| --- Log opened Thu Sep 26 00:00:13 2013 | ||
| stekern | _franck_: the picture seems to be removed. But to answer your question. no, we shouldn't de-assert cyc nor stb after receiving an ack. | 02:15 |
|---|---|---|
| stekern | If we have data to send/receive, why waste a cycle with that? | 02:17 |
| stekern | it's perfectly according to the wishbone spec to behave like that. if a slave can't handle that, it's buggy. | 02:20 |
| stekern | it's again completely possible that changing DBUS_WB_TYPE to CLASSIC cover up some unrelated bug, just because the timing change | 02:27 |
| stekern | juliusb: that sounds like a good idea | 02:29 |
| stekern | sometimes you really miss VHDL variables... | 03:14 |
| Powermaniac | Howdy! | 04:39 |
| poke53281 | On question stekern. Probably you know it. I have tried to compile strace and I was surprised that there is already one folder called or1k. | 04:54 |
| poke53281 | https://github.com/bnoordhuis/strace/tree/master/linux/or1k | 04:55 |
| poke53281 | So, does it work? I have compiled it but it does not work. More or less the same strange behavior like I had with X. | 04:56 |
| stekern | yes, I think blueCmd upstreamed that | 04:59 |
| stekern | I have never tried it though | 04:59 |
| poke53281 | The syscall numbers are correct. | 05:00 |
| poke53281 | After some time the kernel shoots itself into hell. Syscall handler starts with number 260 executes the ptrace entry part and returns later after a context switch with syscall number 360. This syscall does not exist, so it tries to execute address 0x0. After an ITLB miss the emulator finally had mercy and stops with an unknown opcode :) | 05:11 |
| powermaniac | O_o | 05:11 |
| poke53281 | But I think the error happens a few million instructions before this. | 05:11 |
| poke53281 | I need strace to analyze this ;( | 05:13 |
| stekern | a classic catch 22 | 05:14 |
| powermaniac | Let me guess you need all of that to work to be able to have strace work...Right? | 05:14 |
| poke53281 | Yep | 05:15 |
| powermaniac | So yeah can anoyne confirm my suspicion that I need to learn verilog to further contribute in any significant way? | 05:16 |
| stekern | powermaniac: if digital design is the path you want to go, yes | 05:17 |
| powermaniac | Well it sort of part of the path I want to go down | 05:18 |
| powermaniac | See let me explain further | 05:18 |
| powermaniac | My one huge goal is to make the technological singularity happen. Currently the path to get there seems to be become an electronics engineering and then go into computational neuroscience | 05:18 |
| powermaniac | And then that should hopefully take me where I want to go | 05:19 |
| powermaniac | As it currently looks hardware isn't really built for making a artificial brain in. Or well current hardware/computers. | 05:20 |
| powermaniac | Although it seems IBM atleast is working to fix that | 05:20 |
| powermaniac | (IBM) This is what I'm talking about: https://plus.google.com/100296114230478191916/posts/2cdekM2Pwy8 | 05:34 |
| powermaniac | stekern: So what do you think of all of that? | 05:40 |
| stekern | to go into computional neuroscience, you'll probably need to be polyskilled. | 05:44 |
| powermaniac | stkern: Yeah I was talking to someone who works in the field recently (well via email) and he was the one who told me to go in via electrnic engineering and then pick up bilogy facts along the way | 05:44 |
| powermaniac | As I think that is what he has done | 05:44 |
| powermaniac | Going to see if I can contact that guy in the video somehow, and ask about the Synapse project | 05:47 |
| olofk | Hi | 05:53 |
| powermaniac | Hey! | 05:53 |
| olofk | I read up on the failing uart-simple test | 05:53 |
| stekern | olofk: any input? | 05:54 |
| olofk | The UART in or1200-generic is my dirty, quick wb_bfm based UART model | 05:55 |
| olofk | So TX doesn't move. It just prints out the value on stdout | 05:55 |
| stekern | is that what _franck_ is using? | 05:56 |
| olofk | Also, I'm not even sure that the bfm handles Wishbone classic cycles correctly | 05:56 |
| olofk | Couldn't find any master to test it with | 05:56 |
| powermaniac | I wonder if IBM will let me talk to Dharmendra Modha...Heh...Random student inquiry asking to talk to him haha | 05:56 |
| olofk | And are you really, really sure that cyc can stay asserted between accesses? I'm not sure it handles that either | 05:57 |
| olofk | All in all, this turned out to be a good test for wb_bfm_slave if nothing else :) | 05:57 |
| stekern | yes, I'm sure | 05:57 |
| powermaniac | Did I somehow unexpectedly help again? If so yay! | 05:57 |
| stekern | powermaniac: sorry, no ;) | 05:57 |
| olofk | :) | 05:57 |
| stekern | we moved into another discussion | 05:58 |
| powermaniac | Ahh okay | 05:58 |
| olofk | The UART model hasn't all functionality either. I just did the minimum effort to pass uart-simple and make linux like it | 05:58 |
| stekern | yeah, if that's what causing it, no one is blaiming you ;) | 05:59 |
| olofk | There's an UART wrapper in or1200-generic. The idea was to be able to switch between implementations with a parameter as verilator doesn't handle the model either | 05:59 |
| olofk | But I haven't finished that part yet | 06:00 |
