IRC logs for #openrisc Monday, 2016-11-14

--- Log opened Mon Nov 14 00:00:42 2016
olofkShentino: Not sure what's needed for virtualization, but I would expect that the spec misses some features01:58
olofkVirtualization probably wasn't a focus point when the spec was written01:59
wallentoShentino: I think unfortunately not01:59
wallentoIt does not fulfill the requirements for trap-and-emulate02:00
wallentothat is something we should maybe fix in the next architecture revision02:00
olofkwallento: How the fuck does the signal search work in Icarus?04:14
wallentoI never use Icarus04:42
olofkAh ok. I thought you told me how to do that once04:43
olofkOh, I mean gtkwave of course04:49
olofkBut still, this is the most unusable thing in the program. It's completely impossible to understand how to use it04:49
stekernit's completely backwards, but it works when you learned how it works05:02
stekernI have unfortunately forgot how it worked :(05:03
olofkFound it05:09
olofkI also found that our read function is empty in newlib, so stdin doesn't work05:09
wallentoyeah, that is right05:37
wallentothere is only uart_*05:37
olofkShould fix that06:02
olofkBut I'll poll the uart directly for now. That's all I need for my application06:02
olofkAha. Just found an example in or1k_support.h06:11
shornestekern: Im about to send some patches again for review08:28
shornefor the TLS change if Jonas naks it Ill just complain its already implemented everywhere, and we will get around to the architecture change when we do08:29
shornestekern: interestingly there seems to be some linkage between use and Linus now after keysigning08:39
shorneAnd for wallento its a bit more due to simon08:40
wallentostekern: sorry, I missed your mail08:40
wallentoI will meet some kernel devs this week I hope08:41
stekernmy mail?08:41
wallentoyou asked me if I got my keys signed08:41
shorne_wallento: my mail08:41
stekernah ;)08:41
wallentoah, yeah, sorry :)08:42
stekernI guess s-tab ;)08:42
wallentoyeah, but the brain-tab ;)08:42
shorne_wallento: if you do let it would be great, I was thinking maybe another way is with Video Conf to some people we might know08:43
shorne_i.e. Guenter has been pretty helpful on the list08:43
wallentothe guys here in Munich are all 2nd level08:43
wallentothat will be a great bump08:43
shornehmm, why am I logged in two times08:44
wallentoseems we need pgp for irc ;)08:44
shorneyeah, it will be good help.  Any cool event you are going to meet them at?08:46
wallentoshorne: no, just dropped a mail09:33
wallentoso we will have a beer and exchange keys I suppose09:33
shornewallento: cool I was talking to gregkh right now on linuxnewbies channel.  He said if all fails we can ask him to get patches in.09:39
wallentoah, thats great09:40
shorneAlso the guys over there suggested going through the linux/MAINTAINERS list an looking for .jp accounts09:40
shorneI went through them all, found only one guy in the PGP net09:40
shornemaybe I can ask him for beers09:41
stekernasking to buy beers for him may be a better approach ;)10:03
shorneok, I sent off a mail to another guy in japan10:09
olofkWhat was that pointer alignment problem
olofkI was wondering if it was just a simple endianness problem, but I didn't look at it enough to understand the whole thing13:26
olofkhmm... running the example code in or1ksim, it just exits halfway in through the code13:39
olofkAha! An exception13:58
olofk0x600, so an alignment exception. I assumed we would get some kind of visible warning for that13:59
wbxshorne: gcc 5.4.0-musl , binutils 2.26.1, musl 1.1.15, busybox ash, linux version we talked about. and now my problem is back again.14:00
wbxshorne: the toolchain and the system is build with OpenADK.org14:09
wbxshorne: aha. set -x in my startup script shows it does bootup. just login does not appear on telnet anymore.14:12
olofkwbx: Is openadk some kind of light weight alternative to Yocto or buildroot?15:16
shornewbx, I get up to "System initialization ..."16:56
shorneis that what you see?16:56
shorneI see something strange in my traceL17:40
shornewbx, I see something strange happning with an iTLB lookup. maybe the binary is missing something?17:47
shorneplace its jumping to bad memory location is17:49
shornec0007aa4:       04 02 01 14     l.jal c0087ef4 <find_vma>17:49
shornec0007a98:       04 07 f9 a7     l.jal c0206134 <down_read>17:50
shorneinside kernel do_page_fault17:50
shornebut when it gets there it has...17:53
shorneS c0206134: 00000000 l.j 0x017:53
shorneI can see in your image17:54
shornec0206134:       d7 e1 4f fc     l.sw -4(r1),r917:54
--- Log closed Tue Nov 15 00:00:43 2016

Generated by 2.15.2 by Marius Gedminas - find it at!