**** BEGIN LOGGING AT Sun Oct 20 02:59:58 2019 Oct 20 03:00:44 your expectations? Oct 20 03:01:07 that this is a good entry point to programming TI MCU's...? Oct 20 03:01:15 no, these are not MCUs Oct 20 03:01:23 it said so on the front page Oct 20 03:01:47 I suppose programming some of the auxiliary cores has some similarities to programming MCUs, but... also not really Oct 20 03:01:50 where? Oct 20 03:02:20 I think I was mistaken Oct 20 03:02:27 sorry! Oct 20 03:04:55 afaik the original definition of a microcontroller was a microprocessor with integrated non-volatile storage to allow it to function autonomously... but it's also generally understood to be much smaller and easier to understand processors, typically programming baremetal or with an rtos... generally speaking, if you run linux on it, it's not a microcontroller Oct 20 03:06:07 I would Agee with that Oct 20 03:07:36 having said that, I've done baremetal programming on TI SoCs and it is kinda fun having all these immense resources at your disposal without a general-purpose OS getting in your way ^_^ Oct 20 03:09:19 I had written a Forth system for the DM8148 (an ancestor of the AM335x).... and even though I had the DDR3 ram fully working I rarely had any use for it since my stuff would fit quite comfortably in its 128KB of internal SRAM Oct 20 03:10:07 but yeah, everything you want to do takes a lot of time Oct 20 03:11:39 Hey! Oct 20 03:12:06 Could a sound sensor be used instead of the button easily on the Xbee idea? Oct 20 03:12:33 zmatt: did you not find writing Forth to be pleasurable? Oct 20 03:13:23 jkridner[m]: any idea why TI seems to be tauting the AM5749 (2 EVE + 2 DSP) as their fastest platform for TIDL when the older AM5729 has 4 EVE + 2 DSP ? Oct 20 03:13:54 ullbeking: I love Forth Oct 20 03:15:12 actually this might be a good time to ask... Oct 20 03:15:36 ullbeking: having an interactive Forth running on a target is also the best way I can imagine to play with many peripherals and subsystems available on these SoCs and really get to know how they work Oct 20 03:15:50 nice if this was standard Oct 20 03:16:38 I'm working on a project to implement a remote control module of a server over IP, e.g., serial over LAN Oct 20 03:17:01 it would be an SBC-like board (e.g., BBB) plus firmware (e.g., OpenBMC) Oct 20 03:17:09 there's a lot of work already done Oct 20 03:17:19 but I'm taking it in a different direction Oct 20 03:18:00 what in your experience is the "most libra" SBC? I saw an NXP i.MX mentioned above, which are very good at this, and this is what prompted the question. Oct 20 03:18:05 any ideas? Oct 20 03:18:16 s/libra/libre/ Oct 20 03:20:49 TI is very open with a lot of stuff, both in terms of documentation and software. I think the only part of current TI SoCs that requires anything closed-source is the (rarely-used) 3D GPU, whose firmware and userspace libraries are closed source (but freely redistributable) and TI doesn't own the rights so they couldn't open it if they wanted to Oct 20 03:21:29 overall I'm not really a "libre"-obsessed guy though Oct 20 03:22:12 obviously I prefer open source, but I'm not into religious purging of anything closed source Oct 20 03:22:47 especially firmware seems like a non-issue to me, I view it as part of the hardware. (indeed they could have baked that firmware into ROMs in the hardware, and I'd much rather they didn't) Oct 20 03:31:59 zmatt: could the AM5729 be running the EVE slower for thermal reasons ? Oct 20 03:35:09 AM572x datasheets specs for DSP/EVE: 600/535 MHz at OPP_NOM, 700/650 MHz at OPP_OD, 750/650 MHz at OPP_HIGH Oct 20 03:36:08 vs what speeds on the AM574x? Oct 20 03:36:57 same Oct 20 03:37:10 any errata limiting speeds ont he 572x? Oct 20 03:37:21 not that I remember Oct 20 03:38:25 iirc am574x 1.0 is basically equivalent to am572x 2.0 as far as errata goes Oct 20 03:45:51 ok no they did fix some stuff Oct 20 03:47:57 I think I got the mic working w/ the xbees and a LED! Oct 20 03:48:40 Noise makes my LED turn on! boy! Oct 20 03:51:20 Who knows? Maybe one day I can talk in morse code to my BBB via Xbee S1. Oct 20 03:57:15 ds2: https://pastebin.com/raw/wsr5tTgw Oct 20 03:57:41 (of course the am572x trivially lacks CSI-related errata since it lacks CSI) Oct 20 05:22:16 the EMIF-FW one looks annoying Oct 20 05:25:00 I don't think linux even has any code to support the firewalls, afaik they're typically mostly used on HS devices (and there set up by the secure-world code)... regardless, it's easy enough to deal with Oct 20 05:29:04 like the workarounds say, if you want to reconfigure the EMIF firewall from the cortex-a15, just make sure only core is running at that time (by doing the config before the second core is started, or by using an inter-processor interrupt) Oct 20 05:31:29 ah wait it doesn't matter from where the reconfiguration is done, the cortex-a15 must simply be prevented from accessing EMIF Oct 20 05:35:23 okay I need to sleep I think... for some reason my head blended this erratum with a different one that's specific to requests from cortex-a15, and it's silly how many times I read it before realizing this erratum isn't mentioning the A15 at all Oct 20 05:36:12 okay that does make this erratum a fair bit more serious Oct 20 05:45:43 can be used to protect the system from the PRU Oct 20 05:48:30 ah! it seems I was thinking of omap4470 erratum i677, which also concerns lockup as a result of accessing EMIF doing reconfiguring the EMIF firewall, but that erratum is specifically to concurrent access by the two cortex-A cores Oct 20 05:48:42 *during Oct 20 05:48:51 yeah true Oct 20 05:50:45 you could statically reserve some physical memory for PRU and setup the firewall in u-boot Oct 20 05:51:15 (though even that with a bit of care, e.g. via a helper function placed in SRAM) **** BEGIN LOGGING AT Sun Oct 20 09:03:25 2019 Oct 20 15:41:51 hello guys Oct 20 15:53:29 bye! :p Oct 20 16:04:34 rcn-ee[m]: just spied https://www.kernel.org/best-way-to-do-linux-clones-for-your-ci.html on the kernel.org pages. Maybe worth incorporating into your build scripts perhaps? Oct 20 18:03:10 veremitz: never though of that.. Oct 20 18:03:35 certainly seems interesting Oct 20 18:04:14 when was git bundle first added? Oct 20 18:09:40 idk Oct 20 18:13:02 okay ancient git: 2.1.4 (jessie) has it.. that's minimal in my scripts.. Oct 20 18:16:34 I'da thought it would make clones faster .. Oct 20 18:20:10 ah now why was I trying to do git over nfs .. this would be a clean solution .. Oct 20 18:20:23 something lacking disk space I suspect :/ Oct 20 18:21:17 yeah that was a lot faster.. Oct 20 18:21:56 rcn-ee[m]: hmm there was something I was gonna send you via patch, but I can't remember what RN :/ lol Oct 20 18:22:26 oh .. updates to your wiki .. is there a repo for it or anything? Oct 20 18:22:43 I found a shortcut/better method for something .. Oct 20 18:24:39 hmm need to finish setting up uboot+grub2 .. :D Oct 20 18:24:56 I need to revisit the opensuse guy's presentation >,< **** ENDING LOGGING AT Mon Oct 21 02:59:58 2019