| 00:04.55 | WarheadsSE | Turl arokux_h hno busy day in here |
| 00:05.18 | WarheadsSE | ccssnet: hah. 4 hours of server + modem & router on a 825VA pack |
| 00:05.31 | WarheadsSE | runs too much arm |
| 00:05.37 | ccssnet | hmm |
| 00:07.33 | hno | WarheadsSE ? |
| 00:08.07 | ccssnet | hno: see above where i talked about my router uptime compared to my modem on battery |
| 00:10.00 | ccssnet | WarheadsSE: i am using raw 12v dc out though for battery backup. not using a AC battery backup that uses VA ratings |
| 00:10.26 | hno | arokux_h, that's the hack that hardcodes machid & atags location ignoring what the bootloader says. |
| 00:10.27 | ccssnet | 7500mAh battery |
| 00:11.17 | arokux_h | hno, will you add it to bootloader? |
| 00:11.40 | lundman | ok back at work, so what did I miss |
| 00:11.41 | hno | the bootloader do set machid and atags location. This hack is just stupid. |
| 00:12.06 | arokux_h | hno, so there is no need for it? |
| 00:12.35 | hno | only needed to support loading kernel directly from allwinner bootloader without using a 2 line glue to set up registers correctly before starting kernel. |
| 00:13.52 | hno | it's imho an ugly hack that should die a painful death. Those that need it need to solve it elsewhere (i.e. by using a 2 line wrapper to start the kernel) |
| 00:14.43 | hno | that hack is one of the things that can never even be shown to upstream. |
| 00:15.01 | hno | see comment above on what may be done here. |
| 00:15.10 | arokux_h | hno, the comment at the beginning of head.S says no machine crap should be added here, so one should follow the advice :) |
| 00:15.21 | arokux_h | :D |
| 00:15.29 | hno | yes. |
| 00:15.30 | arokux_h | yes, was thinking the same. |
| 00:15.49 | arokux_h | hno, so do you have this 2 line glue already? |
| 00:16.58 | hno | It's just to cut those lines out into a separate source file, and add a jump to the kernel location at the end. The nassemble it separately and ask the bootloader to load that one as first image and kernel as second image. |
| 00:17.44 | hno | the allwinner bootloader can load any number of images into memory, and jump to the first. |
| 00:18.02 | arokux_h | hno, I thought its U-Boot bootloader? |
| 00:18.03 | hno | those using u-boot do not need to care about this. |
| 00:18.30 | hno | as u-boot do set machid and atags per kernel boot specifications. |
| 00:19.20 | hno | that hack only exists for those who insist on using allwinner original bootloader without u-boot. |
| 00:20.35 | arokux_h | hno, didn't know that such a think as aw-bootloader exist. |
| 00:21.29 | ccssnet | arokux_h: a few companies fork u-boot out, wondermedia has one also |
| 00:21.48 | arokux_h | it's not a fork as I understood..? |
| 00:22.06 | ccssnet | WonderMedia Technologies, Inc. |
| 00:22.08 | ccssnet | W-Load Version : 0.19.00.03 |
| 00:22.21 | arokux_h | hno, I'm using http://hands.com/~lkcl/mele_uboot.img so I could safely delete this hack? |
| 00:22.25 | ccssnet | hmph. no idea |
| 00:22.48 | arokux_h | hno, just not sure what is really aw-bl |
| 00:23.33 | Turl | hno: "allwinner bootloader" is uboot too |
| 00:23.51 | hno | arokux_h, allwinner do have their own proprietary bootloader, Have nothing to do with u-boot. |
| 00:24.18 | arokux_h | hno, I see. so with mele_uboot.img I'm safe to delete the hack? |
| 00:24.47 | hno | I honestly have no idea what that u-boot img contains. |
| 00:25.40 | arokux_h | hno, this answer is already enough. I then should compile and run an up to day bootloader, so I know *what* it contains |
| 00:26.24 | hno | Allwinner had not requested a machid when sources was originally released. |
| 00:27.36 | hno | Turl, allwinner bootloader is boot0 + boot1+ boot.axf, plus a bit more. |
| 00:28.20 | arokux_h | hno, "AW had not requested a machid" <--- what does it mean? |
| 00:28.52 | hno | arokux_h, that there is a lot of confusion on what machid to use in older sources. |
| 00:30.22 | hno | they did not have a assigned number (getting one is one email away) so they just picked one, which is not the same they later got assigned. |
| 00:31.16 | Turl | hno: that chainloads an allwinner uboot |
| 00:32.35 | hno | Turl, it does not need to chainload u-boot. It can boot anything including Linux. |
| 00:33.17 | Turl | yeah, I know, but they chainload their own uboot on most if not all cases |
| 00:33.23 | Turl | to boot boot.imgs |
| 00:33.24 | hno | yes. |
| 00:33.59 | hno | I said the override for supporting the allwinner bootloader is an ugly hack that deserves to die a painful death. |
| 00:34.26 | hno | it's all wrong, and not even needed for loading the kernel from the allwinner bootloader. |
| 00:35.14 | ccssnet | hno is correct, chainloading one bootloader from another is pointless |
| 00:43.17 | Marex | catches fire |
| 00:43.29 | Marex | hno: all machineID must die altogether! |
| 00:43.41 | Marex | device tree is the one and only rightous way to go |
| 00:43.44 | Marex | (TM) |
| 00:48.52 | arokux_h | hno, are you sure the instructions on your wiki are correct? in section "How to make a bootable sunxi SD-card" it says u-boot.bin should be dd'ed, but there is no u-boot.bin, there is u-boot-spl and u-boot-spl.bin after compilation. should u-boot-spl.bin be used? |
| 01:04.21 | *** join/#arm-netbook hipboi (~hipboi@221.4.213.95) |
| 01:11.08 | hipboi | arokux_h: SW prefix stands for softwinner |
| 01:11.24 | hipboi | I have complained that misleading prefix |
| 01:11.33 | hipboi | many times |
| 01:11.43 | arokux_h | thanks, so hno was right. |
| 01:12.04 | arokux_h | hipboi, how is cubieboard? |
| 01:12.24 | hipboi | you need to ask cubieboard |
| 01:12.31 | hipboi | is him here? |
| 01:13.39 | *** join/#arm-netbook cubieboard (dd04d55f@gateway/web/freenode/ip.221.4.213.95) |
| 01:13.52 | lundman | yes |
| 01:14.05 | cubieboard | hi |
| 01:14.50 | Turl | cubieboard: are you hipboi? :) |
| 01:14.59 | arokux_h | actually I was asking about http://cubieboard.org/ |
| 01:15.48 | cubieboard | arokux_h: cubieboard will be available next week |
| 01:16.44 | Turl | IP gives you away, just saying :) |
| 01:16.53 | arokux_h | want somebody to buy Mele A1000? :) |
| 01:16.57 | Turl | cubieboard: nice, do you have an updated picture of it? |
| 01:17.02 | arokux_h | almost unused! |
| 01:17.15 | Turl | arokux_h: it makes a nice server if you ask me :) |
| 01:17.31 | arokux_h | Turl, want it? |
| 01:18.22 | Turl | arokux_h: I already have a A2k |
| 01:21.28 | *** join/#arm-netbook drynish (~drynish@modemcable233.94-177-173.mc.videotron.ca) |
| 01:21.34 | arokux_h | cubieboard, is cubieboard enough for kernel hacking? |
| 01:21.44 | hno | arokux_h, there is u-boot.bin in top level u-boot directory. |
| 01:22.34 | hno | arokux_h, and spl/sun4i-spl.bin is the spl. You need both. |
| 01:22.46 | arokux_h | hno, indeed. was looking in spl only. |
| 01:22.48 | arokux_h | thanks |
| 01:23.41 | *** join/#arm-netbook cubieboard (dd04d55f@gateway/web/freenode/ip.221.4.213.95) |
| 01:24.08 | arokux_h | cubieboard, cubieboard seems to have everything that mele a1000 has |
| 01:24.20 | cubieboard | arokux_h: yes, it's enough for kernel hacking |
| 01:24.44 | cubieboard | and also has what mele doesn't have |
| 01:25.06 | cubieboard | like user gpio pins |
| 01:25.42 | arokux_h | cubieboard, so it is even better then mele?! |
| 01:25.47 | cubieboard | like some sensor boards in the future |
| 01:26.11 | cubieboard | no case for cubieboard currently |
| 01:26.28 | arokux_h | cubieboard, I do not need one. I use mele without a case |
| 01:26.30 | arokux_h | :) |
| 01:26.47 | arokux_h | and mele has only 300Mb of RAM! |
| 01:27.06 | hno | arokux_h, for hacking cubieboard is nicer. |
| 01:27.07 | Turl | arokux_h: eh? mele has 512 |
| 01:27.27 | hno | yes it has 512, of which ~300 is available in normal usage. |
| 01:27.48 | Turl | not really |
| 01:28.08 | Turl | emilio@laptop:~$ ssh mele free |
| 01:28.08 | Turl | <PROTECTED> |
| 01:28.09 | Turl | Mem: 510956 469192 41764 0 31972 407236 |
| 01:28.12 | Turl | :) |
| 01:28.19 | hno | I said normal usage. |
| 01:28.30 | Turl | you mean android? |
| 01:28.49 | hno | I mean with framebuffer, gpu and video decoder engine enabled. |
| 01:29.10 | arokux_h | cubieboard will have 1Gb. |
| 01:29.13 | Turl | well that makes the cubieboard have ~800M if you want to compare then |
