**** BEGIN LOGGING AT Tue May 17 02:59:58 2016 May 17 07:22:29 good morning May 17 13:11:33 ~morning May 17 13:11:34 Good morning! May 17 15:44:39 hey hrw good morning May 17 15:47:27 ndec, dragon is booting, need to find an hdmi cable for my display and see if gqrx works May 17 15:47:43 it took forever to write the 1G ext4 file though May 17 15:48:50 Crofton|work: use sparse file system May 17 15:49:11 and bmap, apparently May 17 15:49:16 though i prefer non-shit storage May 17 15:49:29 flashboot May 17 15:49:43 I just added meta-qcom and ran bb gnuradio-demo-image :) May 17 15:49:53 you can flash just the sparse image, but bmap makes it easier to not lose that sparseness through compression, remote transfers, etc, afaik :) May 17 15:50:05 i like IMAGE_FSTYPES += "wic.bz2 wic.bmap" May 17 15:50:20 Crofton|work: add 96boards-tools to IMAGE_INSTALL May 17 15:50:21 but i don't think you can use that with fastboot. May 17 15:50:31 and then there is this https://github.com/96boards/96boards-tools/blob/master/resize-helper.service May 17 15:50:40 I just left it flashing and went to bed :) May 17 15:50:40 which will resize emmc May 17 15:50:58 i really need to play with resize-helper one of these days May 17 15:51:09 that'd be nice for varying sd card sizes, at least with r/w rootfs May 17 15:51:10 kergoth: it worked really great May 17 15:51:17 right now we have a different wks for each common sd card size :) May 17 15:51:31 my dragon I flash core-image-minimal May 17 15:51:35 though if wks.in gets merged, that'd be slightly less annoying, could use a variable May 17 15:51:39 then I set up feeds May 17 15:52:16 Crofton|work: the ext4 image from OE is not 'sparse' (as far as fastboot sparse format). so you would end up writting '0's as well, this is what takes time. May 17 15:52:22 ok, I am going for cable May 17 15:52:34 sure, just letting you about wierd things I see May 17 15:52:39 go for it Crofton|work :) May 17 15:52:51 as khem said, you need to make the smallest image as possible, so that there are less 'empty' space, and use resize-helper. May 17 15:52:57 fastboot has a special sparse format? May 17 15:53:11 why couldn't 96borads be minnow compatible? May 17 15:53:13 oh, you're using the ext4 image directly, not via wic, gotcha May 17 15:53:17 right May 17 15:53:32 we should really make those sparse too if possible May 17 15:53:55 or you can run ext2simg to sparse the image. May 17 15:54:18 https://github.com/96boards/documentation/wiki/Dragonboard-410c-OpenEmbedded-and-Yocto May 17 15:54:34 Crofton|work: thats like asking why day and night cant be same May 17 15:54:36 kergoth: yes, fastboot has its sparse format: http://www.2net.co.uk/tutorial/android-sparse-image-format May 17 15:54:42 a friend of mine tried writing an image iwth dd with conv=sparse.. not good. loses the *important* series of zeroes :) May 17 15:54:43 I'm just following ndec's instructions :) May 17 15:54:53 bbiab May 17 15:55:20 i wish we could get ext2simg into oe-core.. so that we could create fastboot 'sparse' image out of the box. May 17 15:56:47 ndec: there has to be more than 1 board using it for making a case for oe-core May 17 15:57:05 i am sure there is more than 1 board that comes with fastboot.. May 17 15:57:10 what does ext2simg do? May 17 15:57:31 takes extX image and create a fastboot sparse image May 17 15:57:41 it's in android-fsutils right now. May 17 15:57:49 whats a fastboot sparse image? :) May 17 15:57:50 and it's in meta-android May 17 15:58:06 rburton: http://www.2net.co.uk/tutorial/android-sparse-image-format. May 17 15:58:11 thans May 17 15:58:35 the file format that they primarly use for Android images, but it's not stricly something specific to android, of course. May 17 15:59:00 basically a ext4 but with the sparse bits transfered as sparse events instead of 000000000p May 17 15:59:02 right? May 17 15:59:21 so its basically bmap but for a transfer protocol May 17 15:59:35 huh, interesting May 17 16:23:24 kergoth: ndec: mainline u-boot has support for it as well (fastboot with sparse) May 17 16:24:16 great.. one more reason to have a 'etx4.img' image type ;-) May 17 16:24:48 Erk, this is strange. I have a case where parsing is giving 'unable to get checksum' errors due to not being able to find these file:// files.. but do_unpack manages to find and use them just fine May 17 16:24:55 Anyone seen that before? May 17 16:29:21 rburton: I saw there is a new version of weston upgrade in oe/master-next. Should I rebase the patches and resend? there is a conflict. May 17 16:29:45 kergoth: " classes/base: get_lic_checksum_file_list imporve validaton of url's"? May 17 16:29:51 otavio: please! May 17 16:30:14 otavio: actually i think there'll be another update from jku so i'd wait for a new series from him first May 17 