00:15.45 | *** join/#arm-netbook mikey_w (~mike@pool-74-110-218-2.rcmdva.fios.verizon.net) |
00:16.08 | *** join/#arm-netbook mike_ (~mike@pool-74-110-218-2.rcmdva.fios.verizon.net) |
00:35.48 | *** join/#arm-netbook hipboi (~hipboi@221.4.213.95) |
01:06.36 | *** join/#arm-netbook A2Sheds (~ly@unaffiliated/l84supper) |
01:14.55 | toxicpsion | hno: usb dump@http://www.mediafire.com/?5b57786ltrkr5zn |
01:15.11 | toxicpsion | FES 1-2 sent starting packet 225 |
01:15.24 | toxicpsion | FES_0 @ 397-427 |
01:15.33 | toxicpsion | FES_2 starting @ 583 |
01:16.01 | toxicpsion | then the device disconnects, and reconnects @ rev ffff for the rest of the flash. |
01:17.40 | toxicpsion | > http://www.mediafire.com/?5b57786ltrkr5zn ... sorry |
01:19.44 | toxicpsion | probably will illustrate flow control fairly well as windows is in a VM(VirtualBox) on LinuxMint 11 |
01:23.24 | toxicpsion | dumped by wireshark 1.4.6 |
01:36.06 | hno | thanks! |
01:36.28 | toxicpsion | np... let me know what you find out. |
01:46.43 | CIA-46 | rhombus-tech: Dja master * rb0b51ac94d51 /allwinner_a10/orders/Djavdet.mdwn: |
01:46.46 | hno | Looks pretty trivial. |
01:48.45 | toxicpsion | yup AWUS -->AWUC-->FF ->>00-->AWUSBFEX-->START |
01:57.39 | hno | 159 ie fex_1-1, 190 is a response from fex_1-1 |
01:59.34 | hno | hmm. what is in 291? |
01:59.54 | toxicpsion | yes. |
02:05.23 | toxicpsion | i don't know. |
02:05.24 | toxicpsion | 293 is probably sechond chunck |
02:06.10 | toxicpsion | and probably 312, 314 also |
02:29.17 | hno | seems to do some ping-pong of the same data between host & a10. strange. |
02:30.46 | toxicpsion | if you colorize 'usb.capdata contains "AWUC"' and "AWUS" it shows the start/end a bit better. |
02:34.16 | hno | 291,293 sends some data to A10, 312,314 returns same(?) data to host. |
02:36.08 | *** join/#arm-netbook QingPei (~qingpei@inet-hqmc02-o.oracle.com) |
02:39.14 | hno | don't quite understand the framing. Where is it told how much data is coming? Or where this data should go? |
02:41.19 | toxicpsion | probably just a predetermined order for this stage... the flow control in the actual image transfer is much more robust. |
02:42.01 | toxicpsion | it uses FET_MAGIC_DE and FET_MAGIC_CRC a whole lot... |
02:42.26 | toxicpsion | i'm intrigued by 70. |
02:43.06 | toxicpsion | why all the c's -> binary "11001100" |
02:47.40 | hno | Length field identified in AWUC. Actually it's given twice for some reason. |
02:50.56 | hno | 70 is a response, host reading something from the A10. |
02:51.24 | toxicpsion | all it is is alternating 2 byte stripes tho. |
02:51.46 | hno | 67 is the request for 0x0100 bytes. |
02:52.35 | hno | length seen at offset 0x48 and 0x52. |
02:53.08 | hno | type of request at 0x50. 11 for reading, 12 for writing. |
02:53.13 | toxicpsion | 'doh forgetting endianness :X |
02:54.38 | hno | yes, it's little-endian. |
02:54.57 | toxicpsion | also remember the first 0x40 bytes are capture overhead from my usb subsystem/flow control... the actual data starts at 0x40 |
02:55.38 | toxicpsion | s in the actual software the offset ofor request type would be 0x10. |
02:56.40 | toxicpsion | i think? |
03:04.06 | hno | yes |
03:07.31 | hno | nothing obvious telling what the current request is. Have a feeling we may need to extract the BROM to make sense of the protocol. |
03:08.10 | hno | but should be trivial to replay injecting your own initial fes code. |
03:08.23 | hno | fes_1-1, 1-2 and 2. |
04:14.19 | toxicpsion | everything is identical up to 157 after the rev FFFF switch too. |
04:17.45 | toxicpsion | sry... not 157.... 141. |
08:03.53 | zenitraM | hipboi: hi! |
08:04.00 | hipboi | hi |
08:04.00 | zenitraM | can you access pastebin? |
08:04.05 | hipboi | yes |
08:04.14 | zenitraM | i just tried the kernel with AXP |
08:04.16 | zenitraM | this is what happens |
08:04.18 | zenitraM | http://pastebin.com/NbnCY25b |
08:04.42 | zenitraM | going to try the other one now |
08:04.56 | hipboi | which kernel is this |
08:05.10 | hipboi | the one with axp ? |
08:05.14 | zenitraM | yes |
08:05.23 | zenitraM | uImage |
08:07.39 | hipboi | could you load the script.bin as well? |
08:08.16 | hipboi | load script.bin to 0x43000000, the load the uImage |
08:09.47 | zenitraM | mmm do you have the .bin somewhere? |
08:09.58 | hipboi | ok |
08:10.03 | zenitraM | fatload didn't work from nand i think.. |
08:10.11 | hipboi | yes |
08:10.11 | zenitraM | or else i can build it from the .fex that is on the wiki |
08:10.37 | zenitraM | this is without axp: http://pastebin.com/X0Cry7TQ |
