IRC logs for #openrisc Monday, 2014-11-24

--- Log opened Mon Nov 24 00:00:42 2014
olofkpoke53282: I saw your wayland discussion in the backlog. Wayland has been used in production for some time now actually. My Jolla phone from last year uses it07:45
olofkstekern: I like the gource video.07:46
olofkstekern: Oh, by the way. You have experimented a bit with MIDI over UART, right? I'm working on adding MIDI support for my Korg Monotron via an Arduino. How do I make linux send MIDI through the UART connected to the Arduino? Just enable some midi serial driver and set that in Jack?07:48
stekernolofk: possibly, don't know if the midi-serial driver is hw-specific though08:17
olofkmaxpaln: I don't like your BFM. It is mean to my SDRAM controller08:17
olofkstekern: Guess I just have to try then08:18
stekernwhy don't you just make the Korg Monotron midi a proper midi-interface though?08:19
stekernand use a midi-usb converter08:20
stekernto hook it up to your linux machine08:20
olofkI'm planning to do that too, but I need an opto-coupler, and I though it would be handy to use USB directly08:20
olofkThe opto-coupler is the main reason now08:20
stekernremember to use a fast one08:21
olofkDo you know how fast? The bit rate is only ~31250bps, but I might need some headroom to get nice edges I guess08:22
olofkIt's quite easy to buy common electronics stuff like Arduinos, resistors, capacitors and stuff now here in Sweden, but as soon as you need something a little more specific it gets very awkward08:23
stekernyes, it's the edges you have to worry about08:24
stekernI've used an EL817 once, and it was slightly too slow08:24
stekernI could get it to work, but I had to carefully adjust the resistor08:25
olofk10MBit, TTL08:26
olofkAh no. They have closed their store here in Gothenburg, and I'm not going through the hassle of putting in a mail order for one component :)08:28
olofkI'll ask a friend to sneak one out from a lab instead08:29
stekernI checked my old notes, I suggested to myself to replace the EL817 with a HCPL-0701 there08:29
olofkIf I buy a two-channel opto-coupler I can send you half the IC and we can split the cost08:31
stekerngood idea ;)08:32
stekernOT: I found this gem while visiting my sons school:
stekernI would have loved to be on the meeting when decided on the name08:34
olofkYeah. I mean, that even looks like it's a european company.08:36
olofkSo they should have known08:36
olofkWhere do you usually get your cables btw? I always use these guys
olofkI have a photo transistor and some LEDs. I wonder if that would be fast enough08:39
olofkOr maybe I should find some broken MIDI equipment and raid that08:40
stekernman, that's not safe for work...08:40
stekern...people might suspect I'm not doing work when they hear me laugh ;)08:41
mboehnerthey there! is somebody here who is responsible for the opencores homepage & especially the user management?09:57
wallentoORSoC is still in charge there, but I am not sure who excatly10:17
olofkmboehnert: They tend to be quite busy. Anything a non-responsible can help you with?10:18
mboehnert\msg olofk I want to change my email-address in the profile. Two emails within the last weeks got no response at all ...10:48
maxpalnolofk: if its any consolation - it is mean to mine too! :-)11:27
olofkmaxpaln: As long as it's not just my controller I'm ok :)11:36
olofkand great work. I got a few bugs sorted out. Especially read-after-read was completely untested11:36
olofkI want to merge it into the orpsoc-cores tree when I can find time for that. Are you ok with releasing your changes under the LGPL license?11:38
maxpalnyep - of course.11:39
maxpalnANd yeah, the back to back transactions were something that just couldn't be tested with the old BFM. It was where most of my bugs cam from11:40
maxpalnthere were others related to no wait states too11:40
maxpalnfrom my debugging of the linux boot process, this new BFM should be much more representative of the types of transactions that can occur in the real world I think.11:41
olofkmaxpaln: I think that's the wrong way to do it. We should change Linux memory accesses to look like my initial BFM instead11:49
maxpalnlol - ah, I like you're thinking!11:50
olofkI'll send a patch :)11:50
olofkIt's always easier to change sw than hw11:50
amsdepends on the hardware and software..11:54
amsmicrocoded machine vs. non-free software...11:54
olofkysionneau: Are you up for som DFI discussions? :)12:13
ysionneauplease bear in mind that's I'm struggling at the moment with dfi/my ddr controller :p13:03
ysionneauso I'm not an expert of the topic :p13:03
ysionneaubut I managed to do something which works for sdram13:03
poke53282olofk: Those are special environments with only a limited number of features. But I guess, they didn't use an unpatched wayland.17:37
poke53282At the moment it looks like, that the hype is gone.17:41
LimbIs there a specification on the wishbone bus that I can use to help implement the SRAM controller interface with or1k?18:14
LimbOr something I can use to see what I need to do18:15
LimbOh look at that, the wishbone specs have an example18:42
olofkLimb: You should probably also use the wishbone BFM to test your implementation19:51
stekernLimb: in addition to the wishbone spec -
Limbolofk: Can the wishbone BFM be used on the actual boad? or is it for simulation only?22:52
stekernit's for simulation23:01
LimbAh ok23:02
Limb Would that cause the following line where rcr_mode_state==0 is checked to never execute? The register would get updated after, no?23:04
LimbIt's the only thing I can find that might cause julius' code not to work correctly23:04
stekernno, it wouldn't, the reg is updated the next clock23:06
LimbOk, I assumed so but wasn't 100% sure23:07
LimbI wonder if openOCD is causing problems with its clock speed23:11
LimbBut then again, I don't see why that would cause it to work with a lower value of ram space and not a higher one23:11
--- Log closed Tue Nov 25 00:00:43 2014

Generated by 2.15.2 by Marius Gedminas - find it at!