**** BEGIN LOGGING AT Thu Dec 22 02:59:57 2011 Dec 22 04:47:50 JaMa|Zzz: hmm, edit wiki to use "source" then? Dec 22 05:48:49 moo Dec 22 06:25:16 JaMa|Off: moin... sorry was already off Dec 22 06:35:22 DocScrutinizer: IIRC I've put it there already in text just the example was wrong and people are not reading text.. Dec 22 06:35:42 mrmoku: no problem.. I've found right username to try later :) Dec 22 06:36:06 mrmoku: and it's still building .. Dec 22 06:36:19 :-D Dec 22 06:36:42 mrmoku: there was also cornucopia bump, so I'll rebuild images too today Dec 22 06:39:37 JaMa|Zzz: ok, good Dec 22 06:40:06 JaMa|Zzz: I will try to do some UI fixing today Dec 22 06:40:24 JaMa|Zzz: and from tomorrow to 31st I'm away in italy Dec 22 06:40:33 back in time for the release :) Dec 22 06:41:38 ok :) Dec 22 07:49:32 SHR: 03Martin.Jansa 07shr-chroot * r4ae559f8c9cf 10/ (826 files in 18 dirs): system upgrade Dec 22 08:48:11 blah :( http://www.zdnet.com/blog/hardware/apple-granted-patent-on-using-apps-during-calls/17407 Dec 22 08:49:50 not sure if we should fix shr_elm_softkey :) Dec 22 09:12:57 JaMa|Wrk, hi Dec 22 09:13:12 libpam perl gcc libav mysql5 fails, I'm trying to cleansstate + rebuild Dec 22 09:13:20 for libpam it's a gettext issue Dec 22 09:13:40 (it doesn't find config.rpath ) Dec 22 09:17:18 GNUtoo: hi Dec 22 09:17:59 GNUtoo: rebuild from scratch finished again yesterday on jenkins Dec 22 09:20:25 freesmartphone.org: 03mickey 07cornucopia * r8a7f3d56b323 10/fsosystemd/src/bin/.gitignore: fsosystemd: add .gitignore Dec 22 09:21:40 hmmm Dec 22 09:22:38 how can I see form which package is a file in sysroot ? Dec 22 09:23:13 .../work/armv7a-vfp-neon-oe-linux-gnueabi/libpam/libpam-1.1.4-r2/Linux-PAM-1.1.4/conf/pam_conv1/pam_conv_l.c:871: undefined reference to `yywrap' Dec 22 09:23:22 let me look Dec 22 09:25:45 hmmm from libfl Dec 22 09:25:55 let me clean that Dec 22 09:26:58 maybe it's part of flex Dec 22 09:36:44 ok libpam succedded Dec 22 10:17:10 Project shr-jansa-test-crespo-aurora-image build #8: SUCCESS in 42 min: http://norman-schleicher.de:8080/job/shr-jansa-test-crespo-aurora-image/8/ Dec 22 10:51:44 Project shr-jansa-test-crespo-shr-image build #1: FAILURE in 34 min: http://norman-schleicher.de:8080/job/shr-jansa-test-crespo-shr-image/1/ Dec 22 10:54:36 JaMa|Wrk, should I rebuild from scratch again??? Dec 22 10:54:46 still some stuff fails because of a problem with libx11 Dec 22 10:55:47 and lacking libreadline, I'll look Dec 22 11:15:03 JaMa|Wrk, what's the cause of all that packaging mess? Dec 22 11:15:23 I mean you have to cleansstate stuff and rebuild stuff because some headers do not get staged Dec 22 11:15:38 should I turn off sstate? Dec 22 11:25:37 Project shr-jansa-test-nokia900-aurora-image build #9: FAILURE in 33 min: http://norman-schleicher.de:8080/job/shr-jansa-test-nokia900-aurora-image/9/ Dec 22 11:38:27 Project shr-jansa-test-nokia900-shr-image build #2: STILL FAILING in 12 min: http://norman-schleicher.de:8080/job/shr-jansa-test-nokia900-shr-image/2/ Dec 22 11:58:30 GNUtoo: sorry I don't know more that what I've already reported on oe-core ML.. Dec 22 12:17:35 ok Dec 22 12:17:37 I'll go Dec 22 12:17:42 thanks Dec 22 13:05:01 hmmm Dec 22 13:05:09 the regulator changed in 3.2 Dec 22 13:07:07 for gta03 Dec 22 13:07:10 *gta04 Dec 22 13:13:41 is that documented somewhere? Dec 22 13:28:25 SHR: 03Martin.Jansa 