| *** tpb has joined #timvideos | 00:00 | |
| *** rohitksingh_work has joined #timvideos | 02:47 | |
| *** samsagaz__ has joined #timvideos | 04:09 | |
| *** samsagaz_ has quit IRC | 04:10 | |
| *** CarlFK has joined #timvideos | 04:17 | |
| *** ChanServ sets mode: +v CarlFK | 04:17 | |
| shivm28[m] | shorne: ping? | 06:06 |
|---|---|---|
| *** sb0 has quit IRC | 06:20 | |
| shorne | shivm28[m]: hello | 11:15 |
| shorne | I am not sure how much time I have now, kid just went to sleep, but wife is coming home soon | 11:15 |
| shorne | Did you get what I meant about the du_* signals on the mor1kx? | 11:16 |
| shorne | DO you agree those map to the CDM spec signals? | 11:16 |
| shorne | I think the system interface section needs some help in : http://opensocdebug.readthedocs.io/en/latest/02_spec/07_modules/cdm/ | 11:17 |
| tpb | Title: Core Debug Module (CDM) Open SoC Debug 0.1 documentation (at opensocdebug.readthedocs.io) | 11:17 |
| *** nancy_ has joined #timvideos | 11:22 | |
| *** nancy_ has quit IRC | 11:34 | |
| *** rohitksingh_work has quit IRC | 12:03 | |
| shivm28[m] | shorne: Yes, those signals are same as or1k debugging signals. | 12:21 |
| shivm28[m] | In CDM module, each or1k debug registers is mapped into two 16 bit wide OSD register. Now, I can set breakpoints in my program using GDB. Every time, a breakpoint is to be set up, GDB asks for register access. | 12:22 |
| shivm28[m] | This is accomplished using osd_regaccess_layer (internal system of Opensocdebug). Now, the address of the breakpoint is loaded into the module debug registers over the signals defined in system interface. | 12:25 |
| *** rohitksingh has joined #timvideos | 12:38 | |
| shorne | shivm28[m]: have you deloped the CDM already? | 12:47 |
| shorne | *developed* | 12:48 |
| *** rohitksingh has quit IRC | 13:14 | |
| shivm28[m] | shorne: Nope, that's the first evaluation. | 13:15 |
| shivm28[m] | I am working on specification of the module: https://github.com/opensocdebug/osd-doc/pull/8 | 13:20 |
| tpb | Title: CDM specification content and structure by shivmgg · Pull Request #8 · opensocdebug/osd-doc · GitHub (at github.com) | 13:20 |
| shorne | shivm28[m]: undersstood. | 13:20 |
| shivm28[m] | shorne: cool!! | 13:21 |
| shorne | shivm28[m]: this is a good diagram of the current debug itnerface | 13:21 |
| shorne | http://www.rte.se/blog/blogg-modesty-corex/debugging-openrisc-1200/2.13 | 13:21 |
| tpb | Title: Debugging the OpenRISC 1200 | Realtime Embedded (at www.rte.se) | 13:21 |
| shorne | Notice how the debug itnerface also gives access to memory | 13:22 |
| shorne | will CDM do that? A lot of GDB functionality is actually just 1. stall cpu, 2. read/write to memory | 13:22 |
| shorne | i.e. in GDB we dont use debug registers usually, we just write an l.trap instruction into memory | 13:23 |
| shorne | well, also 3. read/write registers | 13:23 |
| shorne | so the du_* interface helps with 1 and 3 | 13:24 |
| shivm28[m] | I think the overall OSD system will do that. | 13:24 |
| shivm28[m] | Is there any document which explains how GDB interacts with or1k in adv_dbg_sys? | 13:25 |
| shivm28[m] | Yup. And MAM module in OSD helps in accessing memory. | 13:25 |
| shorne | OK, so we will need to use that as well | 13:26 |
| shivm28[m] | So after first two GSOC evaluations, we will be able to have all the GDB functionality. | 13:26 |
| shorne | The link I just sent you is pretty good explaining adv_dbg_sys | 13:26 |
| shorne | But its missing one thing, the current gdb uses openocd | 13:27 |
| shorne | so we have | 13:27 |
| shorne | GDB <-> OpenOCD <-jtag-> [adv_debug_sys] <--> [mor1kx] | 13:27 |
| shivm28[m] | shorne: I will look at it and probably till next, we will have complete specification for CDM module. | 13:29 |
| shorne | this is what the openOCD interface looks like | 13:30 |
| shorne | https://github.com/ntfreak/openocd/blob/master/src/target/openrisc/or1k_du_adv.c | 13:30 |
| tpb | Title: openocd/or1k_du_adv.c at master · ntfreak/openocd · GitHub (at github.com) | 13:30 |
| shivm28[m] | shorne: this is how OSD works: http://opensocdebug.readthedocs.io/en/latest/01_overview/getting_started/index.html#osd-for-run-control-debugging | 13:32 |
| tpb | Title: Open SoC Debug Primer Open SoC Debug 0.1 documentation (at opensocdebug.readthedocs.io) | 13:32 |
| shivm28[m] | more specifically, run control debugging | 13:32 |
| shorne | https://github.com/ntfreak/openocd/blob/master/src/target/openrisc/or1k.c#L1429 | 13:33 |
| tpb | Title: openocd/or1k.c at master · ntfreak/openocd · GitHub (at github.com) | 13:33 |
| shorne | i.e. in openocd, we just define this target_type struction with hooks like .read_memory, .write_memory, .halt, .resume, .step | 13:34 |
| shorne | openocd abstracts the need to write the gdb server too | 13:34 |
