IRC logs for #openrisc Monday, 2016-03-14

--- Log opened Mon Mar 14 00:00:33 2016
wallentoshorne: I tend to move away from the wiki and put everything into one repo03:08
wallentoI didn't change the pins for now and updated the software03:09
wallentoand put everything into one repo03:09
wallentoThe plan is to have release tgz of the repo with bitstream and elfs included to allow for a real quick start03:10
wallentoFor a first release it would be nice to include your Linux work03:10
-!- Netsplit *.net <-> *.split quits: stekern, heroux, kaaliakahn03:15
-!- Netsplit over, joins: heroux, stekern03:16
olofkwallento: 15:36 < wallento> olofk: What's the plusarg in other simulators?05:03
olofkCan you explain?05:03
LoneTechplus arguments are a way to pass parameters to a verilog simulation model at runtime, so you don't need to recompile for every different run.
wallentoI mean the name of it05:17
wallentocurrently I named it +parameter+05:17
wallentobut if you say other simulators have it, I was wondering about the name05:22
wallentoLoneTech: I think its compile time here05:23
wallentoRuntime parameter passing is impossible05:23
wallentofor verilator of course05:23
LoneTech ... looks like you have to feed them through a systemc function call05:30
shornewallento: cool, do you plan to link to that from who runs that page?05:31
wallentoactually me05:32
shorneAlso, my work in linux kernel is not much, just merging and build  the work stekern has done.05:32
wallentookay, would be great to get everything into openrisc/linux05:32
shorneok, cool so thats a good place to link it in :)05:32
shorneyes, actually I would like to work on getting it into linux upstream too including de0_nano dts05:33
wallentovery cool05:33
wallentoI wil finish up the current stuff, currently work on the Nexys4 DDR board05:33
shorneI see patches were submitted before, so I was going to ask if its ok if I follow up and try to get them in05:33
wallentoLoneTech: Yes, but you cannot use them to overwrite parameters05:34
wallentoI try to implement it currently, but the Verilator structure has to be changed massively to support this05:34
wallentohence it needs to be compile time05:34
shornewallento: if you are doing to de0 nano tutorial in markdown/git repo ill do the same for a openrisc/gdb tutorial (not including building it though)05:36
wallentoyes, that would be great05:37
wallentoI will just add you to the repo05:37
shorneDo we have recent guide to the toolchain build?05:38
wallentoI just updated it yesterday05:38
shorneok cool, I followed the guide on the mediawiki05:39
wallentoyeah, we should make it clear that this is deprecated05:39
wallentothere were a few bug reports recently for newlib, but to me it seems they were with outdated builds05:39
shornelinux 4.5 was released today... just merged it with openrisc here:
shornetesting right now05:44
shornedoes Jonas still maintain the kernel?05:50
wallentovery good question, maybe blueCmd or stekern know05:53
shornein MAINTAINERS its still saying Jonas (  but it seems that domain is lost05:55
shorneBilling Street: ToproomShinagawa 1015, Kitashinagawa 1-9-705:55
shorne(whois data)05:55
shorneinteresting, that address is like right nextdoor to me05:56
shorneregistered to ultra-domain (sqwaters that is)05:58
wallentoyeah, was shutdown06:14
wallentomaybe we can update maintainers to stekern and you and point it to
wallentoor so06:14
wallentoI have updated with a photo:
wallentoBut maybe it makes sense to permanently move the pins to the top06:17
shorne_what do you mean permanently? the picture looks good06:18
-!- shorne_ is now known as shorne06:19
shornewallento: do you know if the mailing lists at are active?06:22
shorneI havent been able to subscribe06:22
poke53281Yeah, a student who wrote me said something similar. He sent questions to the mailing list but no one received it.06:23
wallentoI am not sure if they are moderated06:23
wallentomaybe also a good chance to move to a new list under librecores.org06:24
wallentoshorne: Where do you put the pins? In the ORCONF2013 tutorial they are moved to the top of the board06:24
shornewallento: I have the pins like in your picture on openrisc/tutorials06:25
wallentookay, cool06:25
wallentothen we stick with this. I think the idea is that people can use the expansion headers for other stuff06:25
shornethats good06:26
shorneI havent even been able to open thst, so might be good06:26
shorneI guess we should wait for a few other people feedback on06:29
shorne1. linux upstreaming maintainer06:29
shorne2. mailing list (do we have one?  do we need one?)06:30
shorne_franck__: thanks, I didnt see that.06:35
shorneit looks like the maintainers entry was not updated06:36
shorneInteresting discussion here too:
_franck__we shouldn't wait to submit patches in stekern's github upstream06:42
wallentookay, to summarize, we should prepare an upstream patch for 4.5 that includes an update to the maintainers file, right?06:54
_franck__wallento: stekern is the one to be asked07:00
_franck__since we want to put his name :)07:01
wallentohehe, of course :)07:01
wallentoby the way _franck__07:01
wallentowhat about gdb upstreaming?07:01
wallentoyou did the majority of the work there07:01
_franck__there was a lot of error in regression testing, I think we should re-run it and see haw bad it is07:02
_franck__I reduced it a bit07:02
shornewallento: yes need to check with stekern, also it will be interesting to see who will be accepting the patches07:03
wallentookay, great, I will also run the regression tests and see07:04
_franck__wallento: great. There is also this patch:
_franck__it touch non target code07:05
_franck__so may be I'll be not that easy to merge07:05
wallentomaybe we can try to get this patch in first?07:05
wallentoat least separate it from target support07:06
_franck__jeremybennett might know if it is a least acceptable07:06
_franck__but yes, as a first step we should do that07:06
_franck__eat time now ! brb07:07
wallentoready for testing, but only decimal parameters allowed currently07:59
wallentoand it would be awesome to allow for hierarchical names07:59
wallentojust build and test verilator +parameter+name=value07:59
olofkRegarding linux maintainer, stekern has tried to get jonas to sign his key several times to officially have him take over as a kernel maintainer08:38
olofkBut it's been hard to reach Jonas lately08:38
olofkI guess the other way would to talk to whoever is coordinating the arch maintainers and tell them that we want to switch maintainer08:39
olofkAnd do we really want to put up the dts files upstream? I would prefer to keep them in each fusesoc system08:39
olofkThat way we can just add or remove i2c controllers and stuff freely, without having to send a kernel patch for it08:40
shorneI see, I was reading about the pgp signing. I guess thats the tricky part08:40
olofkI mean, stekern's de0_nano board is a bit more pimped up than mine :)08:40
olofkFor the mailing lists, I say we put something on librecores so that we can have control over it. I'm dead tired of other people running our mailing lists, and neglect to take care of them08:41
olofkIt would be great if we could get the archives from opencores and though08:41
shorneolofk: I was thinking that as well in the orpsoc-cores would the the other ideal place08:41
shorneLinux openrisc 4.5.0-de0_nano-00142-g63f1d7c #57 Mon Mar 14 18:43:54 JST 2016 openrisc GNU/Linux08:43
shornerunning 4.508:43
stekernshorne: btw, I synced the openrisc/linux repo to 4.408:44
shorneon de0 nano08:44
stekern(since that is the most current release)08:44
shornestekern: I noticed that. I merged it08:44
shorne4.5 was released today :)08:44
stekernaha ;)08:44
shornestekern: what exactly is the pgp key for? linus only accepts pull request it mail is pgp signed?08:45
stekernbasically yes08:46
stekernand you have to have your pgp key signed by other kernel devs08:47
shorneyes, so who signed Jonas in?08:49
shorneI have a few (2) friends/contacts of maintainers08:49
shornemaybe I can ask them for help08:50
olofkIt would be nice to get it going. I want to have my memset patch upstreamed :)08:51
shornefyi, jonas key
shorneI dont know those 209:02
shornestekern: do you have a link to your pgp ket?09:06
shorneHave we tried to just submit patches to Andrew Morton as Jonas mentioned in a mail?09:33
shorneolofk: have you tried to send any patches to the list recently?09:33
shorneanyway, I sent a mail to my friends, lets see what they say09:39
shorneboth have keys signed by greg kh09:39
jeremybennett_franck_: shorne: wallento: It's a useful patch, adding some introspection to the Target Description. There was a good talk on how much more could be done with Target Descriptions at GNU Cauldron 2014.12:12
jeremybennettWorth submitting, if only to see what people think.12:12
jeremybennettI can give Andrew Burgess a heads-up to take a look at it. Not sure if this is an area he concerns himself with.12:13
jeremybennettLet me know when you submit it.12:13
jeremybennettBTW - if there is any of my work remaining in GDB, I'm quite happy to assign it to the FSF. But I think all the real work is from _franck_.12:13
_franck__jeremybennett: most of the target file is from you. I mostly implemented tdesc support. Look at my contribution here:
shorneolofk: do you mind if I do a test of kernel patch submitting with your memset patches? With you in Signed-off-by? I will search before if you have already tried.18:52
--- Log closed Tue Mar 15 00:00:35 2016

Generated by 2.15.2 by Marius Gedminas - find it at!