IRC logs for #openrisc Friday, 2012-11-16

@stekernthe flag == 'x' assert is catching bugs 'like a boss'05:58
_franck_my openocd jtag tcp client stuff already exist :)08:57
_franck_however, I think it is much slower as it sends a packet for each bitbang operation08:59
_franck_every time the JTAG lines are moving there is a packet transfered08:59
-!- Netsplit *.net <-> *.split quits: derRichard11:29
-!- Netsplit over, joins: derRichard11:34
@olofkams, Hi there11:46
@olofk_franck_: jtag_vpi works fine in your orpsocv2. I have probably some connection problem in my orpsocv311:48
_franck_ok, great to hear (for me, not for you :))12:20
@stekernolofk: tried this:
@olofkstekern: That worked great. It seems like I had some dirt in my sockets and tck was a bit rusty, so it didn't move every time13:35
amsolofk: Ciao!13:47
@stekernolofk: just as I suspected!14:38
_franck_stekern: why don't you ever pushed your orpsoc boards ports to opencores svn ?14:38
@stekernI did push the atlys board14:39
_franck_yeah but you also have de0 and de214:40
@stekernyeah, I should probably put that there14:41
@stekernand I've done a lot of changes to the atlys board14:41
_franck_ok, no problem, just wondering14:42
@stekernI should clean up some things before pushing de0 nano14:43
@stekernlike rename that versatile mem define, but IIRC that needs some global changes to the sim scripts14:44
@stekernetc etc14:44
_franck_that why I also don't push the de1 board15:02
_franck_I did try the sim of the de0 board and the process get stopped while modelsim is doing some optimization (AFAIR)15:03
_franck_does it works for you ?15:03
@stekernbut there is some cruft there too15:08
@stekernI might have fixed that, but earlier the adv_debug_sys module was in vhdl and that always failed the first time when compiling15:09
@stekernI've translated it to verilog now15:10
_franck_ok, I'll take another look at this15:11
* stekern is playing exceptions in control stage sudoku16:53
@stekerna bit annoying, I get the following instruction after a lsu error partly executed16:54
_franck_LANG=en && cat test | sed -r "s/CONFIG_CROSS_COMPILER_PREFIX/$CROSS_COMPILE/"16:54
_franck_sed: -e expression #1, char 33: unknown option to `s'16:54
_franck_anyone knows sed ? ;)16:54
@stekernworks here17:26
_franck_export CROSS_COMPILE=franck17:28
_franck_can I tell sed to replace something with a variable value ?17:28
_franck_juste remove backslah before $17:29
@stekernah, I thought you wanted to actually change the text to "$CROSS_COMPILE" ;)17:30
@stekernbah, I'm stuck, I need to prevent execute_valid when lsu is generating exceptions, but I can't do that since then I will prevent padv_execute and the exception will not propagate17:35
@stekernand I can't bypass that for the lsu exceptions, since then there'll be a logic loop17:36
@stekernI'll probably have to create some more special exception handling for the lsu exceptions17:37
@stekernneed a break to digest this...17:39
_franck_works in a shell where export CROSS_COMPILE=bla17:40
_franck_but doesn't when executed in a makefile and CROSS_COMPILE set in the makefile17:41
_franck_oups, it is openrisc channel ! :) ok, I stop spaming now17:49
@stekerndoes it work if you change it to $(CROSS_COMPILE)17:50
_franck_no it says error17:51
@olofkI fucking hate verilog'17:58
@olofkjtag_vpi failed because of a misspelled clock17:58
@olofkThe good news is that it works now in orpsocv3 :)17:58
@olofkDoes anyone know which debug unit that works best? A feature comparasion table would be nice18:01
_franck_good question18:04
-!- aura is now known as Guest3153819:57
Guest31538need some help here please19:58
Guest31538i was using minsoc to implemennt in on spartan 3e19:59
Guest31538there is one unrouted conenction19:59
Guest31538how am i supposed to manually ruote it ?19:59
Guest31538here is the minsoc_par.unroutes file
Guest31538have one more question, how can i view there results in ISE GUI interface? what files should i open with ISE20:00
Guest31538or how do i start the manual routing software from command line ?20:04
@olofk_franck_: I have included jtag_vpi in orpsocv3 now, but I did a few changes to your code.23:57
_franck_ok, tell me23:58
@olofkThe tree is available at if you want to see23:58
@olofkI added two inputs. One is called enable, and controls if the VPI stuff should be run at all23:59
@olofkThe other is called init_done and replaces the hardcoded wait before entering main23:59

Generated by 2.15.2 by Marius Gedminas - find it at!