10:07 | Artox | good morning everyone |
10:07 | Artox | a bit offtopic, but is anyone aware of a simple algorithm for object recognition using a point cloud? |
10:11 | topi` | do I need to somehow "set up" a correct .dtb file before booting the kernel via u-boot's "boot" command? |
10:11 | topi` | bootm |
10:11 | Artox | yes |
10:11 | topi` | all I just did was ext4load mmc 0:2 0x10800000 /boot/uImage |
10:11 | topi` | maybe that's not enough |
10:11 | Artox | well, .dtb can be embedded in uImage |
10:11 | Artox | but, we are not doing it |
10:12 | topi` | if I run ext4ls, there are some .dtb's on that boot directory |
10:12 | topi` | I don't really remember exactly what I wrote on this SD, and I cannot get it to boot :) |
10:12 | Artox | maybe you will find this helpful: http://pastebin.com/kk6ETwRp |
10:13 | Artox | basically, check line 13 to 16 |
10:13 | Artox | and if tehre is a zImage, use it |
10:20 | topi` | grr. the parameter for ramdisk comes before that of fdt, and I have no ramdisk :/ |
10:20 | topi` | help bootm |
10:22 | topi` | I can't just give it a random addr, since it says "ramdisk image is corrupt or invalid" :( |
10:22 | topi` | OK I give up, and go back home to fetch the "right" SD card so that I can boot the board. |
10:22 | topi` | I wish I could boot from the mSATA... |
10:36 | jnettlet | topi`, why are you using bootm? |
10:36 | jnettlet | the proper way to boot any modern kernel is to use a zImage and dtb and use the bootz command. |
10:40 | jnettlet | ultimately you can boot whatever way suits you but using a uImage takes additional steps and is less flexible. |
11:00 | Artox | topi`: if you dont have ramdisk, the bootz command looks like this: bootz ${kerneladdr} - ${fdtaddr} |
11:00 | Artox | thats hard to get by accident, but easy when known |
11:00 | Artox | and I guess bootm takes same arguments |
11:01 | topi` | Artox: thanks :) |
11:02 | Artox | sorry it took me so long; i was off buying some tea |
11:05 | topi` | OK the kernel works :) now I just need to find a root somewhere... |
11:06 | topi` | I'm under-equipped in the way of SD card readers and such... |
11:06 | Artox | I see |
11:06 | Artox | those are damn useful |
11:06 | topi` | I think I just need to call it quits for the day |
11:06 | Artox | if you got a floppy drive slot, go put a cardreader in there :) |
11:06 | topi` | we moved to a different place and I need to unpack all my professional stuff at home |
11:06 | Artox | ah |
11:07 | topi` | and the corporate Lenovo laptop is useless - it has a SD reader, but since it runs windows, and Virtualbox cannot access SD readers (it's not on the USB bus), I cannot do much with it |
11:07 | topi` | unless Windows starts supporting ext4... |
11:07 | Artox | hmmmmmmmmmmmmmmm |
11:07 | Artox | not.happening :) |
11:07 | topi` | I requested a Macbook from my boss :) it'll arrive soon |
11:07 | Artox | aww :/ you should have gone with tuxedo |
11:07 | Artox | but a macbook will do ofc |
11:08 | topi` | at least with a Mac, I can create a suitable .vmdk file that I can mount on Virtualbox and I'll have the MMC reader in /dev/sdx |
11:08 | Artox | yeah |
11:08 | topi` | even though OSX itself won't support ext4 |
11:08 | topi` | the sad part about big corps is how they're so tied to Windows workstations |
11:09 | Artox | yes |
11:09 | topi` | all processes, etc, go through windows apps and windows firewall software and so on |
11:09 | Artox | though I guess its inevitable they will all switch to thinclient terminals |
11:09 | Artox | one day |
11:09 | topi` | most of the apps are already on servers, behind HTTP |
11:09 | topi` | but there's always that one (1) app that cannot be upgraded! |
11:10 | Artox | yes |
11:10 | bencoh | and The Domain ... ;) |
11:10 | topi` | especially if it's something the salespeople or the money handlers use... |
11:10 | Artox | thats, btw, a use-case of reactos that they wrote na article about a while ago |
11:10 | topi` | the closer it is to Money, the harder it is to migrate! |
