<xiangfu>
lekernel: Hi is this the correct file for add more keymap: http://pastebin.com/yhpgW60e  mtk/lib/keymap.c
<lekernel>
yeh
<lekernel>
now you also need API calls to change the keymap at runtime
<xiangfu>
lekernel: yes.
<xiangfu>
one question. I fond the '`' and '-' not working in german layout.
<xiangfu>
I mean when I press those two key. no keycode output. (I add one line output all keycode)
<xiangfu>
when I press  '`'    '-'    in german layout, us layout, nothing output.
<lekernel>
ok... maybe there's a missing entry in the keymap somewhere
<lekernel>
'-' definitely works on my keyboard. '`' not thoroughly tested
<lekernel>
just add entries so it works on your keyboard...
<wolfspraul>
lekernel: do you think 1500 mA is enough for m1?
<lekernel>
on 5V? no
<lekernel>
this wouldn't allow high current USB gadgets to be connected
<lekernel>
2A is better
<wolfspraul>
ah I see. you basically have 2*500mA over USB host already, right?
<lekernel>
yes
<wolfspraul>
actually then even 2A may not be enough :-)
<wolfspraul>
but anyway, I got it.
<lekernel>
in some extreme cases, maybe not. but I believe it is enough for all reasonable cases.
<xiangfu>
lekernel: one small thing. do you think we should change the osc port from 7777 --> 4444, I found 4444 and 7777 are both common, but I found 4444 is more then 7777. (in my android device, all of those three app the default port is 4444)
<lekernel>
hmm I don't know
<lekernel>
we can make it configurable too
<xiangfu>
lekernel: yes. also the osc 'address' that will be great.
<xiangfu>
lekernel: if we don't have plan on osc. I would like change it to 4444. (just for common fit all other apps :)
<lekernel>
ok
<roh>
isnt there a iana assigned port?
<xiangfu>
we may have a lot of work on osc, like forward osc message to DMX etc, we can full control the whole m1 by using osc (by an iPAD or Android-pad :)
<xiangfu>
roh: seems no document on that. :(
<roh>
it seems the iana people dont really have a clue how ip works (my assumption by reading the ml)
<roh>
atleast not that one can get ports assigned for free, when asking properly
<roh>
and that its stupid not to have a proper port... i can only find reports that there isnt one, that there should be one, but not that somebody did the work. instead there was a discussion how and if to use mdns (for autodiscovery) and which code to take for it
<roh>
m-/
<CIA-48>
flickernoise: Xiangfu Liu master * r70b7807 / src/osc.c : osc: change default port to 4444 - http://bit.ly/dEmNPG
<lekernel>
you should have access... can you move your ethernet issue?
<lekernel>
btw I think it has little to do with the minimac core
<xiangfu>
moving now.
<kristianpaul>
hmm 10base-T i think you will need MDIO so tell tranceiver do 10base-T, but is just a guess..
<wolfspraul>
xiangfu: did ethernet work with that router before?
<xiangfu>
works fine before.
<kristianpaul>
hmm
<wolfspraul>
maybe if you have some time and if your build system works well, you can find the commit that caused the regression?
<wolfspraul>
not sure how much has happened everywhere and whether this is worth it.
<kristianpaul>
curios, because at i cant nenver get the mm1 ethernet to link at 10Mpbs
<wolfspraul>
but typically if something worked before one would hope it's relatively easy to track down why it stopped working...
<kristianpaul>
(commit that caused the regression?) well, i'm aware minimac is a new core
<lekernel>
xiangfu: are you sure you're not simply running into the intermittent PHY reset bug?
<lekernel>
the MAC should not interfere with the LEDs on the ethernet connector ...
<kristianpaul>
indeed
<kristianpaul>
also the rtems driver changed.. but i dont remenber it did initialization or something related before..
<xiangfu>
hmm... I test on three device. I don't know much about ethernet. but at least the 'red' light should be on when I connect m1 and hubs.
<kristianpaul>
xiangfu: can you show us some debug messages from your router/swich?
<lekernel>
the only significant thing that changed is that the PHY reset line is asserted longer (and deasserted by software instead of being connected to the system reset)
<xiangfu>
kristianpaul: the boot log from m1?
<kristianpaul>
xiangfu: try another cable? :-)
<lekernel>
I tried to do that to help solve the reset bug, but it didn't work
<kristianpaul>
xiangfu: no, log from your router os
<lekernel>
otoh we now know that the reset bug is only because of the PHY: pulsing it in software when ethernet is broken fixes it
<xiangfu>
kristianpaul: tried two cables
<xiangfu>
kristianpaul: we are luch one of then have serial output :), I will try that
<xiangfu>
s/luch/luck
<kristianpaul>
oh, i tought the router ran openwrt or something.. but np just asking in case was easy for you
<kristianpaul>
in anycase will be nice to know why the negotiation fail..
<roh>
lekernel: some phy are always weird
<xiangfu>
(I install the openwrt in it, but for ddns. I change it back to origin image, but after you tell me the freeDNS, maybe it's the time to flash openwrt again :)
<roh>
ive seen drivers resetting them twice by default on linux
<kristianpaul>
xiangfu: can you use ethtool on your laptop and force 10 base T mode, and compare if got same behaveior with the router/switch ?
<xiangfu>
lekernel: any things is when after reflash. after 'pld reconfigure' then press #2 for poweron the m1. the ethernet in m1 will not working.
<kristianpaul>
ah
<kristianpaul>
know bug ;-)
<xiangfu>
(even direct connect to my laptop), I have to unplug replug again for make the ethernet work.
<roh>
so expect some 'bugs' which need not-so-nice workarounds. but yes.. when there i no autonegot. from the mm1 side.. its up to the logic in the switch to guess and nail it correctly
<kristianpaul>
xiangfu: (freedns) why?. no... well a crontab with wget where just enought in openwrt as well :-)
<xiangfu>
kristianpaul: before freedns I using the buildin DDNS in openwrt. yes. since we have freedns. we can using the 'wget'
<kristianpaul>
xiangfu: yup
<xiangfu>
kristianpaul: do you think the NIC in laptop support 10 base-T, I thought it's only support 100M/1G, it's not phy, it's software ?
<xiangfu>
s/any things/another thing/
<kristianpaul>
xiangfu: ah, well, you need check specs first
<xiangfu>
s/DDNS in openwrt/DDNS in my router/
<kristianpaul>
lekernel: had you measure power compsuption of mm1 running a patch?