| 01:29.21 | hno | yes |
| 01:29.37 | arokux_h | hm.. I need to find somebody to sell my mele to :$ |
| 01:31.33 | hno | cubieboard do not have PB JTAG, right? Or is those pins available on GPIO headers? |
| 01:32.05 | arokux_h | cubieboard, also what about wifi and the size of nand? |
| 01:32.09 | *** join/#arm-netbook cubieboard_ (dd04d55f@gateway/web/freenode/ip.221.4.213.95) |
| 01:32.57 | cubieboard_ | hno: no jtag pins |
| 01:33.07 | cubieboard_ | hno: jtag on pf |
| 01:33.34 | cubieboard_ | arokux_h: no wifi |
| 01:33.48 | cubieboard_ | arokux_h: no nand by default |
| 01:34.08 | cubieboard_ | arokux_h: two mmc slot |
| 01:34.25 | cubieboard_ | one for os, one for storage |
| 01:34.53 | *** join/#arm-netbook Boulet (~sid6581@184.22.183.18) |
| 01:35.09 | lundman | needs 5-6 sata for raid :) |
| 01:35.35 | hno | lundman, you won't find that in a tablet oriented soc. |
| 01:37.24 | arokux_h | pf? |
| 01:37.35 | arokux_h | (jtag on pf) |
| 01:38.25 | cubieboard_ | arokux_h: mux with sdcard pins |
| 01:42.18 | arokux_h | good night guys. |
| 01:43.32 | hno | cubieboard_, which pins will be available on the gpio extension headers? |
| 01:44.51 | cubieboard_ | hno: will be available as a diagram |
| 01:45.21 | cubieboard_ | hno: besides the schematic will be pulic |
| 01:45.50 | hno | nice! |
| 01:46.09 | cubieboard_ | not sure what's the definition of "open hardware" |
| 01:46.47 | cubieboard_ | so we don't regard cubieboard as open hardware now |
| 01:47.03 | cubieboard_ | but will open what we can open |
| 01:49.21 | hno | open hardware requires schematics and preferably PCB design as well to be published under a license that allows reuse and customization. |
| 02:13.02 | cubieboard_ | another concept is copyleft hardware |
| 02:13.06 | cubieboard_ | interesting |
| 02:21.22 | *** join/#arm-netbook tuliom (~tuliom@186.214.90.201) |
| 02:43.49 | *** join/#arm-netbook hipboi (~hipboi@221.4.213.95) |
| 02:46.50 | drynish | is it the good place to talk for the Mele? |
| 02:48.01 | Marex | yes |
| 03:05.18 | hipboi | drynish: here is even a mele tv box seller |
| 03:39.40 | Turl | hipboi: the newer mali libs have been working perfectly so far :) |
| 03:39.58 | hipboi | Turl: good : ) |
| 03:42.11 | Turl | hipboi: the "Spirits" game from Humble Bundle works really nice on A10+Mali https://www.youtube.com/watch?v=xsiwROHN35c&feature=player_embedded |
| 03:54.47 | lundman | crikey, allwinner took a bit of a beating from xbmc i see |
| 04:11.35 | furan | as they should have |
| 04:16.22 | furan | hipboi why is allwinner giving the xbmc guys the runaround anyway |
| 04:17.18 | hipboi | furan: i am not sure |
| 04:17.44 | furan | hey I figured out how the nand table entries are defined :) |
| 04:17.57 | hipboi | furan: congraduations |
| 04:18.01 | furan | upgraded a device to 64gb nand |
| 04:18.06 | hipboi | can you add it to the wiki? |
| 04:18.13 | furan | working on it |
| 04:18.13 | hipboi | really great |
| 04:18.27 | L84Supper | anyone know which PCB layout software was used for the Cubieboard layout? |
| 04:19.03 | hipboi | about xbmc apparently the market of allwinner doesn't have a good communication with the software |
| 04:19.10 | Turl | http://forum.xbmc.org/showthread.php?tid=126995&pid=1171315#pid1171315 :x |
| 04:19.22 | Turl | L84Supper: pretty sure hipboi knows |
| 04:19.54 | furan | oh I see. so the xbmc guys are mainly talking to marketing |
| 04:20.35 | hipboi | furan: maybe one cedarx lib developer |
| 04:20.40 | hipboi | i am not sure |
| 04:20.57 | hipboi | but the fact is in allwinner nobody want to do extra job |
| 04:21.26 | hipboi | if something can work, they don't want to improve it, update it |
| 04:21.40 | L84Supper | the reference PCB from alwinner used an ancient version of PADS IIRC |
| 04:22.28 | hipboi | allwinner is very product oriented |
| 04:22.28 | furan | yeah. xbmc will sell more allwinner cpus, which will result in more orders. but I am not sure that is evident enough, or if allwinner maybe already has too many customers right now |
| 04:22.48 | hipboi | furan: yes |
| 04:23.08 | hipboi | xbmc will sell more chips, but they can't see like |
| 04:23.12 | L84Supper | too many customers? FAB problems? |
| 04:23.18 | hipboi | they can't see it |
| 04:23.26 | hipboi | they only see orders |
| 04:23.26 | furan | makes sense |
| 04:23.34 | hipboi | how many orders do you have? |
| 04:23.55 | furan | someone should make a petition where people sign it saying they'll buy an allwinner device if allwinner helps the xbmc devs |
| 04:23.57 | hipboi | if you have enough like 100k, you can get support |
| 04:24.04 | furan | then they could see that they will at least get a percentage of that |
| 04:24.21 | furan | don't think it would be 100k though |
| 04:28.43 | Turl | OMX will be less maintenance, no merging and fixing weird player code into android |
| 04:29.06 | Turl | I hope they actually see that one and it's not marketing speak from Eva |
| 04:29.20 | Turl | even if its only for sun6i or w/e |
| 04:30.13 | furan | what is OMX? |
| 04:30.20 | Turl | OpenMAX |
| 04:30.31 | furan | oic |
| 04:30.42 | Turl | furan: https://www.khronos.org/openmax/ |
| 04:31.02 | Turl | from the same people doing OpenGL, OpenCL etc |
| 04:31.07 | furan | nods |
| 04:32.32 | hipboi | Turl: the android source code of allwinner breaks other platforms |
| 04:34.06 | Turl | yeah it needs a lot of #ifdef hipboi |
| 04:34.23 | hipboi | they don't like to change to fit the standard |
| 04:34.41 | Turl | if they used OMX on the other hand, they wouldn't even need to touch the android framework |
| 04:34.47 | L84Supper | http://www.hardkernel.com/renewal_2011/products/prdt_info.php?g_code=G133999328931 has anyone gotten their boards yet? |
| 04:34.55 | furan | guessing marketing sells current product and dev is completely focused on new product |
| 04:35.06 | furan | e.g. no dev support really |
| 04:35.20 | Turl | L84Supper: I think I heard of some people getting it |
| 04:35.29 | Turl | not sure tho :| memory is crap these days haha |
| 04:36.06 | L84Supper | Turl: just wondering if the boards actually work or are still buggy |
| 04:36.25 | Turl | L84Supper: why would they be buggy? |
| 04:36.34 | furan | has dev boards still in boxes |
| 04:36.49 | L84Supper | bad silicon, poor board design, etc etc |
| 04:37.02 | L84Supper | just like the imx6 dev boards |
| 04:37.16 | furan | fell off the back of the silicon truck |
| 04:37.45 | Turl | L84Supper: http://com.odroid.com/sigong/blog/blog_list.php?bid=126 |
| 04:37.51 | Turl | looks professional :P |
| 04:38.00 | furan | is there a good photo of the back an allwinner chip that charts out the bga pins? |
| 04:38.19 | Boulet | furan, there is a datasheet for that, no ? :) |
| 04:38.21 | furan | I want to change the voltage input for one of the usb ports from 3.3v to 5v |
| 04:38.30 | furan | there is but it does not have the pretty picture I am asking for |
| 04:38.50 | Boulet | a picture would just show you a lot of dots hahahaha |
| 04:38.53 | L84Supper | there's an A10 shortform data sheet floating around with pinouts |
| 04:39.04 | furan | Boulet: I stare at bga pictures of fpgas all the time |
| 04:39.11 | furan | so, no, it wouldn't |
| 04:39.39 | Boulet | furan, i am curious what you would see other than the balls themselves ? |
| 04:40.05 | Boulet | i have some A13 chips with me, but no A10 anyways |
| 04:40.46 | furan | I am talking about a picture that shows pin names instead of balls |
| 04:41.54 | Boulet | ah, yeah i thought the datasheet would do... |
| 04:42.33 | Boulet | they have all the pins in 2 pages |
| 04:42.42 | Boulet | 4 quadrants |
| 04:43.16 | furan | yeah hunting through it :) |
| 04:43.24 | Boulet | ah good :) |
| 04:43.36 | furan | I will probe the board to see if I've got it sorted |
| 04:43.37 | furan | thanks |
| 04:45.36 | Boulet | probably not too easy to do mods, you have good soldering skills i suppose ! :o) |
| 04:47.20 | furan | I have good tools, a fancy microscope with a camera and fine iron |
| 04:48.16 | Boulet | nice |
| 04:48.33 | Boulet | i am considering to buy one of those microscope with camera and display |
| 06:27.12 | furan | they're not too costly. I got a stereo boom one with a camera port, and then got a 11? megapixel camera that has a little tiltable lcd attached |
