IRC logs for #openrisc Monday, 2013-03-18

glowplugWe all have an inner Stallman.  8)00:23
GentlemanEngineeQuite.00:24
glowplugI'm not quite sure about the working on the first try thing.  That is even more optimistic than me. And I am *very* optimistic.00:25
GentlemanEngineeI had not so observed...00:25
GentlemanEngineeA further examination of the Memory Controller in the Milkymist Project appears to indicate that, documentation notwithstanding, it is designed for the Spartan 6.00:27
GentlemanEngineeIt is using the primatives: IDDR, ODDR, & IDELAY.00:30
GentlemanEngineeI am uncertain if Altera's offerings have similar primatives available.00:30
glowplugSo it's using interfaces that only exist in Spartan-6 FPGA's?00:30
GentlemanEngineeI do not know if these are only for the Spartan 6. The documentation indicates it is only for the Virtex 4.00:31
glowplugInteresting.00:31
GentlemanEngineePerhaps someone with more experience on a wider subset of FPGAs could shed some light upon this...00:32
glowplugHopefully I can design a board that will work with DDR at all.  :/00:34
GentlemanEngineeI would think that this could be something that others are not offering.00:35
GentlemanEngineeAt least not at that price point.00:36
glowplugI dont know of any boards with DDR.  So I think it has some potential.  =)01:10
GentlemanEngineeOf course, there could be a reason for it.01:10
glowplugHopefully not.  =)01:12
GentlemanEngineeI think it is worth the old college try...01:15
GentlemanEngineeHello simoncook.01:27
GentlemanEngineeHello rolloTomasi.05:46
stekernbah... this crap works perfectly in the simulations... and I'm simulating the exact soc that I load to the atlys board06:08
stekernHAHA, just when I said that, I manage to bring out something that didn't work ;)06:09
stekernmore precisely, running dhry06:10
GentlemanEngineeI had a circuit simulation once that would run on one machine, and not another.06:26
GentlemanEngineeVariety is the SPICE of life, after all.06:34
stekernheh06:44
stekernseems like that was an simulation issue, the program was larger than the amount of data that is loaded into DDR209:52
stekernI just stumbled on a very interesting thread regarding open hardware: http://marc.info/?l=openbsd-misc&m=119935499328116&w=210:33
stekernwhere Richard Stallman is participating, basically saying that Verilog/VHDL should be considered as software10:34
stekernjeremybennett: have you seen that?10:34
stekernit is 5 years old after all ;)10:42
jeremybennettI asked him about OS HW at GNU Cauldron last year. He doesn't get it - he just answered about software. As a counter-point, take a look at: http://www.embecosm.com/2013/03/08/a-license-to-build/11:09
jeremybennettBTW Andrew Katz talk to OSHUG is on YouTube on the Embecosm channel.11:12
juliusbstekern: what bug are you tracking? something in the cappuccino on the board?12:29
juliusbmaybe I read that post by Stallman incorrectly (he's replying, right?) but he seems to understand things pretty well and I agree with his analysis12:31
--- Log closed Mon Mar 18 13:18:00 2013
--- Log opened Mon Mar 18 13:18:16 2013
-!- Irssi: #openrisc: Total of 27 nicks [0 ops, 0 halfops, 0 voices, 27 normal]13:18
-!- Irssi: Join to #openrisc was synced in 20 secs13:18
stekernjeremybennett: I'll read that13:19
stekernjuliusb: yeah, I'm investigating why cappuccino doesn't seem to work _at_all_ on the atlys board, not even the most simple software will print anything on the terminal13:21
stekernit seems like it goes into 0x200 when attaching gdb13:22
stekernall simulations so far works perfectly, now I'm running with the SPI loading bootrom13:23
stekernjeremybennett: but what's interesting is his definition of software (or at least definition of software that should be free), it only applies to devices where the user 'typically' would change what's programmed13:25
stekerni.e. it doesn't matter if it's written in C, verilog, VHDL, basic, whatever. If it's burnt into a ROM, it's not "software".13:37
stekernthen it's a "circuit"13:37
juliusbmmm, RTL simulation, though - try gatelevel13:41
juliusb(gatelevel sucks :-/)13:41
stekernI am trying to avoid that ;)13:42
stekernI rather debug on HW than doing gatelevel13:42
juliusbhaha13:42
juliusbeveryone does13:42
juliusbbut you get limited visibility13:42
juliusbone option is to gatelevel just the mor1kx and plumb that in13:42
juliusbor gatelevel just a part of it13:42
juliusbcheck synthesis logs or warnings13:43
juliusbfor13:43
stekernthe visibility in signaltap is actually pretty good, unfortunately everything works fine on the de0-nano :/13:43
stekernbut I have my homebrewn logger, that's essentially a signaltap light13:44
stekernjust have to dust it off, and this might be a good oppurtunity to write a nice dump-to-vcd converter for it13:45
juliusb:)13:53
stekernbut it has to be something simple and stupid, like some load getting stuck or something like that13:55
juliusbya13:55
juliusb:-/13:55
juliusbah so it's something to do with the memory subsystem you think?13:55
stekernI have no idea what it is at this point, but 0x200 sounds like databusish13:56
stekernthat's the only indicator from the hw that I have at this point13:57
stekernor at least busish ;)14:03
stekernI keep forgetting that ibus can cause bus error exception14:03
jeremybennettstekern: That's the problem - he only deals with "programmable" stuff. Yet I can have a "free" design for a car, in the sense that anyone can take the plans and build their own or a variant.14:06
