#milkymist IRC log for Wednesday, 2013-03-06

GitHub91[misp] sbourdeauducq pushed 1 new commit to master: http://git.io/1DKgww10:21
GitHub91misp/master ef45d06 Sebastien Bourdeauducq: Use new compiler-rt path10:21
GitHub144[milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/yOPqrw10:21
GitHub144milkymist-ng/master 2059592 Sebastien Bourdeauducq: software/libcompiler-rt: add ctzsi210:21
GitHub198[migen] sbourdeauducq pushed 1 new commit to master: http://git.io/xURQmg10:42
GitHub198migen/master 6fa3005 Sebastien Bourdeauducq: fhdl/verilog: tristate outputs are always wire10:42
lekernel/* The core i2c driver will automatically retry the transfer if the14:16
lekernel * adapter reports EAGAIN. However, we find that bit-banging transfers14:16
lekernel * are susceptible to errors under a heavily loaded machine and14:16
lekernel * generate spurious NAKs and timeouts. Retrying the transfer14:16
lekernel * of the individual block a few times seems to overcome this.14:16
lekernel */14:16
lekernelah, linux ...14:16
Fallenouahah yes I saw that :)14:17
Fallenoui2c in linux has a few quirks14:17
lekernelah and the nouveau driver loves broken DDCs. typically, it will disable the port until you reboot or suspend...14:27
Fallenoufantastic !14:44
_florent_just for info, after the de0 nano port21:06
_florent_I'm starting working with milkymist-ng with the KC70521:06
_florent_Nothing very insteresting to say for the moment because nothing is tested yet...21:07
_florent_but I'll tell you when it will start working21:07
_florent_for the moment there won't be a DDR3 controller21:08
_florent_but we will have to discuss about that if some design can be share with your future M^321:09
lekernelgreat :)21:10
_florent_Hi lekernel21:19
_florent_for info the LM32 seems to  work at at least 156MHz on the kintex21:20
_florent_I've tried 200MHz but I was having timings errors....21:21
_florent_156 is probably not the max, but I haven't played with the mmcm for the moment21:21
Fallenouvery nice :)21:31
lekernelhmm, weird, I'm quite sure I had 20021:32
lekernelhave you tried disabling some options (multipliers, divider ...)21:33
_florent_in fact 200 was ok for the cpu alone21:40
_florent_but it must be close to 200, I was only having small timings issues on the multiplier21:41
_florent_haven't try without21:41
_florent_On the kc705 I'm going to use the milkymist-ng frame buffer + some color conversion module + the ADV7511 on the board21:43
_florent_Do you have developped some simple code to test the frame buffer on the ng?21:44
_florent_If so I'll be interested to test the frame-buffer --> color conversion --> ADV7511 chain21:45
larsc_florent_: what are you running on the lm32?21:49
_florent_for the moment , only the bios ;)21:51
_florent_but I only wanted to know if some simple test code for the frame buffer existed for the bios21:51
Fallenouyou can flash the PONG demo instead of the BIOS for instance21:54
Fallenouit makes simple use of the framebuffer21:54
Fallenouthis is working on the M1 board21:55
_florent_that's what I need I think21:55
Fallenouyou can play 1-player pong with the push buttons and VGA output21:55
_florent_only a simple code that generate video21:56
Fallenouyou're welcome :)21:56
_florent_I have to go21:58
_florent_I'll  send you a capture of the pong demo in 1080p when it will be working ;)21:58
Fallenouhehe very good :)21:59
Fallenoua video ;)21:59
Fallenoubye !21:59
lekernelFallenou, iirc the pong demo still uses the old SoC interface and should be loaded with netboot/flterm/flashboot (not flashed instead of the bios)22:32
Fallenouthe pong demo for the ml401 maybe22:34
Fallenoubut the pong demo for M1 is a replacement for the BIOS22:34
Fallenouit initializes ddr controler etc22:35
Fallenouit calls _sdram_init22:35
--- Thu Mar 7 201300:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!