**** BEGIN LOGGING AT Fri Nov 27 03:00:53 2015 Nov 27 03:01:40 So what is r1.1 and r2 of? Nov 27 03:01:51 What design? Nov 27 03:02:02 03:53 < zmatt> in this case I was referring to the processor of course Nov 27 03:02:28 see http://www.ti.com/lit/er/sprz429g/sprz429g.pdf for the silicon revision 1.1 errata Nov 27 03:04:47 so we're still left somewhat guessing which ones will be fixed in 2.0, although this git commit entry mentions some of them: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/arch/arm/mach-omap2/id.c?id=81032e34e184a8d63598d215a6d4a3352018ffea Nov 27 03:12:20 I'm not sure why it's called a silicon revision. Nov 27 03:14:18 Also Wikipedia hasn't got an article on it: https://en.wikipedia.org/wiki/Special:Search?search=%22silicon+revision%22 Nov 27 03:14:22 :-( Nov 27 03:14:33 it's just what's it's called Nov 27 03:14:54 just read it as "chip revision number" Nov 27 03:15:57 Yep. Nov 27 03:16:22 I don't know how it is different to the design revision, other than being finer. Nov 27 03:17:55 coarser Nov 27 03:19:14 the chip's design probably has seen many revisions before an actual prototype is made Nov 27 03:19:32 What? Nov 27 03:19:37 considering that fabricating an IC is a rather involved process Nov 27 03:20:04 think of any software project: its git revision history will have many more commits than the are released versions of that software Nov 27 03:20:46 Isn't it more like ‘TI AM5728.r1.1’? Surely the silicon revision is the finest and last change. Nov 27 03:21:18 I don't understand what you're saying Nov 27 03:21:37 So is it more like ‘TI r1.1.AM5728’? Nov 27 03:21:42 I mean, I don't know much details of the whole design process, but presumably they'll also have a devel branch Nov 27 03:21:53 no, it's r1.1 of the AM572x Nov 27 03:22:16 Ah yeah… Nov 27 03:22:51 So is it more like ‘TI AM572.r1.1.8’ then, sort of thing? :-P Nov 27 03:23:43 (more generally of "Vayu", which is the name of the actual design. The broadmarket version gets an AM572x part code stamped on, while for automotive infotainment the same chip is known as DRA75x/74x) Nov 27 03:23:49 It's finer than the family but there are other models with the same silicon revision. Nov 27 03:24:18 usually silicon revisions don't have a third component (but exceptions exist) Nov 27 03:25:01 The ‘8’ was the x in ‘TI AM5728’. Nov 27 03:25:35 for example Subartic (aka AM335x, the processor on the beaglebone) has three revisions so far: 1.0, 2.0, and 2.1) Nov 27 03:25:45 the 8 is a feature indication Nov 27 03:25:45 It didn't really make sense though. Nov 27 03:25:52 Yep. Nov 27 03:25:54 that's unrelated to revision Nov 27 03:26:01 Indeed. Nov 27 03:26:15 they're all exactly the same chip, but due to production variability, some parts may not pass factory testing Nov 27 03:26:24 hence they become a lower speed-bin or feature-bin Nov 27 03:26:50 some features may also be disabled because they're only intended for specific markets Nov 27 03:27:08 Yep. Nov 27 03:27:37 for example the automotive-specific peripherals present on the Subarctic due are disabled on the AM335x but will presumably be enabled on its automotive sibling (DRA60x/61x I think) Nov 27 03:28:13 So the silicon revision is very much a fine/minor adjustment to the model. Nov 27 03:29:19 it's just a revision, same as with software revisions Nov 27 03:29:23 I don't know why it would be coarser. Nov 27 03:29:29 Surely it's finer. Nov 27 03:29:55 I think we just miscommunicated and meant different things with "design" Nov 27 03:30:05 Okay. Nov 27 03:30:40 Also, what are the classification names? Like family, model, etc.? Nov 27 03:31:39 I meant the "source code" of the chip, which will probably see many internal updates before it's pushed out into the world as a new silicon revision Nov 27 03:31:40 I'm never sure what to call things like the ‘TI AM5728’ versus the ‘TI AM572x’ versus broader collections and such. Nov 27 03:32:04 most such classifications are done by marketing people and have no technical value Nov 27 03:32:19 Hmm. Nov 27 03:32:43 there are some clear ways to group them though Nov 27 03:35:02 Is it generally like series/family/model/revision? Nov 27 03:35:07 for example AM572x series (and automotive DRA74x/57x and TDA2) are all the same die, just different settings burned into eFUSE and possibly different package Nov 27 03:36:15 many dies fortunately have names ("Vayu" in this case), which I personally consider more convenient than trying to memorize similar-looing part numbers Nov 27 03:36:40 Hmm, it's 