08:10.40 | hipboi | there is no nand driver of the mmc-boot branch |
08:11.03 | zenitraM | it seems to work to a later point.. although half of the drivers aren't initalized |
08:11.07 | hipboi | good |
08:11.14 | hipboi | the second one looks normal |
08:11.29 | hipboi | just no script.bin |
08:11.33 | hipboi | i mount root failed |
08:11.44 | hipboi | let me find the sys_config of mele box |
08:11.51 | zenitraM | okay! |
08:15.27 | zenitraM | what would be disabled by having no axp? suspend? |
08:16.17 | zenitraM | btw voltage here is 220v, checked with a multimeter and the adapter works fine at 5V |
08:18.33 | hipboi | http://dl.dropbox.com/u/65312725/sys_config1.mele.fex |
08:18.59 | zenitraM | that is uncompiled right? |
08:19.02 | zenitraM | i'll compile it |
08:19.08 | zenitraM | to get the .bin |
08:19.25 | hipboi | yes |
08:21.55 | CIA-46 | rhombus-tech: chris master * r543d65d289ab /allwinner_a10/orders/Impactcs.mdwn: |
08:25.53 | zenitraM | okay, here it is: http://pastebin.com/QysejG3Z |
08:27.01 | zenitraM | it seems to boot fine..and ignore the nand root I set (is it nandb?) |
08:28.38 | hipboi | sun4i#set mmc_root /dev/nandb |
08:28.54 | hipboi | you just set the variable mmc_root |
08:29.11 | zenitraM | oh, setargs is set at startup? |
08:29.18 | hipboi | [ 0.000000] Kernel command line: mem=448M@0x40000000 console=ttyS0,115200 |
08:29.26 | hipboi | not passed to the kernel |
08:29.29 | zenitraM | i thought it was set when loading the kernel. that explains a lot |
08:29.50 | zenitraM | right, will try.. although i'm not sure if android will like the 3.0.8 kernel |
08:30.02 | hipboi | ok |
08:31.24 | zenitraM | [ 1.940000] VFS: Cannot open root device "nandb" or unknown-block(0,0) |
08:31.28 | zenitraM | it is not detecting.. the nand? |
08:32.28 | hipboi | are you sure the nand is not corrupted? |
08:32.58 | zenitraM | i tried using a normal u-boot |
08:33.20 | zenitraM | and it detected the nand partitions and booted from the nand |
08:33.32 | zenitraM | i'll try it again just in case... |
08:33.44 | hipboi | ok |
08:36.26 | zenitraM | here: http://pastebin.com/DXmsaXJD |
08:36.43 | zenitraM | it finds the partitions, reads the kernel, and then reboots |
08:36.50 | zenitraM | s/reads/starts/ |
08:38.08 | zenitraM | [ 0.790000] [NAND] nand_set_module_clk fail |
08:38.08 | zenitraM | [ 0.790000] [NAND]init_blklayer fail |
08:38.33 | hipboi | mmm... |
08:38.46 | hipboi | maybe the clk has somethin with the axp |
08:39.08 | hipboi | btw, can you try the kernel again with sys_config.bin ? |
08:39.17 | hipboi | the one with axp |
08:39.24 | zenitraM | ok gonna try |
08:45.16 | zenitraM | mmm seems to work too |
08:45.17 | zenitraM | http://pastebin.com/bAT7Mjgn |
08:45.29 | zenitraM | but still [ 0.790000] [NAND] nand_set_module_clk fail |
08:45.53 | hipboi | let me look into it |
08:46.16 | zenitraM | some axp errors.. |
08:46.24 | hipboi | yes |
08:52.04 | *** join/#arm-netbook QingPei (~qingpei@inet-hqmc02-o.oracle.com) |
08:54.43 | hipboi | i think the tv box can work well without axp |
08:55.00 | zenitraM | probably |
08:55.40 | zenitraM | but then the problem is nand |
08:55.46 | hipboi | yes |
08:56.58 | zenitraM | it fails to set the clock even with axp enabled |
08:57.37 | zenitraM | mmm when using nand u-boot, it still doesn't load sys_config, right? |
08:58.14 | hipboi | you can do some test on nand with the nand u-boot |
08:58.15 | zenitraM | i could try to boot the nand kernel loading sys_config first |
08:58.27 | hipboi | i think the nand is ok |
08:58.44 | *** join/#arm-netbook lkcl (~lkcl@host86-142-136-60.range86-142.btcentralplus.com) |
09:01.04 | zenitraM | ohh |
09:01.15 | zenitraM | i loaded sysconfig from u-boot |
09:01.16 | zenitraM | it boots |
09:01.23 | zenitraM | i think |
09:03.38 | *** part/#arm-netbook QingPei (~qingpei@inet-hqmc02-o.oracle.com) |
09:04.28 | zenitraM | yes, it boots android… and something keeps crashing on android |
09:05.10 | zenitraM | but i now have a serial console at least |
09:05.20 | zenitraM | (linux shell i mean) |
09:08.54 | zenitraM | this is the android log: http://pastebin.com/fD89p25g |
09:13.24 | hipboi | [ 0.780000] [ccmu] set nfc clock rate to 60000000 failed! |
09:13.37 | zenitraM | nfc? |
09:13.38 | hipboi | i was wondering if this happens everytime |
09:14.01 | hipboi | nfc clock shouldn't be set to 60M |
09:14.11 | hipboi | usually 20M |
09:14.39 | zenitraM | boot log when using the built-in kernel (on recovery mode): http://pastebin.com/XKvDtvCY |
