| 01:34 | rabeeh | SPL support added - http://imx.solid-run.com/forums/viewtopic.php?f=2&t=333&start=10#p1885 |
| 01:34 | rabeeh | finally - no need for 4 images |
| 01:59 | Thecko | hi, anyone here? |
| 02:22 | mhoney_work | hello |
| 06:51 | _dab_ | rabeeh does ping on your uboot work for you? It fails for me with the uboot at http://dl.dropboxusercontent.com/u/72661517/tbr/spl/u-boot.img |
| 06:51 | _dab_ | I am trying to get an nfs boot to work |
| 06:55 | _dab_ | jnettlet? |
| 10:09 | MikeSeth | rabeeh: oh awesome |
| 11:33 | xraxor | great stuff about the SPL |
| 11:38 | rabeeh | xraxor: yes. this will save uploading four different images |
| 11:38 | rabeeh | there is a way also to detect if it's HB or CuBox-i; so afterall we will have a single u-boot for both families of machines :) |
| 11:41 | hste | rabeeh: so the ram settings for higher speed etc will now be ok? |
| 12:11 | _rmk_ | well, I guess I now need to sort out a DT file for the cubox-i |
| 12:24 | dv_ | has anyone tried out the imx6 wayland support yet? |
| 12:32 | rabeeh | dv_: i'm not familiar with anyone that tried it |
| 12:36 | xraxor | sounds good rabeeh. |
| 12:43 | xraxor | man this internet has got some issues :( |
| 12:50 | _rmk_ | xraxor: maybe freenode is under attack yet again |
| 12:50 | _rmk_ | it seems to be a regular thing |
| 12:52 | _rmk_ | I once asked one of the freenode staff what the nature of an attack they announced was... the reply was along the lines of "we don't know" |
| 12:55 | rabeeh | _rmk_: i was wondering that too |
| 12:55 | rabeeh | i mean who would even want to attack freenode; besides making some noise |
| 12:57 | _rmk_ | attacking irc networks is unfortunately one of those things that has been going on for some time... some people derive fun from depriving others of stuff |
| 13:00 | xraxor | shit |
| 13:01 | xraxor | _rmk_: thats worrying, why would they hack? what can they get? |
| 13:02 | _rmk_ | xraxor: try a different server - I'm on leguin.freenode.net |
| 13:02 | _rmk_ | xraxor: from what I've seen it seems to be a case of trying to overload a server to cause it to split off from the rest of the network |
| 13:05 | _rmk_ | as for what can they get - irc has had a history of issues with sorting out netsplits - by that I mean sorting out what should be the correct channel modes. |
| 13:06 | _rmk_ | for instance, if you get stuff just right, it used to be possible to gain channel operator status by causing a netsplit... |
| 13:06 | _rmk_ | you could also do evil stuff like cause a netsplit, change your nick, when the server rejoins, you might get your target killed off the network |
| 13:06 | rabeeh | _rmk_: this is something worth an attack then. |
| 13:08 | _rmk_ | yea. there's been many attempts to plug these holes in the inter-server protocols, but the problem is far from trivial |
| 13:11 | xraxor | thanks will do the same as you |
| 13:13 | xraxor | ok seems like im in leguin |
| 13:14 | _rmk_ | the hybrid 8 and ratbox ircds get around some of this by assigning users a "UID" when they connect - the first three characters are server specific. These remain constant for the duration that the client is connected, so nick changes and netsplits have no effect on that. Inter-server messages use the UIDs as targets rather than the nicknames, the messages only get converted to nicknames when being sent to a non-UID supporting server or clients. |
| 13:18 | _rmk_ | the channel stuff is slightly more complex - at the simple level, a channel has a network wide timestamp associated with it. when two servers split and rejoin, the timestamps are compared, and the server with the earliest timestamp wins the "election" and gets to impose the channel modes. |
| 13:19 | _rmk_ | so... if you can trick a server into creating an already present channel while split, but with an earlier timestamp (remember, first nick in a new channel gets chanops), it wins the election... |
| 13:20 | MikeSeth | most ircds have protection against collision attacks, though |