2-dimensional, isn't it. Nov 27 03:38:56 I prefer the numbers/codes than strange names that sometimes I can't pronounce. Nov 27 03:39:37 well, people confuse the AM335x with the AM35xx way too often Nov 27 03:39:48 even though they're quite dissimilar Nov 27 03:40:23 on the other hand "subarctic" has hard to confuse with anything else Nov 27 03:40:24 * JamesJRH shrugs. Nov 27 03:42:03 unfortunately it seems minor derivatives of a chip (like the AM571x / DRA72x) don't seem to get their own name, and get wiped under the name of the parent device it seems Nov 27 03:42:25 another useful classification is by tracing the lineage of a processor Nov 27 03:42:57 Vayu is a very clear omap 5 derivative for example (even though substantial modification has been done) Nov 27 03:44:10 subarctic is part of a family that started with Netra (dm816x/am389x), then Centaurus (dm814x/dm812x/am387x) and many Nov 27 03:44:45 minor variants (dm811x, dm38x) Nov 27 03:45:24 then subarctic, and finally aegis (am437x) Nov 27 03:45:35 all have many structural similarities Nov 27 03:46:32 aforementioned family is all omap4/5-generation (this in contrast to the am35xx which is an omap3-derivative) Nov 27 03:47:36 By “2-dimensional”, I mean that it's kind of like series/family/(revision×model) or equivalently series/family/(model×revision). I.e., model and revision are 2 dimensions of family; they are both a subclassification of it. Nov 27 03:48:19 except the part codes that split that nicely Nov 27 03:48:35 about the only standard part is that the last digit is a feature-level indication Nov 27 03:49:17 That's what I mean. Nov 27 03:49:20 and the "am" prefix is for Sitara ARM processors, but that's itself just a marketing decision Nov 27 03:49:38 Maybe ‘part’ is what I meant by ‘model’. Nov 27 03:50:57 Not sure. Nov 27 03:50:59 but look at the "evolution" of "am" processors: Nov 27 03:51:55 (freon) AM18xx with last digit even (same die is also omap-L138 and C674x with x even) Nov 27 03:52:25 (primus) AM17xx with last digit odd (same die is also omap-L137 and C674x with x odd) Nov 27 03:53:04 am35xx (omap3-derived, but with substantial changes) Nov 27 03:53:22 am37xx (just an omap36xx with some features disabled) Nov 27 03:54:25 am389x (netra; definitely shared some development with the omap4, but still quite different from an omap) Nov 27 03:54:35 Wait, what happened to Omap? Nov 27 03:54:47 I'm just listing AM processors here Nov 27 03:54:58 actually Nov 27 03:55:38 I mean, the BeagleBoards were Omap weren't they? Now Am. Nov 27 03:56:00 the bb was omap35xx, which was the "broadmarket" version of the omap34xx Nov 27 03:56:47 for the omap36xx TI decided on a naming change and call its broadmarket version the "dm37xx" instead of "omap37xx" (although you can still occasionally find references to "omap37xx" in old docs) Nov 27 03:57:01 am37xx is a downfeatured version thereof Nov 27 03:57:12 But Omap is ARM too, so why do you say?: 03:49:20 < zmatt> and the "am" prefix is for Sitara ARM processors, but that's itself just a marketing decision Nov 27 03:57:44 most TI SoCs have ARM processors Nov 27 03:58:03 the AM prefix is just for those labeled "Sitara" Nov 27 03:58:10 while DM parts are "DaVinci" Nov 27 03:58:38 Omap? Nov 27 03:58:45 omap is omap Nov 27 03:58:49 Lol. Nov 27 03:58:50 :-) Nov 27 03:59:33 So why did the BeagleBoard go from Omap to Sitara for the X15? Nov 27 04:00:06 I really have no idea however why they placed Vayu under Sitara instead of DaVinci... it has everything you'd expect to find on a DM and far more Nov 27 04:00:36 BB went from omap to davinci (dm37xx) to sitara Nov 27 04:00:50 but like I said, that's just the marketing department at work Nov 27 04:01:45 at least I'm assuming dm37xx was labeled davinci, considering its DM prefix Nov 27 04:02:08 "DM37x DaVinci Video Processors" bingo Nov 27 04:02:30 even though the same die as the omap36xx Nov 27 04:02:47 but different audience, different marketing, different part number :P Nov 27 04:03:12 But, it all ended up at the same audience. Nov 27 04:03:24 The marketting department failed there. Nov 27 04:04:04 Yes, I forgot about most of them being DM, I just had another look now. Nov 27 04:04:14 yeah, well, and look at the DM816x and DM814x whose less-featured Sitara counterparts are the AM389x and AM387x Nov 27 04:04:47 (At https://en.wikipedia.org/wiki/BeagleBoard#Specifications_2 .) Nov 27 04:05:06 also, the part numbering within the DM81xx series: DM816x is oldest (Netra), then came DM814x (Centaurus) which is also available