07meta-smartphone * r0de1f2f9c54f 10/meta-openmoko/recipes-graphics/drm/ (libdrm_2.4.27.bbappend libdrm_2.4.29.bbappend): libdrm: rename bbappend to match new version Dec 22 13:41:58 gnutoo, What I need to do with the first patch in patchwork? mark it as rejected? Dec 22 13:42:54 I'm not at home sorry Dec 22 13:43:01 so I cannot look now Dec 22 13:43:03 ping me later Dec 22 13:43:11 gnutoo, ok Dec 22 14:15:33 * mrmoku understands the bogus no free slots message on startup... Dec 22 14:15:35 will fix that now Dec 22 14:23:28 mrmoku: yeah! squash that bug! :) Dec 22 14:27:15 JaMa|Wrk: :) Dec 22 14:31:05 btw images still building.. Dec 22 14:31:19 ok Dec 22 14:31:29 including the phonefsod fix? Dec 22 14:31:33 yes Dec 22 14:31:36 good :) Dec 22 14:31:38 but not shr_elm_softkey :/ Dec 22 14:31:46 bad :P Dec 22 14:32:02 * JaMa|Wrk hopes again that he will arrive home earlier then in last 3 days Dec 22 14:32:04 I can take a look when I finished squashing this one Dec 22 14:32:28 * mrmoku won't do any daywork anyway today :/ Dec 22 14:32:44 I've read efl commits between those 2 bumps and haven't seen anything really suspicious Dec 22 14:32:53 ok Dec 22 14:33:11 and gdb shows that it's running just fine (waiting in select()) Dec 22 14:33:17 hmm Dec 22 14:33:44 GNUtoo has the theory it might just be resized wrong Dec 22 14:33:47 so it's probably just not shown Dec 22 14:33:52 he's seeing a small white line Dec 22 14:34:02 (which my eyes don't see... but that's just my eyes :P) Dec 22 14:34:13 ah could be (but I cannot see it too) Dec 22 14:34:19 ok Dec 22 14:34:39 I will take a look soon Dec 22 14:34:43 that is quite a blocker Dec 22 14:34:45 maybe the top bar and everything is really on top Dec 22 14:34:59 yeah Dec 22 14:40:53 freesmartphone.org: 03morphis 07cornucopia * r015e1da66a85 10/fsodeviced/src/plugins/herring_quirks/plugin.vala: Dec 22 14:40:53 freesmartphone.org: fsodeviced: herring_quirks: use correct name for configuration section Dec 22 14:40:53 freesmartphone.org: Signed-off-by: Simon Busch Dec 22 14:42:22 JaMa|Wrk: i can also see a withe line right under the topbar Dec 22 14:42:37 it also goes away Dec 22 14:42:54 s/withe/white/ Dec 22 14:43:27 behaves like shr_elm_softkey Dec 22 14:43:29 ahh, maybe that depends on the scaling Dec 22 14:43:53 probably Dec 22 14:43:58 * mrmoku does not select 2.0 Dec 22 14:44:15 one smaller is what I do Dec 22 14:44:22 too small for me ;) Dec 22 14:44:47 doesn't look like its behind the topbar though Dec 22 14:44:49 yeah, but the toolbar runs out of space otherwise :P Dec 22 14:45:44 what extra stuff are you putting there? Dec 22 14:47:47 the connman gadget Dec 22 14:47:56 to play with i Dec 22 14:48:08 t Dec 22 14:48:14 ah :) Dec 22 14:48:27 and I need some space to tap for the softkey Dec 22 14:48:53 right Dec 22 14:50:10 I'm rebuilding from scratch... again Dec 22 14:50:16 let's see how it goes Dec 22 14:59:35 hmm Dec 22 14:59:41 my bet is this is not good anymore: Dec 22 14:59:42 evas_object_move(win, zx, (zy + zh - finger)); Dec 22 14:59:42 evas_object_resize(win, zw, finger); Dec 22 15:02:28 why this one? Dec 22 15:02:42 I'm also looking at @66170 Dec 22 15:05:06 just a feeling Dec 22 15:05:13 added a printf there and building now Dec 22 15:12:42 heh, there was not even a trac ticket for the bug I fixed :/ Dec 22 15:12:47 SHR: 03mok 07phonefsod * r58cd417800f8 