<wpwrak>
kristianpaul: no need to get half the USRP again, when i already have one ;-)
<wpwrak>
kristianpaul: alas, only RX, not TX,
<wpwrak>
i like the frequency range, though
<wpwrak>
kristianpaul: (mm) why, anything interesting happening there today ?
<wpwrak>
kristianpaul: (HDPE) reusable as in ... all it takes is a trip to the supermarket and you'll have supply for free ? (grocery bags)
<wpwrak>
kristianpaul: i think plastified wax may be an interesting option to look into. that may work with low pressure while still having many benefits of plastics - ideal for prototyping.
<wpwrak>
kristianpaul: i already did a try (in a old wooden counterweight mold) with pure stearine. that stuff is as liquid as water and stays liquid for a very long time (minutes). unfortunately, it adheres better to the wood than to itself, so the parts breaks up already when opening the mold. still, with a bit of plastic added, this could get interesting.
<wpwrak>
kristianpaul: (there are many wax+plastic procedures and recipes under the name of "machineable wax")
<kristianpaul>
wpwrak: (milymist) not not today but somedays happen :)
<kristianpaul>
wpwrak: (epiq) i pointed tou you because frequency range
<wpwrak>
kristianpaul: (mm) ah, i think the news propagate also here when something fun happens. most of the daily fpga stuff would be a bit obscure for me anyway :)
<wpwrak>
kristianpaul: (epiq) yeah, the frequency range is nice. ettus should steal some ideas and make an xwbx ;-) (extra-wide)
<kristianpaul>
wpwrak: (mm) for mee too, but i still learning , i heard you we're interested in a Ben based on milkymsit SoC
<wpwrak>
(mm-based ben) i think it would be a good platform for disseminating the mm. make it something people can relate to. not just yet another black box with cables coming out.
<wpwrak>
so i'm thinking of it more in terms of PR and attracting developers than in terms of something immediately useful
<wpwrak>
i think that MM holds quite some promise for the future, thought. the project just mustn't run out of steam :)
<kristianpaul>
sure MM One VJ Stuff is lekernel focus and MM SoC is for any want use it
<kristianpaul>
plus lm32 is beeing used for other things right now
<kristianpaul>
>
<kristianpaul>
> What is your LM32 SoC doing by the way?
<kristianpaul>
Mobile telephony layer 1 processing mostly.
<rafa>
or maybe several guys do not use git and qi-bot is not so smart enough to realize their work ;)
<kristianpaul>
s/way/why
<wpwrak>
retiring for lunch ans siesta soon. the hay fever is bad today :-( wish those plants would get their mating rituals over with ...
<kristianpaul>
rafa: i dont use git yet
<kristianpaul>
but i'll do soon
<kristianpaul>
i started with github :)
<rafa>
wpwrak: have a good rest
<kristianpaul>
next qi git next my own git at home :)
<kristianpaul>
rafa: do you ise git or CVS o anithing?? btw
<rafa>
kristianpaul: svn.. at home
<kristianpaul>
i see
<rafa>
kristianpaul: maybe git if I do something with kernel linux
<kristianpaul>
git is cool i realized that now i want it at home :)
<kristianpaul>
and a simple web interface that do not require apache and/or php
<kristianpaul>
:)
<rafa>
kristianpaul: ..well.. svn is nice as well.. and I do not use a lot of features or complex situations to learn a lot more git :)
<rafa>
web interface? .. for what?
<kristianpaul>
to see my comming history
<kristianpaul>
changes
<kristianpaul>
easilly !
<rafa>
git log ? git diff ?
<kristianpaul>
ahh
<kristianpaul>
hehe
<kristianpaul>
you know git more that me for sure :)
<Ayla>
hi there
<Ayla>
larsc: hi there too
<Ayla>
I have a question, regarding the RTC driver
<Ayla>
the RTC device is not initialized with the "wakeup" flag,
<Ayla>
is that an omit, or maybe it's not working yet?
<larsc>
Ayla: it's just not implemented. but should be easy to do
<larsc>
Ayla: do you want to try to add support for it?
<Ayla>
larsc: okay, but AFAIK, the functions are already there, that's why i'm asking
<Ayla>
I could try, but I'm a kernel newbie
<larsc>
it's really easy
<larsc>
take a look how the other rtc drivers do it
<larsc>
it's basically adding suspend and resume callbacks and then call irq_wake_enable, irq_wake_disable there
<Ayla>
yes, in theory it's simple
<Ayla>
but I'm browsing the code since 6pm
<Ayla>
:)
<Ayla>
I did add a call to device_init_wakeup() on jz4740_rtc_probe() right before the "return 0"
<Ayla>
it should mark the wakeup as available, and so create the "wakealarm" file on /sys/class/rtc
<larsc>
yes
<larsc>
and then in the suspend callback you need to check whether it is set to tru
<larsc>
e
<larsc>
and if it is call enable_irq_wake on the rtc irq
<Ayla>
oups
<Ayla>
got disconnected...
<Ayla>
hmm, if I compare the two rtc_class_ops between davinci and jz4740, jz4740 misses irq_set_freq and irq_set_state callbacks
<Ayla>
meh, the wakealarm file is not created even when calling "device_init_wakeup" !(
<Ayla>
:(
<larsc>
where did you add the device_init_wakeup?