09:16.50 | hipboi | very strange |
09:16.57 | hipboi | it looks normal |
09:18.09 | zenitraM | it seems to work.. but something makes android crash |
09:18.11 | hipboi | can youecho fantasy > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor |
09:18.33 | hipboi | echo fantasy > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor |
09:18.44 | hipboi | to see if axp works |
09:21.37 | zenitraM | i did.. nothing seems to happen |
09:22.20 | hipboi | any print from kernel? |
09:22.22 | hipboi | echo 8 > /proc/sys/kernel/printk |
09:22.35 | hipboi | should be a lot of print |
09:22.39 | zenitraM | i got the normal kernel to print stuff now |
09:22.39 | zenitraM | http://pastebin.com/xeW9Fj4H |
09:23.52 | zenitraM | you can see the fantasy there |
09:23.55 | zenitraM | i tried twice in fact |
09:24.25 | zenitraM | looks like.. it is suspending? |
09:24.45 | zenitraM | [ 62.990000] wlan0: no IPv6 routers present |
09:24.45 | zenitraM | [ 72.180000] request_suspend_state: wakeup (0->0) at 72721517436 (2010-01-01 00:02:04.109901180 UTC) |
09:24.45 | zenitraM | [ 72.230000] +rtw_drv_halt |
09:24.45 | zenitraM | [ 72.230000] ##########rtw_suspend_lock_uninit########### |
09:24.45 | zenitraM | [ 72.280000] usbcore: deregistering interface driver rtl8192cu |
09:24.46 | zenitraM | [ 72.280000] ###=> usb_read_port_complete => urb status(-108) |
09:24.46 | zenitraM | [ 72.290000] +rtw_dev_remove |
09:24.47 | zenitraM | [ 72.290000] rtw_set_ps_mode(): Busy Traffic , Leave 802.11 power save.. |
09:26.05 | hipboi | i am afraid you have to send it back |
09:26.38 | zenitraM | :( oh well |
09:26.49 | zenitraM | any idea of why is this happening? |
09:26.56 | hipboi | no... |
09:27.21 | hipboi | can you run a system on mmc? |
09:27.58 | zenitraM | it is indeed quite strange... |
09:28.27 | zenitraM | i can, i think.. but i'd prefer if everything worked... |
09:28.50 | zenitraM | because with all those random failures.. if the axp is broken/outputting wrong voltages |
09:29.06 | zenitraM | it might go kaboom some time |
09:29.42 | hipboi | ok |
09:29.48 | hipboi | you can send it back |
09:29.54 | hipboi | i will send you a new one |
09:30.27 | zenitraM | thank you :) |
09:30.37 | zenitraM | i'll try to send it tomorrow.. do you need the whole box? |
09:30.43 | hipboi | now |
09:30.55 | hipboi | just the box is ok |
09:31.06 | hipboi | lighter |
09:31.28 | zenitraM | yes.. |
09:31.32 | zenitraM | have to go to class now |
09:31.35 | hipboi | ok |
09:31.37 | zenitraM | back in 15 min from there |
09:31.46 | hipboi | i will send you my address |
09:31.55 | zenitraM[brb] | thanks for all your help tom :) |
09:32.08 | hipboi | i have to say sorry |
09:32.33 | hipboi | for the quatity of made in china |
09:38.14 | *** join/#arm-netbook eflatun (~eflatun@109.228.222.95) |
09:47.30 | zenitraM | hipboi: don't worry, everything can have failures… |
09:53.16 | zenitraM | if they didn't then there wouldn't be engineers to fix it :) |
09:57.45 | *** join/#arm-netbook kadaki01 (~textual@wifi-82-167.uc3m.es) |
09:59.30 | zenitraM | thank you again for all the help.. and sorry for the hassles |
10:50.59 | *** join/#arm-netbook gordan (~gordan@178.100.190.102) |
11:27.33 | CIA-46 | rhombus-tech: arnaud master * r27142f2c9836 /allwinner_a10/orders/baso.mdwn: |
12:01.12 | *** join/#arm-netbook QingPei (~qingpei@123.118.99.239) |
14:04.59 | *** join/#arm-netbook hipboi_ (~hipboi@123.89.249.135) |
14:05.22 | hipboi_ | i have a question |
14:06.15 | hipboi_ | how to pronounce eoma? |
14:07.29 | hipboi_ | should i just pronounce e-a-m-a? |
14:07.48 | hipboi_ | s/e-a-m-a/e-o-m-a |
14:09.05 | hipboi_ | or əʊmə |
14:11.14 | MaDMaLKaV | quick question, the A10 only supports one SATA device right? |
14:14.07 | menomc | i pronounce it e-oma |
14:14.24 | menomc | like grand mother in german :p |
14:23.32 | hipboi_ | i will pronounce like that tomorrow |
14:32.30 | A2Sheds | EOMA is only the name for the standard, we can think of cute and fuzzy names for the actual cards |
14:33.04 | lkcl | :) |
14:33.27 | lkcl | hipboi: like nee ho maah except without the "N" :) |
14:33.48 | lkcl | nee ho maah this is eo ohh maa :) |
14:34.04 | A2Sheds | maybe the A10 card should be called "da da ba ba" |
14:34.11 | QingPei | who will talk about EOMA tomorrow? |
14:34.12 | A2Sheds | big bi daddy |
14:34.56 | A2Sheds | QingPei: daddy = baba? |
14:35.12 | A2Sheds | what kids called their fathers |
14:35.30 | QingPei | bari: yeah... |
14:35.35 | QingPei | daddy == baba |
14:36.40 | A2Sheds | the "baba" board |