11:10 | Artox | so it can be moved into a cheap vm one day |
11:10 | topi` | VMs aren't cheap, not even on AWS |
11:11 | Artox | if you run them on local hardware they are |
11:11 | topi` | we have a Medium AWS instance and it's already costing a lot |
11:11 | Artox | they jsut dont cost |
11:11 | Artox | qemu is free |
11:11 | topi` | true |
11:11 | topi` | I wish there'd be a Cortex-A15 (or A7) version of i.MX6 so that I could virtualize it |
11:11 | topi` | could run 4 small machines on a Quad core cpu :) |
11:11 | Artox | me too |
11:11 | topi` | there's already some working code for ARM virtualization in the mainline kernel |
11:12 | Artox | kvm *is* working on arm already |
11:12 | topi` | and I've heard it's decently fast, it beats Intel on the I/O handling speed |
11:12 | Artox | the opensuse project performs all arm builds inside kvm on arm |
11:12 | topi` | evidently when you virtualize an Intel, you need to go thru some loops to do proper I/O |
11:12 | bencoh | do we get proper iommu ? |
14:45 | ido_ | hey, can anyone recommend an IR remote that works with openelec+cubox out of the box? |
18:22 | topi` | I installed firmware-brcm80211 (from debian/jessie), is that supposed to work the the builtin wifi? |
18:23 | topi` | I also rmmod'ed brcmfmac and modprobed it again, but still nothing in /proc/net/dev |
18:23 | topi` | [ 743.418331] brcmfmac: brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl 0x50 |
18:23 | topi` | this is what I get in dmesg... |
18:26 | topi` | funny thing - lsusb doesn't list any wifi devices at all! |
18:26 | topi` | however, I ordered the damn SoM with wifi+bt |
18:26 | topi` | I only got this (aside of the two linux foundation hubs): Bus 002 Device 002: ID 04b4:6570 Cypress Semiconductor Corp. |
19:21 | bencoh | topi`: do you have the nvram config file ? |
19:22 | bencoh | (oh and, it's a sdio wifi/bt chip, not a usb one) |
19:54 | topi` | bencoh: no, how do you config the nvram? |
19:54 | topi` | I just installed the 'firmware-brcm80211' debian package |
19:54 | topi` | nothing called "nvram" there |
19:55 | topi` | I'm also thinking that maybe I'm missing some bits from the .dtb file |
19:55 | topi` | and that's why it's impossible to detect the BCM4330 |
19:55 | topi` | (how do I know if it's a BCM4329 or BCM4330?) |
20:05 | bencoh | topi`: there's a link on the forum |
20:09 | topi` | hmm, ok |
20:09 | topi` | i'll try to search |
20:18 | bencoh | topi`: http://www.solid-run.com/community/viewtopic.php?f=10&t=766&hilit=wifi+3.10&start=20#p5966 |
20:19 | topi` | woohoo! |
20:19 | topi` | wlan0 exists!! |
20:19 | topi` | now, I'll try to force that to master mode. let's see... |
20:20 | bencoh | I couldnt set it to master mode, please tell if you succeed :) |
20:20 | bencoh | (I think we have no support for it) |
20:20 | topi` | damn, then that's a real show stoopped |
20:20 | topi` | stopper |
20:21 | bencoh | you can still use adhoc (ibss) mode |
20:21 | bencoh | but most android systems wont like it |
20:21 | topi` | yeah, that's a real problem |
20:21 | topi` | we're trying to develop a wireless router |
20:21 | topi` | so it's a real requirement |
20:21 | bencoh | anyway I only tried ap mode with iwconfig, not with iw/hostap, so ... |
20:22 | topi` | there are wifi cards as mPCIe, but then there's only one mPCIe slot in hummingboard... :( |
20:22 | topi` | yeah, hostapd would be my next step |
20:22 | bencoh | considering that the driver has different behavior when used with (old) iwconfig and iw ... |
20:23 | topi` | there's a comparison list of open source wireless drivers |
20:23 | topi` | and it lists brcmfmac NO master(AP) mode |
20:23 | topi` | brcmsmac does support... |
20:23 | topi` | it's either a driver or firmware issue |
20:24 | topi` | it seems b43 driver supports master mode, but does b43 work with the chip in hummingboard? |
20:27 | topi` | also, the FreeBSD bwi driver does not support master mode... so there's a pattern here |
20:29 | topi` | even the BCM4330 data sheet does not mention master mode in any way |