#qi-hardware IRC log for Thursday, 2012-12-27

Patero-ngis anyone here from UT01:10
qi-bot[commit] Werner Almesberger: components/tps6216x.lib: add 9th pin for center pad (AGND) (master) http://qi-hw.com/p/kicad-libs/435a33901:16
qi-bot[commit] Werner Almesberger: tornado/cpu/: new version of the CPU board (WIP) (master) http://qi-hw.com/p/UNNAMED PROJECT/87b1fbb01:16
qi-bot[commit] Werner Almesberger: tornado/cpu/cpu.brd: improve 3V3 routing (master) http://qi-hw.com/p/UNNAMED PROJECT/0e9b09301:16
qi-bot[commit] Werner Almesberger: tornado/cpu/cpu.sch: split overly crowded sheet into two (master) http://qi-hw.com/p/UNNAMED PROJECT/ed61eab01:16
qi-bot[commit] Werner Almesberger: tornado/cpu/: rearrange things and fix ERC problems (master) http://qi-hw.com/p/UNNAMED PROJECT/9081d9101:16
kristianpaulrearrange sounds big01:50
wpwraknaw, just making it look better. the hard one was the split01:54
kristianpaulsplit is not just drag and drop then? ;-)01:55
wpwrakit's cut and paste, followed by converting local labels to global labels, giving them a direction, moving them to the right place (local labels and global labels go to different locations with respect to the wire), and then fixing the component references01:56
wpwrakmessy work. not difficult but many little things01:57
kristianpaulyeah need to track of all of then..01:58
Patero-ngis anyone here from UT02:10
xiangfuqi-hardware.com don't have ipv6. right? ping6 give 'unkonwn host'03:44
xiangfuthe 'miredo - Teredo IPv6 tunneling for Unix' works just fine in my ubuntu system.03:45
kristianpaulseems dig aaaa return nothing ipv6 indeed03:50
kristianpauldig aaaa qi-hardware.com i meant03:54
kyakhow long do you think it takes kernel to set the desired speed of serial port (cfset(o|i)speed)?05:42
kyakor, how i can measure it?05:42
kyaki suspect that my application start communicating with device connected via serial before the communication speed has settled05:43
wpwrakwhich side is changing the speed ? and what is your port like - via usb ?05:44
kyakso i set 115200 for example, and my external device does a simple echo. And i see that send/receive soem crap.. If i do "stty 115200 -F /dev/ttyACM0" before communications, everything is fine05:44
kyakthe host side is changing the speed, and yes it's serual over USB05:45
kyakthe target side has a fixed UART speed - 11520005:45
wpwrakand then, what tcsetattr actions argument to you use ?05:45
kyak(well, i said it so)05:45
kyakone se05:45
wpwrakTCSAFLUSH should be what you want05:46
wpwrakin case there's garbage in the receiver05:46
kyaki use tcsetattr(fd, TCSANOW, &sTermios);05:47
kyakok, i'll have a look whats different between TCSANOW and TCSAFLUSH05:48
wpwrakTCSANOW is the worst choice :) that gives you no synchronization at all05:48
wpwrakstill shouldn't cause data sent after the tcsetattr to be messed up, though05:48
kyaki'll definitely try with TCSAFLUSH to see if it makes any difference, thanks for the hint!05:49
kyakbtw, i even implemented the "suck off all the garbage from UART before sending actual data" feature on target side, but it doesn't seem to help05:50
wpwrakmaybe the garbage is already sitting in the receive buffer at the host side05:50
kyakthis is possible05:51
kyakin fact, it is very likely...05:51
wpwrakthen a good flush will take care of that :)05:53
kyaki'm also using cfsetispeed and then cfsetospeed instead of just a single cfsetspeed. Could it be a problem?05:53
kyakor everything takes effect after tcsetattr?05:53
wpwrakdo you set different rates or are they the same ? in the latter case, there shouldn't be a problem05:54