14:36.51 | A2Sheds | it's cute and catchy! |
14:41.58 | lkcl | baba bing, baba boom. |
14:42.26 | lkcl | QingPei: hipboi, at beijing LUG |
14:44.33 | QingPei | oh, cool...i will join that... |
14:52.00 | CIA-46 | rhombus-tech: Dietmar master * rebb4109f6b9b /allwinner_a10/orders/robertogecco.mdwn: |
14:58.33 | A2Sheds | how many preorders are there? |
15:16.24 | lkcl | A2Sheds: about 400. |
15:16.59 | lkcl | lkcl@teenymac:~/src/rhombus/allwinner_a10$ ls orders|wc |
15:16.59 | lkcl | <PROTECTED> |
15:17.09 | lkcl | A2Sheds: about 474. |
15:17.31 | A2Sheds | almost 500 |
15:17.37 | lkcl | yeah. |
15:18.14 | A2Sheds | thats more embedded boards than some OEM's sell in a year :) |
15:18.21 | lkcl | even at alpha-grade if the price is $30 there are 356 units |
15:18.31 | lkcl | (assuming zero NREs that is) |
15:18.45 | lkcl | at $35 it drops to 269 people |
15:19.00 | lkcl | at $50 it's still 180 alpha-grade units |
15:20.02 | lkcl | for "stable", if we can get down to $20 it's 2228 units for goodness sake (that includes the 1,000 you wanted) |
15:20.24 | lkcl | even at $30 it's 1840 units |
15:21.02 | menomc | any idea when (timeframe) you'll be ready to start making the alpha boards? |
15:21.06 | lkcl | if you git checkout a copy of the web site you can do cd..... |
15:21.11 | lkcl | menomc: connectors! :) |
15:21.32 | lkcl | f****g a****e f*****g connectors! |
15:21.39 | menomc | how long will you keep searching for them? until 2020? |
15:21.40 | A2Sheds | by july at the very latest |
15:21.44 | lkcl | jumps up and down |
15:21.51 | menomc | A2Sheds: thanks :) |
15:21.51 | lkcl | hell no. |
15:22.11 | lkcl | at some point i'll get pissed off and just say "rrright, that's it, we're using MOLEX". |
15:22.22 | A2Sheds | it's like kids in the back seat "are we there yet?" |
15:22.42 | A2Sheds | after the 37th time you just start making things up |
15:22.46 | gordan | Err... July??? |
15:22.48 | lkcl | but the MOLEX Micro HDMI Type D price is 3x higher ($0.80 or so) |
15:22.53 | gordan | That's another 4 months! |
15:22.58 | lkcl | february 2012. |
15:23.00 | A2Sheds | end of march |
15:23.02 | lkcl | no, 2013 |
15:23.10 | lkcl | end of march 1984 |
15:23.14 | A2Sheds | April 15th |
15:23.17 | A2Sheds | May1 |
15:23.24 | lkcl | lol |
15:23.28 | gordan | gives up and goes away again |
15:23.44 | A2Sheds | april 5th, 6:30 local paris time |
15:24.02 | lkcl | it'll happen when it happens. bari can get the PCB done within a few days. ellen can get the first samples out in about a week. |
15:24.09 | A2Sheds | we are only looking for the last 2 connectors |
15:24.19 | gordan | What I fear is that by the time this gets to the point of availability it's going to be sufficiently deprecated that nobody will care. :( |
15:24.28 | A2Sheds | we might already have everything |
15:24.48 | lkcl | QingPei has found the micro-sd, ellen's found usb-ab (mid-mount) and the PCMCIA |
15:24.53 | menomc | gordan: sadly agree |
15:24.55 | A2Sheds | gordan: yeah, that bad part of having open discussions on this |
15:25.03 | lkcl | that just leaves micro-hdmi and the audio connector |
15:25.07 | A2Sheds | people ask the same questions all the time |
15:25.15 | lkcl | gordan, menomc : help find parts! |
15:25.28 | menomc | I already lost 80% of my interest in this project |
15:25.30 | lkcl | look around for 3.5mm mid-mount 4-pole audio jacks |
15:25.34 | A2Sheds | even after explaining the situation several times |
15:25.56 | A2Sheds | menomc: and irritated other devs here along the way |
15:25.58 | lkcl | menomc: i don't like it either! |
15:26.00 | gordan | OK - what is the actual problem? Availability of parts that will match the physical requirements or finding them cheap enough? |
15:26.11 | lkcl | gordan: both. |
15:26.31 | A2Sheds | all for what 100 boards? |
15:26.39 | menomc | A2Sheds: you mean that guy called memleak? |
15:26.41 | gordan | If it's about the price, then I think adding $1-2 to the BoM is not exactly the end of the world. |
15:26.47 | lkcl | i've done a diagram explaining why SMT (3mm high) Micro-HDMI D isn't ok |
15:26.59 | lkcl | gordan: yes, we're coming to that conclusion. |
15:27.06 | A2Sheds | menomc: for one |
15:27.20 | gordan | Once you have some alpha/beta boards that people can actually use and play with, you can continue searching for better/cheaper parts for Mk2+ boards. |
15:27.21 | menomc | A2Sheds: if you read his mails you'll notice HE wasn't interested at all in sharing his work porting the kernel |
15:27.38 | A2Sheds | he posted his entire tree!!!!!!!!!!!!!!!!!!!!!!!! |