| 06:27.22 | furan | so I can just stare at the lcd above the scope while I solder |
| 06:47.32 | *** join/#arm-netbook sspiff (~quassel@94.103.155.103) |
| 06:47.32 | *** join/#arm-netbook sspiff (~quassel@unaffiliated/yukito) |
| 06:58.39 | *** join/#arm-netbook rellla (~chatzilla@p5B0792FE.dip0.t-ipconnect.de) |
| 07:16.14 | *** join/#arm-netbook popolon (~popolon@2a01:e35:2e1e:a6c0:992:840f:c6d0:7bd1) |
| 07:38.22 | *** join/#arm-netbook Quarx (~Quarx@178.74.72.15) |
| 07:39.42 | *** join/#arm-netbook popolon (~popolon@2a01:e35:2e1e:a6c0:2926:6d2:a57b:6917) |
| 09:06.52 | *** join/#arm-netbook tavish (~tavish@unaffiliated/tavish) |
| 09:09.23 | *** join/#arm-netbook Quarx (~Quarx@46.233.248.191) |
| 09:42.43 | *** join/#arm-netbook Vayu (~yundin@fw2.nbi.dk) |
| 10:42.40 | Boulet | furan, seems very convenient indeed |
| 10:53.00 | *** join/#arm-netbook sspiff (~quassel@unaffiliated/yukito) |
| 11:33.26 | *** join/#arm-netbook TomNL (~TomNL@82-169-150-89.ip.telfort.nl) |
| 11:34.20 | TomNL | hey guys, anyone familiar with android_external_cedarx |
| 11:35.46 | TomNL | CedarXPlayer.cpp -> this is a basic player interface for cedarx? |
| 11:55.40 | *** join/#arm-netbook hipboi (~hipboi@123.64.148.23) |
| 12:19.10 | *** join/#arm-netbook sspiff (~quassel@94.103.155.103) |
| 12:19.10 | *** join/#arm-netbook sspiff (~quassel@unaffiliated/yukito) |
| 12:58.20 | *** join/#arm-netbook furan (ian@timbits.ript.net) |
| 12:59.10 | *** join/#arm-netbook CIA-16 (cia@198.71.88.9) |
| 13:20.11 | rz2k | https://github.com/allwinnerwk is "official" android repo from allwinner, right? |
| 13:27.40 | WarheadsSE | wow that account is a flipping mess. |
| 13:28.49 | rz2k | they've just pushed jb-exdroid |
| 13:29.08 | *** join/#arm-netbook uwe_ (~uwe_@dslb-088-064-209-125.pools.arcor-ip.net) |
| 13:29.17 | rz2k | exdroid-ics was ly-f1 android in sources, so... |
| 13:29.23 | WarheadsSE | ah |
| 13:29.56 | WarheadsSE | looks like they are pushing a lot atm |
| 13:30.00 | rz2k | i remember sztupy from XDA did normal kitchen and mirrored all that from allwinnerwk |
| 13:30.26 | rz2k | casts Turl as local android guy |
| 13:30.39 | WarheadsSE | looks like they are pushing JB to all the major branches? |
| 13:30.49 | rz2k | yeah |
| 13:31.03 | rz2k | I'm subscribed to that and I have 3 pages of mess at github, lol. |
| 13:31.09 | WarheadsSE | lol |
| 13:31.17 | WarheadsSE | cait wait for that notification email.. |
| 13:32.47 | specing | Wait, is allwinner opening up? Unprecedented. |
| 13:33.56 | Mazon | I'm confused too |
| 13:34.53 | rz2k | allwinnerwk came up when 3.0.8 leaked I think |
| 13:36.27 | specing | What is the "wk"? |
| 13:37.17 | rz2k | dont know, maybe hipboi knows? |
| 13:37.21 | hipboi | wk |
| 13:37.38 | hipboi | the name of one market guy of allwinner |
| 13:37.54 | hipboi | the pm of A10 |
| 13:38.38 | specing | pm? |
| 13:39.37 | rz2k | product manager? |
| 13:39.43 | specing | wait |
| 13:39.47 | specing | marketing |
| 13:39.51 | specing | released |
| 13:39.55 | specing | the source?! |
| 13:40.14 | specing | is confused |
| 13:40.17 | specing | <PROTECTED> |
| 13:40.21 | rz2k | they probably have pro-marketing guys |
| 13:40.32 | rz2k | :D |
| 13:41.09 | specing | Or maybe they finaly figured out why it is good to outsource work |
| 13:41.56 | hipboi | maybe he just made a mistake |
| 13:41.59 | hipboi | :) |
| 13:44.23 | TomNL | well if he made a mistake, we should clone fast? |
| 13:44.29 | Mazon | I'm hoping for the former ... would be a shame to disregard allwinner |
| 13:44.57 | Mazon | I'l already looking at amlogic |
| 13:45.03 | Mazon | seems to be the best bet |
| 13:45.15 | specing | Hehe, Im not sure it is possible to uncommit anything on github |
| 13:45.16 | Mazon | ignoring lukes warnings, ofc |
| 13:45.33 | Mazon | isn't that kinda the purpose of VCS ? |
| 13:45.46 | TomNL | well he could close down the account...right? |
| 13:45.55 | hipboi | clone fast first |
| 13:46.56 | specing | hipboi: :) |
| 13:47.05 | hipboi | about xbmc, my workmate Chen xiaochuan |
| 13:47.23 | hipboi | made a mistake about the #ifdef |
| 13:47.39 | hipboi | which leads the lib uncompiled |
| 13:47.50 | hipboi | so the checksum is the same |
| 13:48.05 | specing | prepares the gallows |
| 13:57.47 | ccssnet | that really is a horrible github account... |
| 14:00.07 | TomNL | so hipboi: they got newly compiled libs now |
| 14:00.40 | hipboi | i am not sure if the new lib is sent to gimli now |
| 14:01.10 | TomNL | ah ok, but there is still progress ongoing :) |
| 14:05.28 | TomNL | let's hope j1nx and gimli still have some energy left to work with allwinner ;-) |
| 14:07.08 | ccssnet | looking once at that github account... its enough to drain all the energy from the nile |
| 14:11.08 | phh | there is nothing interesting in that github account |
| 14:11.34 | phh | arf, all repos aren't shown |
| 14:13.18 | rz2k | https://github.com/allwinnerwk/repositories |
| 14:13.42 | phh | yeah well ctrl-f in that is not easy :p |
| 14:16.40 | L84Supper | http://www.logicpd.com/products/system-on-modules/dm3730-torpedo-wireless-som/ I forget if these came up here or not before. DM3730, 512MB, 15mm x 33 x 3.8 module |
| 14:17.31 | L84Supper | single core 1GHz cortex A8 |
| 14:19.49 | WarheadsSE | if it werent so expensive :p |
| 14:20.08 | WarheadsSE | ~ $300USD |
| 14:21.11 | L84Supper | yeah, no real winner yet |
| 14:22.04 | phh | two years later, best is still ac100 ... |
| 14:22.20 | L84Supper | http://www.digikey.com/product-detail/en/SOMDM3730-20-1780AGIR/460-3513-ND/2708829 |
| 14:22.25 | phh | wait what. |
| 14:22.40 | phh | ok i thought it was a netbook not a som -_-' |
| 14:23.20 | rm | haha |
| 14:23.25 | L84Supper | a netbook is easy enough to make but I've been busy with other things |
| 14:23.34 | rm | from the price it could've been a netbook yes |
| 14:23.44 | rm | an expensive one too |
| 14:24.06 | L84Supper | maybe a kickstarter for an ARM netbook done right....hhmmm ? |
| 14:25.33 | L84Supper | I really don't know if there is a market for more than a few K's of developers that would eve care |
| 14:26.14 | RITRedbeard | i want to start a kickstarter to get my college education finished |
| 14:26.21 | L84Supper | nobody seems to mind having a closed ecosystem in exchange for bragging rights for a status symbol mostly used to waste time |
| 14:27.17 | L84Supper | RITRedbeard, it's worth a shot, that NJ grandmother raised >$500K for a vacation |
| 14:27.45 | RITRedbeard | seriously? |
| 14:28.12 | RITRedbeard | sooooo these guys are gonna get more time than Madoff and Eben Upton put together I would guess |
| 14:28.19 | L84Supper | http://www.complex.com/tech/2012/06/crowd-fund-for-bullied-bus-monitor-exceeds-500k |
| 14:28.28 | RITRedbeard | both ponzi scheme masters extraordinares |
| 14:28.58 | L84Supper | make a vid of you getting bullied at college or something |
| 14:29.31 | L84Supper | http://www.indiegogo.com/loveforkarenhklein HC! $703K |
| 14:43.05 | rm | http://www.cnx-software.com/2012/08/20/codethink-launches-the-baserock-slab-arm-server/ |
| 14:43.08 | rm | ...REALLY? |
| 14:43.13 | rm | an ARM server with 11 fans? |
| 14:45.47 | RaYmAn | and eight quad core SoM's, lol |
| 14:45.52 | rm | sorry, 9 fans :) |
| 14:46.04 | RaYmAn | if you count PSU fan, you get 10 |
| 14:46.44 | L84Supper | I don't see any conductive cooling plates on the SOC's so it looks like forced convection cooling only |
| 14:47.42 | L84Supper | thermal management is probably a mystery to them anyway, so they just slap on tiny fans that get clogged with dust in 6 months |
| 14:48.19 | L84Supper | the SOC's run hotter and die sooner, probably just after the warranty is up |
| 14:52.28 | L84Supper | so maybe they do understand and just build in the obsolescence |
| 14:55.52 | *** join/#arm-netbook chz_ (~chz@121.18.213.64) |
| 15:03.14 | *** join/#arm-netbook xxiao (~xxiao@ec2-50-112-57-59.us-west-2.compute.amazonaws.com) |
| 15:14.57 | *** join/#arm-netbook tavish3 (~tavish@59.177.0.179) |
| 15:27.40 | ccssnet | L84Supper: lol |
| 15:27.52 | ccssnet | i do agree that the fans will be an issue |
| 15:28.20 | ccssnet | would much rather see a big alluminum or copper plate |
| 15:28.25 | ccssnet | with fins |
| 15:28.31 | ccssnet | and 1 or 2 fans on that |