| stekern | I was actually thinking that the uart-test went unexpectedly fast in or1200-generic | 06:00 |
| stekern | what did you mean "not sure if bfm handles wb classic cycles"? how can it not? | 06:02 |
| olofk | hmm | 06:05 |
| olofk | Thinking of it, that code is in wb_bfm_master | 06:05 |
| olofk | The slave should probably be fine | 06:05 |
| olofk | Thinking some more, I had some trouble with a few parameter settings in mor1kx when I tried it. Seem to remember that it was related to DBUS_WB_TYPE | 06:07 |
| stekern | I can't help silently translate bfm to "big f*cking <something>" in my head | 06:08 |
| olofk | :) | 06:08 |
| stekern | like in bfd | 06:08 |
| olofk | I have to consider that if I ever want to write a Bus Functional Generator then ;) | 06:08 |
| stekern | olofk: define "trouble" ;) | 06:10 |
| olofk | Can't remember exactly, but I think it just sent a few wb transactions and then got stuck or something like that | 06:11 |
| olofk | It's probably somewhere in the logs :) | 06:11 |
| olofk | I can see if I find any clues in my local mor1kx-generic system that I was playing with | 06:11 |
| stekern | it's working like a champ on sockit now at least | 06:12 |
| olofk | It might also be that you actually fixed the issue, so I'm not sure | 06:12 |
| stekern | but yeah, there was this bug that _franck_ helped me track down yesterday, where wb_dbus_adr_o would go X | 06:14 |
| stekern | olofk: do you have a plan for passing parameters down to orpsoc_top when running simulations | 06:16 |
| stekern | or can you do that already? | 06:17 |
| stekern | or pass defines or something else | 06:17 |
| stekern | well, not only when running simulations | 06:17 |
| stekern | I'm thinking how we should handle configuration of cores like mor1kx | 06:18 |
| stekern | right now I've just "hardcoded" the parameters in orpsoc_top, but that's suboptimal | 06:19 |
| olofk | I have thought a lot about that, but haven't got a clear strategy yet | 06:19 |
| olofk | Defines can be specified with vlog_options and iverilog_options | 06:19 |
| olofk | But that's about it, right now | 06:20 |
| stekern | I think defines in a *top* top level isn't that bad | 06:20 |
| olofk | Me neither | 06:20 |
| olofk | And simulation parameters are easy. You just set them to false in your orpsoc_top and override them in the test bench | 06:21 |
| olofk | mor1kx configurations are trickier | 06:21 |
| olofk | One way could be to write a wrapper generator, like the one for wb_intercon | 06:22 |
| stekern | as long as you don't name them something stupid like "DEBUG" or "TEST" or something like that | 06:22 |
| olofk | Yes. But we will get rid of those with patches :) | 06:22 |
| stekern | yeah, I'm thinking about how we should transform or1200-generic into or1k-generic | 06:22 |
| powermaniac | Excuse muah but could I connect this to 2x these: https://www.sparkfun.com/products/9947 | 06:22 |
| powermaniac | This: http://saanlima.com/store/index.php?route=product/product&product_id=51 | 06:22 |
| powermaniac | (Yes I'm still trying to come up with ways to make it a very underpowered computer...) | 06:23 |
| olofk | stekern: I think we will have to live with a defines-based specification for or1200 and parametrized for mor1kx for some time | 06:24 |
| olofk | I think that an easy way to start would be to just hard code most settings and have a parameter to select which core to instantiate | 06:26 |
| stekern | olofk: yes, that's not the problem I'm thinking about, I'm mostly thinking how we would go about to make it possible to choose between mor1kx-cappuccino, mor1kx-espresso and mor1kx-prontoespresso | 06:26 |
| powermaniac | LOL at the names | 06:26 |
| olofk | Yes, it's a bit unfair to mor1kx, since that has much more fun stuff to configure | 06:26 |
| stekern | I think it's ok to just have a "hardcoded" default config for each of those three | 06:27 |
| olofk | And it would be a bitch to maintain a common wrapper that exposes all functionality. Especiallly since or1200 defines it's stuff with defines | 06:27 |
| olofk | stekern: That's not a bad idea. I mean, you can still use mor1kx directly if you want all the special features | 06:29 |
| stekern | powermaniac: actually, the best choice as a board for you would probably be the milkymist projects m1 board. But I don't know if they have any to sell anymore | 06:29 |
| powermaniac | stekern: Oh really, and yeah I don't think they sell them anymore =( as I've been on the Milkymist site before | 06:30 |
| powermaniac | But I shall look around some more | 06:30 |
| olofk | But that would make it easy to run comparing regression tests against all four pipelines | 06:30 |
| stekern | olofk: why "but" ;) | 06:31 |
| olofk | s/but/and ;) | 06:32 |
| stekern | ?=) | 06:33 |
| stekern | oh, weird smiley... | 06:33 |
| stekern | powermaniac: if you can live without USB, I'd suggest the atlys board, since that we have orpsocv2 ports for | 06:33 |
| olofk | Yep. It's not a regexp | 06:33 |
| olofk | juliusb: I know, the ORPSoCv3 wiki page is badly out of date. Lack of time :( | 06:35 |
| powermaniac | The atlys might be an option if I wait a while | 06:38 |
| powermaniac | But so will a whole lot more if I wait a while | 06:38 |