15:27.53 | menomc | A2Sheds: an luke commented on his "I, I I, me, me, me" style mails |
15:27.57 | lkcl | http://hands.com/~lkcl/oops.pcb.2.png |
15:28.26 | A2Sheds | yes "he" wanted to help |
15:28.51 | lkcl | menomc: it's really important that people in free software projects avoid personal pronouns. |
15:29.13 | lkcl | i may not have explained that too well, but... |
15:29.17 | lkcl | anyway. |
15:29.19 | menomc | :) |
15:29.27 | A2Sheds | that so ridiculous, anal |
15:29.28 | gordan | Is on-board HDMI really that important on a PCMCIA card? |
15:29.57 | lkcl | gordan: *hand-waving*.... *sigh* well think about it: we've said to people that this can operate as a stand-alone computer. |
15:30.01 | gordan | Will anyone actually use this with HDMI out (bearing in mind there is serial console) without putting it on an expansion header? |
15:30.18 | lkcl | if it's a stand-alone computer that doesn't have HDMI out, it's gonna be a bit limited. |
15:30.19 | gordan | Yeah, just add PSU, case, USB hub.... |
15:30.19 | A2Sheds | he's a kid that wanted to help, but on the good side he got to see what people act like in the real world |
15:30.52 | gordan | By the time you add PSU and USB hub, well, it's not really standalone. |
15:31.00 | lkcl | gordan: ok, ok :) |
15:31.17 | gordan | And if you're adding a USB hub, you don't even need HDMI to make it a full computer - you can use a MIMO screen. |
15:31.21 | lkcl | but you don't need to buy special parts. |
15:31.30 | A2Sheds | stand alone, without display is still stand alone |
15:31.34 | lkcl | ok, all right :) displaylink etc. |
15:31.39 | gordan | Apart from a PSU and USB hub? |
15:32.04 | gordan | Most of my servers have a serial console, and they're pretty standalone. :) |
15:32.23 | lkcl | A2Sheds: actually, gordan's got a good point. how about doing the first boards just with pads for a breakout board to HDMI? |
15:32.41 | lkcl | forget HDMI for the first revision (and audio) |
15:32.50 | lkcl | just to get.... something! |
15:32.52 | gordan | \o/ |
15:33.01 | gordan | Exactly! |
15:33.02 | A2Sheds | it's unfortunate the personality disorders and mental illnesses get in the way of discussions and cooperating |
15:33.14 | lkcl | we've got the micro-usb AB, we've got the ... thingy... hoojamaflip. |
15:33.17 | gordan | I mean seriously, HDMI? Main use cases are probably closer to the LVDS use-case. |
15:33.43 | lkcl | hey! who are you calling wotsit jiggumy memory-retarded, mr whoever-you-are? :) |
15:33.47 | A2Sheds | or pads to a higher priced HDMI that can be changed later |
15:33.53 | lkcl | teehee |
15:34.31 | gordan | How about a "different HDMI that doesn't delay the project indefinitely"? |
15:35.03 | gordan | If we need more pins, what about using a smart-card style connector set on top? |
15:35.03 | A2Sheds | RMS stopped in the coreboot channel once, the place went absolutely crazy with school yard behavior |
15:35.03 | lkcl | gordan: http://hands.com/~lkcl/oops.pcb.2.png |
15:35.36 | gordan | lkcl: I saw that, hence my balking at HDMI |
15:35.47 | lkcl | you have to reduce the PCB thickness to about 0.8mm and that's starting to get expensive |
15:35.50 | lkcl | ahh ok |
15:36.05 | A2Sheds | didn't qingpei send a lower profile HDMI? |
15:36.13 | lkcl | A2Sheds: *grin*.... no :) |
15:36.22 | gordan | My point is that you can just add smart-card style top facing pads. Cover the whole card with them if you must. |
15:36.27 | lkcl | the factories kept on sending him standard SMT ones |
15:36.37 | gordan | If we need more pins than PCMCIA provides. |
15:36.59 | lkcl | A2Sheds: what do you think of that idea? i'm for it. |
15:37.09 | lkcl | +1 to gordan |
15:37.30 | lkcl | yeah i know, it means a 2nd set of NREs on the PCBs. |
15:37.46 | gordan | That way if the recipient slot has top connectors, great, it can hook those features. If not, as long as things don't short out, it doesn't matter. |
15:37.53 | lkcl | but to be honest there's enough people so that the costs would be covered. |
15:38.35 | lkcl | i can feeeel your tiiny braiiin munching this ideaaa, A2Sheds :) |
15:39.07 | A2Sheds | "smart-card style top facing pads" ? do you mean smt pads without any parts stuffed? or something else? |
15:39.07 | lkcl | i'm not sure yet if it's spat it out whole, having found it crunchy and sour |
15:39.36 | A2Sheds | pokes lkcl in the eyes 3 stooges style |
15:39.36 | lkcl | something that people could potentially solder to (easily enough) |