| 15:29.13 | *** join/#arm-netbook Quarx (~Quarx@46.233.248.191) |
| 15:35.25 | Mazon | does anyone know if the Geniatech ATV1200 or clones have come out yet ? |
| 15:55.13 | popolon | looks like DVI is reall dead ? lot of device with VGA & HDMI |
| 15:55.41 | traeak | hdmi -> dvi-d adapters are cheap and available |
| 15:55.56 | traeak | just lose the sound |
| 15:56.44 | popolon | oh :) |
| 15:57.24 | popolon | ok, there are generally (but on PC-on-a-stick) rca output on those little devices |
| 15:58.29 | popolon | http://www.mycablemart.com/store/cart.php?m=product_detail&p=1409 |
| 15:59.04 | popolon | with audio |
| 15:59.35 | traeak | ouch pricing isn't exactly great |
| 15:59.41 | RaYmAn | hacky audio :P Not all tv's etc support getting audio seperate from HDMI signal |
| 16:01.24 | popolon | in fact I d'ont have TV, so that's not a problem for me :)) |
| 16:02.59 | popolon | Mazon, do you know wich SoC is in the Geniatech ? |
| 16:03.21 | popolon | Cortex A9@1.5GHz as CPu after their site, no more information |
| 16:11.13 | WarheadsSE | Mazon: idk if they are available yet, unsure of their CPU. Arch Linux ARM does have the ATV310B |
| 16:11.25 | traeak | and these soc's are small package, few connectors, etc. if you want gazillions of connectors you'll have to get a PC |
| 16:11.41 | *** join/#arm-netbook simonckenyon (~kvirc@195.7.61.8) |
| 16:20.50 | popolon | does someone know if there is this kind of device with video-in connectors ? |
| 16:29.48 | *** join/#arm-netbook simonckenyon (~kvirc@195.7.61.8) |
| 16:31.45 | Guest43356 | sup |
| 16:37.58 | rm | finally built a booting kernel again :) |
| 16:38.37 | furan | :D |
| 16:39.50 | rm | I tried to enable USB serial dongle support and multimedia (for USB webcams), also a couple of other changes, some sunxi-related |
| 16:39.53 | rm | likely a bad idea |
| 16:39.59 | rm | 3 rebuilds and none booted |
| 16:40.38 | rm | and I don't really want to open this particular MK802 for serial console |
| 16:40.41 | furan | are you monitoring serial? |
| 16:40.54 | furan | oh, why not? super easy to open |
| 16:41.06 | rm | I opened one, and I think it's really hard to avoid visibly-damaging the case during opening |
| 16:41.13 | furan | I just ordered two because I killed one soldering new nand to it |
| 16:41.14 | rm | maybe I'm just way too clumsy :) |
| 16:41.19 | furan | been there |
| 16:41.45 | RaYmAn | rm: what, you actually expect any kind of warranty on it? :P |
| 16:41.46 | rm | sooo, as a start I decided to just take my working config and rebase from 3.0.31 to 3.0.39 |
| 16:41.48 | rm | that worked |
| 16:41.50 | rm | http://media.romanrm.ru/a10/server/3.0.39-rm1/ |
| 16:41.52 | furan | with the one I have I snapped off the usb port when I was taking it out of its case because I wasn't paying attention. soldered it back on and didn't kill any traces thankfully |
| 16:42.00 | furan | but getting 2 more today |
| 16:42.05 | rm | RaYmAn, I might want to sell it, dunno |
| 16:42.14 | furan | one I'll keep pristine, the other I'll solder a nand socket to |
| 16:43.27 | rm | I have two MK802s, one is in "production" use and generally works (unopened); and the other has problems booting from SD, and I opened that one |
| 16:44.30 | furan | haha wrong choice! |
| 16:44.37 | furan | you want sd to save your butt |
| 16:44.41 | furan | imho |
| 16:45.02 | rm | furan, wat |
| 16:45.02 | RaYmAn | as long as you have usb you're mostly fine (assuming your nand works ;) |
| 16:45.09 | rm | the second one does not boot from SD, period |
| 16:45.17 | rm | read my saga on the mailing list :p |
| 16:45.35 | rm | the only way to get it to boot Linux was to put my kernel image into NAND |
| 16:46.08 | rm | well, the SD works for root FS |
| 16:46.13 | furan | what I meant was, I would have used the device with working sd as the one you crack open, because you will be working on that one more and booting from sd can save you butt |
| 16:46.18 | rm | just the uBoot SPL can't detect SD to load uBoot |
| 16:46.24 | furan | but I am just getting into this stuff still |
| 16:46.36 | furan | you/your |
| 16:59.47 | *** join/#arm-netbook gimli (~gimli@xbmc/staff/gimli) |
| 17:09.37 | TomNL | gimli: did you get any update from allwinner, about wrong ifdef's in lib compiling? |
| 17:13.43 | furan | grep the source code for @allwinner.com and just start emailing everybody |
| 17:15.01 | gimli | TomNL: they where just lying to us. they clamed to make changes and then they send me old crap |
| 17:15.10 | gimli | TomNL: fucking assholes from Allwinner |
| 17:15.41 | furan | probably just marketing guys maaan |
| 17:15.43 | TomNL | gimli: hehe...but hipboi said today his colleague made a mistake while compiling..forgot to turn on/of ifdef |
| 17:16.10 | TomNL | <hipboi> about xbmc, my workmate Chen xiaochuan |
| 17:16.15 | TomNL | <hipboi> made a mistake about the #ifdef |
| 17:16.23 | TomNL | <hipboi> which leads the lib uncompiled |
| 17:16.24 | furan | yeah but wouldn't it have a different crc due to different timestamp |
| 17:16.28 | TomNL | <hipboi> so the checksum is the same |
| 17:16.32 | furan | like not exactly the _same_ timestamp |
| 17:17.01 | TomNL | well not if not compiled at all..and just send old libs thinking it got compiled.... |
| 17:17.10 | gimli | i still believe that they where lying. the file date and md5sum was the same |
| 17:17.20 | RaYmAn | furan: crc is contents only...File date is another matter of course. |
| 17:17.24 | gimli | quite unbelievable that there was something new compiled |
| 17:17.42 | rm | don't assume to malice what can be explained by incompetence :P |
| 17:17.53 | furan | I thought file date was part of the file |
| 17:17.59 | xxiao | gimli: it might be just a stuipd mistake, don't over-estimate some eningeer's capability |
| 17:18.00 | gimli | if allwinner did something wrong they are wellcome to excuse themselfe. |
| 17:18.11 | TomNL | true true...i was just wondering if they contacted you again...now hipboi gave this info |
| 17:18.22 | RaYmAn | furan: it's metadata stored by the fs. |
| 17:18.44 | furan | RaYmAn: and archives store that metadata? |
| 17:18.45 | gimli | xxiao: but i can over-estimate being ignored when asking questions |
| 17:18.52 | RaYmAn | furan: yes |
| 17:18.58 | furan | got it, thx |
| 17:19.14 | RaYmAn | well, not *all* archives obviously, but they often do. |
| 17:19.15 | RaYmAn | :) |
| 17:19.20 | furan | sure |
| 17:19.25 | xxiao | gimli: for that i'm sure if you ask them questions you will be more likely getting answers |
| 17:19.45 | xxiao | i was told by some chinese developers, that it took long time for them to reply email in english |
| 17:20.08 | xxiao | imagine you have to reply their chinese emails, |
| 17:20.28 | xxiao | s/questions/questions in chinese/ |
| 17:20.42 | gimli | and, i'm austrian. my native language is german |
| 17:21.17 | xxiao | i think language remains to be one major barrier when certain chinese IC tries to embrace OSS |
| 17:21.19 | furan | I've been helping make some allwinner based products better by patching drivers, so the guy selling these products contacted me and offered free hardware. so I told him here's what I need to know (he offered to talk to the companies since he gets a lot of their stuff). he listened one thing and responded about it and completely ignored the rest. the one thing was for his most profitable device. |
| 17:21.32 | furan | it was really upsetting, he has done that with every mail I've had with him |
| 17:23.01 | gimli | also the question for a ChangeLog and getting none is more than strange. they should know what they changed |
| 17:23.11 | gimli | all in all it doesn't look like a simple mistake |
| 17:23.44 | thefrog | we had hardware manufactuered in china for a while.. we'd always get these strange lies about something that was clearly not what they said... |
| 17:23.49 | phh | i understand making a changelog could be complicated ... |
| 17:23.54 | thefrog | it was like a stalling tactic |
| 17:24.12 | furan | I really like a lot of the hardware I get from china and don't want to apply a stereotype |
| 17:24.21 | thefrog | we'd have to start over and explain the problem like the lie had never happened |
| 17:24.23 | furan | but I do not have good experience in communication |