| powermaniac | What can you expand with Pmod conenctors? | 06:42 |
| powermaniac | This would've been perfect if they were both USB Host ports: http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=167&No=830&PartNo=3 | 06:47 |
| powermaniac | Also what does a USB Blaster Port do? | 06:49 |
| stekern | that doesn't have ethernet | 06:50 |
| powermaniac | Why is ethernet so important? | 06:50 |
| stekern | usb blaster is jtag connection | 06:50 |
| powermaniac | Okay what is jtag used for? | 06:51 |
| stekern | mainly debugging, but if you search the global network internet you might find more info ;) | 06:52 |
| powermaniac | Ahh okay | 06:52 |
| stekern | ethernet is important because it's an easy way to get fast access to the outer world | 06:53 |
| stekern | why is USB so important? | 06:53 |
| powermaniac | Yes but I'm weird and want to connect to the net via usb wifi adapter | 06:53 |
| powermaniac | Unless you can get an ethernet wifi adapter...? | 06:53 |
| stekern | there exist that kind | 06:54 |
| powermaniac | O_O | 06:54 |
| powermaniac | That opens up more options O_O | 06:54 |
| powermaniac | Wait maybe not =\ | 06:56 |
| stekern | getting a usb wifi adapter to work in for example a bootloader is going to be a lot more painful than just plain ethernet | 06:56 |
| powermaniac | Oh well that is good then, as ethernet is the way to go! | 06:56 |
| powermaniac | Now to cipher through this list of boards: http://tristesse.org/FPGA/CheapFPGADevelopmentBoards | 06:56 |
| stekern | but I know LoneTech have used the usb host core together with an usb wifi adapter under linux, so it should be possible there at least | 06:57 |
| LoneTech | yes, it worked, though the usb core was rather slow. Things like lack of busmaster and only 8-bit interface, certainly fixable. | 07:25 |
| hansfbaier | powermaniac: Why the singularity? I don't think it's a good idea. You will make humans obsolete. | 08:20 |
| hansfbaier | powermaniac: forget about those boards. | 08:21 |
| hansfbaier | powermaniac: I mean the list | 08:22 |
| hansfbaier | it utterly confused me | 08:22 |
| hansfbaier | powermaniac: there are only a few really good deals: | 08:22 |
| hansfbaier | the de0_nano / the cyclong V GX starter kit / the SocKit / the Parallela | 08:23 |
| hansfbaier | s/cyclong/cyclone/ | 08:23 |
| hansfbaier | IMHO | 08:23 |
| hansfbaier | And Waveshares EP4CE10 + modular system maybe | 08:23 |
| powermaniac | hansfbaier: Well I like the idea of the singularity, I think it could answer all of our questions and more...And if it goes along the ideas of Ray Kurzweil's then we might be able to 'become one' with the singularity thus we don't exactly become obsolete we become soemthing else | 08:25 |
| blueCmd | poke53281: strace worked for me, but I never tried to compile it to uclibc | 08:26 |
| powermaniac | hansfbaier: My parents think I like the idea of the singularity because I've suffered from a lot of different illnesses over the 18 years I've been alive, and the idea that it might solve them. I just think that is a bonus | 08:27 |
| hansfbaier | powermaniac: The No. 1 top problem the world has is that we have learned to master the outside world, but not even remotely master our inside world. All current big world problems derive from that. | 08:27 |
| blueCmd | I used a couple of non-upstreamed kernel patches, but they should not be relevant as long as you are not using the atomic syscalls | 08:27 |
| powermaniac | hansfbaier: I'm not disagreeing with that, haven't really thought about it that way yet anyway, but I think the singularity maybe the quick answer to everything although probably very dangerous | 08:28 |
| hansfbaier | powermaniac: the best thing the singularity could teach us is to master our lusts, and get along with each other really well. | 08:28 |
| powermaniac | hansbaier: Did you read this when I posted it? http://forums.parallella.org/viewtopic.php?f=23&t=639&sid=057a65e4a5af2351b417310304ed926f | 08:28 |
| powermaniac | hansbaier: I think it sort of rules out the Parallella for my idea, I'm not sure now whether they did...Think they just said it would be difficult... | 08:29 |
| hansfbaier | stekern: how fast does the or1200/mor1kx run on the sockit? | 08:33 |
| hansfbaier | stekern: Q2: Has somebody done a Timequest analysis of the cores? | 08:34 |
| hansfbaier | powermaniac: Yes the answer of the singularity may be as quick as to dispose of obsolescence | 08:35 |
| hansfbaier | powermaniac: Or keep us around as we do with our cattle/lifestock | 08:35 |
| stekern | hansfbaier: I haven't pushed them, just run it at 50 MHz for now | 08:35 |
| hansfbaier | powermaniac: To build something more intelligent than mankind is a very bad Idea if we cannot even remotely master ourselves | 08:36 |
| hansfbaier | powermaniac: the collective use of Billions of brains is extremely powerful | 08:36 |
| hansfbaier | powermaniac: the main problem is only billions are malfunctioning because of hunger | 08:36 |
| stekern | I kind of agree with hansfbaier, I want my machines to be dumb, efficient and do what I tell them | 08:37 |
| hansfbaier | powermaniac: other billions are malfunctioning because of our (mal)education system and/or mass media stupidification and consumerism | 08:37 |
