**** BEGIN LOGGING AT Fri Dec 08 03:00:02 2017 Dec 08 05:08:27 I am using a BBB wireless board for a project, where I need to run the old Angstrom OS (2013). Can someone provide Wireless driver for this OS? Dec 08 05:57:03 someone sure can. Dec 08 06:05:30 I'm guessing that would mean updating the board file on a pre-DT kernel Dec 08 06:06:23 and probably backporting the driver Dec 08 06:33:09 easy Dec 08 08:13:43 for you anyway :) and probably me too, if I cared even the slightest bit about supporting ancient kernels Dec 08 09:46:46 I'm trying to work out if I can get the spi controller to add some time between words Dec 08 09:46:57 I've checked all the registers documented in spruh73p.pdf Dec 08 09:47:09 and can't find anything - so I'm guessing no. Dec 08 09:47:19 anyone know different? Dec 08 10:18:45 hmm Dec 08 10:18:55 * zmatt checks his notes Dec 08 10:21:49 yeah it can add a few cycles between CS assertion and the first word, but not between words Dec 08 10:23:17 so the only option for that is by delaying in software Dec 08 10:31:20 keeping turbo mode disabled and using an tx+rx transfer (even if you don't care about the rx data) also causes some inter-word gap Dec 08 10:41:47 (adds about 146 us delay between words) Dec 08 10:41:55 uh, ns sorry Dec 08 10:42:49 (tested at 12 MHz and 24 MHz, fifo enabled) Dec 08 11:12:55 thanks @zmatt Dec 08 11:13:33 how can I change these options? Dec 08 11:42:40 are you interacting directly with the spi peripheral in some way or do you using the linux driver? Dec 08 11:42:43 *are you Dec 08 11:43:39 using the python wrapper for the linux driver Dec 08 11:45:33 it would require changing the driver... although I'm not even sure it uses turbo mode currently Dec 08 11:47:04 yeah it doesn't Dec 08 11:47:17 I think Dec 08 11:48:02 it definitely doesn't (even though I'd consider that to be a bug) Dec 08 11:48:17 I'm using this driver https://github.com/doceme/py-spidev# Dec 08 11:48:47 so I've checked out that c code Dec 08 11:48:54 and I can see the spi transfer structure getting set up Dec 08 11:48:54 that's not a driver, that's a python wrapper for the spidev api Dec 08 11:49:31 yep Dec 08 11:49:40 this is what I@m looking at on the linux side Dec 08 11:49:41 https://github.com/spotify/linux/blob/master/include/linux/spi/spidev.h Dec 08 11:50:01 there's no way to influence the turbo-flag via the spidev api Dec 08 11:50:04 and I don't see any reference to turbo mode there, only in the chip's datasheet Dec 08 11:51:06 that's because it has nothing to do with the spi protocol, it's a configuration bit specific to the mcspi peripheral Dec 08 11:52:30 right Dec 08 11:52:36 https://github.com/RobertCNelson/linux-stable-rcn-ee/blob/4.9.67-ti-r82/drivers/spi/spi-omap2-mcspi.c#L1178-L1182 Dec 08 11:52:37 looks like by default turbo mode is off Dec 08 11:52:53 this piece of code is responsible for the decision to enable turbo mode Dec 08 11:53:18 cd will be NULL in this case Dec 08 11:54:41 also, even if cd && cd->turbo_mode were true, apparently it only enables turbo mode for rx-only transfers.... I have no clue why they'd impose that restriction Dec 08 11:55:14 because of the t->tx_buff == NULL Dec 08 11:55:15 ? Dec 08 11:55:32 yeah that means no tx data, i.e. an rx-only transfer Dec 08 11:55:36 right. Dec 08 11:56:02 anyway, turbo mode is clearly off, so you already have the highest inter-word gap you can get purely with hardware flags Dec 08 11:56:08 right Dec 08 11:56:38 so you'll have to manually insert delays Dec 08 11:56:55 if the python wrapper supports that. (I haven't looked) Dec 08 11:58:01 I had a look at the cs_change option Dec 08 11:58:15 so I could send one word, then wait, then send the rest of the words Dec 08 11:58:18 that would need to be set for all but the last transfer Dec 08 11:58:24 to ensure cs remains asserted Dec 08 11:58:33 uh no wait Dec 08 11:58:54 I mean, if you do one multi-part transfer, the defaults are right Dec 08 11:59:00 it's because I'm transferring data from an fpga and I want to get the data I specifiy in the first 7 bits (the address) to come straight back in that read Dec 08 11:59:22 which means transferring the data to the spi register on the fpga in the time between the first and second word sent from the fpga Dec 08 11:59:56 but maybe I'll just set the address on one write, and then read the result on a second Dec 08 12:02:15 @zmatt. thanks for the help Dec 08 12:02:46 hmm, looks like py-spidev might not support it but the code isn't entirely clear to me Dec 08 12:05:39 you're welcome Dec 08 12:09:58 is this group quite active? Dec 08 12:10:16 this irc channel? varies Dec 08 12:10:24 I'm fairly new to irc and theres a lot of channels with very low activity Dec 08 12:11:19 it obviously varies even more per channel Dec 08 12:57:38 and you're likely get answers on 'interesting' stuff quicker than for boring 'I cannot connect via usb on my windows 7/8/10)' Dec 08 13:08:50 hehe, indeed Dec 08 13:10:06 can u plz send me teh exact step i need to make opencv work on beagle my projet is due monday plz help my mail is (INSERT HER) Dec 08 13:10:40 ^^^^^^^ this is a full-on no-answer guaranteed question, just as a counter example. Dec 08 13:13:28 lol Dec 08 13:14:33 (all typos were intended and part of the artistic perforemance, just for the record.) Dec 08 13:31:03 I try to research as far as I can before bothering other people, but often I've over looked something straight forward Dec 08 13:31:32 Dec 08 13:35:42 mattvenn: it showed. your remark about having checked the TRM and not found anything at least helped you get my attention more quickly Dec 08 13:35:56 unfortunately, the same isn't true for everyone who drops in here Dec 08 13:43:01 mattvenn: yeah, the pun was really not on you. Dec 08 16:57:44 can i buy beagle and use texas instrument coupon code for paying money Dec 08 17:00:58 anyone Dec 09 02:29:56 Hi I am trying to get a GUI installed on my BBG for x11 forwarding Dec 09 02:30:00 I am failing miserably Dec 09 02:30:28 dpkg: error processing archive /var/cache/apt/archives/chromium_62.0.3202.89-1~deb9u1_armhf.deb (--unpack): trying to overwrite '/usr/lib/chromium/natives_blob.bin', which is also in package chromium-common 61.0.3163.100-2rcnee0~stretch+20170927 dpkg-deb: error: subprocess paste was killed by signal (Broken pipe) Selecting previously unselected package chromium-l10n. Preparing to unpack .../chromium-l10n_62.0.3202.89-1~deb9u1 Dec 09 02:30:51 so i am giving up on chromium but lxde does not work either **** ENDING LOGGING AT Sat Dec 09 03:00:00 2017