| 17:24.40 | TomNL | I understand all this, but untill now hipboi mostly told info that is true |
| 17:25.10 | furan | hipboi might be getting lied to as well |
| 17:25.10 | gimli | from my point of view Allwinner fucked it up. extrem unprofessional. maybe i reconsider my standpoint when they excuse themselfe ;) |
| 17:25.45 | gimli | since last week when i stated that they did nothing and lyed to us, the got silent again and no message from them. |
| 17:26.37 | furan | it makes me wonder if you are working with the wrong people there |
| 17:26.38 | gimli | also, the wrong #ifdef. the filedate muast cahnge. the #ifdef means source canges. but it didn't |
| 17:26.41 | furan | marketing or something |
| 17:28.08 | Turl | I heard jelly in here? :P |
| 17:28.46 | xxiao | on the other hand, what's alternatives for A10 |
| 17:28.51 | thefrog | I donno. when we worked with chinese manufactueres.. they lied all the time. we only made progress by asking over and over for what we needed. they never owned up to any of the lies. |
| 17:29.08 | furan | I really like the a10 stuff because even though they are secretive, it feels more open and hackable |
| 17:29.17 | furan | it's easy to modify a10 hardware to do cool stuff |
| 17:29.25 | xxiao | i used to check am335x unitl i saw A10, but am disappointed by A10 as well |
| 17:29.31 | xxiao | s/unitl/until/ |
| 17:29.32 | thefrog | I ordered a pivos .. shipped today. |
| 17:29.44 | *** join/#arm-netbook Quarx (~Quarx@46.233.248.191) |
| 17:29.45 | thefrog | not sure what to do with the A100 under my TV |
| 17:29.47 | furan | that amlgic was interesting |
| 17:29.48 | furan | amlogic |
| 17:30.30 | xxiao | amlogic can not compete with A10 by price, meanwhile AML used to power problems |
| 17:30.52 | furan | power problems? as in cpu? |
| 17:30.58 | traeak | and i've seen comparative reviews, the a10 and amlogic cortex a9 are about equal on performane |
| 17:31.32 | xxiao | right, if A8 can do the job fine I don't care if it's a quad-core-A9 which has large leakages |
| 17:32.36 | xxiao | Rockchip is even more closed, AML is no better, while A10 is not perfect but I have yet to find alternatives, hope i missed something |
| 17:34.12 | gimli | fuck allwinner for xbmc, they are no reliable company for development |
| 17:38.06 | traeak | cedarx still jacked, nice |
| 17:40.45 | jelly-home | gimli: which hw vendor is, in that price range? |
| 17:41.33 | Turl | xxiao: if you want more source you end up needing to go with the 'big guys' (TI, QCom to a lesser extent) |
| 17:43.55 | gimli | jelly-home: amlogic |
| 17:48.57 | xxiao | Turl: Qcom is a no-no, TI still looks a good option, Freescale might think dealing with small players is a waste oftime |
| 17:49.11 | xxiao | avoids Qcom/Bcom/Marvell at all costs |
| 17:49.55 | Mazon | popolon: WarheadsSE : afaik, the geniatechs are amlogic mx |
| 17:50.23 | Mazon | at least atv1200 and atv1600 |
| 17:50.40 | WarheadsSE | yeah, idk about the one specific to the 1200/1600 though |
| 17:50.48 | WarheadsSE | the 310B is an amlogic single |
| 17:51.20 | Turl | xxiao: another option to consider is Ericcson |
| 17:51.27 | *** join/#arm-netbook j1nx (~j1nx@163-62-223.ftth.xms.internl.net) |
| 17:51.52 | Turl | (ST Ericcson that is) |
| 17:52.07 | j1nx | evening |
| 17:52.08 | Turl | or Ericsson or however you spell it :P |
| 17:52.14 | Turl | evening j1nx |
| 17:52.19 | xxiao | Turl: not sure about that, i heard that portion is IC biz was for sale? |
| 17:52.31 | xxiao | s/is/of/ |
| 17:52.40 | Turl | xxiao: no idea |
| 17:52.45 | Turl | they have http://www.igloocommunity.org/ though |
| 17:53.14 | xxiao | Turl: they really don't have the tech-know-how in house, i think they just contracted the devel to Linaro |
| 17:53.29 | xxiao | unlike TI/FSL/Qcom/etc who has its own R&D team |
| 17:53.50 | j1nx | does anybody have a good link to check out the differences between 8726-M1,M3,MX ? |
| 17:53.53 | Turl | might've been, linaro is good anyway :P |
| 17:58.21 | arokux | hi, which file do you have on your first fat partition? |
| 17:58.26 | arokux | s/file/files |
| 17:59.25 | Turl | usually kernel, script.bin |
| 17:59.31 | Turl | and the boot.scr file for uboot |
| 18:04.50 | arokux | so 3 files? |
| 18:04.57 | j1nx | initrd |
| 18:05.04 | j1nx | makes 5 |
| 18:05.14 | j1nx | 4 |
| 18:05.15 | arokux | should I have initrd? |
| 18:05.28 | j1nx | depends on you rootfs |
| 18:05.43 | arokux | and this one: mele-a1000.bin? |
| 18:06.05 | RaYmAn | hmm.. Next linaro connect is just around the corner from me |
| 18:06.12 | j1nx | rename to either evb.bin or script.bin |
| 18:06.19 | *** join/#arm-netbook zoobab_ (zoobab@vic.ffii.org) |
| 18:06.21 | zoobab_ | hi |
| 18:06.29 | Turl | RaYmAn: allwinner people went to the previous one :P |
| 18:06.34 | RaYmAn | so has to see if he can sort it out with $workplace to go there |
| 18:06.36 | RaYmAn | heh |
| 18:06.58 | Turl | hi zoobab_ |
| 18:07.40 | arokux | Turl, they couldn't speak english anyway |
| 18:07.59 | Turl | true that |
| 18:07.59 | RaYmAn | minor details =P |
| 18:10.50 | arokux | does anyone here want to buy a mele 1000? |
| 18:11.10 | zoobab_ | @arokux why not |
| 18:12.23 | arokux | shit, I've booted to the root console: sun4i# |
| 18:13.12 | arokux | oh no, it's u-boot console |
| 18:14.52 | arokux | it expects mele_stock.bin, not script.bin... |
| 18:21.11 | arokux | hm.. I've renamed script.bin to mele_stock.bin, but it still cannot find it. |
| 18:21.17 | arokux | Unable to read "mele_stock.bin" from mmc 0:1 |
| 18:24.49 | furan | xxiao: I thought rockchip and allwinner were the same |
| 18:24.57 | furan | e.g. rockchip a10 = allwinner a10 |
| 18:24.59 | furan | oh no |
| 18:25.01 | furan | boxchip |
| 18:26.16 | arokux | that's funny! my file manager hasn't renamed the file... |
| 18:31.21 | *** join/#arm-netbook Tsvetan (5e48971e@gateway/web/freenode/ip.94.72.151.30) |
| 18:31.24 | Tsvetan | hi |
| 18:31.33 | arokux | zoobab_, you want to buy a mele? |
| 18:31.48 | WarheadsSE | arokux: ? |
| 18:32.02 | WarheadsSE | just not into the A1000 ? |
| 18:32.09 | arokux | WarheadsSE, cubieboard should be enough for kenel hacking. |
| 18:32.16 | WarheadsSE | true |
| 18:32.22 | arokux | and it costs the half |
| 18:32.36 | WarheadsSE | but you already have the A1000 :p |
| 18:32.36 | arokux | without wifi, though |
| 18:32.57 | zoobab_ | how much is it? |
| 18:33.14 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 18:33.21 | zoobab_ | I have 3 A10 devices now |
| 18:33.21 | arokux | I've got it for 85 EUR |
| 18:33.33 | zoobab_ | how much do you sell it back? |
| 18:34.26 | arokux | zoobab_, 70 (blush) |
| 18:34.59 | arokux | zoobab_, but only after I have a cubieboard |
| 18:35.16 | xxiao | furan: you're right, they're getting rid of boxchip naming though |
| 18:35.22 | arokux | zoobab_, which country are u in? |
| 18:35.30 | jelly-home | it's a second-hand electronic device. Going above 70% of original price is unusual |
| 18:35.33 | furan | I confused rockchip with boxchip |
| 18:36.01 | xxiao | that means the marketing guys did an excellent job |
| 18:36.01 | arokux | jelly-home, well, I haven't included 15 EUR import tax. |
| 18:36.20 | xxiao | furan: boxchip -- rockchip, A8 with A10 name, etc |
| 18:36.41 | xxiao | the similarity in names, i believe, is intentional |
| 18:38.37 | ccssnet | soooo many new embedded devices and chips. so hard to keep up |
| 18:38.44 | ccssnet | googles a few key terms |
| 18:39.51 | jelly-home | arokux: it somehow feels nice that .de is charging exhorbitant taxes for imports just like .hr here does |
| 18:39.58 | jelly-home | "nice" |
| 18:41.12 | hno | gimli, who have you spoke to in Allwinner? Is all communication going via Eva as usual? |
| 18:42.29 | Turl | hipboi said something about a work mate failing hard with an #ifdef |
| 18:42.45 | gimli | hno: they gave us "weng jun hong" as contact. tryed to contact Eva but she didn't anwser direct. |
| 18:43.06 | hno | Better not talk to Eva once you get another contact.. |
| 18:43.40 | hno | The only dealings with Eva should be to get the approval of talking. |
| 18:45.03 | Tsvetan | I got several times incorrect information from Eva, it's not that she do not want to help she just couldn't |