jeremybennettAnd RepRap is free, but not covered by Stallman's concept.14:06
stekernoh, I completely agree14:06
stekernI'm mostly interested in licensing of "source code" though, and we have had discussions that gpl for example isn't suitable for VHDL/Verilog designs, but the real issue is what kind of end device the compiled source code goes into14:08
juliusbI'm a bit lost after the "normal to install different software" thing, though. I'm not sure what my view is on that. Just because it's not normal to install software on it doesn't mean that a) I shouldn't be able to and b) anything like that which uses open source source shouldn't be exempt from any open source license-type effects of making them release the source code14:54
juliusbRTL design which has gone into an ASIC is exactly the "might as well be circuits" thing. What does he mean by that? That it doesn't matter because he can't then change it (the bit of the chip the RTL affected?)14:55
stekernjuliusb: yeah, that's exactly my view too14:58
stekernwhat I'm interested in is protecting the source, not what is done with it after it is compiled15:01
stekernperhaps that's why I'm not a huge fan of gplv315:01
juliusbI'm not sure why anyone would pay attention to RMS on these matters anyway. The GNU guys haven't ever really focussed on hardware before.15:02
juliusbwell, hardware designs, be they schematics and the like15:03
juliusbalthough, I'm not an expert on the matter, so I could be wrong, but whatever, we see the main efforts for a license in this area coming from people who have, or at least used to have, very little to do with the software open source people like GNU, FSF, and of course I'm talking about the Tuscon guys, the CERN guys (they now talk a lot to the FSF guys I believe), the SolderPad license which was put togeher by Andrew Bac15:08
juliusbmyself, as well, who did a find/replace on the MPL :)15:09
stekernyour previous sentence was cut off at "Andrew Bac"15:13
juliusbah15:15
juliusbwasn't much more15:15
juliusb....the SolderPad license which was put togeher by  Andrew Back who does OSHUG in London15:15
juliusbbut my point is the existing open source institutions seem to be very software-centric (who can blame, them, it's what they were set up to look at), so you can't be surprised that someone like Stallman doesn't understand all of the aspects of hardware design15:21
juliusbHe doesn't speak for all, though, obviously. The guys from GNU and FSFE I meet at things like FSCONS are very encouraging regarding extending the work towards software freedom to hardware15:22
juliusb... but it's still a relatively nascent idea/movement/effort so there's no widespread understanding of what the goal is, I think15:23
amsit is also a bit of "this is my forte"; stallman is a software dude, that is what he does and understands best.  someone else, within the hardware community would be a better leader for freeing hardware.15:24
ams (oh yeah, i hate software today)15:26
amsjuliusb: beer?15:28
juliusbhis view is still valuable, I agree with a bit of his analysis, but I think, as you say, it's not his forte, so we probably shouldn't be too concerned if his view is a little off15:30
juliusbams: certainly! I met up with a mate yesterday with the intentino of drinking a lot of craft beer but we got stuck in London's oldest wine bar for several hours15:30
juliusbI had the obligatory guiness earlier in the day, but no craft beer unfortunately, so I have a defecit of it in my life at present15:30
amsnice15:32
amsi was mostly removing rust from a bicycle i'm fixing up for my gf15:33
juliusb:)15:40
juliusbsounds like you could do with some craft beer, too15:40
juliusbwhy do you hate software today, btw?15:40
amsjuliusb: the web15:52
amsjuliusb: trying to write one of those fancy pancy web "applications" ... failing horribly.15:57
juliusbI don't even know what one of those things are. Way too high up the stack...16:35
glowplugIt seems to me that a license that properly defends rights for any "idea" which is what we defend in patent law would work equally well for music, playing music (physical representation of an idea of music), high level software, RTL, or written RTL (physical representation of verilog in a circuit).17:30
glowplugDrawing a line between an idea and an ideas physical manifestation doesn't seem necessary.17:31
glowplugIf I can sue for playing Free Bird on my guitar in public (possible) then I can sue for creating a circuit based based on a design I saw when I worked at Intel.  If the song or design are under a free license that protects any ideas then they should defend both cases equally.17:33
glowplugIt was my understanding that Patents were the major issue however, not copyright.17:35
glowplugI do like the SolderPad license but it seems that they just added the word "object" to the Apache license.  O_O17:40
stekernhmm, interesting, disabling caches and all seems to be hunky-dory21:26
stekernI assume that disabling icache is enough, since the commit that appears to be guilty only touches that21:28
stekernhmm, no, then a bus error occurs21:48
stekern..on the ethernet mac wb address, and I have disabled the ethernet mac, so that's all normal21:50
stekernwonder why that didn't happen when dcache was disabled though..21:51

Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!