#qi-hardware IRC log for Thursday, 2010-05-27

kyak06:29:42 < xakh> for instance it's 2230 here <-- hehe02:29
shevekurandom__: Having any luck compiling Iris?10:59
urandom__shevek not yet but did not spend much time trying it, sory11:21
shevekurandom__: No problem, I was just curious. :-)11:23
urandom__shevek maybe more people would try it if they would not need this pypp stuff :P11:26
shevekurandom__: Hm, perhaps.  I can put a generated tree of c++ source somewhere, but it gets outdated fast.11:28
shevekurandom__: And currently the usb server is required for testing anyway, so libshevek is still needed.  That's a bigger problem than pypp, I expect.11:28
urandom__so we need better docs about compiling it, you could start a wiki page about it so more people would become interested in it11:32
shevekGood idea.11:33
mthlarsc: usbboot/src/Makefile.am sets prefix to /usr, overriding the value set by configure11:34
mthshould I report this as a bug somewhere?11:35
larscprobably yes11:54
shevekurandom__: I started http://en.qi-hardware.com/wiki/Iris.  If anything is unclear, please modify it. :-)12:00
mthlarsc: report it as a bug on the Ben NanoNote, or elsewher?12:18
larscmth: http://projects.qi-hardware.com/index.php/p/xburst-tools/issues/12:21
mthah, I was scanning the page for the word "bug" and forgot to actually read the tabs...12:22
mthI couldn't get it to dump the A320 NAND, by the way12:22
mthoob and raw mode don't work at all, normal mode seems to work on some pages but not all12:23
mththat might be because the first pages of A320 have a different size because the primarly boot loader expects 2K pages12:23
Action: FrankBlues waves14:05
larscmth... hm indeed, the doc says so. but the ingenic codes suggests that writing to the pending register clears the interrupt. i'll do some tests18:20
mthis it a problem that Linux thinks it has 1 shadow register set and the xburst data book says there are none?18:24
mthI accidentally found that difference when I was debugging the audio problem18:25
larschm, looks like they are only important for vector interrupts18:31
larscwhich we don't use anyway18:31
larscthe code looks like the number of shadow register sets is set to 1 if it's not supported18:31
mthso it's the number of register sets then, shadow or non-shadow?18:37
mthI mean, it's a confusing name if you count the main thing as shadow too18:38
larscthat is at least of i understood it18:38
larscthe code won't work if the value is 018:39
larscand shadow registers are new in R218:39
mthdo all interrupts go through the INTC or are there interrupts that go directly to the CPU or are even generated inside the CPU?18:40
mthI'm wondering why enabling interrupts on the CPU would hang Linux18:41
mththe hang must happen immediately, since directly after I put an instruction to disable backlight18:41
mthand it is not disabled18:42
larscnope. all irqs go through intc. (execept the intc irq)18:44
mthif the INTC would generate an interrupt and the CPU has interrupts disabled, then all interrupts are masked inside the INTC, would the CPU still consider the interrupt pending or not?18:45
larscnot sure18:47
larscbut i guess yes18:47
mthhowever, jz4740_cascade does check for ffs() to return 0, so it should be robust against this situation18:47
larsci've seen intc where no bits where set in pending18:47
mthhmm, jz4740_cascade is called, but intc_irq is not 018:49
mthlet's apply binary search...18:50
mthintc == 2418:52
mthso that's actually bit 23 then, TCU018:53
mthaccording to the docs, the pending register will have bits set for interrupts that are both pending and not masked18:57
mthif that is true, something must have unmasked TCU018:57
mthah wait, that's interrupt 23, not INTC interrupt 2318:58
larschm, 15 is gpio iirc?19:00
mthJZ4740_IRQ_BASE is 8, so it would be INTC bit 15, the RTC19:00
mthbut can the RTC interrupt be enabled before the driver is?19:00
larscsure. if something else enables it ;)19:01
mthok, let me rephrase that :)19:01