| 18:45.09 | gimli | hno: fuck it. when they started to mess with us last week, i put the allwinner support email in the CC. which should be read by Eav. i hate it when people think im a kid with no clue |
| 18:45.58 | jelly-home | watch out for cultural impedance mismatches |
| 18:46.52 | Tsvetan | it's more probably that people who know techncal details do not speak English and there is broken translation from the marketing |
| 18:56.15 | xxiao | Tsvetan: agreed |
| 18:56.58 | xxiao | i was told by two very good chinese engineers that sometimes it took hours for them to reply English emails, they want to make the syntax/grammar etc correct and it takes time |
| 18:58.00 | xxiao | just imagine you're a western small IC company and now has a few chinese customers to support :) |
| 18:58.01 | Tsvetan | my advice is to ask hipboi to check what is the situation as he know English and have technicall background |
| 18:58.56 | furan | maybe find a translator service |
| 19:01.00 | penguin42 | hmm, I do know a couple of chinese tech guys; their English isn't too bad - they might not get the grammar perfect, but I can understand what they mean which is more important (but there again as an English only speaker, my grammer is pretty awful) |
| 19:04.48 | penguin42 | http://www.baserock.com/servers is an interesting ARM box |
| 19:05.25 | phh | price's missing |
| 19:05.31 | penguin42 | yeh I noticed that :-) |
| 19:05.36 | phh | and erm, 10Gbps ? seriously ? |
| 19:05.43 | phh | i'm not even sure it can handle 1Gbps -_-' |
| 19:06.03 | penguin42 | phh: Why? that's 32 1.3GHz ARMs boiling down to a 10Gbps link |
| 19:06.10 | phh | oh, i misread, it's 5Gbps between cpus and a global 10 Gbps |
| 19:06.10 | phh | ok |
| 19:06.22 | phh | i was pretty sure i read it was 10 Gbps for every cpu |
| 19:06.25 | penguin42 | anyway, a quad 1.33GHz ARM should manage 1Gbps easy |
| 19:07.01 | penguin42 | they seem to use these http://cogcomp.com/csb_csb1726.htm modules, which I'm curious how much they are individually |
| 19:08.39 | phh | i wouldn't have thought cogent makes their own pieces ... |
| 19:08.48 | phh | sounds totally logical though |
| 19:09.52 | popolon | furan, Rockchip = Fuzhou Rockchip (Fuzhou municipality in Fujian province, near taiwan) |
| 19:10.07 | popolon | Allwinner => Zhuhai (juste beside Macao) |
| 19:10.28 | popolon | HiVision (for Huawai) => Shenzhen (just beside Hongkong) |
| 19:10.35 | popolon | Huawei |
| 19:11.33 | popolon | the last one do the more powerfull SoC on paper, but first devices should be sold in 1 or 2 month |
| 19:11.46 | popolon | cortex A9 quadcore :) |
| 19:12.25 | phh | i'd rather take a dual core cortex A15 |
| 19:12.44 | popolon | with mali T6xx :) |
| 19:13.28 | Tsvetan | Huawei probably will use the SoCs for their own equipment and do not bother to sell to some smallish western companies |
| 19:14.13 | phh | popolon: yeah whatever, current gpus are really quite good |
| 19:15.54 | xxiao | Huawei will never deal with small players |
| 19:16.08 | xxiao | you will be redirected to some distributors immediately |
| 19:16.30 | xxiao | most of them can not spell linux fully |
| 19:17.22 | xxiao | http://www.baserock.com/servers ... are those fans? |
| 19:17.42 | phh | lol yes |
| 19:17.45 | Turl | looks like i xxiao |
| 19:17.48 | Turl | it* |
| 19:17.56 | penguin42 | xxiao: It's obvious to me which way the air is supposed to go |
| 19:18.01 | penguin42 | <PROTECTED> |
| 19:18.06 | xxiao | check freescale's P5040, one chip does 20Gbps easily |
| 19:18.20 | Turl | it's a half depth 1U |
| 19:18.48 | penguin42 | Turl: Supposed to be able to go back to back toget twice of them into 1U |
| 19:18.51 | phh | xxiao: it's just not a poor arm -_-' |
| 19:19.35 | Turl | yep 64 cores/1U |
| 19:19.47 | xxiao | anyway the fans make me dizzy, nice try |
| 19:19.49 | thefrog | is this a hadoop setup? |
| 19:20.00 | penguin42 | Turi: But then I really don't understand which way the air is supposed to go |
| 19:20.03 | phh | xxiao: well, they wanted it to be as noisy as usual servers :D |
| 19:20.29 | Turl | penguin42: to the bottom though the heatsink |
| 19:20.32 | penguin42 | phh: Interesting way to build a quadcopter |
| 19:20.36 | phh | :) |
| 19:20.57 | xxiao | another hi-tech oven |
| 19:21.04 | thefrog | not sure how to pass a hot/cold audit with that setup |
| 19:21.14 | phh | there are top 500W to evacuate + hard drives |
| 19:21.17 | penguin42 | thefrog: yeh, especially back to back |
| 19:21.18 | phh | it's obvious it needs so many fans ! |
| 19:21.32 | xxiao | hard-drive is of no concern, it's said using SSD |
| 19:21.52 | phh | well, 32 SSD |
| 19:21.56 | thefrog | i had a hard enough time with my network switches that were left-right airflow (sides). |
| 19:21.56 | phh | that still heats. |
| 19:22.17 | phh | oh wait, 30GB ssd * 32, that thing is going to cost a kidney. |
| 19:22.18 | xxiao | SSD is pretty cool all the time from my experience |
| 19:22.25 | Gumboot | Is that eight separate logical machines, then? |
| 19:22.35 | popolon | on most 1U server there is only heat exchanger on cpu |
| 19:22.37 | phh | Gumboot: yes |
| 19:22.38 | Turl | Gumboot: yes |
| 19:22.44 | xxiao | closes that tab, brain hurts |
| 19:22.46 | popolon | and fans are in front and rear of the computer |
| 19:22.48 | Turl | each one with a quad core cpu |
| 19:23.40 | Gumboot | So, eight separate computers, each a fairly fast quad-core A9. So what's just one of those made out of? Whose SoC? What specs? |
| 19:23.55 | penguin42 | Gumboot: Armada XP |
| 19:24.20 | phh | wow, it even has ecc |
| 19:25.07 | *** part/#arm-netbook Tsvetan (5e48971e@gateway/web/freenode/ip.94.72.151.30) |
| 19:25.10 | penguin42 | it's a little mean on RAM/node, but other than that and the air flow it's quite nice |
| 19:25.17 | penguin42 | (oh and the missing price tag) |
| 19:25.26 | Turl | <PROTECTED> |
| 19:25.36 | phh | i'm expecting something like 5k€ |
| 19:26.36 | Gumboot | Hm. |
| 19:26.50 | Gumboot | What's the latency on DDR3 compared to DDR2? |
| 19:27.07 | Gumboot | Or is the protocol the same? |
| 19:30.44 | Gumboot | Needs more L2. |
| 19:31.13 | traeak | i think ddr3 gets its rate through block transfers |
| 19:31.56 | Gumboot | Don't like the sound of that. |
| 19:33.07 | penguin42 | if you can persuade page zeroing or a good memcpy to kick that into life it would be good |
| 19:34.01 | Gumboot | Is page zeroing a notable concern? |
| 19:34.09 | Gumboot | I've always wondered, but never tested. |
| 19:34.36 | phh | Gumboot: ram latency hasn't changed since sdram ... |
| 19:34.43 | phh | oh well, perhaps it has been /2 |
| 19:35.41 | Gumboot | phh: There's always the risk of it getting worse. The number of cycles is increasing, though, and that may be a liability in itself. |
| 19:36.16 | Gumboot | And if things start getting optimised for bulk transfers, but there's nothing inside the chip to cache those, then it gets worse again. |
| 19:36.32 | popolon | 2*8 = 16 GB of RAM, today there are 192 GB of ram servers with 24 cores Xeon (2*(6cores + hyperthreading)) @2,3 GHz and virtualization capabilities and you can put 2 of this blades on 1U, if there wasn't energy efficency that wouldn't be interesting at all. |
| 19:37.25 | Gumboot | popolon: The complete split between clusters is probably a big performance benefit. |
| 19:37.46 | Gumboot | Wasn't it in here someone was on about the virtualisation tax? |
| 19:37.51 | popolon | 2GB of ram, hope there will not be lot of date |
| 19:37.53 | popolon | data |
| 19:38.43 | popolon | but that's interesting for little low cost servers |
| 19:38.58 | phh | or just have fun :D |
| 19:39.08 | penguin42 | popolon: Yeh seems a bit tight |
| 19:39.08 | phh | hum no, way to costly actually |
| 19:40.08 | Gumboot | Well, this is one of not-so-many ARM SoCs with SATA for a high-speed swap interface, right? |
| 19:40.56 | popolon | and no hot plug on the boards |
| 19:40.57 | penguin42 | Gumboot: The SoC will take upto 8GB |
| 19:41.04 | xxiao | Gumboot: more and more ARMs are equiped with SATA |
| 19:41.15 | popolon | so if one is dead, you need to stop the whole computer to replace him |
| 19:41.26 | penguin42 | Gumboot: http://www.dell.com/armserver is a different build using the same SoC |
| 19:41.27 | popolon | (and to open it) |
| 19:41.29 | popolon | lot of work |