10/src/phonefsod-fso.c: phonefsod-fso: fix check for free SIM slots Dec 22 15:12:50 lol Dec 22 15:12:53 * mrmoku is missing the satisfaction to close one :P Dec 22 15:13:00 SHR: 03mok 07phonefsod * rd838a6010d14 10/src/phonefsod-fso.c: phonefsod-fso: use correct callback for display dialog Dec 22 15:13:06 close one as fixed is what I mean :-) Dec 22 15:13:13 yes I understood Dec 22 15:13:25 maybe put fixed that bug in the news then Dec 22 15:14:00 nah... not that bad Dec 22 15:14:05 I'm happy it's fixed Dec 22 15:14:10 that has to suffice Dec 22 15:14:21 * mrmoku looking at the softkey now Dec 22 15:15:38 * JaMa|Wrk will bump SRCREV Dec 22 15:15:49 thx Dec 22 15:18:52 SHR: 03Martin.Jansa 07meta-smartphone * r46645e9c0d8a 10/meta-shr/recipes-shr/shr/phonefsod_git.bb: phonefsod: bump SRCREV, SIM is full message finally solved Dec 22 15:24:50 JaMa|Wrk: SOFTKEY: zx=0, zy=0, zw=1, zh=1, finger=70 Dec 22 15:34:07 mrmoku: weird, maybe wrong theme so that icons are not there? Dec 22 15:35:17 hmm... I think that has nothing to do yet with the theme Dec 22 15:35:23 it is the size of zone[0] Dec 22 16:33:38 gah, why does that damn local build overwrite my changes now :/ Dec 22 16:34:41 hmm Dec 22 16:36:29 013 closed Dec 22 16:36:37 nice :) Dec 22 16:36:51 ok Dec 22 16:36:56 we need to build for gta04 also Dec 22 16:37:14 indeed Dec 22 16:37:27 ok 014 will have it Dec 22 16:37:32 ok nice Dec 22 17:17:59 nschle85: Hi Dec 22 17:18:11 hi :-) Dec 22 17:18:22 nschle85: I've cleaned and synchronized configs for all shr- tasks :) Dec 22 17:18:49 nschle85: now they differ only in name/description and actual task Dec 22 17:19:53 nschle85: and how differ they from current local.conf ? i saw you changed the makefile Dec 22 17:20:21 local.conf files should be also fine Dec 22 17:20:52 hope we do not run out of disk space Dec 22 17:21:54 i think i now what mistake happened deleting the job :-) Dec 22 17:22:16 ? Dec 22 17:22:31 i know why the workspace was deleted Dec 22 17:22:41 70G free and we can remove older build results now Dec 22 17:25:05 JaMa|Wrk: next time ill add a line which sets the right distro_feeds pointing to the last successful collected artefacts Dec 22 17:25:35 JaMa|Wrk: and do we change to 3.1 kernel for gta02 ? Dec 22 17:26:56 not before release I guess Dec 22 17:28:07 JaMa|Wrk: so now ill re-partition my sd card for freeing a staging partition :-) Dec 22 17:29:21 great Dec 22 17:30:00 I'll go to buy some presents after buildhost is cleaned and new build started.. :) Dec 22 17:30:28 JaMa|Wrk: hope you have success :-) Dec 22 17:47:35 014 building.. time to go from work, bbl Dec 22 17:58:26 hi mrmoku Dec 22 17:58:50 is the gta04 able to go to 1GHz without overclocking? Dec 22 17:59:06 because the fsodeviced.conf has 1GHz maximum Dec 22 17:59:15 we should not overclock Dec 22 17:59:23 that is very dangerous Dec 22 18:07:58 GNUtoo-netbook: no cpufreq anyway Dec 22 18:08:13 the plugin does not load Dec 22 18:08:21 ok Dec 22 18:08:30 but once it loads we have a problem Dec 22 18:08:51 yeah, we should change that according to what's appropriate Dec 22 18:08:58 no idea though Dec 22 18:09:30 GNUtoo-netbook: datasheet of DM3730 says 1GHz and also 800MHz etc Dec 22 18:10:18 mrmoku: hello, please read http://shr-project.org/trac/ticket/1584 Dec 22 18:10:23 ok Dec 22 18:11:02 beagleboard has same chip and runs on 1GHz too. so should be fine i guess