mthis it expected that an interrupt of a device is enabled outside the driver?19:02
mthand drivers are not initialized until the board-X.c code runs, correct?19:03
larscyou could add debuging output to unmask19:04
larscor a BUG(); to get a stack trace19:04
mthI don't have any debug facility except the backlight though19:08
mthoh, I misinterpreted the code, it is TCU0 after all19:11
mthsince the timer does not use a platform driver, it might be inited earlier19:13
mthif I disable all timers initially (0x100fc -> 0x100ff), the boot gets slightly further but still hangs (maybe because it has no timer)19:15
mthis the timer configured before it is enabeld?19:18
mthmaybe if it is set at a ridiculously high frequency, it starves the CPU19:19
larscwell, the timer should still be disabled. only the clock is not stoped19:20
mthdisabled in what way?19:21
mthusing the TER register?19:22
mthgot it to boot :)19:28
mthby disabling the half irq for timer 0/119:29
mth(probably timer 0 was the essential one)19:29
mthnot the right approach?20:06
larscno. jst uhm-ing about my own stupidity. i removed similar code when doing the refactroing for pwm20:08
mthah, that explains why earlier versions did boot through usb20:17
mthunfortunately I nuked the git history when rebasing master several times, so I couldn't bisect20:17
xakhhey, guys, I'm trying to follow these instructions20:51
xakhand I got to "creating a package for OpenWRT"20:51
xakhand got confused.20:51
rafasomebody knows how to use mkfs.ubifs and ubinize (command line arguments for nn) to create a rootfs for nn?21:30
rafathat is why there is wiki : http://en.qi-hardware.com/wiki/Ubifs :D21:31
rafasorry for the question before to read the answer ;)21:31
kristianpaulhola rafa21:55
qi-commitsCarlos Camargo: Fixing plasma files.. Now works :) http://qi-hw.com/p/nn-usb-fpga/01e672d22:28
qi-commitsCarlos Camargo: Adding ucf file to plasma demo http://qi-hw.com/p/nn-usb-fpga/c3ff3a322:33
rafakristianpaul: hola23:01
rafacomo estas?23:01
rafawolfspraul: hey man.. we are starting to make our oe builds ;) (in jlime)23:03
wolfspraulrafa: please blog about it a bit23:03
wolfspraulJon is totally right we need to connect the culture of copyleft23:03
wolfspraulfor example when I think about you, I think about the drinks you mixed on Werner's rooftop :-)23:03
wolfspraulhe he23:04
wolfspraulnot about jlime (sorry)23:04
wolfspraulrafa: do you have a blog (rss/atom) we can aggregate into the qi hardware planet?23:05
xakhI noticed that my nanonote looks significantly different from Zear's23:31
xakhthe one he used for demoing quake23:31
wolfspraulxakh: what is the difference?23:32
xakhlemme take a picture23:32
xakhit's easier just to show you guys23:32
wolfspraullong ago we have several times of prototypes and stuff, but everything that was sold since the official launch should be exactly the same23:33
xakhsorry, had to line up the shot23:34
rafawolfspraul: no sorry. I am old style :)23:43
rafawolfspraul: we often use our forum, but until we have something useful to show we would not create a new forum for this work.23:44
wolfspraulhe he23:44
rafawolfspraul: we can let you know when things start to look useful though ;)23:44
wolfspraulthat's a start23:45
wolfspraulrafa: which kernel are you using and how do you build the kernel?23:47
rafawolfspraul: I have built the jlime nanonote via oe. And, also, I did the minimal bootstrap rootfs. But no luck yet.23:50
rafaIt does not boot. Well, it does, but something inside does a reset. SO now, I need to talk with kristoffer to know how to kill the problem23:51
rafajlime nanonote kernel*23:52
rafawolfspraul: I also converted the rootfs jffs2/tar.gz rootfs to ubi format, so I tried to test it with current Qi kernel, but no luck either.23:53
rafawolfspraul: But well, it was my first day working on that.. we will go further surely ;)23:54
qi-commitsXiangfu Liu: [gcal] cleanup Makefile http://qi-hw.com/p/openwrt-packages/7aa1d3a23:59
--- Fri May 28 201000:00

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