| 19:41.33 | xxiao | i think each ARM core can only handle 2GB? |
| 19:41.35 | Gumboot | xxiao: But _real_ SATA, right? Not just SATA over USB? |
| 19:41.41 | penguin42 | popolon: I think the hope is you'll have an entire rack of them |
| 19:41.43 | penguin42 | xxiao: No |
| 19:41.44 | xxiao | Gumboot: real SATA |
| 19:41.58 | penguin42 | xxiao: The Armada XP can deal with 8GB visible |
| 19:42.01 | xxiao | check Freescale i.MX6, TI's DM814x/816x etc |
| 19:42.42 | Gumboot | popolon: Also, it may be that with the lower power consumption the risk of failure is reduced accordingly. |
| 19:43.02 | xxiao | for ARM to get out of portable domain, it's not about SATA, it's about interfaces such as pcie |
| 19:43.11 | popolon | yes, but there is only a AC adapter, if it fail, the 8 computers fail |
| 19:43.16 | penguin42 | xxiao: The Armada XP has PCIe |
| 19:43.24 | popolon | hand the adpater is not hotplug too |
| 19:43.37 | popolon | lot of point of failure |
| 19:43.47 | xxiao | i believe that, Armada and Calxeda(?) are the only two ARM-server-trials nowadays |
| 19:44.22 | xxiao | i think HP is using Calexda? while Dell is using Armada, or the other way around, for server |
| 19:44.28 | popolon | I wouldn't use it in production and have to run fast on the night to change it :) |
| 19:44.55 | penguin42 | popolon: If you've got a few racks of them it doesn't matter if the whole 8 board chassis goes pop |
| 19:45.02 | penguin42 | xxiao: Correct |
| 19:45.26 | popolon | this mean the system will shutdown and have to be started on another computer :) |
| 19:45.32 | popolon | good for loadbalanced clusters |
| 19:45.38 | popolon | only |
| 19:45.39 | penguin42 | yeh |
| 19:45.59 | popolon | this really reduce the field of usage |
| 19:46.04 | xxiao | somehow i feel server is off-topic in netbook channel :) |
| 19:46.19 | popolon | probably :) |
| 19:46.24 | phh | why ? :D |
| 19:46.25 | RaYmAn | people talk about netbooks here? |
| 19:46.31 | RaYmAn | I've never seen that happen :P |
| 19:46.35 | xxiao | i will use AWS to do that and don't care about what hardware it's running on, as long as it's not too expensive |
| 19:47.27 | xxiao | RaYmAn: netbook, i guess it means light weight arm stuff |
| 19:47.40 | phh | xxiao: i'm not sure this server is really light :p |
| 19:47.48 | Turl | this should be #arm-devices or something :) |
| 19:48.13 | Gumboot | I guess PCIe would do just fine, if only I saw boards that connected SATA to it. |
| 19:48.20 | xxiao | arm-tablet, arm-devices, arm-servers... |
| 19:48.27 | phh | #arm-stuff |
| 19:48.28 | popolon | :) |
| 19:48.29 | Turl | Gumboot: well these have mSATA |
| 19:49.13 | xxiao | arm-ultrabook |
| 19:49.22 | phh | :) |
| 19:49.24 | phh | nice troll |
| 19:49.30 | Turl | #armtastic |
| 19:51.06 | Gumboot | I can't actually figure out from the wiki page what mSATA is. |
| 19:51.41 | Gumboot | It sounds kind of like it's neither SATA nor PCIe, but associates itself with both. |
| 19:51.41 | phh | just a different sata connector ? |
| 19:51.42 | popolon | https://en.wikipedia.org/wiki/Serial_ATA#mSATA |
| 19:51.54 | penguin42 | Gumboot: My reading is it's SATA over an edge connector |
| 19:52.49 | penguin42 | Gumboot: Actually over a connector the same as a PCIe mini-card |
| 19:54.24 | Gumboot | That would imply that all I needed was a passive adapter. That can't be right, because then the SoC's specifications wouldn't be qualified as mSATA because that would be down to what you soldered to the board. |
| 19:56.17 | penguin42 | shrug |
| 19:59.54 | Gumboot | So how many SoCs offer PCIe, then? Any good ones? Anything from Samsung? |
| 20:00.12 | penguin42 | Gumboot: Some of the exynos chips from Samsung do |
| 20:00.29 | penguin42 | Gumboot: Ditto some of the nvidia tegras |
| 20:01.38 | RaYmAn | iirc trimslice went usb sata because they couldn't get any PCIe sata working on tegra2 |
| 20:01.51 | Gumboot | What went wrong there? |
| 20:02.17 | *** join/#arm-netbook arokux_h (~arokux@HSI-KBW-46-223-181-3.hsi.kabel-badenwuerttemberg.de) |
| 20:02.17 | penguin42 | RaYmAn: Yeh I can't remember seeing anything that used tegras pcie |
| 20:02.36 | Gumboot | Actually, I'm starting to suspect there are serious difficulties with SATA. |
| 20:02.42 | Gumboot | But I don't know what. |
| 20:03.28 | penguin42 | <PROTECTED> |
| 20:03.50 | Gumboot | I don't know what. It's like some kind of Bermuda triangle of connectivity. |
| 20:05.27 | Gumboot | I heard that i.mx5 had it, then it went away again. |
| 20:07.10 | Gumboot | I guess firewire would be fine as well, but I don't see that happening. |
| 20:07.28 | WarheadsSE | Gumboot: i think it is there on the imx6 |
| 20:07.33 | WarheadsSE | I'll have to double check |
| 20:07.49 | Gumboot | Double-check what bus it's hung off of while you're at it. |
| 20:07.50 | WarheadsSE | many of the marvel chips have sata or pcie that is used for it/usb3/etc |
| 20:08.00 | WarheadsSE | when the n6x gets here, I will |
| 20:08.27 | Gumboot | How far away is that, then? |
| 20:08.48 | WarheadsSE | ~2 days |
| 20:08.52 | Gumboot | I'll stay quiet about it until then, then. |
| 20:09.53 | WarheadsSE | http://www.linuxfordevices.com/c/a/News/Freescale-iMX-6/ |
| 20:09.58 | WarheadsSE | looks like integrated PHY |
| 20:10.19 | WarheadsSE | I kow the armada platofmr has integrated sata as well |
| 20:10.30 | Gumboot | Yeah, but integrated _how_? |
| 20:10.35 | WarheadsSE | I know that from working with them on some projects. |
| 20:11.36 | WarheadsSE | http://cache.freescale.com/files/graphic/block_diagram/IMX6Q_BD_IMG.jpg |
| 20:12.01 | WarheadsSE | inthe SoC w/ PHY SATA 3.0G |
| 20:12.14 | penguin42 | Gumboot: I don't think it's sata that's magical bermuda triangle stuff - it's most high speed serial interfaces; the guys who know how to wiggle pins at gigabit speeds are special |
| 20:12.19 | WarheadsSE | agreed, still don't know which bus it is hanging off. |
| 20:13.15 | Gumboot | I think it's hanging things off the wrong bus that gave me such terrible experiences with Dell so many years back. |
| 20:13.49 | xxiao | i'm scared by marvell, how friendly are they to the oss or small players? |
| 20:14.10 | xxiao | for their media-centric(e.g. tablet) chips, i guess it will be fully closed |
| 20:14.24 | xxiao | for kirkwood/server-arm chips, are they open sourced? |
| 20:14.29 | Gumboot | can ignore that, because he wants a headless box anyway. |
| 20:14.56 | WarheadsSE | xxiao: I've had little issue with them so far, but then I am generally not dealing with gpu/etc |
| 20:15.33 | xxiao | WarheadsSE: recall their first PLUGs had heat issues, are those a cpu issue or board/enclosure design issue? |
| 20:15.57 | xxiao | nobody wants a server-fanless-device died of heat silently at home |
| 20:17.14 | WarheadsSE | it all depends on a # of things |
| 20:17.41 | WarheadsSE | but often, lets over-crank this and leave too-little dissipation |
| 20:17.52 | WarheadsSE | But, for most consumer uses, they never hit that envelope |
| 20:18.02 | xxiao | guess the first gen mk802 did something similar |
| 20:18.11 | WarheadsSE | I wouldnt rule it out |
| 20:18.31 | WarheadsSE | If you have something liek that, then crank near 100% CPU for extended periods, it need dissipation |
| 20:18.46 | xxiao | best chip for me: lots usb port + sata + pcie |
| 20:19.03 | xxiao | low power is good, fanless is enough |
| 20:20.12 | WarheadsSE | then I would say you will very much like the plug that the SMILEplug is based on |
| 20:20.26 | WarheadsSE | I can't talk much about it, but.. zippy |
| 20:20.40 | xxiao | WarheadsSE: does it have pcie? |
| 20:21.07 | WarheadsSE | pcie, 2xGbE. plus mmc/usb2&3/probably sata |
| 20:21.25 | L84Supper | the conversation about allwinner engineers having problems with English is silly, they are making plenty enough profit to hire a tech translator |
| 20:21.50 | L84Supper | allwinner is making the profit, maybe not the engineers |
| 20:22.51 | WarheadsSE | And it's not just allwinner |
| 20:22.52 | Mazon | WarheadsSE: http://www.marvell.com/solutions/cloud-computers/assets/Marvell-SMILE-Plug-Platform-Brief.pdf doesn't mention sata |
| 20:22.56 | WarheadsSE | marvells aren't that much better |
| 20:23.06 | Turl | "hire? will it make more monies? no? nowai" |
| 20:23.07 | WarheadsSE | Mazon: ;) I know it doesn't |