kyakthe same, of course05:54
kyakanother idea could be to disable receive and send buffer on host side (if the driver allows to), is it a bad idea? :)05:55
wpwrakyes, the cf... only change the parameters you'll later apply with tcsetattr?05:55
wpwraknot sure you can do that. in any case, a flush eliminates these issues. after that, the buffers are transparent.05:56
kyakok.. now that you gave me ideas i will try that later in the evening. Thanks again!05:56
wpwrakno problem. happy experimenting ! ;-)05:58
kyakbtw, serial programming howtos (like this one http://www.tldp.org/HOWTO/Serial-Programming-HOWTO/x115.html) all give examples with TCSANOW, i think this is where i took it from without thinkinh05:58
wpwrakdepends a bit on what you do. in general, if you're starting from an unknown state, you want to flush. then, when done and you want to restore the original settings, you can use TCSANOW if you're sure you're synchronized. otherwise drain or flush.06:05
wpwraksynchronized = tx and rx buffer are empty. e.g., if the other side acks the data you send, including the last packet/whatever you just sent.06:06
kyakyeah, this is definitely not my case - i'm not synchronized06:06
kyakbtw, what would be the default speed of serial device when it's plugged in? i noticed that when i call "stty 115200 -F /dev/ttyACM0", and then replug the device, then "stty -F /dev/ttyACM0" indicates the speed is 11520006:08
kyaki'm not sure if i can trust that and if it persists across reboots of host06:09
wpwrakhmm, not sure there's a single default speed06:11
kyakas a side topic, i discovered that there is no way to debug baud rate issues using a loopback jumper on my USB-serial converter :)06:11
wpwrakin any case, it's easy enough to set the speed you want. so you should do that :)06:12
kyaksince the baud rate i set is always "correct"06:12
wpwrakyou can try an oscilloscope06:12
wpwrakor a very very long jumper cable :)06:12
kyakheh, right :)06:13
GNUtoolarsc, hi17:29
larscGNUtoo: hi17:29
GNUtoodoes the xf86-video-glamo do rotation in software or in hardware17:29
GNUtoobecause I've that: http://paste.debian.net/219597/17:30
GNUtoofor xf86-video-omapfb17:30
larscbut the last time I lokked at xf86-video-glamo was about 3 years ago17:31
GNUtooand the way that works is flawed....it change the resolution17:31
GNUtoothe other way, the image is not rotated17:31
GNUtoolike the image is 640x480 but goes far over the screen on the right17:31
GNUtoothanks a lot for the answer btw17:33
qi-bot[commit] Werner Almesberger: tornado/led/: update schematics for new power distribution (master) http://qi-hw.com/p/UNNAMED PROJECT/d278ebb17:37
qi-bot[commit] Werner Almesberger: tornado/led/led.cmp: update for new components and changed references (master) http://qi-hw.com/p/UNNAMED PROJECT/bf2e2c317:37
qi-bot[commit] Werner Almesberger: tornado/led/: layout for new LED board version (master) http://qi-hw.com/p/UNNAMED PROJECT/352779a17:37
qi-bot[commit] Werner Almesberger: tornado/led/led.brd: clean up layout (master) http://qi-hw.com/p/UNNAMED PROJECT/12fc6cc17:37
qi-bot[commit] Werner Almesberger: tornado/cpu/cpu.brd: include subsystem name (CPU) in text label (master) http://qi-hw.com/p/UNNAMED PROJECT/b75570a17:37
larsciirc, it would not work otherwise17:55
GNUtoolarsc, ok18:50
GNUtooI'll write to xorg devel mailing list for my issue18:50
qi-bot[commit] Werner Almesberger: cameo/templates/mkmk-simple: use "set size ratio -1", not "set size ratio square" (master) http://qi-hw.com/p/cae-tools/9a6772f22:54
--- Fri Dec 28 201200:00

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