Dec 22 18:11:06 mrmoku: and then evaluate my change, should the recipe also changed ? http://shr-project.org/trac/attachment/ticket/1556/fix_nmon13g_missing_-ltinfo.patch Dec 22 18:11:44 why do the website says 800Mhz then? Dec 22 18:12:13 nschle85: will do when back to real kbd Dec 22 18:12:38 maybe a kernel limitation? Dec 22 18:14:46 or ram Dec 22 18:14:54 GNUtoo-netbook: talking about kernel... for the release we should remove the extensive debugging from kernel config Dec 22 18:14:54 gta02 could run on 500Mhz too Dec 22 18:15:17 mrmoku, ok Dec 22 18:15:31 maybe that's something you could take care off? Dec 22 18:15:42 I decided to rebuild from scratch and it's at: Dec 22 18:16:00 NOTE: Running task 2301 of 7011 (ID: 3988, /home/gnutoo/embedded/oe/oe-core/repos/openembedded-core/meta/recipes-support/gdbm/gdbm_1.10.bb, do_package) Dec 22 18:16:17 you have time for the rest of the year :) Dec 22 18:17:51 lol Dec 22 18:18:06 I must also do stuff for morphis Dec 22 18:18:11 like make him an initramfs Dec 22 18:18:14 to boot crespo Dec 22 18:22:49 ok will take alook for myself when neil updates again Dec 22 18:23:55 ok Dec 22 18:26:05 nschle85: ahh, that one... slyon tried that and it failed at runtime Dec 22 18:26:14 nschle85: we have to examine it Dec 22 18:33:17 hmm... heh, I think I know what's going on Dec 22 18:35:43 mrmoku: which one failed ? Dec 22 18:37:10 is it 100% sure that it's caused by that patch? Dec 22 18:37:22 not because of other changes? Dec 22 18:37:39 #1556 patch is also correct Dec 22 18:38:27 JaMa|Off: I explained my thinking in the ticket Dec 22 18:38:40 will try if that is true and apply it if so Dec 22 18:39:16 nschle85: phoneuid failed at runtime after applying that pathc Dec 22 18:39:26 see my comment (and slyons comment) Dec 22 18:39:37 * JaMa|Off reloading Dec 22 18:39:46 JaMa|Off: local building is playing games with me :/ Dec 22 18:40:02 it usually works very nice... but with shr-e-gadgets I have a strange problem Dec 22 18:40:13 whenever I build my changes get overwritten Dec 22 18:40:37 vim asking me that file changed and if I want to reload Dec 22 18:42:04 JaMa|Off: makes sense? (my comment) Dec 22 18:44:42 * JaMa|Off has to open whole Makefile.am Dec 22 18:45:20 should i open ticket about gps not turning on because of sys path beeing wrong? Dec 22 18:45:41 jake42: what machine and why do you think so? Dec 22 18:45:49 gta02 Dec 22 18:46:04 frameworkd.log? Dec 22 18:46:18 mom Dec 22 18:46:32 jake42: no, I mean why do you think the path is wrong? Dec 22 18:46:38 did you see some error somehwere? Dec 22 18:46:43 jake42, with current shr-core? Dec 22 18:46:48 jake42, the patch isn't wrong Dec 22 18:47:00 the error is because it doesn't find a deprecated path Dec 22 18:47:03 from an old kernel Dec 22 18:47:06 yup Dec 22 18:47:19 it looks for neo1973 and should be gta02 Dec 22 18:47:20 and it does find paths from the new kenrel Dec 22 18:47:31 it looks for both Dec 22 18:47:33 yes so look if it works nevertheless Dec 22 18:47:44 it does something like that: Dec 22 18:47:48 oh i see ;) Dec 22 18:47:54 if (old path) Dec 22 18:48:00 //warning that you see Dec 22 18:48:05 if (new path) Dec 22 18:48:06 JaMa|Off: any idea what could cause my local build problem? Dec 22 18:48:07 //works Dec 22 18:48:20 so it goes trough the 2 pathes Dec 22 18:48:25 jake42: if you prefer different order of paths