**** BEGIN LOGGING AT Wed Jan 11 02:59:56 2023 Jan 11 03:11:27 Dang cameras. Nice articulation of ideas...I could not find the darn cam to save my life. Blah. Jan 11 08:18:43 BeagleBone® AI Board, indian dealers list? Jan 11 16:12:29 What controls which gpios get created automatically under /sys/class/gpio/ ? Jan 11 16:13:18 I know I can create additional directories/gpios myself after the fact, but I'm asking about what mechanism creates the ones that are already there. Jan 11 17:27:16 Guest4494: gpios can be configured and exported via device-tree (by creating a gpio-of-helper device). cape-universal (if enabled, which it is by default) does this for all gpios Jan 11 17:27:38 all gpios corresponding to P9/P8 expansion header pins that is Jan 11 17:35:30 zmatt, that seems more complicated than making gpio settings in uEnv.txt.  Or am I misunderstanding? Jan 11 17:37:48 I don't know what you mean by that Jan 11 17:39:57 there's no way to "make gpio settings in uEnv.txt" Jan 11 17:41:01 gpios can either be configured at runtime (which is what most people do), or they can be given a default config via device-tree Jan 11 18:36:36 (Sorry, got pulled away.) I just meant that one can, say, set a gpio to output using u-boot's 'gpio set' command and placed in uEnv.txt. Jan 11 18:53:54 (Sorry, got pulled away.) I just meant that one can, say, set a gpio to output using u-boot's 'gpio set' command and placed in uEnv.txt. Jan 11 19:53:01 you can add that sort if thing to the uenvcmd variable, if still there. Jan 11 19:58:55 HI!   How can I contact your sales team? Jan 11 20:04:30 hi am using beagle bone xm its a old board lying with me can anyone help me with putting os in sdcard where to download and flash with etcher and start using the board Jan 11 21:17:54 Hi, I believe this is a community channel instead MAri Jan 11 21:18:25 Slaes at the beaglebone site porobably if you need contact Jan 11 21:29:44 I'm trying to create a bootable disk image for a beagleboard x-15; I've compiled u-boot and dd'd MLO and u-boot.img to an sd card but it's not booting Jan 11 21:30:30 to compile, I did "make CROSS_COMPILE=arm-linux-gnueabihf- am57xx_evm_defconfig" followed by "make CROSS_COMPILE=arm-linux-gnueabihf- -j20" Jan 11 21:33:17 to dd, I did "sudo dd if=MLO of=/dev/sde conv=notrunc count=2 seek=1 bs=128k" followed by "sudo dd if=u-boot.img of=/dev/sde conv=notrunc count=4 seek=1 bs=384k" Jan 11 21:33:29 any ideas why it might not be working? Jan 11 21:37:05 nevermind, I've seen the problem: the u-boot.img is bigger than (count=)4 * (bs=)384k Jan 11 21:50:29 thank you, cardboard developer Jan 11 21:59:56 Guest4494: oh, that's what you meant... ew? are you using some kind of ancient image? I feel like I haven't seen u-boot commands in an uEnv.txt the last decade or so :P anything you do to gpios in u-boot would normally be undone by the kernel if the gpios are exported by DT, which on current images by default happens to all pins Jan 11 22:01:16 so I'm guessing you're either using an image that predates cape-universal or you've explicitly disabled it Jan 11 22:02:16 (in retrospect I should have guessed that from your question about automatically exporting gpios) Jan 11 22:17:31 hmm, when I zero the first 10M of the SD card and write MLO/u-boot.img, the card boots fine; when I have a partition table and root FS and write ML/u-boot.img, the card does not boot; any ideas why? Jan 11 22:32:58 rah: can you hexdump the first 257 sectors of the non-working case? i.e. dd if=DEVICE count=257 status=none | hexdump -C Jan 11 22:33:23 btw you can just omit the count= argument to dd to avoid the problem you had earlier Jan 11 22:34:16 also instead of messing with the blocksize I'd just specify a proper seek value, i.e. "seek=256" instead of "seek=1 bs=128k" and "seek=768" instead of "seek=1 bs=384k" Jan 11 22:35:25 or if you really want to write in 128k blocks then at least "seek=3 bs=128k" instead for the u-boot.img ... using a blocksize of 384k is just kinda gross :P Jan 11 22:36:54 anyway, the only thing I can think of is... if there's crud in the initial bytes of sector 0 (the space for the MBR) then maybe bootrom mistakenly thinks it's an MLO and tried to load and execute it Jan 11 22:38:17 this is what it should preferably look like: https://pastebin.com/8F0a3yiR Jan 11 22:44:08 zmatt: here's my image: https://pastebin.com/YCvBVm8C Jan 11 22:45:45 yep, exactly what I guessed Jan 11 22:47:27 pretty sure that's an x86 MBR Jan 11 22:50:38 presumably whatever tool you used to create the partition table thought that was a good idea... fdisk doesn't do that Jan 11 22:56:54 you can clean this up e.g. with head -c $((0x1b0)) /dev/zero 1<>FILENAME **** ENDING LOGGING AT Thu Jan 12 02:59:56 2023