| powermaniac | hansfbaier: But currently there seems no way to combat the problem caused by whoever really has the most control over the globe | 08:37 |
| hansfbaier | powermaniac: the single most destructive force on this planet is the addiction to wealth, power and status. If untamed it will destroy everything | 08:38 |
| hansfbaier | powermaniac: and we are well on that way | 08:38 |
| powermaniac | hansfbaier: Exactly, I'm sort of of the opinion do whatever I can to combat it, but I myself am not enough | 08:38 |
| hansfbaier | powermaniac: actually the world has only one problem: It is addiction | 08:39 |
| powermaniac | hansfbaier: So did you read my forum post? And if so what did you think as I'm not certain now whether it would work for my idea of using OpenRISC as an as open source as possible computer | 08:39 |
| hansfbaier | powermaniac: If we could find a solution for that I am sure 99% of all our problems will go away | 08:40 |
| powermaniac | hansfbaier: By the way, I love these kinds of conversations, barely ever find people willing to have them =\ | 08:40 |
| hansfbaier | powermaniac: Most of our world today is run by psychopaths/addicts | 08:40 |
| powermaniac | hansfbaier: Don't look at me (I'm bipolar...xD) | 08:40 |
| hansfbaier | powermaniac: Consider this: The money spent on Iraq war would have ended world hunger for 10Years+ | 08:41 |
| powermaniac | hansfbaier: Yeah, I've heard similar things about all the money being spent on spying | 08:41 |
| powermaniac | Sad really | 08:41 |
| hansfbaier | powermaniac: openrisc is as opensource as you can get, if you want to use linux | 08:41 |
| hansfbaier | powermaniac: look at this: http://www.informationisbeautiful.net/visualizations/the-billion-dollar-gram/ | 08:42 |
| hansfbaier | powermaniac: and you know how sick our world really is | 08:42 |
| hansfbaier | my statement about world hunger was an understatement | 08:42 |
| hansfbaier | powermaniac: this picture says it all | 08:44 |
| powermaniac | hansfbaier: Linux is fine, actually was the plan to use Debian although still got to get use to Debian as only recently started using Linux | 08:44 |
| powermaniac | hansfbaier Woah that infographic is interesting | 08:46 |
| hansfbaier | powermaniac: Yes it says a LOT | 08:46 |
| hansfbaier | powermaniac: money makes the world go round / In German: "Geld regiert die Welt" (money rules the world) | 08:47 |
| powermaniac | hansfbaier: True, I always found it odd when I was younger how money worked, well the idea of coins and notes being worth something when it's just a mineral or plastic...or paper | 08:48 |
| hansfbaier | powermaniac: basically it's only a system of keeping score | 08:49 |
| powermaniac | hansfbaier: Well yeah basically | 08:49 |
| hansfbaier | powermaniac: but it doesn't always work that way | 08:49 |
| hansfbaier | quite a bunch of people try to get something for nothing | 08:49 |
| hansfbaier | and those who had this legalized are the biggest players in that game: | 08:50 |
| hansfbaier | The banks | 08:50 |
| hansfbaier | They lend money that does not exist | 08:50 |
| hansfbaier | And earn interest on that from money that exists | 08:50 |
| hansfbaier | It is a legal huge-scale ponzi scheme that will explode some day | 08:50 |
| powermaniac | hansfbaier: Yeah, reminds me of a study some university did where they basically followed the money to discover all the coorporations they control the money flow... | 08:50 |
| hansfbaier | powermaniac: would be interesting to have some infographic on that | 08:51 |
| powermaniac | I think there is a web type thing, let me find the link, I think I still have it | 08:52 |
| mor1kx | [mor1kx] wallento opened pull request #5: DCache: Non-disrupting invalidation (master...master) https://github.com/openrisc/mor1kx/pull/5 | 08:52 |
| hansfbaier | powermaniac: that's a good one oo: http://www.informationisbeautiful.net/visualizations/what-are-wallst-protestors-angry-about/ | 08:53 |
| hansfbaier | stekern: sweden, impressive | 08:54 |
| hansfbaier | powermaniac: The US is a banana republic | 08:54 |
| hansfbaier | ranks below Nigeria | 08:54 |
| powermaniac | This link is always hard to find | 08:57 |
| stekern | can't find finland in there | 08:59 |
| hansfbaier | stekern: Forgive me for bothering you, but I could not find something proper in the internet. Do you know a way to make quartus synthesize parallel pipelind adders instead of this: http://i.imgur.com/IzC6AIb.png | 08:59 |
| hansfbaier | s/pipelind/pipelined/ | 08:59 |
| powermaniac | To find the link I'm looking for I may need to log into my Windows 7 boot | 09:01 |
| powermaniac | So be right back | 09:02 |
| hansfbaier | stekern: finland is near the top | 09:03 |
| Powermaniac | hansfbaier: Here is the link: www.reddit.com/r/science/comments/lif3w/study_finds_that_a_superentity_of_147_companies/ | 09:08 |
| Powermaniac | Now excuse me as I go and have dinner | 09:08 |
| hansfbaier | Powermaniac: thanks. that's interesting | 09:08 |
| Powermaniac | Well this is the link inside that link if you didn't realise: http://www.newscientist.com/article/mg21228354.500-revealed--the-capitalist-network-that-runs-the-world.html | 09:16 |
