| 00:06.33 | Gumboot | I guess I haven't seen anything with eSATA, so a tablet would still never be the computer I need. |
| 00:06.51 | penguin42 | well it depends what you want to do with it doesn't it |
| 00:08.36 | Gumboot | wants dual gigabit ethernet and SATA. |
| 00:09.48 | Gumboot | Pretty much just that, actually. That with enough compute to operate those components efficiently. |
| 00:09.54 | penguin42 | I think some of the Marvell chips can do that - e.g. the Sheeva plug stuff - or their newer chips |
| 00:10.42 | Gumboot | The other thing I wanted was a portable ssh client. An ARM netbook. |
| 00:10.58 | penguin42 | ASUS Transformer |
| 00:11.24 | Gumboot | That's been recommended to me before, actually. |
| 00:12.14 | penguin42 | not cheap though |
| 00:12.23 | Gumboot | I nearly did, but then I saw it was only Tegra 2. No NEON. At the time that kind of offended me. |
| 00:13.23 | penguin42 | nod; the Transformer Primes have the newer Tegra - I assume they have Neon? |
| 00:14.05 | Gumboot | I believe they do, yes. |
| 00:19.44 | Gumboot | I spent a long time looking for netbooks. At the time there were a lot being announced, but they all fell through. Then I started to wonder about a tablet and a chorded keyboard. |
| 00:22.55 | penguin42 | I've seen one add chording onto mobiles |
| 02:29.01 | xvilka | hi! anyone know cheap avr or pic tiny pcb with gpio pins - i need upgrade some locomotive toy? |
| 02:37.52 | A2Sheds | how small does the board need to be? |
| 02:39.54 | A2Sheds | xvilka: the stm32 dev boards are only $10 and up |
| 02:40.49 | A2Sheds | http://pjrc.com/teensy/index.html or these for $16 |
| 02:42.44 | A2Sheds | http://www.futurlec.com/ET-AVR_Stamp.shtml |
| 02:43.16 | A2Sheds | http://www.futurlec.com/ET-PIC_Stamp.shtml |
| 02:45.08 | A2Sheds | http://www.mouser.com/ProductDetail/STMicroelectronics/STM32VLDISCOVERY/?qs=sGAEpiMZZMuqBwn8WqcFUl6U5N3WXcwphmvSj8wh2%252bI%3d |
| 02:45.19 | A2Sheds | $9.88 |
| 02:52.51 | A2Sheds | Slow smartphone? It's not the network, it's NAND flash http://www.computerworld.com/s/article/9224352/Slow_smartphone_It_s_not_the_network_it_s_NAND_flash |
| 02:53.53 | A2Sheds | http://static.usenix.org/events/fast12/tech/full_papers/Kim.pdf Revisiting Storage for Smartphones |
| 03:17.28 | xvilka | thx |
| 11:11.46 | *** join/#arm-netbook gordan (~gordan@78-86-195-86.zone2.bethere.co.uk) |
| 12:53.55 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 14:42.07 | *** join/#arm-netbook mikey_w (~mike@pool-74-110-218-2.rcmdva.fios.verizon.net) |
| 17:19.45 | lkcl | http://www.telechips.com/eng/Product/mobile_pro09.asp |
| 17:20.03 | lkcl | anyone got any contact with telechips? |
| 17:25.44 | lkcl | the TCC8901 looks extremely good |
| 17:25.53 | lkcl | sorry |
| 17:25.54 | lkcl | 8801 |
| 17:25.56 | lkcl | TCC 8801 |
| 17:36.34 | mnemoc | wants one a10 eoma68 module... |
| 17:44.18 | mnemoc | lkcl: what gpu does the ttc8801 have? kernel and library sources? |
| 17:58.22 | penguin42 | lkcl: Ooh that does look nice |
| 17:58.46 | penguin42 | It's got PCI-e |
| 18:00.40 | penguin42 | lkcl: https://www.telechips.com/technical_support/kor/opensource/opensource_list.asp |
| 18:05.23 | penguin42 | looks like it's connected to the internet via a 9k6 line |
| 18:13.06 | A2Sheds | I tried to contact them last year, they are another co that is picky about who they trade parts for money with |
| 18:13.32 | A2Sheds | there are some SDK's floating around out there as well |
| 18:14.03 | A2Sheds | and weren't they another co on the list of GPL violators? |
| 18:15.58 | A2Sheds | http://wenku.baidu.com/view/e7bd54f79e314332396893c8.html |
| 18:16.57 | A2Sheds | password protected android and linux git server |
| 18:18.41 | A2Sheds | wasn' the top speed something like 700mhz for cortex a8? |
| 18:19.49 | A2Sheds | sorry, 1.2GHz for the 8801 |
| 18:20.23 | A2Sheds | http://hi.baidu.com/liulin61914517/blog/item/8cca5bd594c95b0962279855.html |
| 18:25.08 | A2Sheds | http://www.xh-hope.com/at002.html |
| 18:40.00 | penguin42 | that tar has a TCC88xx in |
| 18:44.02 | penguin42 | and a bunch of other tcc stuff - tcc92xx, 892x, 88xx, 93xx |
| 18:46.43 | penguin42 | careful, it has a gpu/mali |
| 18:47.47 | penguin42 | hmm mind you - some of the (c) in that mali directory is (c) 2011 ARM and has a GPLv2 header |
| 18:52.46 | penguin42 | some files are (c) all rights reserved telechips in other parts of the tree |
| 19:49.15 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 20:25.40 | lkcl | A2Sheds: yes strictly speaking they have completely lost all rights under the GPL to even operate as a SoC company, and are now classified as a Criminal Organisation under Copyright Law. |