it tries (to show that error only with older kernel not with current) send patch :) Dec 22 18:48:47 mrmoku: it happens if you have some .c file opened and build from 2nd shell? Dec 22 18:49:01 yeah Dec 22 18:49:15 ah those changes are really overwritten? Dec 22 18:49:18 yup Dec 22 18:49:23 JaMa|Off: wouldn't care though gps doesn't seem to work nevertheless Dec 22 18:49:27 it reverts back :/ Dec 22 18:49:56 could be that package.bbclass stuff which packages sources to -dbg but not sure why and from where it would take "older" sources Dec 22 18:50:17 indeed strange Dec 22 18:50:17 mrmoku: isn't it because it unpacks SRC_URI sources in your S dir? Dec 22 18:50:26 mrmoku: try to empty SRC_URI Dec 22 18:50:31 hmm, ok Dec 22 18:50:55 mrmoku: just to confirm it's from SRC_URI not package.bbclass Dec 22 18:51:26 JaMa|Off: it should not use it at all, right? (SRC_URI) Dec 22 18:51:41 fatal: http://git.foo.org/repo/shr-e-gadgets.git/info/refs not valid: is this a git repository? Dec 22 18:51:44 hmm Dec 22 18:52:12 INHERIT_append_pn-shr-e-gadgets = "srctree gitpkgv" Dec 22 18:52:12 SRCREV_pn-shr-e-gadgets = "${GITSHA}" Dec 22 18:52:12 S_pn-shr-e-gadgets = "/OE/source/${PN}" Dec 22 18:52:25 is just copied from the other entries which work Dec 22 18:52:30 and can't see no error there Dec 22 18:55:20 and srctree.bbclass empties the SRC_URI Dec 22 18:56:41 mrmoku: does it do this only in last few days? Dec 22 18:56:51 mrmoku: so I'll check bitbake logs Dec 22 18:57:00 ahh Dec 22 18:57:03 gitkpkgv is using it Dec 22 18:57:04 src_uri = bb.data.getVar('SRC_URI', d, 1).split() Dec 22 18:57:24 JaMa: no, I built phonefsod this way just fine today Dec 22 18:57:37 heh, at least I hope so :P Dec 22 18:57:45 and you're not using gitpkgv for phonefsod? Dec 22 18:57:47 well.. the bug was fixed, so yes Dec 22 18:57:52 sure I do Dec 22 18:58:05 INHERIT_append_pn-phonefsod = "srctree gitpkgv" Dec 22 18:58:05 SRCREV_pn-phonefsod = "${GITSHA}" Dec 22 18:58:05 S_pn-phonefsod = "/OE/source/${PN}" Dec 22 18:59:31 let me check if I have the same with shr-e-gadget Dec 22 18:59:38 cool, thx Dec 22 18:59:42 +s Dec 22 19:01:45 Project shr-core-nokia900-shr-image build #24: SUCCESS in 1 hr 29 min: http://norman-schleicher.de:8080/job/shr-core-nokia900-shr-image/24/ Dec 22 19:07:08 mrmoku: seems to be working here Dec 22 19:07:22 mrmoku: just -c cleansstate and -c build, both with -b Dec 22 19:08:15 * JaMa trying once more Dec 22 19:08:34 hmm... ok maybe I should try with -b Dec 22 19:08:59 saving between do_fetch and do_compile should work too ;) Dec 22 19:10:02 mrmoku: ah reproduced here Dec 22 19:10:19 mrmoku: it doesn't overwrite ie TODO file but with main.c it "works" Dec 22 19:10:38 hehe good :) Dec 22 19:10:57 * mrmoku couldn't care less when it would be TODO that get's overwritten :-P Dec 22 19:13:38 hehe :) Dec 22 19:13:56 but what it appends to it? :P Dec 22 19:14:27 appends? Dec 22 19:15:20 adds more items to TODO.. :) Dec 22 19:15:28 uhh nooooooooooo :P Dec 22 19:15:59 well from main.c it removed all the nice stuff I added... so I had the same hope for TODO too :-) Dec 22 19:16:03 brb Dec 22 19:23:25 freesmartphone.org: 03felix.huber 07zhone * rc4a4debe917b 10/src/zhone: Fix strange indentation mess Dec 22 19:30:02 mrmoku: nschle85: I've checked log.do_compile from phoneuid, before and after that patch and it looks right, no idea why it should cause