| --- Log opened Sat Sep 06 00:00:40 2014 | ||
| poke53281 | sstekern: Indeed xcb-proto was at the wrong position | 00:56 |
|---|---|---|
| sb0 | by the way, if someone feels like putting mor1kx on a serious diet, I have some budget for that | 03:57 |
| rjo | ;) | 04:19 |
| poke53281 | stekern: freera works, but only with around two fps. And it does not like very stable. | 04:46 |
| poke53281 | also it takes forever to load. But Ok, this is kind of expected. Around 30-40MB currently. | 04:47 |
| poke53281 | 'B+ | 04:48 |
| poke53281 | But I guess that this freera version is not very optimized. | 04:48 |
| poke53281 | Or they use floating point math | 04:50 |
| poke53281 | toppler runs also with 2 fps. Damn, this is a remake of a game from the C64. | 05:24 |
| poke53281 | Now I really wonder why prboom is so fast. Something in SDL? | 05:25 |
| poke53281 | And dillo runs theoretical too. It starts, but complaints about missing fonts :) | 06:11 |
| poke53281 | fluid from fltk has the same problem. | 06:11 |
| poke53281 | Wow, this were productive 3 hours. Could compile and run 4 programs. Three of them really big. | 06:12 |
| stekern | poke53281: and in the meanwhile, I haven't been able to compile a working X | 06:33 |
| stekern | I got mouse and kb working at the same time with your build though | 06:34 |
| stekern | the only thing that I do differently is that I used xcb 1.10, maybe that's the problem | 06:36 |
| hesham | FYI: I am able to run some RTEMS applications successfully on or1ksim, verilator, icarus and qemu (with clock driver and UART console) | 12:25 |
| hesham | Two issues: For QEMU can not it be renamed to qemu-system-or1k? And for Verilator, how can I manage the clock rate? It's too slow (like for ever) to run simple hello world RTEMS sample. | 12:27 |
| stekern | hesham: how do you mean manage the clock rate? | 12:55 |
| hesham | For example I configure or1ksim to run at 100MHz, what clock rate does mor1kx-generic runs at? | 12:56 |
| stekern | 50 MHz iirc | 12:57 |
| hesham | I got weird ASCII symbols from Verilator in result of running hello world character, can this be something related to the clock? | 12:58 |
| hesham | I configured the SW code to run at 100MHz | 12:58 |
| stekern | but increasing that to 100 MHz (for instance) will not make verilator run faster | 12:58 |
| stekern | yes, that might be related | 12:59 |
| hesham | I just need to get the string character right, no matter it's slower or faster. I will build RTEMS again with 50MHz and see what I will get. | 12:59 |
| stekern | this is the only place (AFAIK) where the 50 MHz assumption is made: https://github.com/openrisc/orpsoc-cores/blob/master/systems/mor1kx-generic/rtl/verilog/orpsoc_top.v#L268 | 13:05 |
| hesham | From SW I have to set the baud rate with an equation including Clock frequency, but even with 50MHz I got the same behavior. RTEMS app is 32MB memory map and from OpenOCD/GDB, it reaches the end of the program as expected. | 13:15 |
| hesham | This is what I got from Verilator: "�������..." | 13:15 |
| olofk | Looking for some basic info on the or1k arch.I know I have had slides that tells it's load/store, 32-bit, and describes the ISA classes, but I can't find them | 13:36 |
| poke53281 | stekern: I think I forgot to tell you a tiny detail. | 14:24 |
| poke53281 | I have never run X in the sysroot. | 14:24 |
| poke53281 | The problem is, what you install and what I don't. I don't install any fonts or encodings. | 14:25 |
| poke53281 | If I install a font, the booting of X takes forever. | 14:26 |
| poke53281 | The best way would be, to make a sysroot out of those .tar.bz2 files. That's the only tested way. | 14:27 |
| poke53281 | But I have to solve the font problem anyhow. Otherwise dillo won't start. | 14:28 |
| poke53281 | You should also take a look in filesystem.make. Sometimes there are also some necessary patches included, like copying of files, touching files and directories. | 14:30 |
| poke53281 | I exchange startx with my own file for example. | 14:30 |
| poke53281 | and the starting script forthe X desktop is in /etc | 14:30 |
| poke53281 | I should change my script in order to make the sysroot consistent with jor1k. But at the moment this is just a tar -xvjf *.tar.bz2 away | 14:32 |
| poke53281 | Ahh, finally *dillo* works | 15:21 |
| poke53281 | and not that slow | 15:21 |
| stekern | poke53281: SDL under X works fine at least (e.g. scummvm) | 16:54 |
| poke53281 | is it faster than the direct framebuffer solution? | 17:12 |
| poke53281 | Dillo is really fast. Definitely usable. | 17:13 |
| poke53281 | I have updated my official version one minute ago | 17:13 |
| stekern | it's probably about as fast | 17:17 |
| poke53281 | the solution to the font problem is btw to copy any small font in the TTF directory. | 17:21 |
| poke53281 | Then, everything works. | 17:21 |
| poke53281 | for example decker.ttf from directfb | 17:21 |
