**** BEGIN LOGGING AT Mon Nov 23 03:00:35 2015 Nov 23 03:16:57 hi. i have a systemd question. i am overriding a recipe which inherits systemd. i only want to disable *some* of the systemd services that are in the original recipe. how do i specify that? Nov 23 03:18:04 i have tried SYSTEMD_AUTO_ENABLE-blah = "disable", where blah is the name of the service file i want to disable, but that doesn't work Nov 23 03:18:33 if i did SYSTEMD_AUTO_ENABLE_${PN} = "disable" it would disable all of them... Nov 23 08:12:46 good morning Nov 23 08:23:36 hi Nov 23 08:24:02 I need help Nov 23 08:24:25 that's a mighty introduction Nov 23 08:24:37 I did yocto core minimal image but there is no i2c port under /dev Nov 23 08:25:30 how can I put i2c port under /dev Nov 23 08:25:46 could you help me? thanks Nov 23 08:26:23 * rink thinks you need to add it to your kernel config Nov 23 08:26:55 poyraz: that depends on your kernel features, which machine are you using? Nov 23 08:27:29 microzed Nov 23 08:29:22 I try to "bitbake linux-yocto -c menuconfig" but after I added i2c driver, I am doing bitbake core-image-minimal Nov 23 08:29:32 am I right? Nov 23 08:30:10 how to open kernel config rink? Nov 23 08:30:28 how can open kernel config rink? Nov 23 08:30:36 * rink does not know Nov 23 08:31:19 you can look at /proc/config on the device to see if the option you enabled, is enabled Nov 23 08:34:10 poyraz: you'd better launch bitbake virtual/kernel -c Nov 23 08:34:30 bitbake virtual/kernel -c compile Nov 23 08:35:17 poyraz: or bitbake virtual/kernel -c deploy Nov 23 08:35:49 poyraz: redoing bitbake core-image-minimal won't recreate your kernel binary Nov 23 08:36:23 ok I will try Nov 23 08:36:26 thanks Nov 23 09:22:56 poyraz: microzed doesn't have a default setup for i2c due to pinmuxing via fsbl. So its disabled in the device-tree by default (at least in the meta-xilinx layer) Nov 23 09:25:13 nrossi_:I changed device-tree. I added i2c in device-tree Nov 23 09:28:29 poyraz: ok, if you are after /dev/* nodes for the linux-xlnx you should be good, check with cat /proc/devices to see if they are available Nov 23 10:10:28 nrossi: I found i2c under /proc/devices. later I did " mknod /dev/i2c-0 c 89 0", but now when I enter "i2cdetect -y 0" Nov 23 10:11:14 nrossi: answer is error: couldnt open file /dev/i2c-0 : no such device Nov 23 10:11:34 nrossi: what should I do now? Nov 23 10:12:25 poyraz: interesting, are you sure the i2c probed correctly? does it appear in the kernel log? like 'cdns-i2c e0004000.ps7-i2c: 100 kHz mmio e0004000 irq 168' Nov 23 10:21:04 poyraz: another thing to check is sysfs, run 'cat /sys/bus/i2c/devices/i2c-0/name', it should exist and have the value of 'Cadence I2C at e0004000' (or similar) Nov 23 11:16:15 nrossi: i can see i2c-0 under dev but I didnt see "cdns-i2c e0004000.ps7-i2c: 100 kHz mmio e0004000 irq 168" smth like that. when I entered "cat /sys/bus/i2c/devices/i2c-0/name" there is an error. no such file or directory. Nov 23 11:20:09 poyraz: yep but you said you manually made the /dev/i2c-0 node yes? Nov 23 11:20:16 yes Nov 23 11:20:57 nrossi: I can see /dev/i2c-0 Nov 23 11:20:57 ok, how did you setup your device tree? are you using the ones in the meta-xilinx layer? or did you manually modify it? Nov 23 11:21:24 nrossi: meta-xilinx layer Nov 23 11:21:24 yep you can see it, but the chardev is not actual attached to any backing device Nov 23 11:21:53 I modified for i2c Nov 23 11:22:27 ok did you remove the lines in the -board.dtsi file that force the compat string to invalid? (http://git.yoctoproject.org/cgit/cgit.cgi/meta-xilinx/tree/conf/machine/boards/microzed/microzed-zynq7-board.dtsi#n63) Nov 23 11:23:04 yes Nov 23 11:23:32 i removed that lines and added ps7_i2c_0 like others Nov 23 11:26:48 nrossi: i removed that lines and added ps7_i2c_0 like others Nov 23 11:29:29 nrossi : i think i found my error thanks Nov 23 11:33:31 poyraz: ok cool, was about to point you at checking the device-tree in /proc/device-tree, to make sure it matched what you expected. ;) Nov 23 11:38:44 where could I find more detais about the syntax "${@...