15:40.01 | lkcl | arm swings round on auto-pilot reflex and bangs gordan on the head |
15:40.28 | lkcl | oops sorry about that gordan :) |
15:40.55 | A2Sheds | let me know what you come up with |
15:41.05 | A2Sheds | bbl couple hours |
15:41.24 | lkcl | A2Sheds: ack |
15:41.42 | A2Sheds | so we just ned the low profile HDMI right now? |
15:44.09 | A2Sheds | http://imagebin.org/203038 |
15:44.38 | A2Sheds | if this pic isn't clear enough to explain HDMI midmount.... |
15:44.45 | A2Sheds | change professions |
15:45.03 | QingPei | i have a pic which would explain the mid-mount things clear.. |
15:45.04 | QingPei | wait |
15:45.43 | A2Sheds | hey, that HDMI he sent is mid mount |
15:46.13 | A2Sheds | takes lkcl's ME license away |
15:46.23 | QingPei | http://imagebin.org/203039 |
15:46.30 | A2Sheds | I thought it was |
15:46.44 | QingPei | take a look at the USB mid-mount pic.. |
15:46.57 | A2Sheds | yeah, that should spell it out pretty clear :) |
15:47.12 | QingPei | it would help to understand the difference between normal SMT and mid-mount things |
15:47.22 | A2Sheds | ok, so we have a HDMI midmount |
15:47.58 | A2Sheds | I'll look over all the rest to be sure |
15:51.15 | A2Sheds | http://imagebin.org/203043 |
15:52.20 | A2Sheds | now really have to run.............. |
15:57.47 | zenitraM | really needs to buy a monitor with HDMI. |
16:04.29 | *** part/#arm-netbook QingPei (~qingpei@123.118.99.239) |
16:09.25 | hno | interesting times. board getting ready, fel getting reversed, mali getting reversed, |
16:15.27 | gordan | A2Sheds: By smart-card style top facing pads, I actually mean the sort of thing you see on the chip on a smart card. |
16:15.58 | gordan | The way things interface to it is just by applying contacts directly to the pad. |
16:16.07 | gordan | Like on a SIM card. |
16:16.56 | gordan | You make a bunch of such pads exposed on top, for the host slot to optionally interface with. |
16:17.13 | gordan | You could probably get 40+ pins out that way quite easily. |
16:17.41 | gordan | Just have to make sure that touching all the pads together doesn't short out anything, e.g. if the slot is lined with tin. |
16:32.58 | *** join/#arm-netbook eflatun (~eflatun@109.228.222.95) |
16:41.11 | specing | zenitraM: Suspend works the way it should, I think |
16:41.25 | zenitraM | nice |
16:41.29 | specing | I just did echo mem > /sys/power/state (kernel suspend interface) |
16:41.56 | specing | and then you have to hold the power button for about 3s after which the android locked screen comes up normaly |
16:43.17 | zenitraM | mmm that seems a *deeper* suspend that the android one |
16:43.47 | specing | Yes, it suspends everything except external interrupts and DRAM controller |
16:43.58 | zenitraM | nice |
16:44.01 | specing | It is the version desktop computers use |
16:44.27 | specing | There is another mode in $(cat /sys/power/state) called "suspend" |
16:44.31 | specing | I'll try it now |
16:47.06 | specing | err |
16:47.11 | specing | Its nammed "standby" |
16:51.38 | *** join/#arm-netbook mikey_w (~mike@pool-74-110-218-2.rcmdva.fios.verizon.net) |
16:51.49 | *** join/#arm-netbook mike_ (~mike@pool-74-110-218-2.rcmdva.fios.verizon.net) |
17:02.06 | specing | works like mem |
17:04.00 | specing | Hmm |
17:04.16 | specing | Could this tablet have GPS? |
17:04.25 | specing | It has a gps module in lsmod |
17:06.19 | menomc | specing: the a10 soc has built-in gps, but not even allwinner has access to the kernel module source |
17:08.26 | gordan | Surely if we have access to all the registers and the hardware spec we don't need the source, we can write our own driver. |
17:14.44 | specing | menomc: Does that mean it is usable or not? |
17:14.56 | specing | puts his tablet out in the open |
17:15.22 | menomc | specing: only with violator blessed kernel builds |
17:15.35 | menomc | specing: or as gordan suggested, reverse engineering it |
17:17.09 | specing | mnemoc: Im just testing android atm |
17:17.15 | specing | to see what Im up against |
17:18.04 | zenitraM | what? the A10 has GPS?? |
17:18.21 | zenitraM | now that's something I didn't know |
17:18.46 | specing | gps 112874 0 - Live 0xbf0f5000 |
17:18.50 | specing | straight from lsmod |
17:18.59 | specing | There is also /dev/gps |
17:19.40 | *** join/#arm-netbook gordan (~gordan@178.100.190.102) |
17:19.44 | zenitraM | i'll check with my Novo, i didn't remember that module |
17:20.57 | mnemoc | hipboi said the violators, http://www.highvision.com.cn/, check out their git tree, compile the module and send the .ko to them |