| 20:23.13 | WarheadsSE | i said the plgu it is _based_ on |
| 20:23.17 | Mazon | aha! |
| 20:23.30 | L84Supper | Turl, even by the hour |
| 20:24.14 | xxiao | everytime i search marvell i got a 2-page marketing brief shit, TI/FSL gave you a 5000-page TRM instead |
| 20:24.53 | WarheadsSE | xxiao: I can't give you or Mazon a tonne of information. It's a non-released platform & NDA's |
| 20:25.10 | WarheadsSE | I can talk around it, but not really about it |
| 20:25.11 | xxiao | WarheadsSE: i know that. they don't even sign NDA with small players |
| 20:25.20 | WarheadsSE | is small player |
| 20:25.29 | arokux_h | how do you "turn off" your mele to write new kernel to the mmc? (not sudo halt or so) |
| 20:25.40 | xxiao | probably you can talk about how to sign NDA with them :) |
| 20:25.48 | WarheadsSE | arokux_h: you shouldn't have to.. |
| 20:25.52 | Mazon | WarheadsSE: we could talk about whats NOT on it ;) |
| 20:26.06 | Turl | arokux_h: do it on the mele itself via ssh or w/e then reboot :) |
| 20:26.13 | WarheadsSE | ^ |
| 20:26.19 | Mazon | anyhow, itching to get a hold of some more general purpose arm gear |
| 20:26.24 | Turl | shutdown/halt work too |
| 20:26.29 | WarheadsSE | It's not like it is live-reading the kernel every cycle arokux_h |
| 20:26.33 | Mazon | with proper specs, that is |
| 20:26.45 | L84Supper | Turl: nowai for Laowai? :) |
| 20:26.50 | arokux_h | no, I mean if I do not have access to ssh. |
| 20:26.55 | Turl | L84Supper: ? |
| 20:27.10 | WarheadsSE | arokux_h: try the power button |
| 20:27.14 | Turl | arokux_h: unplug and hope fs don't get corrupted? :P |
| 20:27.15 | L84Supper | Laowai (Chinese: 老外; pinyin: lǎowài) is one of several Chinese words for foreigner |
| 20:27.17 | WarheadsSE | whats the OS on it at the moment |
| 20:27.43 | arokux_h | WarheadsSE, you should know ;) |
| 20:27.45 | Turl | L84Supper: :) |
| 20:27.45 | arokux_h | Arch |
| 20:28.27 | WarheadsSE | yeah, what happened to ssh? |
| 20:28.31 | WarheadsSE | shrugs |
| 20:29.22 | arokux_h | nothing, I'm going not to have ethernet for the near future |
| 20:30.14 | WarheadsSE | ah |
| 20:30.27 | WarheadsSE | should be able to get wifi up on there for yourself |
| 20:30.47 | WarheadsSE | IDK if I put a watch onthe power button, but the AXP drivers are there, so the power button might do it |
| 20:32.08 | WarheadsSE | it ought to survive a less-than pleasant shutdown anyways, due to fsck & ext4 |
| 20:34.46 | *** join/#arm-netbook ZaEarl (~malmrose@203-97-164-178.cable.paradise.net.nz) |
| 20:57.47 | *** join/#arm-netbook newbie (~kvirc@220-136-107-171.dynamic.hinet.net) |
| 21:37.17 | *** join/#arm-netbook raver2046 (raver2046@mer90-1-88-166-249-240.fbx.proxad.net) |
| 21:37.41 | raver2046 | hello ! |
| 21:38.18 | raver2046 | i need a little help somebody can help me ? |
| 21:38.39 | raver2046 | it is verry simple a need a binary or c sources |
| 21:38.43 | raver2046 | for android |
| 21:38.54 | raver2046 | anyone here ? |
| 21:41.22 | xxiao | normally you go directly to the question and see if anyone will respond |
| 21:45.02 | WarheadsSE | especially when its that vague |
| 21:48.56 | raver2046 | ok anybody can compile me that for android http://gitorious.org/android-transmission |
| 21:49.22 | raver2046 | blog article that explain the tool here : http://i-miss-erin.blogspot.fr/2010/09/use-transmission-in-android.html |
| 22:16.22 | Turl | raver2046: I'll give it a try |
| 22:18.29 | raver2046 | thanks a lot |
| 22:21.31 | *** join/#arm-netbook netchip (netchip@loota.org) |
| 22:32.56 | *** join/#arm-netbook j1nx (~j1nx@163-62-223.ftth.xms.internl.net) |
| 22:33.00 | Turl | raver2046: build fails to link on some openssl stuff |
| 22:33.12 | *** part/#arm-netbook j1nx (~j1nx@163-62-223.ftth.xms.internl.net) |
| 22:33.40 | *** join/#arm-netbook j1nx (~j1nx@163-62-223.ftth.xms.internl.net) |
| 22:34.04 | raver2046 | do you compile on debian? |
| 22:35.25 | raver2046 | hi j1nx |
| 22:35.46 | raver2046 | Turl look at this : https://forum.transmissionbt.com/viewtopic.php?f=1&t=10494&p=61286&hilit=android#p61286 |
| 22:35.48 | Turl | raver2046: on an ubuntu chroot fwiw |
| 22:36.02 | raver2046 | Did some manual hacking into the curl_config.h file to disable SSL and it builds on Froyo git source as well now.. |
| 22:36.09 | Turl | I used a gingerbread tree though |
| 22:36.15 | Turl | i don't have any froyo one handy |
| 22:36.30 | raver2046 | ok build with gingerbread no problem |
| 22:36.52 | raver2046 | can you manualy hack curl_config.h to disable ssl ? |
| 22:42.59 | Turl | raver2046: disabled it on the .h but it still fails to link http://paste.debian.net/184798/ |
| 22:43.06 | Turl | (and more that don't fit on my term window) |
| 22:45.04 | raver2046 | Turl is that help you : apt-get install libssl-dev |
| 22:50.26 | Turl | doubt it |
| 22:50.29 | Turl | it's a linking problem |
| 22:50.37 | Turl | installing whatever ssl stuff on my host won't help |
| 22:52.38 | raver2046 | hoo so sad |
| 22:53.18 | raver2046 | i just triing to have a seedbox with webui admin torrent |
| 22:53.29 | raver2046 | software on my android net top |
| 22:53.41 | raver2046 | and this is the only one i know |
| 22:55.28 | Turl | raver2046: can't you just pick any of the lots of torrent apps available? |
| 22:55.30 | raver2046 | another question , i have a mele a1000 and i formated my hdd with my pc in ext3 but when i put the hdd in the a1000 in the paramters i see in the sata section : click to install sd card, (if i click that doesn't do nothing) |
| 22:55.31 | Turl | for example https://play.google.com/store/apps/details?id=com.bittorrent.client |
| 22:56.06 | Turl | or https://play.google.com/store/apps/details?id=hu.tagsoft.ttorrent.lite |
| 22:56.07 | raver2046 | i cant connect remotely with my pc to administrate it |
| 22:56.28 | Turl | or https://play.google.com/store/apps/details?id=com.mobilityflow.torrent |
| 22:56.35 | Turl | well you can administrate it on the device itself |
| 22:57.17 | raver2046 | i know that but it is a server the screen is not always on, and i want to monitor it remotely with my pc |
| 22:58.33 | Turl | can't you install a linux (headless even) distro and the real transmission there? |
| 23:00.08 | raver2046 | no , in fact that was my first goal but i have an allwinner a10 vpu and the harware acceleration for video decoding isn't supported on linux with the a10 |
| 23:00.22 | raver2046 | so i am forced to stay on android |
| 23:01.09 | Turl | you could reboot to android when you need it and run headless the rest of the time |
| 23:14.39 | raver2046 | well not very handy solution |
| 23:15.12 | raver2046 | ho i think i have compiled successfully tranmission hihiihihih :-) |
| 23:21.12 | raver2046 | no this is a compilation for a debian subsytem on my android |
| 23:21.28 | raver2046 | that work on the subsystem but not on the android |
| 23:31.18 | popolon | about allwinner tech that didn't speak english, that's often the case in china, no need to speak english as there is already billion+ of customers that speak the same language, often european customer (and I suppose american, african, other) pay a local/chinese translator that go to China |
| 23:31.57 | popolon | only some commercials (or technico-commercial) people speak english |
| 23:32.03 | popolon | and young students |
| 23:32.40 | popolon | and SoC customers are mother board maker that are often based or have some office in china |
| 23:34.59 | *** join/#arm-netbook tzafrir_laptop (~tzafrir@bzq-218-155-148.cablep.bezeqint.net) |
| 23:35.10 | popolon | good night |
| 23:38.53 | WarheadsSE | Gumboot: n6x in hand |
| 23:39.00 | WarheadsSE | will work on it tomorrow though |
| 23:49.47 | tzafrir_laptop | Any idea how I can tell Aliexpress to not display my purchaces on-line? |
| 23:49.59 | tzafrir_laptop | See my feedback at http://www.aliexpress.com/product-fm/583304547-Smallart-Uhost-mini-pc-Allwinner-A10-hackable-device-wholesalers.html |
| 23:50.22 | tzafrir_laptop | I tried renaming myself, but sadly that did not catch on |
| 23:51.32 | tzafrir_laptop | I'm still named there "Tzafrir" and not "aliexpress-is-bad-for-privacy" |
| 23:52.10 | tzafrir_laptop | I could not find any way on their site to contact them |
| 23:52.50 | *** join/#arm-netbook bsdfox_ (~Bob@unaffiliated/bsdfox) |