| 13:20 | MikeSeth | and in general ircds with first class services do not have this problem |
| 13:20 | MikeSeth | on efnet if all your bots lose ops and you don't know an ircop that knows you.. it's a nightmare |
| 13:21 | MikeSeth | otherwise, DDoS is about the only attack modern ircds are vulnerable to |
| 13:21 | MikeSeth | at least if you exclude dns poisoning and theft of ssl keys |
| 13:23 | _rmk_ | ... but only if they're running something like the TS6 protocol between servers which fixes a set of netsplit/rejoin issues |
| 15:48 | pepedog | Cooee |
| 15:49 | pepedog | This new SPL uboot, what toolchain is being used? |
| 15:51 | pepedog | Native I get complaint that at LD vfp is used, chile SPL doesn't |
| 17:03 | rabeeh | pepedog: tomlohave-away had the same issue |
| 17:03 | rabeeh | oh; pepedog left |
| 17:04 | rabeeh | anyhow; the SPL code uses division '/' which breaks since it uses some libgcc.a that requires multilib |
| 17:04 | rabeeh | i'm trying to replace those division with do_div() |
| 17:05 | rabeeh | dv_ / tomlohave-away: what is the best place to gather the gpu binaries from? |
| 17:05 | jnettlet | rabeeh, oh I have a patch for that. rather than using do_div you can just change one of the u64's to u32 |
| 17:05 | rabeeh | jnettlet: great |
| 17:05 | rabeeh | btw - please look at my git |
| 17:06 | rabeeh | https://github.com/rabeeh/u-boot-imx6 |
| 17:06 | rabeeh | ok to pull request? |
| 17:06 | rabeeh | shortly i will be working on auto detection board and then combine HB / CBi together |
| 17:07 | jnettlet | rabeeh, I was reviewing that earlier. I need to break up my branches as I am currently working in the 01/14 HEAD, trying to keep us up to date. |
| 17:07 | rabeeh | the SPL support is mainly two big patches gathered from Wandboard u-boot - |
| 17:07 | rabeeh | https://github.com/rabeeh/u-boot-imx6/commit/080f0b65edba894e99af416249da53ed048bb31c |
| 17:07 | rabeeh | https://github.com/rabeeh/u-boot-imx6/commit/d3ca03b5ea9f5dc850bfbba77ca9e06eae048ab2 |
| 17:08 | jnettlet | rabeeh, take a look at this patch and see if you agree. https://dl.dropboxusercontent.com/u/736509/0001-mx6-fix-build-due-to-divide-u64-u32.patch |
| 17:11 | rabeeh | did you test? |
| 17:12 | rabeeh | there were two places with the division issue (tomlohave-away reported) |
| 17:13 | jnettlet | yep that has been working for me. The other place I think fixed itself once I fixed this function |
| 17:15 | rabeeh | ok; works for me too |
| 17:16 | rabeeh | btw - with u-boot.img on a fat partition |
| 17:16 | rabeeh | so that one works too |
| 17:21 | tomlohave-away | rabeeh: for gpu binarie, we use the same as the Yocto project. |
| 19:46 | rabeeh | https://github.com/rabeeh/u-boot-imx6/commit/920ea0f20276614e11961924b1990b6c922c8d0e |
| 19:47 | rabeeh | code that auto detect HB / CBi and accordingly sets machine ID (and enables / doesn't enable the front LED) |
| 20:08 | zecrazytux | hi |
| 20:43 | zecrazytux | is there any document that shows whats mainline and whats not ? |
| 21:48 | pixelad0 | Hello!! |
| 21:49 | pixelad0 | i nedd help |
| 21:49 | pixelad0 | i can't boot my cubox-i 4 |
| 21:50 | pixelad0 | I followed the steps in the wiki |
| 21:50 | pixelad0 | http://cubox-i.com/install-os-on-micro-sd-flash-card/ |
| 21:51 | pixelad0 | download GeexBox XBMC and copy to microSD |
| 21:54 | pixelead0 | can someone help me? |
| 21:54 | cbxbiker61 | pixelad0, in my testing the carrier-one was sensitive to the brand of sd card, sandisk works well |
| 21:56 | pixelead0 | ok |
| 21:57 | pixelead0 | I'll try a scandisk memory |
| 21:57 | pixelead0 | thnks |
| 21:58 | pixelead0 | I have to make an additional partition or something similar? |
| 21:58 | cbxbiker61 | no |
| 21:58 | pixelead0 | o ?nicamente con copiar la imagen a la memoria? |
| 21:58 | pixelead0 | or just copy the image to the memory? |
| 21:59 | cbxbiker61 | sd card, not memory |
| 22:00 | pixelead0 | I'll try, thank you very much. |