03-18T10:59:31 *** core-ix has quit IRC 2011-03-18T11:27:12 *** dr__house has joined #rtems 2011-03-18T11:27:12 *** dr__house has joined #rtems 2011-03-18T11:45:58 *** sebhub has quit IRC 2011-03-18T11:48:12 *** mmu_man has quit IRC 2011-03-18T12:10:17 rokka: ping 2011-03-18T12:29:56 *** antgreen has joined #rtems 2011-03-18T13:44:55 *** dr__house has quit IRC 2011-03-18T13:51:33 *** DrJoel has quit IRC 2011-03-18T13:59:40 *** arvind_khadri has joined #rtems 2011-03-18T14:07:46 *** arvind_khadri has quit IRC 2011-03-18T14:08:03 *** arvind_khadri has joined #rtems 2011-03-18T14:11:37 Congrats :) 2011-03-18T14:11:55 Maybe it is nothing new for you guys :P 2011-03-18T14:44:10 *** mmu_man has joined #rtems 2011-03-18T15:07:03 arvind_khadri: that's a good news each year :) 2011-03-18T15:07:17 some organizations have been in gsoc for several years and get sometime rejected 2011-03-18T15:07:29 like enlightenment this year for example : 2011-03-18T15:07:30 :( 2011-03-18T15:43:06 *** core-ix has joined #rtems 2011-03-18T15:57:50 *** dr__house has joined #rtems 2011-03-18T15:57:50 *** dr__house has joined #rtems 2011-03-18T17:28:52 *** dr__house has quit IRC 2011-03-18T22:09:57 *** mmu_man has quit IRC 2011-03-18T22:20:00 *** antgreen` has joined #rtems 2011-03-18T22:21:35 *** antgreen has quit IRC 2011-03-18T22:43:06 *** antgreen` has quit IRC 2011-03-18T23:32:16 *** arvind_k has joined #rtems 2011-03-18T23:32:58 *** arvind_khadri has quit IRC 2011-03-18T23:33:14 *** arvind_k is now known as arvind_khadri 2011-03-18T23:33:26 *** arvind_khadri has joined #rtems 2011-03-19T02:05:21 *** arvind_khadri has quit IRC 2011-03-19T02:05:41 *** arvind_khadri has joined #rtems 2011-03-19T03:04:41 *** arvind_khadri has quit IRC 2011-03-19T03:18:04 *** arvind_khadri has joined #rtems 2011-03-19T03:18:05 *** arvind_khadri has joined #rtems 2011-03-19T04:41:56 *** arvind_khadri has quit IRC 2011-03-19T05:13:00 *** arvind_khadri has joined #rtems 2011-03-19T07:17:19 *** methril has quit IRC 2011-03-19T07:19:44 *** methril has joined #rtems 2011-03-19T08:00:26 *** arvind_khadri has quit IRC 2011-03-19T08:07:19 *** arvind_khadri has joined #rtems 2011-03-19T08:21:56 *** mmu_man has joined #rtems 2011-03-19T09:12:28 *** arvind_khadri has quit IRC 2011-03-19T09:45:28 *** arvind_khadri has joined #rtems 2011-03-19T10:44:13 *** antgreen has joined #rtems 2011-03-19T10:52:09 *** arvind_khadri has quit IRC 2011-03-19T11:11:19 *** mmu_man has quit IRC 2011-03-19T11:25:12 *** gedare has joined #rtems 2011-03-19T11:27:25 *** gedare has quit IRC 2011-03-19T11:27:48 *** gedare has joined #rtems 2011-03-19T11:44:32 *** antgreen has quit IRC 2011-03-19T14:03:24 *** gedare has quit IRC 2011-03-19T19:14:45 *** dr__house has joined #rtems 2011-03-19T20:58:00 *** zwj has joined #rtems 2011-03-19T22:05:37 *** dr__house has quit IRC 2011-03-19T22:49:30 *** wenjiez has joined #rtems 2011-03-19T22:51:58 *** zwj has quit IRC 2011-03-20T01:36:07 *** xiangfu has joined #rtems 2011-03-20T02:30:29 *** xiangfu has quit IRC 2011-03-20T02:45:05 *** wenjiez has quit IRC 2011-03-20T02:51:45 *** xiangfu has joined #rtems 2011-03-20T02:54:15 *** zwj has joined #rtems 2011-03-20T04:13:02 *** tamirtw has joined #rtems 2011-03-20T05:42:48 *** methril has quit IRC 2011-03-20T05:51:33 *** tamirtw has quit IRC 2011-03-20T06:16:53 *** wenjiez has joined #rtems 2011-03-20T06:17:41 *** zwj has quit IRC 2011-03-20T07:06:39 *** wenjiez is now known as zwj 2011-03-20T07:10:32 *** wenjiez has joined #rtems 2011-03-20T07:11:07 *** zwj has quit IRC 2011-03-20T07:13:28 *** wenjiez is now known as zwj 2011-03-20T07:22:19 *** zwj has quit IRC 2011-03-20T07:24:14 *** core-ix has quit IRC 2011-03-20T07:32:31 *** zwj has joined #rtems 2011-03-20T08:22:58 *** methril has joined #rtems 2011-03-20T08:39:06 *** tamirtw has joined #rtems 2011-03-20T09:18:34 *** mmu_man has joined #rtems 2011-03-20T09:20:14 *** xiangfu has quit IRC 2011-03-20T09:33:39 *** xiangfu has joined #rtems 2011-03-20T09:42:59 *** zwj has quit IRC 2011-03-20T09:57:32 *** zwj has joined #rtems 2011-03-20T11:26:53 *** core-ix has joined #rtems 2011-03-20T11:46:19 Alo 2011-03-20T11:48:41 *** gedare has joined #rtems 2011-03-20T12:05:10 *** zwj has quit IRC 2011-03-20T12:07:16 *** gedare has quit IRC 2011-03-20T12:18:43 *** xiangfu has quit IRC 2011-03-20T12:29:41 *** xiangfu has joined #rtems 2011-03-20T13:06:41 *** xiangfu has quit IRC 2011-03-20T14:30:28 *** dr__house has joined #rtems 2011-03-20T14:53:10 *** dr__house has quit IRC 2011-03-20T16:32:51 *** gedare has joined #rtems 2011-03-20T16:33:19 *** gedare has joined #rtems 2011-03-20T16:59:27 *** dr__house has joined #rtems 2011-03-20T17:30:53 *** _Sergey_ has joined #rtems 2011-03-20T17:47:54 *** projectgus has joined #rtems 2011-03-20T18:09:08 *** _Sergey_ has quit IRC 2011-03-20T18:42:11 *** mmu_man has quit IRC 2011-03-20T18:49:43 *** antgreen has joined #rtems 2011-03-20T20:35:28 *** gedare has quit IRC 2011-03-20T22:23:57 *** dr__house has quit IRC 2011-03-20T23:17:07 *** antgreen has quit IRC 2011-03-20T23:18:01 *** xiangfu has joined #rtems 2011-03-20T23:52:32 *** xiangfu has quit IRC