**** BEGIN LOGGING AT Fri Aug 05 02:59:56 2011 Aug 05 03:52:30 is there any document regarding to how to build/generate the oneiric preinstall image? Aug 05 12:04:36 hello Aug 05 12:07:13 building on 11-04 , what is the -mfloat-abi flag for gcc-4.5 ? hardfp or softfp ? Aug 05 12:07:27 I forgot : I'm building on a Pandaboard Aug 05 12:07:49 hey, I'm looking for an official ARM-Repository for Ubuntu 10.04, is there any? I'm on Hercules ecafé with Freescale i.MX515 Aug 05 12:19:56 kki313, arm packages are on ports.ubuntu.com Aug 05 12:20:24 ogra_ thanks! Aug 05 13:15:35 persia, you there? Aug 05 14:16:52 ndec, https://bugs.launchpad.net/bugs/709245 Aug 05 14:16:53 Ubuntu bug 709245 in linux-ti-omap4 "panda: USB disk IO slow" [High,Confirmed] Aug 05 15:29:48 ogra_: about the IO bug, you told me it was affecting SD too, but that's not what I understand from the bug report. are u sure? Aug 05 15:33:06 have you looked at the measurements ? Aug 05 15:33:21 (and at the linked duplicate bug) Aug 05 15:33:40 * ndec looking again Aug 05 15:34:42 ogra_: ok. i see it. 'to a lesser extent' though ;-) Aug 05 16:01:19 <[7]> ndec: and I'm not sure if it's the same cause Aug 05 16:02:03 ok Aug 05 16:03:05 <[7]> the ping doesn't seem to affect SD much, but the speed vs. block size graph shows similar weirdnesses like USB, just not as huge Aug 05 16:03:27 <[7]> that might possibly be some odd behavior of the card though, wouldn't be the first time :) Aug 05 16:04:14 is someone looking at the problem related to SMP? Aug 05 16:04:47 <[7]> ndec: I just performed my tests with SD instead of usb, and pinging brought it from 8.8 to 8.9MB/s Aug 05 16:07:05 <[7]> the same test shows USB accesses skyrocket from 3 to 20 MB/s Aug 05 16:07:34 that's not the same definition of improvements ;-) Aug 05 16:08:20 <[7]> I haven't done nosmp tests for SD, but for USB nosmp will make it behave as if I would be pinging it Aug 05 16:15:24 [7]: I'm running comparative tests now. Aug 05 16:15:35 Give me ~10 minutes. Aug 05 16:19:35 GrueMaster, would be interesting to try different schedulers and see what impact that has Aug 05 16:19:48 with USB Storage that is Aug 05 16:20:26 If I can find a kernel guy that can give me a few respun kernels, sure. Aug 05 16:22:50 http://www.cyberciti.biz/faq/linux-change-io-scheduler-for-harddisk/ Aug 05 16:23:02 no need for a new kernel can change on the fly Aug 05 16:24:05 <[7]> davidm: for my tests it was "noop" Aug 05 16:24:43 <[7]> testing cfq now Aug 05 16:25:41 <[7]> heh, with cfq not even pinging will help Aug 05 16:25:54 <[7]> or at least help a lot less Aug 05 16:26:10 <[7]> *facepalm* Aug 05 16:26:21 <[7]> ignore my last messages, was using my tool in the wrong way Aug 05 16:27:34 Should record results in bug when you are sure you did the test correctly Aug 05 16:27:55 * [7] will do Aug 05 16:31:12 <[7]> wtf. after playing around with these schedulers a bit the problem just disappeared Aug 05 16:31:20 <[7]> i'll reboot and retry Aug 05 17:18:46 <[7]> davidm: uploaded the logs to the bug report Aug 05 17:18:53 <[7]> the results aren't really conclusive to me Aug 05 17:21:13 <[7]> especially the perl -e "while (1) { }" test seems to behave somewhat randomly **** ENDING LOGGING AT Sat Aug 06 02:59:57 2011