**** BEGIN LOGGING AT Fri Jul 13 02:59:58 2012 Jul 13 10:04:19 gm Jul 13 10:04:23 gm ensc|w Jul 13 10:05:55 ensc|w: Re yesterday's i.MX28 FEC discussion. The MODE0:2 pins are pulled-up in the i.MX28, thus when the PHY releases them during RST# asserted low, the MODE0:2 are driven high during RST# going low. Jul 13 10:06:41 ensc|w: I see however that the latest HEAD kernel explicitly drives the signals high using GPIO rather than internal pull-ups. Jul 13 10:07:35 ok, sheevaplug works fine now Jul 13 10:09:40 moin mickey_office Jul 13 10:49:11 morning likewise, hrw Jul 13 10:49:31 hrw: I read your debricking blog post with interest this morning :) Jul 13 10:51:29 morning all Jul 13 10:52:42 bluelightning: I wanted to make it in a way that anyone can recreate all steps Jul 13 10:53:20 I had not realised openocd restore on some devices could be done without a special jtag cable... Jul 13 10:54:09 bluelightning: sheevaplug has jtag+serial on miniusb Jul 13 10:54:28 hrw: is that common? Jul 13 10:54:39 bluelightning: nope ;( Jul 13 10:54:53 and all you need is iirc one FTDI chip to have it working Jul 13 10:55:40 I was suprised that u-boot has working ubi(fs) support Jul 13 11:32:52 likewise: driving them actively high is a little bit critical; lan87xx switches them to output 800ns after deasserting reset which is to fast for gpiolib Jul 13 11:33:10 likewise: external pullups/downs are the best solution Jul 13 11:34:12 ensc|w: but the lan87xx samples the mode pins on reset going high, and certainly before the lan87xx switches the pins to output mode (RX0/1) Jul 13 11:34:58 ensc|w: I'm not driving the pins high actively using GPIO, I am enabling the internal pull-ups in the i.MX28 Jul 13 11:35:14 ensc|w: that should equal external pull-ups. Jul 13 11:38:23 likewise: yes; internal pullups are possible when configuring them as gpio-in Jul 13 11:39:17 but lan87xx seems to go into some strange state when there are no external gpios and requires *two* long resets to be usable again Jul 13 11:39:52 (with "long" == 10-20ms which is much more than the reset time from datasheet) Jul 13 11:41:08 ensc|w: and these device quirks arein the kernel at meta-fsl-arm repo? Jul 13 11:42:12 ensc|w: currently I am facing the same issue Daniele Dall'Acqua saw on http://imxcommunity.org/forum/topics/help-about-fec-problem?id=4103961%3ATopic%3A13423&page=2#comments Jul 13 11:46:54 likewise: https://www.cvg.de/people/ensc/fec-reset.c Jul 13 11:47:14 likewise: it's a custom hardware and too ugly to push it upstream Jul 13 11:53:13 bluelightning, I believe the beaglebione has jtag on usb also Jul 13 11:53:51 also, good to see the updated systemd patchset hit the list Jul 13 11:54:25 hi all Jul 13 12:29:56 Crofton|work: yes, I was happy to see that too Jul 13 12:49:48 When I do "bitbake -c confgure puppa.bb", where are stored puppa's configure options? Jul 13 13:04:08 ensc|w: thanks, so that double reset helped for you? Didn't encounter it anywhere else. Jul 13 13:19:51 likewise: yes; I was able to reproduce the problems nearly to 100% when booting directly into linux without a bootloader. the long double reset solved it Jul 13 13:58:50 silvio__: not sure what you mean by "stored" but the input configure options for an autotools recipe come from EXTRA_OECONF, and we save the configure logs for the output... Jul 13 15:27:32 bluelightning: I mean if i want add or remove same specific option in the ./configure, which is the correct way to proced? Jul 13 15:28:04 silvio__: right, that would be the EXTRA_OECONF variable then Jul 13 15:36:45 bluelightning: thanks Jul 13 16:25:45 bb Jul 13 16:26:34 otavio: I didn't realize you're using repo nowadays. That's cool. We're using that for some stuff at work. The only thing that took a little getting used to was not havinga toplevel repository, instead having the mostly out-of-band manifest Jul 13 20:48:19 kergoth: so in repo you dont have someone who is pulling patches into repo Jul 13 21:20:04 hi all. i've just installed openembedded and bitbake using "getting started" guide. i'm trying to build something but am getting this error: http://pastie.org/4252232 Jul 13 21:20:28 what am i doing wrong here? Jul 13 21:21:10 both oe and bb are from git master Jul 13 23:04:02 khem: not sure what you mean by that. repo is just another way to manage multiple repositories, like git-submodules. the difference is that repo can point to the tip of a branch of a repository, not just a specific commit Jul 13 23:04:28 and there's no toplevel repository, no 'master' repository, instead the 'manifest' which points at the individual repositories is slightly out of band Jul 13 23:05:12 oh I confused it with gerrit Jul 13 23:05:14 sorry Jul 13 23:05:36 ah, yeah, gerritt is the code review system that repo happens to be slightly integrated with **** ENDING LOGGING AT Sat Jul 14 02:59:58 2012