| Powermaniac | Oh and here is a huge archive of things that may or may not be connected to (mainly NSA related stuff): https://pay.reddit.com/r/moosearchive/comments/1hhjnb/archive | 09:16 |
| hansfbaier | Powermaniac: thanks | 09:25 |
| Powermaniac | Anyway, now on Windows so can't really mess around with ORPSoC even if I wanted to | 09:26 |
| Powermaniac | hansfbaier: You never did tell me if you looked at that thread I linked on the Parallella forum and what you thought? | 09:52 |
| hansfbaier | Powermaniac: I did not read every bit of it, but I think it is you who knows best what you want to do, and thus it is you who has to do the choosing. | 09:53 |
| hansfbaier | Your preferences will change as you learn more | 09:53 |
| Powermaniac | hansfbaier: Well the Parallella seems to be the only board price wise and connection wise that will work for my goal | 09:54 |
| hansfbaier | Powermaniac: probably starting with a simple board like the de0_nano would be a good idea | 09:54 |
| hansfbaier | Powermaniac: get the 7020 if you can | 09:54 |
| mor1kx | [mor1kx] skristiansson closed pull request #5: DCache: Non-disrupting invalidation (master...master) https://github.com/openrisc/mor1kx/pull/5 | 09:54 |
| Powermaniac | hansfbaier: Okay that means more GPIO pins as well which was the plan | 09:54 |
| hansfbaier | Powermaniac: But you will need to get some of those lovely Samtec connectors and SMT soldering skills + routing your own breakout boards | 09:56 |
| hansfbaier | Powermaniac: unless there are some | 09:56 |
| hansfbaier | Powermaniac: I canceled my 7020 parallela order and ordered the SocKit | 09:56 |
| Powermaniac | Done some basic soldering before | 09:56 |
| Powermaniac | SocKit? | 09:56 |
| hansfbaier | Powermaniac: Yes it's killer for that price | 09:57 |
| Powermaniac | Can you link me to the SocKit please? | 09:57 |
| hansfbaier | Powermaniac: I would want to avoid Xilinx. Alteras software is much better | 09:57 |
| hansfbaier | Powermaniac: If you live in the western world you can get one on a $99 seminar for free | 09:57 |
| hansfbaier | Powermaniac: it's more powerful than the 7020 | 09:58 |
| Powermaniac | Oh really | 09:58 |
| hansfbaier | 7020 has 75k cells, SocKit has 155k | 09:58 |
| Powermaniac | I'm in Australia actually | 09:58 |
| hansfbaier | 115k sorry | 09:58 |
| hansfbaier | Powermaniac: If you choose to order over FedEx tell them in advance you don't want to use their "customs clearance service". | 09:58 |
| hansfbaier | They will rip you off big time | 09:58 |
| hansfbaier | they did with me. | 09:58 |
| hansfbaier | the order was $250 | 09:59 |
| hansfbaier | their "service" $100 | 09:59 |
| hansfbaier | on top of the shipping cost of $29 | 09:59 |
| hansfbaier | crooks | 09:59 |
| Powermaniac | Still not sure what the SocKit is xD | 09:59 |
| hansfbaier | bandits | 09:59 |
| hansfbaier | they even got the customs declaration wrong. | 09:59 |
| hansfbaier | and didn't even bother to call me to get it right | 09:59 |
| hansfbaier | said my order was too small for that nuisance | 09:59 |
| Powermaniac | =\ | 09:59 |
| hansfbaier | Now I'm going to kick ass | 09:59 |
| hansfbaier | But unfortunately there is not much I can do. | 10:00 |
| hansfbaier | They have the joker: The package | 10:01 |
| _franck_ | openrisc is now part of the openocd mainline ! Finally ! | 10:01 |
| hansfbaier | Powermaniac: I could have bought a two way plane ticket to Jakarta and did the customs myself for their "service charge" | 10:01 |
| hansfbaier | _franck_: wow, good news! | 10:01 |
| Powermaniac | hansfbaier: WOAH WHAT? O_O | 10:01 |
| hansfbaier | Powermaniac: unfortunately I could not cancel the second order while still at Arrow. They said their orders are final and non returnable. | 10:02 |
| hansfbaier | I already told FedEx I'll pick it up myself by plane. | 10:03 |
| hansfbaier | Those crooks don't get another dollar from me. | 10:03 |
| _franck_ | Powermaniac: openocd is kind of a bridge between gdb and the processor | 10:03 |
| Powermaniac | hansfbaier: So can you link me to the SocKit as I don't know it's full name thus can't find it with a google anyway | 10:03 |
| hansfbaier | Powermaniac: Will use the opportunity to go to a jam session at night | 10:03 |
| hansfbaier | http://www.arrownac.com/solutions/sockit/ | 10:03 |
| hansfbaier | http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&No=816 | 10:04 |
| _franck_ | http://www.altera.com/b/arrow-sockit.html | 10:04 |
| olofk | _franck_: Great job with OpenOCD! | 10:04 |
| Powermaniac | Ohhh that one | 10:04 |
| hansfbaier | Powermaniac: has no digital video though. | 10:04 |
| Powermaniac | Yeah the Parallella is still cheaper overall | 10:05 |
| hansfbaier | Powermaniac: You can add it with a nice $400 expansion board :] | 10:05 |
| Powermaniac | Although with the 7020 it gets closer to the SocKit in price | 10:05 |
| hansfbaier | 7020 is 75K LUT and no Chipscope | 10:05 |
| hansfbaier | sockit 115k | 10:05 |
| hansfbaier | with SignalTap | 10:05 |
| Powermaniac | Not sure what all that means really so... | 10:05 |
| Powermaniac | Although for $299 I could get: http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=165&No=502 | 10:06 |