17:21.05 | zenitraM | although it is strange - if it had GPS, all of the tablet manufacturers would be using it |
17:21.31 | mnemoc | antenna problems? |
17:21.52 | mnemoc | even big brands have troubles with proper gps support |
17:22.04 | zenitraM | it's not that they really care about their features working |
17:22.13 | zenitraM | the *wifi* already has antenna problems |
17:22.23 | mnemoc | :) |
17:22.35 | lkcl | zenitraM: yes, there's 10 or 12 pins or so which directly support *certain* types of GPS R.F. front-ends |
17:23.54 | lkcl | i don't know if you're aware of this but GPS is a bitch. you have to use correlators to work out from the *timings* received (in nanoseconds) from the GPS satellites, based on their known orbits (!) to find out where you are |
17:24.06 | lkcl | and the amount of processing power required to do that is just insane. |
17:24.16 | lkcl | but, once you've got a quotes lock quotes |
17:24.44 | lkcl | the amount of processing power required to work out a little bit of variation is really quite small |
17:25.09 | lkcl | so you actually need a really quite good CPU just to get the initial lock! |
17:25.10 | zenitraM | that's why you use external gps chips that just spit out the NMEA |
17:25.12 | lkcl | stupid |
17:25.31 | lkcl | zenitraM: ah, but those external gps chips have to have a really quite powerful on-board CPU! |
17:25.55 | lkcl | which adds to the cost |
17:25.57 | lkcl | _but_ |
17:26.06 | lkcl | if you can just receive the raw GPS data |
17:26.20 | lkcl | and use the ARM Cortex's NEON instruction set to do the calculations.... |
17:26.34 | lkcl | you've just saved .... i dunno... $5 on the cost? |
17:26.37 | lkcl | $3? |
17:26.42 | lkcl | don't honestly know |
17:26.52 | lkcl | but the whole point of the allwinner is to reduce the BOM. |
17:27.08 | zenitraM | but that would have to be done on software… and nobody has bothered to implement it on a working project. i see |
17:27.14 | zenitraM | s/project/product/ |
17:27.19 | lkcl | yes, it would. |
17:27.34 | lkcl | i know someone who would be perfectly competent to do it. |
17:27.37 | lkcl | *sigh*. |
17:27.52 | lkcl | actually i know someone who _has_ done it - about.... 12 years ago! |
17:28.22 | lkcl | he wrote his own FPU routines using an 8-bit processor. took over 2 minutes to get a lock, but it was fine once that happened. |
17:28.27 | lkcl | it's just maths, zenitraM |
17:28.51 | zenitraM | too much maths for chinese programmers. |
17:28.51 | lkcl | richard lightman. bloody genius. and extreme introvert, bless him. |
17:29.12 | lkcl | he disappeared off the face of the earth in about 2008. |
17:29.25 | specing | To where? |
17:31.23 | lkcl | i have absolutely no idea. |
17:31.50 | specing | CIA? |
17:31.54 | lkcl | it's like he doesn't exist any more. his web site, nezumi.plus.com, contained some really valuable code such as depinit |
17:32.39 | lkcl | naah - unfortunately i know that he was working on a project for a friend of mine, and he was struggling to get it sorted. he gave up and disappeared. |
17:33.53 | lkcl | he worked for skyguard technologies for a few years, they did a small security device, hand-held, which contained a panic button, GPRS and a GPS. |
17:34.23 | lkcl | he _was_ going to do the GPS on the on-board ARM processor, but the "manager" didn't trust him. |
17:34.47 | lkcl | actually this "manager" tried to jeopardise the entire project, so that he could "look good" having blamed richard and the team |
17:35.05 | lkcl | and then "rescue" it by bringing in his own people. |
17:35.50 | lkcl | that idiot's decision-making cost them about £450,000 and delayed the project by 2 years, which was long enough to destroy it completely. |
17:36.06 | lkcl | grrr |
17:36.14 | lkcl | but yeah. |
17:36.15 | lkcl | anyway. |
17:36.56 | lkcl | richard was - is - one of the few software engineers who i'd consider to be competent enough to do GPS entirely in software, but i'm sure there are other people who've done it. |
17:38.51 | lkcl | http://www.home.earthlink.net/~cwkelley/ |
17:38.53 | lkcl | there you go. |
17:39.21 | lkcl | soft GPS. |
17:56.45 | specing | Hmm |
17:56.56 | specing | cat-ing /dev/gps produces nothing |
18:20.54 | A2Sheds | http://code.google.com/p/lemoncurry/wiki/main new pic of printer kit |
18:22.16 | A2Sheds | heh sorry wrong channel |
18:22.26 | A2Sheds | anyway, it looks like we have the connectors |
18:22.40 | lkcl | gooood |
18:22.56 | A2Sheds | the gold plated smy pads don't work well |