| shorne | shivm28[m]: looking at the OSD overview doc, I have seen it before | 13:38 |
| shorne | can you clarify how much of it works? | 13:38 |
| shorne | i.e. does the gdb server exist now? | 13:38 |
| shorne | does a working/skeleton CDM module exist? or MAM? | 13:39 |
| shorne | I looked through the git repo a few times but could never find much completed | 13:41 |
| shivm28[m] | There is no GDB server in OSD yet. | 13:49 |
| shivm28[m] | We do have a MAM module and it works as well | 13:49 |
| shivm28[m] | So, to clarify: | 13:49 |
| shivm28[m] | First evaluation: CDM module | 13:49 |
| shivm28[m] | Second evaluation: GDB server bridge | 13:49 |
| shivm28[m] | OSD supports trace debugging as of now. | 13:49 |
| shivm28[m] | There is no GDB server in OSD yet. | 13:49 |
| shivm28[m] | shorne: all the modules: https://github.com/opensocdebug/osd-hw/tree/master/modules | 13:51 |
| tpb | Title: osd-hw/modules at master · opensocdebug/osd-hw · GitHub (at github.com) | 13:51 |
| *** samsagaz__ has quit IRC | 13:53 | |
| *** rohitksingh has joined #timvideos | 14:03 | |
| *** Kripton has quit IRC | 14:11 | |
| *** thaytan has quit IRC | 14:19 | |
| *** Kripton has joined #timvideos | 14:24 | |
| *** thaytan has joined #timvideos | 14:30 | |
| *** ChanServ sets mode: +v thaytan | 14:30 | |
| *** sb0 has joined #timvideos | 14:36 | |
| *** sb0_ has joined #timvideos | 14:53 | |
| *** sb0_ has quit IRC | 14:54 | |
| *** rohitksingh has quit IRC | 15:20 | |
| *** rohitksingh has joined #timvideos | 16:12 | |
| *** CarlFK has quit IRC | 16:35 | |
| *** rohitksingh has quit IRC | 16:48 | |
| *** rohitksingh has joined #timvideos | 17:06 | |
| *** sb0 has quit IRC | 17:14 | |
| felix_ | rohitksingh: might be a bit off-topic here, but what plug types are used in india? next week i'll fly to mumbai and the week after that to hillhacks near the himalaya and wonder if i'll need a type d or type m adapter or if type f will just fit | 17:16 |
| rohitksingh | felix_: Wow! Didn't know you were coming to hillhacks. If I knew, I would have also come. Anyways, type d plugs should work fine | 17:21 |
| rohitksingh | type m are used for high-ampere sockets, which you wouldn't commonly find | 17:22 |
| felix_ | ok, thx. yeah, in the beginning i didn't really plan to go there (i'm a bit worried about getting some bad stomach infection there thb...), but two friends of mine will go to mumbai to do some workshops there and then visit the hillhacks and so maybe a month ago i decides to join them and go there too | 17:24 |
| rohitksingh | felix_: Oh great! and don't worry about bad stomach infection as long as you eat food from good restaurants (as opposed to road-side stalls) and drink packaged water | 17:26 |
| felix_ | also one of the people organizing that event wanted me to go there, but without the two people traveling with me, i probably wouldn't have booked the flights | 17:26 |
| rohitksingh | oh | 17:26 |
| rohitksingh | good! | 17:27 |
| felix_ | yeah, travelling basically to the middle of nowhere alone and not being able to understand or speak the local language would be quite uncomfortable for me. but one of the people i'll be traveling with has been to india a few times and already knows how stuff works there, so i'd guess that it'll be fun | 17:29 |
| *** rohitksingh has quit IRC | 17:30 | |
| felix_ | https://zkm.de/en/magazine/2018/04/code-for-creativity-open-codes-for-mumbai is the exhibition in mumbai where we'll probably do some workshops | 17:31 |
| tpb | Title: Code for Creativity: »Open Codes« for Mumbai | ZKM (at zkm.de) | 17:31 |
| *** rohitksingh has joined #timvideos | 17:31 | |
| felix_ | rohitksingh: https://zkm.de/en/magazine/2018/04/code-for-creativity-open-codes-for-mumbai is the exhibition in mumbai where we'll probably do some workshops. i don't really know what's the plan there yet... | 17:32 |
| tpb | Title: Code for Creativity: »Open Codes« for Mumbai | ZKM (at zkm.de) | 17:32 |
| *** rohitksingh has quit IRC | 18:21 | |
| *** rohitksingh has joined #timvideos | 19:09 | |
| *** CarlFK has joined #timvideos | 19:29 | |
| *** ChanServ sets mode: +v CarlFK | 19:29 | |
| *** CarlFK has quit IRC | 19:29 | |
| *** CarlFK has joined #timvideos | 19:47 | |
| *** ChanServ sets mode: +v CarlFK | 19:47 | |
| shorne | shivm28[m]: sorry, it was late last night and I fell asleep. Thanks for the update, plese let me know if you need help once you start the CDM module. | 20:55 |
| shorne | You should be able to start it now while you work on the spec, have you started on anything yet? | 20:55 |
| *** shorne has quit IRC | 21:38 | |
| *** CarlFK has quit IRC | 23:06 | |
| *** CarlFK has joined #timvideos | 23:25 | |
| *** ChanServ sets mode: +v CarlFK | 23:25 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!