<Ayla>
rtc-jz4740.c, line 286
<larsc>
you need to call it before calling rtc_device_register
<Ayla>
no, I doubt it
<Ayla>
look at rtc-sh.c or rtc-davinci.c
<larsc>
well. if you look at rtc_device_register you'll see that it only creates the device if device_init_wakeup has already been called
<Ayla>
then, I don't understand why it's done after on the other files :s
<Ayla>
whohouu! the wakealarm file is here
<Ayla>
larsc: about the check in the suspend callback, isn't it already done in an upper level?
<Ayla>
in the global part of the RTC driver
<larsc>
Ayla: i don't think so
<tuxbrain2>
Ayla: for even a more newbee in this matters than you... that means that NN should be able to be used as alarm clock?? that would be awesome and also a good motivation to push the suspend/resume instead of full shutdown
<tuxbrain2>
with the power volume NN has can wake up even diazepamed marmot :)
<Ayla>
tuxbrain2: no
<Ayla>
:)
<tuxbrain2>
ouch
<Ayla>
well, it could be set as an alarm clock
<Ayla>
but that's not my goal
<tuxbrain2>
what is your objective then?
<Ayla>
suspend to RAM
<tuxbrain2>
:) I like it too :)
<tuxbrain2>
and indeed is the first steep to be able to transform the NN in such alarm clock :), I'm couriouse to know how much standby time can hold NN suspended :)
<tuxbrain2>
count with  tester here :)
<Ayla>
well, calculate it
<Ayla>
the RTC chip should be less than 50mW IMHO
<wpwrak>
RAM refresh will also have a cost
<tuxbrain2>
nothing like real world "in situ" testing :)
<Ayla>
larsc, what do you call the "suspend callback" ?
<Ayla>
wpwrak: overall, the power cost would be very low
<tuxbrain2>
wpwrak: any stimation on power consumption of this refresh?
<kristianpaul>
Ayla: suspend to ram is a linux feature?
<kristianpaul>
or how are you achive that?
<kristianpaul>
i will kike tu see an hibertante feature
<Ayla>
kristianpaul, yes, it's handled by the kernel
<wpwrak>
tuxbrain2: hmm, self refresh current of the RAM chip is 6 mA
<Ayla>
the battery is 3,3V?
<wpwrak>
Ayla: the RAM is on the 3.3 V regulator
<tuxbrain2>
Ayla: 3,7V
<tuxbrain2>
what's the equivalence of 50mW to mA?
<Ayla>
A*V = W
<Ayla>
so 6mA * 3,3V = 20mW
<tuxbrain2>
so 50mW=15,15mA
<Ayla>
does anybody know how much power the nanonote/dingoo eat on standard use?
<Ayla>
the RAM eats 6mA with its standard clock, I guess
<Ayla>
on suspend, it may be clocked way slower
<tuxbrain2>
if only Ram+RTC powered on NN should stay suspended for 40h aprox... not bad
<larsc>
have you tested whether the alarm fires if the device is not asleep?
<Ayla>
I'm doing it right now
<Ayla>
(compiling)
<tuxbrain2>
ok guy time to sleep for me, Ayla I wish you the best of lucks in your resarch good k8
<Ayla>
thanks
<Ayla>
larsc: I added a printk in the interrupt
<Ayla>
when the alarm timer went out, it did set the "alarm_pending" flag from "no" to "yes" in /proc/driver/rtc
<Ayla>
but my printk didn't appear on dmesg
<Ayla>
and as that handler should clear the "alarm_pending" flag, I suppose that the interrupt is masked somewhere...
<Ayla>
meh
<Ayla>
include/linux/interrupt.h:
<Ayla>
"IRQF_NO_SUSPEND - Do not disable this IRQ during suspend"
<Ayla>
I'll add that flag to the request_irq call
<Ayla>
crap
<Ayla>
it's only for 2.6.36 xD
<rafa>
larsc: are you working with rc 2.6.36 right?.. As you know we try to add latest version.. are you happy with your current kernel version?
<wpwrak>
rafa: btw, would you know anyone around here who may have a RF lab that i could use for free for testing my wireless board, particularly the antenna ? e.g., at a university or such ?
<wpwrak>
rafa: desirable lab features would be a spectrum analyzer that covers the 2-3 GHz band and a reasonably large anechoic chamber (>= 1.5m)
<wpwrak>
rafa: "would you know" = including via recursion :)
<wpwrak>
rafa: "around here" = ideally, in town. but then, i don't mind driving a bit.
<wpwrak>
rafa: (desirable lab features) spectrum analyzer with tracking generator or separate rf generator (also 2-3 GHz band)
<wpwrak>
rafa: of course, the wider the frequency range, the merrier ;-)
<wpwrak>
rafa: measurements would be: 1) antenna tuning, 2) calibration of my USRP2 and my antennas (that way, i can do more things with my own equipment), and 3) check for unwanted emissions
<rafa>
wpwrak: i know people from UBA, UNS and UNcoma.. if some of those universities has I will let you know.. let me ask
<wpwrak>
rafa: kewl, thanks ! they may in turn also know someone, etc.