| stekern | (dillo) that's pretty cool, running a browser in a browser ;) | 17:22 |
| poke53281 | I had this before with "links -g" But dillo is a little bit better. | 17:22 |
| poke53281 | firefox is still far away. :( | 17:23 |
| stekern | do you have a nice recipe for it? | 17:24 |
| poke53281 | links or firefox? | 17:24 |
| poke53281 | or do you mean dillo? | 17:25 |
| stekern | dillo | 17:25 |
| poke53281 | yes, one moment. You will get brand new build scripts. No dri problem anymore := | 17:25 |
| poke53281 | :) | 17:25 |
| stekern | did you get to the bottom of the 'glibc' problem? | 17:26 |
| poke53281 | well, I compile with static libgcc. So far it works very good. | 17:26 |
| poke53281 | so, all files updated | 17:30 |
| poke53281 | you have to install fltk and dillo | 17:30 |
| poke53281 | I made also changes everywhere. Updated some X libraries and so on | 17:31 |
| poke53281 | Look in the readme | 17:31 |
| poke53281 | But if you just want to run dillo, look at fltk and dillo | 17:31 |
| poke53281 | part in X.make | 17:31 |
| poke53281 | fltk is a little bit tricky | 17:32 |
| poke53281 | I haven't wrote an sed line yet. | 17:33 |
| stekern | poke53281: I needed this for prboom: --with-sdl-prefix=$(PWD)/sysroot/usr | 17:33 |
| poke53281 | It's written in X.make | 17:34 |
| poke53281 | you have to remove "test" somewhere in the makefile | 17:34 |
| poke53281 | stekern: That's possible. I have changed my sdl-config file in the global-scope because of another game. | 17:35 |
| poke53281 | but I added the line | 17:36 |
| poke53281 | You have to remove "test" from Makefile DIRS variable in fltk | 17:41 |
| poke53281 | sed -i 's@fluid test documentation@fluid documentation@' Makefile | 17:48 |
| poke53281 | This should do it | 17:48 |
| stekern | poke53281: why did you use scons for freera++? | 17:53 |
| poke53281 | because this was the first I saw in the folder :) | 17:53 |
| poke53281 | Then I tried to change to the Makefile. But sdl makes problems I think | 17:54 |
| poke53281 | if you just want to try I would suggest you take it from my packages folder | 17:55 |
| poke53281 | Before you try it by your own | 17:55 |
| poke53281 | you start with /usr/share/games/freera/freera | 17:55 |
| stekern | I get a "Bus error" after "Loading pallete: SNOW | 18:05 |
| stekern | " | 18:05 |
| poke53281 | ? | 18:05 |
| poke53281 | Didn't check for alignment | 18:05 |
| stekern | when I try to start freera | 18:05 |
| poke53281 | maybe you can remove snow.mix | 18:06 |
| poke53281 | not all .mix files are necessary | 18:08 |
| poke53281 | Is is an alignment issue? | 18:09 |
| stekern | no, bus error is bus error | 18:09 |
| stekern | maybe it doesn't like my screen setup with 640x480-8 | 18:10 |
| poke53281 | change freecnc.ini | 18:10 |
| poke53281 | I changed 480 to 400 for my framebuffer | 18:11 |
| poke53281 | The standard is 16 bpp. This can be also changed in that file. | 18:11 |
| stekern | ok, I'll poke at that and try with -16 | 18:11 |
| poke53281 | I can prove that it works for me :) | 18:17 |
| stekern | I believe you ;) | 18:19 |
| stekern | fluxbox works here | 18:19 |
| poke53281 | at the momen I am fine with twm. Verz easy and it loads fast. | 18:19 |
| stekern | but pressing xterm in the menu didn't work | 18:20 |
| poke53281 | works, without problems | 18:20 |
| stekern | well, fluxbox has a special place in my heart, I used that "for real" for many years | 18:21 |
| poke53281 | if you can reduce to loaded size to let's say two MB then I will install it here | 18:22 |
| stekern | ah, well that's not a problem in my setup ;) | 18:22 |
| poke53281 | When I start X the memory is also already pretty full. | 18:23 |
| poke53281 | the normal install size is 4.1MB | 18:24 |
| poke53281 | for fluxbox I mean | 18:25 |
| poke53281 | the language part takes 1.2MB | 18:26 |
| poke53281 | By the way. Dillo already started on blueCmds Debian port. So I am not the first. | 18:28 |
| poke53281 | stekern: fluxbox works here too. But it looks like. But he seems to fail to read everything. I get a white screen and a few workspaces | 18:38 |
| poke53281 | do you have config files. | 18:39 |
| poke53281 | fluxbox behaves in the worst way when the window is too large for the screen. | 18:42 |
| poke53281 | the title bar is somewhere above the screen and I see the bottom of the window | 18:43 |
| stekern | no, I don't have any config files | 18:43 |
| stekern | alt+left mouse and drag the window? | 18:43 |
| poke53281 | Yes, that seems to work | 18:44 |
| poke53281 | Ok lunchtime. And then I will find out what the guy did who managed it that a port of an old C64 game runs only with 2 fps. | 18:53 |
| -!- Netsplit *.net <-> *.split quits: ssvb_ | 19:16 | |
| poke53281 | stekern. might be that openra was compiled with hard float | 21:57 |
| poke53281 | should I reccompile it? | 21:57 |
| --- Log closed Sun Sep 07 00:00:41 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!