}" ? Nov 23 11:38:59 google didn't help Nov 23 11:39:20 nrossi: I recognized that microzed-zynq7-board.dtsi is wrong. how to change ps7_i2c_0? is there any sample guide? Nov 23 11:45:13 poyraz: what do you mean by wrong? what are you trying to do, do you need info about how to setup the i2c devices in a device tree? Nov 23 11:46:17 mckoan: Not sure there is much more that whats in the bitbake manual -> https://www.yoctoproject.org/docs/current/bitbake-user-manual/bitbake-user-manual.html#inline-python-variable-expansion Nov 23 11:46:44 nrossi: I dont know how to change microzed-zynq7-board.dtsi Nov 23 11:48:07 poyraz: you can just edit the file in the layer (to get things going), or are you refering to not being familar with device-trees? Nov 23 11:50:12 nrossi: 3.1.11. Inline Python Variable Expansion Nov 23 11:50:17 nrossi: thanks Nov 23 12:00:33 Hi there, can anybody help me - I do not know how to add update-rc.t to my rootfs. In my development image I tried CORE_IMAGE_EXTRA_INSTALL += "... update-rc.d ..." Nov 23 12:04:32 Hi there, can anybody help me - I do not know how to add update-rc.d to my rootfs. In my development image I tried CORE_IMAGE_EXTRA_INSTALL += "... update-rc.d ..." Nov 23 12:32:32 <__b4ub> Can't find where to set the size of the serial terminal on my build... Any idea ? Nov 23 16:03:22 anyone attempted to use gstreamer1.0-omx on raspberripi2 ? Looks like that recipe has a hardcoded dependency on bellagio's libomxil, which would warrant a virtual/libomxil of some sort Nov 23 17:55:33 Is it possible to have yocto fetch things with curl instead of wget? Nov 23 18:41:02 morning all Nov 23 18:53:43 Cardoe: yes, there is a way to control the fetch command; we use it to pass arguments to wget, but it can also be used to switch to curl Nov 23 19:04:29 neverpanic: would that be FETCHCMD_wget? Nov 23 19:07:10 I can't check at the moment, but that's possible, yes. Looking at FETCHCMD_wget, it seems that doesn't allow you to use curl, unless you emulate a couple of options of wget using a wrapper around curl. Nov 23 19:08:29 back up a step. Nov 23 19:08:42 what problem are you having that you're trying to address by doing this? Nov 23 19:09:36 I see. I think FETCHCMD_wget would be tricky to use with curl, as some wget arguments are appended to it. Those arguments may not exist for curl or may exist but have different meaning. Nov 23 19:12:21 fledermaus: https://github.com/travis-ci/travis-ci/issues/5156 Nov 23 19:15:32 Cardoe: You can pass --ca-directory or --ca-certificate to work around that Nov 23 19:15:55 ok, actually, no, you cannot :/ Nov 23 19:17:34 what distribution/release is wget failing on, and for what url? Nov 23 19:18:05 12.04 https://static.rust-lang.org/dist/rustc-1.3.0-src.tar.gz Nov 23 19:18:17 Its due to the Comodo RSA certificate chain having two options to go down. Nov 23 19:19:11 Ubuntu 12.04 that is. Nov 23 19:19:18 Unfortunately Travis CI uses 12.04 Nov 23 19:19:38 and you haven't got enough access to add a CA cert? Nov 23 19:20:55 if you do, you should be able to drop the cert into /usr/local/share/ca-certificates/ Nov 23 19:20:55 I don't think so. Nov 23 19:21:07 and run update-ca-certificates Nov 23 19:21:11 And then run update-ca-certificates as root. Nov 23 19:21:17 right. Nov 23 19:30:51 Cardoe: if you only need to do it for this one thing, capture the options that are being passed to wget and use a wrapper that does the appropriate curl call instead? (just for this recipe) Nov 23 19:32:03 or override the fetch command entirely for this recipe. or have a pre-fetch command that does the actual fetching to a known location then use a file:/// url for wget. Nov 23 19:32:09 or something like that. Nov 23 19:32:38 (thinking out loud) Nov 23 19:33:36 That could work Nov 23 19:33:38 I'll try that. Nov 23 19:33:40 Thanks Nov 23 19:33:58 That whole recipe is screwy as it is. **** ENDING LOGGING AT Tue Nov 24 03:00:52 2015