as DM812x Nov 27 04:05:25 I was thinking Omap because I have a PandaBoard ES which I'm pretty sure is Omap. Nov 27 04:05:42 the even newer Centaurus-derivative DM811x exists but was ultimately never released for broadmarket.... continues to exist as DRA62x though Nov 27 04:05:46 pandaboard is omap4 Nov 27 04:06:24 Yup: https://en.wikipedia.org/wiki/PandaBoard Nov 27 04:06:34 DM810x is netra again I think but I'm not sure Nov 27 04:06:45 DM813x was renamed DM38x Nov 27 04:07:33 (which, to be fair, was done for consistency with other DM part numbers, since the DM38x has no DSP core) Nov 27 04:09:41 JamesJRH: https://docs.google.com/spreadsheets/d/1rxDGr72OfNYQ8-7zwzO5GsoykrbMpmKxvWkEVy862K4/pubhtml?gid=733558152&single=true Nov 27 04:10:09 includes most recent SoCs and C6x DSPs, and a sporadic few unrelated parts Nov 27 04:10:55 a revision like "2.1 (C)" means the revision is known (in documentation) as 2.1 and marked "C" on the package Nov 27 04:11:57 Ah! Nice! Nov 27 04:12:14 That makes things a lot clearer. Thanks. Nov 27 04:13:04 a diagram showing the inheritance/evolution of subsystems would also be nice Nov 27 04:13:58 e.g. PWMSS in subarctic is derived from that in freon/primus, which took it from TI's "C2000" series of real-time microcontrollers Nov 27 04:15:39 Is there a way to download that Google spreadsheet as ODS, CSV, or some other common format? Nov 27 04:15:57 should be Nov 27 04:16:16 Yeah, that /would/ be nice. :-D A sort of DAG. Nov 27 04:16:43 https://docs.google.com/spreadsheets/d/1rxDGr72OfNYQ8-7zwzO5GsoykrbMpmKxvWkEVy862K4/edit?usp=sharing you can export from there Nov 27 04:17:18 Thanks. Nov 27 04:33:51 Google's ‘File’ → ‘Download as’ → ‘OpenDocument format (.ods)’ is failing to do anything, yet the other options work. :-/ Nov 27 04:40:27 At least I have a copy incase the online one disappears. Nov 27 04:41:11 Hey... Nov 27 04:41:38 Everyone, Turkey Day is almost over. What projects, if any, happened today? Nov 27 04:41:49 <<<< squat Nov 27 05:19:59 it seems to be common practice to read the ADC from the PRU and then write samples to memory. Why is this better than reading the ADC from the ARM processor? Nov 27 05:22:50 ocamlman.. you mean from userspace? Nov 27 05:22:56 ie .linux Nov 27 05:23:32 under linux I assume the FIFOs will get filled up before you are scheduled, but how about under starterware? Nov 27 05:24:25 I dunno .. :/ Nov 27 05:25:21 I guess most people doing these demos aren't using starterware so thats not a good indication its a good idea Nov 27 05:28:13 Hey...do I need an i2c connect to make my oled display work on a BBG? Nov 27 05:28:29 I am using a Grove connection! Nov 27 05:30:46 ocaml... damn... Nov 27 05:33:07 The software when ran show an output of "Check Your I2C Address." Nov 27 05:33:19 Boo! Nov 27 05:34:04 I am just running "Hello World" first on the Oled for my BBG. Tough times in Oak Grove. Sheesh. Nov 27 05:49:34 The OLED is no joke. It just ruined the entire c9 interface for me. Boo! Nov 27 05:49:56 error code 939 pops up. Yikes... Nov 27 05:50:00 ... Nov 27 12:18:17 anyone heard from rcn-ee lately? Nov 27 12:24:11 probably in a post-turkeyfest food-coma? Nov 27 13:11:18 oh damn this bloody turkey business .. some of us are trying to work .. ahem .. play .. Nov 27 13:11:38 but I may have a cunning plan ... Nov 27 13:11:48 .. which shall tide me over .. Nov 27 13:12:02 there won't be any turkey left for christmas if this continuous Nov 27 20:22:19 Hi. Nov 27 20:22:52 Apparently USB 3 requires nonfree firmware. Nov 27 20:23:01 Is that so? Nov 27 20:25:28 What XHCI controller does the BeagleBoard X15 have? Nov 27 20:27:27 hmm, good question Nov 27 20:37:47 Hello community! :) Is anyone else having trouble downloading the newest Debian images? Nov 27 20:42:13 tbr: Also, is the USB 2 OTG interface an independent controller? Nov 27 20:43:34 tbr: I.e. can it be used even if the nonfree XHCI firmware isn't loaded for the USB 3 controller? Nov 27 20:44:39 JamesJRH: might be better to ask on the X15 mailing list/group Nov 27 23:47:03 What are the best ways to automate system tasks on the Beagle? Nov 27 23:47:30 Just using bash and such? Nov 27 23:55:56 Automate how? Nov 28 01:05:44 Why does my beaglebone keep rebooting when I try to shut it down? Nov 28 01:05:56 I've tried holding the buton for 8 seconds, and also sending it "shutdown -h now" Nov 28 01:06:01 and it still reboots. I want it to shut down Nov 28 02:44:14 JamesJRH: check the TRM and/or the kernel **** ENDING LOGGING AT Sat Nov 28 03:00:46 2015