16:35:27 rburton: RP has made a fixup on master-next; so I based on this May 17 16:36:37 rburton: anyway, v7 is out May 17 17:19:50 khem: hey May 17 17:20:01 khem: I have nios2 qemu support which works for gobject-introspection May 17 17:44:20 ndec, got gqrx running on dagonbaord May 17 17:44:46 once again process is add bsp layer and bb gnuradio-demo-image May 17 17:44:54 this oe stuff really can work May 17 18:17:01 otavio: which recipe provides the librarties for -fsanitize=address May 17 18:17:07 sorry May 17 18:17:11 which recipe provides the librarties for -fsanitize=address May 17 18:35:30 Crofton|work: good, thanks for trying it! May 17 18:35:41 it was designed to work this way ;-) May 17 18:37:08 I likely have some 64 bit neon hacking to do May 17 18:37:16 but should be able to show it at stands May 17 20:21:03 khem: poke ... May 17 22:03:33 Marex: ok May 17 22:03:44 Marex: I am looking into it May 17 22:06:12 khem: hope it helps you May 17 22:06:19 khem: took mew a while to put this together May 17 22:06:41 khem: I actually have those patches kinda separated now and submission to upstream qemu might not be far away May 17 22:08:25 khem: how far are you with this angstrom/nios2 ? May 17 22:09:02 khem: what I see is missing is LTP (that's on my radar) and cross-gdb (I took only a brief look) May 17 22:09:24 khem: libdrm causes ICE with gcc 5.3.0 , but that might be fixed in gcc 6 May 17 22:13:09 Marex: ok. I fixed your 0004 patch to remove pokyness May 17 22:13:19 rest it applied ok May 17 22:14:09 Marex: I see you updated qemu to 2.6, could this patch be also backported to krogoth May 17 22:14:42 khem: probably, does it make sense ? May 17 22:15:33 I dont think we can get a major upgrade backported May 17 22:15:49 but it will be good if nios2 patch is able to apply on 2.5.x May 17 22:16:11 khem: ah May 17 22:16:21 khem: yeah, that should be doable with a few tweaks here and there May 17 22:16:23 Marex: in angstrom I am able to build systemd-image May 17 22:16:36 khem: yeah, systemd has nios2 support in mainline ;-) May 17 22:16:37 Marex: please take that up May 17 22:16:54 Marex: I patched systemd May 17 22:17:18 https://github.com/kraj/meta-altera/commit/63ec985e78889bf2f63897008e6792f467e2ed5f May 17 22:18:11 khem: https://github.com/systemd/systemd/commit/b79660e6acb964f94ddba0ea4e1182215934dfc3 May 17 22:18:19 khem: I upstreamed nios2 systemd support a while ago May 17 22:21:00 https://github.com/kraj/meta-altera/commit/6a1844a415247a389fafa4e0d2890ba349ae5d7a this is wrong, 10m50 uses linux-yocto 4.4 May 17 22:24:15 khem: btw if you look into upstream or upstream bugzilla(s), you will likely find nios2 patches :) May 17 22:29:16 Marex: yeah saw some of it. We are still on 229 May 17 22:30:01 khem: applies without problem to 229 May 17 22:30:11 Marex: do you have links handy beside the above patch for any other patches for nios2/systemd May 17 22:30:21 that's all you need May 17 22:30:27 ok May 17 22:30:33 I had patch for guile 2.0 May 17 22:30:37 that's in the guile bugzilla May 17 22:30:48 there is patch for nspr in mozilla/nspr bugzilla still pending May 17 22:30:54 OK, May 17 22:31:11 khem: should be all I think May 17 22:31:19 I would suggest to put the patches either in oe-core or meta-altera meanwhile the get soaked upstream May 17 22:31:22 khem: jemalloc might need some patch if you come across it, I have one May 17 22:31:43 khem: I put whatever I have into oe-core or meta-altera for last year or so ;-) May 17 22:31:50 ok May 17 22:32:08 khem: I also tried rebootstrap (debian thing) on nios2 May 17 22:32:26 khem: I can compile all the way through (that's where the nspr,systemd,guile patches were needed) May 17 23:09:24 Marex: systemd patch should be backported into oe-core master IMO May 17 23:09:46 btw. have you run into kexec-tools ? May 17 23:10:25 or rather ported it May 17 23:10:39 IIRC it will need klibc for nios2 May 17 23:11:06 khem: regarding systemd stuff, yes it should May 17 23:11:22 khem: regarding kexec-tools, nope, I didn't get to it May 17 23:11:33 khem: what do you use to run the nios2 stuff btw ? May 18 00:25:59 Marex: pixman fails too May 18 00:26:16 Marex: the reason being nios2 does not support all sort of floating exceptions May 18 00:26:46 only seems to support round to nearest May 18 00:31:43 khem: yep May 18 00:33:27 khem: pixman, piglit will also fails for the same reason May 18 00:33:37 khem: boost as well May 18 00:33:59 I have fixed pixman May 18 00:34:28 https://lists.freedesktop.org/archives/pixman/2016-May/004608.html May 18 00:36:00 khem: http://permalink.gmane.org/gmane.comp.graphics.pixman/3954 seems like this a bit May 18 00:36:36 khem: but hrm, wouldn't it make more sense to add support for these exceptions ? **** ENDING LOGGING AT Wed May 18 02:59:58 2016