| Powermaniac | Which is $100 more then the Parallella with the 7020 | 10:06 |
| hansfbaier | Yes but that's only 115k cells | 10:06 |
| hansfbaier | same as SocKit | 10:06 |
| hansfbaier | former is $100 less | 10:06 |
| hansfbaier | $200 less if you attend a seminar | 10:07 |
| Powermaniac | Technically couldn't I buy the 7010 first and then later grab a 7020 board as well, so I could conenct them for more computing power (just for the hell of it thanks to the Epiphany)? | 10:07 |
| hansfbaier | maybe | 10:07 |
| Powermaniac | Where are the seminar's though? | 10:07 |
| hansfbaier | Interfacing is tough work though | 10:07 |
| hansfbaier | https://www.onlineregistrationcenter.com/registerlist.asp?m=135&p=27&group=53&da=1 | 10:07 |
| hansfbaier | https://www.onlineregistrationcenter.com/registerlist.asp?m=135&p=27&group=53&da=1 | 10:08 |
| hansfbaier | http://www.arrownac.com/solutions/sockit/ | 10:10 |
| hansfbaier | or scroll down on last link | 10:10 |
| hansfbaier | it's somewhere there | 10:10 |
| Powermaniac | Okay yeah none of those appear to be hosted in Aus | 10:11 |
| hansfbaier | Powermaniac: see you. Enough time in front of the box for today. | 10:11 |
| Powermaniac | Oh okay | 10:11 |
| Powermaniac | Bye! | 10:11 |
| stekern | I wouldn't recommend the sockit, for the same reasons as parallella. You can't access any of the peripherals from the FPGA | 10:20 |
| stekern | "any", you can access DDR, VGA and AC97 | 10:21 |
| stekern | but not ethernet, sdcard nor usb | 10:21 |
| stekern | I mean, I wouldn't recommend it for what you are looking for Powermaniac | 10:22 |
| stekern | as a development platform for openrisc, I think it's perfect | 10:22 |
| Powermaniac | stekern: Oh okay | 10:25 |
| Powermaniac | stekern: Back to square one then | 10:25 |
| stekern | and there are potential real world uses for it, but that involves running some RTOS on the or1k core performing some time critical tasks | 10:25 |
| Powermaniac | I wonder if there is an arduino board with ethernet | 10:26 |
| Powermaniac | http://arduino.cc/en/Main/ArduinoBoardEthernet + http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=167&No=830&PartNo=3 equals half done | 10:28 |
| Powermaniac | Then I just need to add 2 usb host ports | 10:28 |
| Powermaniac | Which basically brings the cost up to this anyway: http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=139&No=502 | 10:29 |
| Powermaniac | Well the academic cost | 10:29 |
| Powermaniac | That only needs 1 usb port | 10:29 |
| Powermaniac | And a HDMI port | 10:29 |
| Powermaniac | Wait what can GPIO pins be used for? | 10:31 |
| stekern | you'd be better of with the atlys board than the de2-115 (in bang for the buck) | 10:33 |
| Powermaniac | But I thought it was best to go with Altera based boards? | 10:33 |
| stekern | altera have a lot nicer tools, but that's about what's better with altera | 10:35 |
| Powermaniac | So it doesn't really matter then... | 10:35 |
| Powermaniac | Hmm | 10:35 |
| stekern | + I have an atlys board, so you know where to find help getting it up and running ;) | 10:36 |
| Powermaniac | Ohhh | 10:36 |
| Powermaniac | Does the Atlys board have USB Host ports? | 10:36 |
| Powermaniac | As it says Two on-board USB2 ports | 10:36 |
| Powermaniac | But I can't tell | 10:36 |
| stekern | nope, but again, do you *really* need that? | 10:37 |
| Powermaniac | Well only way to connect a mouse and keyboard | 10:37 |
| Powermaniac | I thought I need USB Host for that? | 10:37 |
| Powermaniac | Wait there is a USB-HID port for mouse/keyboard | 10:38 |
| stekern | well, it actually has a PIC32 (or was it PIC24) acting as a USB-HID | 10:38 |
| Powermaniac | So I only need 1 more USB port then | 10:38 |
| Powermaniac | http://www.digilentinc.com/Products/Detail.cfm?NavPath=2,400,836&Prod=ATLYS | 10:38 |
| stekern | but haven't you ever heard of good old trusty PS/2? | 10:38 |
| Powermaniac | On the left of the board in that picture | 10:38 |
| Powermaniac | It has PS/2? | 10:38 |
| stekern | no, but that's a *whole* lot easier to hook up to a pinheader than usb | 10:39 |
| Powermaniac | Oh okay | 10:39 |
| Powermaniac | So is there space to connect a PS/2 connector, as I could use that for keyboard on the Atlys | 10:40 |
| stekern | this is me playing with mouse and keyboard on the atlys board: http://www.youtube.com/watch?v=BRr-BUbBV88 | 10:40 |
| Powermaniac | stekern: You can actually play a LucasArts game on the Atlys board, awesome! | 10:41 |
| stekern | yes, in scummvm under Linux | 10:42 |
| Powermaniac | stekern: So how had you conencted keyboard/mouse in that video to the Atlys? | 10:42 |
| stekern | usb keyboard was connected to the USB-HID connector and a PS/2 mouse was connected to a pinheader | 10:43 |
| Powermaniac | The Atlys board does actually look like the best bang for the buck actually | 10:43 |
| Powermaniac | Oh okay, well I shall get the Digilent Atlys 2 once funds come through | 10:43 |
| Powermaniac | Digilent Atlys* | 10:44 |
| Powermaniac | Hopefully can get it via the academic pricing, as Terasic was going to do that for me atleast | 10:45 |
| Powermaniac | So Digilent I assume will as well | 10:45 |
| Powermaniac | Wait what storage can the Atlys connect to? | 11:06 |
