| 01:09 | Coburn | Well, looks like CuBox is the platform to be |
| 01:10 | Coburn | it may not have 4 core CPU, but at least it's the "open" platform. the ODROID-U2 can't boot from usb due to "uboot limitations" |
| 01:10 | Coburn | and additionally, the bootloader is written into unallocated space on the uSD. |
| 01:10 | Coburn | So hose your uSD, you have a brick, and you need to reflash uboot into the "special" part of the uSD |
| 01:12 | Coburn | Kinda silly if you ask me, would be better with a little NOR/NAND chip that covers the uBoot side of things, but hey, for a competitive price tag, some things had to be left on the cutting room floor |
| 01:12 | rmull | Coburn: The cubox is not an "open" platform, FYI. Don't trick people into thinking they can get schematics and so forth. |
| 01:15 | Coburn | rmull: I mean, it's more open than some |
| 01:15 | Coburn | I should have justified my reasoning |
| 01:16 | Coburn | I have no intention to do hardware mods on the CuBox |
| 01:16 | Coburn | I'm a software guy. |
| 02:10 | dbsx | software wise the cubox is mainly open, except for some marvell drivers. (Respecting marvells proprietary rights) They dont have to be open source, but they could at least publish the code. Marvells contraints on driver code (e.g. SD8787) is a platform killer. |
| 02:27 | Coburn | dbsx: agreed. |
| 02:27 | Coburn | I guess it also applies for other platforms too |
| 02:27 | Coburn | well, arm platforms that is, people gotta protect their secrets |
| 02:28 | Coburn | although I'm sure we can all agree that marvell sucks at releasing code so that we can use it in later kernels |
| 02:29 | Coburn | might also try my own 3.8.0rc4 kernel later today |
| 02:29 | Coburn | btw dbsx |
| 02:29 | Coburn | it seems ODROID also suffered the wifi bug |
| 02:30 | Coburn | I said to them to disable 80211_DEFAULT_PS in kernel |
| 02:30 | Coburn | they asked me "how did you find that out" |
| 02:31 | Coburn | I replied, "I suffered the same problem with another ARM device" |
| 02:31 | Coburn | ;) |
| 03:13 | dbsx | yep it is a BAD kernel default |
| 08:01 | jnettlet | Coburn, what is the bug with 80211_DEFAULT_PS? |
| 08:06 | Coburn | jnettlet: It's powersaving by default |
| 08:06 | Coburn | and as dbsx said, it makes stuff go AWOL |
| 08:06 | Coburn | adapter misses packets and basically takes a nap |
| 08:10 | jnettlet | That is with the Marvel SD8787 chip? |
| 08:43 | Coburn | jnettlet: no |
| 08:43 | Coburn | that setting means ANY adapter |
| 08:44 | Coburn | marvell or not |
| 08:44 | Coburn | realtek, atheos, ralink... etc |
| 10:12 | dbsx | jnettlet: I have tried it with 3 different adaptors and result is not good. You can also disable power saving for a wifi adaptor with iwconfig or iw. |
| 10:14 | jnettlet | dbsx, performance is not good? |
| 11:40 | dbsx | using ssh, if you have not touched the keyboard for 2 secs, it takes 2-3 secs to wake up |