| @juliusb | ok it worked with a fresh ./configure and make | 00:00 |
|---|---|---|
| @juliusb | confused myself with the actual install dir, i have so many toolchain versions sitting around, so need to determine what I did to make it install those libs | 00:00 |
| @juliusb | because I followed the advise of this post: http://lists.opencores.org/pipermail/openrisc/2011-December/000431.html which says to go into bfd/ and run make install-bfdlibLTLIBRARIES | 00:01 |
| @juliusb | and that did something good by the looks of it | 00:01 |
| @juliusb | to get the libbfd.so and libopcodes.so installed | 00:01 |
| @juliusb | got it all working | 00:24 |
| @juliusb | it appears that --enabled-shared installs the libbfd and libopcodes as required | 00:24 |
| @juliusb | or1ktrace all updated | 00:25 |
| @juliusb | and the verilated model now in mor1kx-dev-env works | 00:25 |
| @juliusb | but.. had to do some hackery and trickery | 00:25 |
| @juliusb | will update its docs to reflect this | 00:25 |
| _franck_ | openocd is talking to jtag_vpi :) Now they need to know each other :) | 00:28 |
| @juliusb | _franck_: great to hear :) | 00:29 |
| @stekern | nice job putting together the patch to fix or1ktrace olofk | 05:41 |
| @olofk | _franck_: That sounds great | 07:45 |
| @olofk | Does anyone know how to make a new /dev/tty* device? I've been thinking about making jtag/uart VPI components that present themselves as character devices. One step closer to HW/SW co-simulation | 07:47 |
| @olofk | and of course a /dev/eth* for ethernet | 07:48 |
| @olofk | hmm.. the last one might be weird.. | 07:48 |
| @stekern | olofk: you mean like a "virtual serial port"? | 08:34 |
| @stekern | as for ethernet, tun/tap is probably what you want there | 08:37 |
| jeremybennett | olofk: Take a look at how it is done in Or1ksim | 10:50 |
| jeremybennett | (for Ethernet I mean) | 10:50 |
| @juliusb | jeremybennett: Hi, I have a quick question, at the OpenRISC Project Meeting, you were compiling a list of uses of OpenRISC technology as people mentioned them, was that on the wikipedia page or somewhere on the OpenCores wiki? | 15:12 |
| @juliusb | ah I see it on wikipedia | 15:13 |
| @juliusb | Is that the only place? | 15:17 |
| jeremybennett | juliusb: On Wikipedia as you have found. | 16:51 |
| mor1kx | [mor1kx] skristiansson pushed 5 new commits to master: http://git.io/XUVnlQ | 19:04 |
| mor1kx | mor1kx/master 162dbe7 Stefan Kristiansson: alu: whitespace cleanup | 19:04 |
| mor1kx | mor1kx/master eefeec9 Stefan Kristiansson: alu: avoid nested generate statements... | 19:04 |
| mor1kx | mor1kx/master 9e15261 Stefan Kristiansson: cappuccino/fetch: whitespace cleanup... | 19:04 |
| @olofk | Is it possible to call a verilog function from a VPI callback? | 23:18 |
| @olofk | I guess it should be possible to find the function by traversing the object tree, but is it callable too? | 23:18 |
| @olofk | or1ktrace requires callbacks for getting the gpr and spr values. I can see two workarounds for that for VPI | 23:28 |
| @olofk | 1. Calculate arrays containing values of all GPR and SPR and send them as parameters to the system function. This could work for GPRs, but the SPR map could be pretty big | 23:30 |
| @olofk | 2. Add a few helper functions that returns all the needed SPR and GPR number. Something like src_gpr = $get_src_gpr(current_instruction); get_gpr(src_gpr); or1k_trace(address, instruction, src_gpr, dst_gpr, spr) | 23:33 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!