| Powermaniac | Can you connect to a HDD via ethernet? | 11:08 |
| stekern | short answer - yes | 11:08 |
| Powermaniac | Wait I need that for wifi for some reason thought it had x2 ethernet ports | 11:08 |
| stekern | I use a network share from my workstation as the rootfs on the atlys board | 11:09 |
| Powermaniac | Via ethernet I assume? | 11:09 |
| stekern | yes | 11:10 |
| Powermaniac | Hmm any other way to connect storage to the Atlys? | 11:12 |
| Powermaniac | What about that massive 48 pin GPIO port | 11:13 |
| stekern | mmc via spi is fairly simple | 11:14 |
| Powermaniac | Wait no it says it is a 68-pin VHDC connector according to the Atlys reference manual | 11:15 |
| stekern | there are those pmod connectors that are easy to hook-up to | 11:16 |
| Powermaniac | Okay | 11:16 |
| Powermaniac | What can they be be connected to, sata? | 11:16 |
| stekern | I wouldn't | 11:17 |
| Powermaniac | Okay... | 11:17 |
| Powermaniac | How would you connect it to a HDD? | 11:17 |
| Powermaniac | IDE? | 11:17 |
| stekern | if you need massive amounts of storage, network storage is your best bet | 11:17 |
| Powermaniac | Hmm okay | 11:18 |
| Powermaniac | Wait a second | 11:18 |
| Powermaniac | You can't connect to storage and the internet via ethernet at the same time can you? | 11:18 |
| Powermaniac | Like with some expansion board or something? | 11:19 |
| stekern | of course you can | 11:22 |
| Powermaniac | Aha! | 11:22 |
| stekern | just use a router | 11:22 |
| stekern | or switch | 11:22 |
| stekern | or, if you are connected to internet via wifi on your PC and you have an ethernet connector on that, connect it to that and share a folder on the PC as a network drive | 11:24 |
| Powermaniac | Oh okay | 11:24 |
| stekern | and then use the PC as a router between the ethernet port and wifi adapter | 11:24 |
| Powermaniac | Cool | 11:24 |
| Powermaniac | I wonder if there is an open source router for sale | 11:25 |
| stekern | google openwrt | 11:25 |
| flozn | hello guys! does anyone have some experience with minsoc and the i2c_master_top module of opencores.org? | 11:29 |
| flozn | the simulation works fine (set/read registers, interrupts, scl and sda outputs) but the real system throws | 11:29 |
| flozn | bus errors and fails the self-tests :( . | 11:29 |
| stekern | is that the same i2c core that is used in orpsocv2? | 11:37 |
| flozn | hi stekern, i just checked the two file i2c_master_top.v and i2c_master_slave.v . the info header ("$Id: i2c_master_top.v,v 1.12 2009-01-19 20:29:26 rherveille Exp $") and the verilog ports are exactly the same but the content differs. e.g. the i2c_master_byte_ctrl module has some more entries at orpsocv2 | 11:43 |
| flozn | maybe i try to implement the i2c core of orpsocv2 | 11:44 |
| stekern | I would run a diff between those two, there is a big possibility that there are fixes in the orpsocv2 version that hasn't made it back to the core | 11:44 |
| olofk | I see a boatload of patches for orpsocv3 to cover all the orpsocv2 fixes until they are upstreamed | 11:45 |
| stekern | do we have the i2c core in orpsocv3? | 11:45 |
| olofk | I also see myself co-maintaining a truck load of stuff at OpenCores unless someone else does that | 11:45 |
| olofk | stekern: Not yet | 11:46 |
| stekern | ah, yeah, you meant in general | 11:46 |
| stekern | I'm going to send the vga patches to Richard, I'm pretty sure he'll be responsive | 11:48 |
| stekern | which is larger btw, a boat load or a truck load? | 11:49 |
| flozn | that sounds like some good news ;) . i will try the orpsocv2 i2c core! | 11:49 |
| flozn1 | maybe you two can give a tip to another problem of mine :) ? my ./bld-all.sh call for or32-gcc fails with: | 11:51 |
| flozn1 | checking for suffix of object files... configure: error: in `/mnt/data/FPGA/tools/or-gcc/gnu-stable/bd-elf/or32-elf/libgcc': | 11:51 |
| flozn1 | configure: error: cannot compute suffix of object files: cannot compile | 11:51 |
| flozn1 | the according cmd line is: ./bld-all.sh --force --prefix /mnt/data/FPGA/tools/or-gcc/2013-09 --or1ksim-dir /mnt/data/FPGA/tools/or32-elf-1.0rc1-CentOS-x86_64/or1ksim-0.5.1rc1 --no-uclibc --no-or32-linux | 11:52 |
| stekern | seems like "cannot compute suffix of object files" is the error of the week | 11:52 |
| jonibo | stekern: boat > truck... you don't see ports with a large number of boats lined up to ferry off the contents of a single truck | 11:53 |
| flozn1 | "is the error of the week" ... oh. sorry for that :/ | 11:53 |
| stekern | jonibo: during midsummer in sweden you do, but yeah, you're right ;) | 11:55 |
| jonibo | heh... :) clever! | 11:55 |
| jonibo | i'd qualify it then... a "swedish boatload" may in fact be smaller than a "metric boatload"... kind of like the Swedish mile dwarfs the British one! | 11:56 |
| jonibo | ...and in fact smaller than a "metric truckload" | 11:56 |
| jonibo | anyway... | 11:56 |
| stekern | ok, good we got that sorted out ;) | 11:57 |
| jonibo | yup, glad you brought it up ;) | 12:00 |
| flozn1 | "/mnt/data/FPGA/tools/or-gcc/gnu-stable/bd-elf/./gcc/cc1: error while loading shared libraries: libmpc.so.2: cannot open shared object file" ... where may i place the libmpc.so.2 that the build process is able to find it? (btw. i had to build it manually from source because no libmpc package was available for centos6 :( ) | 12:01 |