18:23.21 | A2Sheds | same problems as sim cards, they always get loose |
18:23.50 | A2Sheds | but it's low cost and if you have no choice then they are fine |
18:24.36 | mnemoc | specing: shouldn't /dev/gps be a serial device? |
18:24.45 | A2Sheds | if you guys want a banana installed or a pez dispenser I'll put it on, just be sure you want to live with it :) |
18:25.10 | specing | mnemoc: I guess |
18:25.29 | specing | A2Sheds: Consider using someone else than google for your project hosting though |
18:26.02 | specing | A2Sheds: They actively prevent people from their banned-country list from accessing it |
18:26.11 | A2Sheds | specing, for the stm32 stuff or?? |
18:26.27 | specing | Whatever you have |
18:26.34 | specing | also sourceforge is bad |
18:26.58 | A2Sheds | yes, you never know when they will pull the plug |
18:27.37 | A2Sheds | we need to have a repo for the openec-32 |
18:27.41 | lkcl | A2Sheds: ask phil (remind him) asking if he'd be happy to host the openec-32 |
18:27.45 | A2Sheds | somebody will get around to it |
18:27.49 | lkcl | yes. send phil a quick reminder |
18:28.09 | A2Sheds | Dear Phil, |
18:56.56 | ka6sox | ... |
19:13.16 | MaDMaLKaV | late to the party, but if the first boards doesn't have hdmi, at least give us some not-to-hard soldering points for it |
19:24.39 | A2Sheds | http://git.hands.com/openec-32 |
19:41.59 | CIA-46 | rhombus-tech: Anton master * r93638dac5991 /allwinner_a10/orders/cortexpower.mdwn: |
19:42.24 | specing | What are these CIA updates? |
19:42.53 | zenitraM | orders made = changes on the wiki = changes on the git repo behind the wiki |
19:43.09 | specing | ah |
20:18.45 | CIA-46 | rhombus-tech: bari master * rca3ff8c8a067 /gsoc2012/ideas.mdwn: |
20:30.54 | mnemoc | http://www.chinavasion.com/hdd-media-players/highdefinition-1080p-hdd-network-media-player/ ... "boxchip f10"... $46+shipping ... I've seen allwinner been called boxchip, and f10 might be a a10 typo |
20:31.28 | mnemoc | quite blackboxed for a devkit but might be good to have after having functional software pieces |
20:34.24 | CIA-46 | rhombus-tech: lkcl master * rde708bbc3e41 /allwinner_a10/orders/ (Impactcs.mdwn robertogecco.mdwn): order processing cleanup |
20:37.43 | mnemoc | my mele is finally in .se :) |
20:49.42 | lkcl | hooraay |
20:50.51 | MaDMaLKaV | mnemoc , not a typo, allwinner have a f10 chip that is a mediaplayer cpu with nothing of ARM of it afaik |
20:51.27 | mnemoc | oh :( |
20:51.40 | mnemoc | too good to be true :) |
20:59.25 | hno | allwinner / boxchip have quite many SoC chips. Only A10 is interesting to us. |
21:01.01 | hno | Oh, if I am lucky there is a mele tomorrow. It's in Stockholm somewhere. |
21:01.49 | mnemoc | oddly no multicore cortex-a9 yet |
21:02.53 | hno | multicore a9 has a bit higher fab requirements, and also licensing cost I think. |
21:04.57 | CIA-46 | rhombus-tech: Jos master * rb729950d2680 /allwinner_a10/orders/CLBBEMP55G.mdwn: |
21:09.52 | MaDMaLKaV | the a10 can only support one sata device, right? |
21:10.34 | CIA-46 | rhombus-tech: Jos master * rd3d1f27c15ff /allwinner_a10/orders/CLBBEMP55G.mdwn: |
21:10.36 | MaDMaLKaV | I mean, without using multipliers and that kind of stuff |
21:20.11 | hno | MaDMaLKaV, I think so yes. |
21:21.33 | hno | which is one more than most of it's competitors. |
21:24.34 | MaDMaLKaV | yeah, sure, just wanted to be sure that in all that shitloads of connections on the soc there were no additional sata ports |
21:48.54 | mnemoc | A2Sheds: from a eoma-module cost perspective, does it make sense to build a card with an fpga? |
21:49.39 | A2Sheds | sure |
21:50.00 | A2Sheds | many FPGA's are low cost |
21:50.21 | mnemoc | from what i've seen here the cost of cpu is a relatively small fraction of the cost of the module |
21:57.40 | ka6sox | A2Sheds, what would you use the FPGA for? |
22:02.28 | A2Sheds | whatever fits, stepper or servo PID controls, test equipment, emulators, IO expansion, controllers of all sorts |
22:02.38 | A2Sheds | http://www.xilinx.com/applications/index.htm some applications |
22:04.30 | mnemoc | what aprox. % of final price difference between the a10-based module and this fpga-based module do you expect? |
23:06.07 | CIA-46 | rhombus-tech: lkcl master * re630eb126788 /faq.mdwn: |
23:15.43 | CIA-46 | rhombus-tech: FrankiR master * ra252e1e5883d /allwinner_a10/orders/frankiR.mdwn: |
23:34.43 | *** join/#arm-netbook ManoftheSea (~dlahouss@72.22.81.52) |
23:55.11 | CIA-46 | rhombus-tech: Giacomo master * rc074e198531a /allwinner_a10/orders/jaimemistus.mdwn: |