| 20:26.10 | lkcl | A2Sheds: TI has Cortex A8 running at 1.5ghz |
| 20:28.56 | lkcl | telechips referring people to dhkip.com patent lawyer in korea?? |
| 20:32.10 | Gumboot | What did they do? |
| 20:33.00 | lkcl | caused people to sign NDAs before allowing them access to GPL linux kernel source code |
| 20:33.39 | lkcl | for GPLv2 software, the penalties are automatic and severe: you *absolutely* have no rights - you lose rights under the terms of the GPL. |
| 20:33.50 | lkcl | thus you end up falling back to "default copyright law" |
| 20:33.57 | penguin42 | lkcl: I've seen one person say that you can immediately get those back just by receiving a new copy of it |
| 20:34.00 | A2Sheds | I tried to work with them last year |
| 20:34.00 | lkcl | and that means you have to ask permission of the copyright holders |
| 20:34.05 | lkcl | no, you can't |
| 20:34.19 | A2Sheds | all we could find was some leaked kernel and android sdk's |
| 20:34.27 | lkcl | and that means you have to ask 1,000+ people for permission |
| 20:34.47 | penguin42 | some of whom are dead |
| 20:34.53 | lkcl | penguin42: precisely. |
| 20:34.56 | penguin42 | lkcl: Still, the telechips code seems to be on their site now |
| 20:35.06 | lkcl | or no longer work for the company ... etc. etc. |
| 20:35.09 | Gumboot | There's piles of binary-only stuff in the Linux kernel. How does everyone else get away with it? |
| 20:35.31 | penguin42 | Gumboot: It's normally firmware to run on the non-main cpu |
| 20:35.39 | lkcl | yes, because they got told off by the SFLC |
| 20:35.52 | Gumboot | And the rest? |
| 20:36.03 | lkcl | Gumboot: if people create .o files (ok .ko) - they *don't* "get away" with it. |
| 20:36.12 | lkcl | that's what the "Taints kernel" thing is about |
| 20:36.15 | penguin42 | Gumboot: I don't think there is any binary in the main kernel tree that runs on the main cpu |
| 20:36.56 | lkcl | but, actually, because nobody (linux kernel copyright holders) has *actually* taken any company to task on this |
| 20:37.19 | lkcl | they're running the risk of *never* being able to tackle this ("estoppel defense") |
| 20:37.30 | penguin42 | does have a slightly odd view on this - if I bought a card which had it's own flash storage on and had it's firmware in there noone would ever complain; but if the same firmware was loaded from the host we do - why? It just implies that a board people are happy to use has to spend an extra $5 on flash |
| 20:38.23 | Gumboot | Presumably the bad thing is to release binaries at a more public level than source. |
| 20:38.33 | lkcl | penguin42: this is why debian has the "nonfree" repository. |
| 20:38.53 | Gumboot | Otherwise, every tweak I do to a kernel to make it work right would make me a violator, because I simply can't be bothered to make all my tweaks in a place where they're automatically publicly accessible. |
| 20:40.15 | Gumboot | Flash is still hazardous, of course. Technically even masked rom is hazardous, but that's normally well controlled. |
| 20:40.20 | penguin42 | Gumboot: it's hard fo rcompanies to organise it right; because in practice they don't want to expose an internal git tree, since they want sign off before releasing things and may have stuff that's on a timed-NDA which the source will get released (hopefully at the same time as the binary) |
| 20:41.41 | Gumboot | And fumbling that gets you shitlisted? |
| 20:42.13 | A2Sheds | there's a bit of a difference between somebody tweaking kernel and an ASIC vendor making people sign NDA's to get a Linux SDK and kernel source |
| 20:42.41 | penguin42 | Gumboot: can do - while I think some of the companies are being shit, others just haven't sorted out how to wrangle their source at the same time as their binary |
| 20:47.45 | Gumboot | This is why I only ever ended up contributing to BSD projects. |
| 21:04.25 | lkcl | Gumboot: and look how well that's working out. apple's proprietary kernel is BSD-based. android is apache2-based. microsoft's proprietary TCP stack is BSD-based. microsoft's successful domination of computing is all thanks to DCE/RPC |
| 21:05.22 | lkcl | ok. i'm too damn cold to sit down here (fire's out). waaargh! :) |
| 21:05.24 | lkcl | night folks |
| 21:07.24 | penguin42 | and that's the problem with low power CPUs |
| 21:58.32 | Gumboot | I thought Microsoft eventually rewrote their TCP/IP stack. |
| 21:58.46 | Gumboot | In any case, do we actually need to see its source code? Is it worth seeing? |
| 21:59.25 | Gumboot | Same for OSX. Is there anything really important in its source code? |
| 23:46.01 | *** join/#arm-netbook gordan (~gordan@78-86-195-86.zone2.bethere.co.uk) |