| olofk | flozn1: Not even libmpc-devel? I'm pretty sure that this problem has come up for centos6 before, and that it was solved | 12:02 |
| jonibo | export LD_LIBRARY_PATH =<whereever you put your libmpc.so.2> | 12:03 |
| jonibo | I think that should do the trick | 12:03 |
| flozn1 | yum search libmpc prints only "musepack audio" ... | 12:03 |
| _franck_ | stekern: so you're telling me you're using the uart16550 core whith mor1kx configured with DBUS_WB_TYPE("B3_REGISTERED_FEEDBACK") and it works ? | 12:03 |
| olofk | _franck_: If you are using or1200-generic you are not using the real uart16550 core. That could be the problem | 12:06 |
| stekern | _franck_: yes | 12:06 |
| _franck_ | olofk: I'm simulating my de1 | 12:06 |
| _franck_ | olofk: I can see tx alive with or1200 not with mor1kx | 12:07 |
| _franck_ | stekern: you're lying :) | 12:08 |
| olofk | _franck_: Ah.. ok. That means I'm innocent (more or less) :) | 12:08 |
| stekern | _franck_: am I? https://github.com/skristiansson/orpsoc-cores/blob/master/systems/sockit/rtl/verilog/orpsoc_top.v#L788 | 12:09 |
| stekern | ;) | 12:09 |
| _franck_ | in uart_wb.v, in the FSM, wre won't come back to '1' unless stb or cyc are '0' that's why on the three consecutive registers write only the first is performed | 12:10 |
| stekern | so, the uart is buggy | 12:11 |
| _franck_ | stekern: ok, you're not | 12:11 |
| _franck_ | it's easy to fix and I'll try tonight | 12:12 |
| _franck_ | http://picpaste.com/pics/Sans_titre-rDtBFfhW.1380197639.png | 12:14 |
| flozn1 | @"(16:03:31) jonibo: export LD_LIBRARY_PATH=.." ... i added the definition to .bashrc and sourced it (error described above occured). now it works :) ! thanks alot :-[ | 12:14 |
| _franck_ | not working on top | 12:14 |
| _franck_ | there should be a picpaste website where you could really paste your clipboard content | 12:15 |
| _franck_ | and not take a screenshot, paste it in your draing program, save as and the send it to picpaste | 12:15 |
| flozn1 | good bye and thanks again! | 12:17 |
| olofk | flozn1: Good lucj | 12:17 |
| flozn1 | thanks :) | 12:17 |
| Powermaniac | Welp it turns out I shall definitely have to wait to get the Digilent Atlys for now as to get it here rounds the cost to $270 | 12:21 |
| Powermaniac | Wait no make that $285 | 12:21 |
| Powermaniac | Hmm, so what other things could I have done with the simulation? | 12:56 |
| Powermaniac | Or is that all so far you can do in the simulation? | 12:57 |
| Powermaniac | Guess I should go learn verilog and come back to this all | 12:57 |
| juliusb | ~Verilog is easy | 13:19 |
| juliusb | you'll be back tomorrow | 13:19 |
| Powermaniac | juliusb: Any recommendations on learning material then? | 13:25 |
| Powermaniac | What can you connect an ethernet port to? | 13:37 |
| Powermaniac | Ahh forget the above | 13:45 |
| Powermaniac | I'll just wait till christmas before buying anything | 13:45 |
| Powermaniac | Are you giving away DE0 Nano's at the Orconf? | 13:51 |
| Powermaniac | Or are they just supplied to use for the conference? | 13:51 |
| Powermaniac | Wait on that note I just noticed something odd | 13:58 |
| Powermaniac | https://orconf2013.eventbrite.co.uk/ | 13:58 |
| Powermaniac | Theres a picture on the right of a DE0 Nano connected to microUSB | 13:58 |
| Powermaniac | What breakout board is that? | 13:58 |
| stekern | uart | 14:11 |
| Powermaniac | Damn | 14:12 |
| Powermaniac | Waiting it is | 14:12 |
| Powermaniac | Thought I found a loophole then | 14:12 |
| Powermaniac | Anyway I think that is all from me tonight | 14:29 |
| Powermaniac | Goodnight everyone! | 14:29 |
| olofk | Time to go home to my disconnected house that will never get a static IP address anyway :( | 14:36 |
| poke53281 | bluecmd and stekern: strace is working with uClibc. | 16:32 |
| poke53281 | I used the newest of the newest kernel patches. And there is a bug somewhere. | 16:33 |
| poke53281 | But it works with the normal branch | 16:34 |
| poke53281 | Thanks blueCmd for the strace implementation :) | 16:34 |
| stekern | poke53281: good to hear, except for the bug in the newest kernel patches | 17:31 |
| stekern | I feel sorry for olofk, it's not gonna feel like a real home until he gets a propoer internet connection | 17:32 |
| poke53281 | The bug should be easy to find for Jonas. | 17:55 |
| blueCmd | poke53281: great! nice that it's useful :) | 18:16 |
| poke53281 | one question to you blueCmd: Is eglibc working? Does it support the same stuff like uClibc at the moment? | 19:03 |
| stekern | poke53281: I think blueCmd brought his patches over to glibc from eglibc and have done some work upon that, since upstreaming would go through that anyway | 19:39 |
| stekern | check his github page | 19:39 |
| blueCmd | poke53281: yeah, glibc sort of works | 19:53 |
| blueCmd | there are a few things left on my todo that are non-trivial, but I'm able to run openssh / apache and so on | 19:54 |
| poke53281 | with shared library support? | 20:28 |
| --- Log closed Fri Sep 27 00:00:14 2013 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!