such runtime issue Dec 22 19:31:55 JaMa: so i was unsure how to add libraries LDADD or LDFLAGS but both are correct true ? Dec 22 19:32:26 LDADD is better Dec 22 19:32:53 but from recipe you don't have LDADD available, that's why LDFLAGS is used sometimes Dec 22 19:37:36 JaMa: well the runtime error is because of g_critical leading to an abort Dec 22 19:37:59 JaMa: which we don't won't and disable by that g_log_set_fatal_mask Dec 22 19:39:01 JaMa: if by reordering stuff in the linking command line G_LOG_DOMAIN is now defined to "phoneuid" and was not before that would explain it Dec 22 19:39:05 (I think) Dec 22 19:39:19 still have to try... wanted to finish fixing the softkey first :/ Dec 22 19:40:21 hmm Dec 22 19:40:28 Btw, what's the deal with that softkey not starting automatically? Dec 22 19:40:47 ohh Dec 22 19:40:50 PaulFertser: it is starting Dec 22 19:40:54 it's just not visible Dec 22 19:41:11 JaMa: there is one big difference between phonefsod and shr-e-gadgets :) Dec 22 19:41:22 one in autorev and the other not Dec 22 19:41:25 wrt overwrites right? Dec 22 19:41:30 both not autorev for me Dec 22 19:41:55 it's overwritten in do_patch or do_configure Dec 22 19:42:15 yeah, do_fetch does not overwrite it Dec 22 19:42:23 does not fetch into git either Dec 22 19:43:17 -c patch it is Dec 22 19:43:45 that's the difference shr-e-gadgets has local patch Dec 22 19:44:11 which I can apply now anyway Dec 22 19:45:24 mrmoku: do you have it locally? patch was added by you so please push it to shr-e-gadgets repo Dec 22 19:45:44 I'll drop it from meta-smartphone Dec 22 19:45:48 JaMa: it's populate lic Dec 22 19:46:03 hmm Dec 22 19:46:34 I thought it is populate lic because that was the moment when vim complained... but that might have a delay Dec 22 19:46:49 it's do_patch for sure :) Dec 22 19:46:50 so if you verified do_patch I will just apply the patch Dec 22 19:46:53 ok Dec 22 19:46:54 :) Dec 22 19:46:58 I was running them one by one Dec 22 19:47:24 I'll look into it later but to let you work please apply to fix local builds now Dec 22 19:51:36 mrmoku: probably if you set PATCHTOOL = "patch" then it will not overwrite it too Dec 22 19:52:24 JaMa: yeah, thx very much Dec 22 19:52:34 and did work too :) Dec 22 19:53:33 do you have SRCREV for me? :) Dec 22 19:53:54 nah, building worked Dec 22 19:54:04 still have to see if it is fixed Dec 22 19:54:44 ok, then don't forget to push that OE patch too as I've already removed it here :) Dec 22 19:54:59 yeah, ok Dec 22 19:55:15 btw with latest EFL it doesn't work too Dec 22 19:55:22 and it even hangs after a while Dec 22 19:55:35 JaMa: does rebuilding the index work with -b too ? Dec 22 19:55:41 I think we did that some time ago Dec 22 19:55:49 it is dead slow Dec 22 19:56:23 -b won't help much, it's dead slow inside Dec 22 19:58:36 hmm ok Dec 22 19:58:49 here is more info, but no solution: http://www.mail-archive.com/openembedded-core@lists.openembedded.org/msg12421.html Dec 22 20:02:25 ok Dec 22 20:03:10 hmm did not work... and devilhorns is not online Dec 22 20:03:41 ~seen devilhorns Dec 22 20:03:48 ahh... apt not there Dec 22 20:03:53 mrmoku: i haven't seen 'devilhorns' Dec 22 20:04:01 not there like in not in e chans Dec 22 20:09:10 mrmoku: did PATCHTOOL setting work for you? Dec 22 20:09:18 mrmoku: for libphone-ui-shr it still fails here Dec 22 20:12:45 ah my fault Dec 22 20:13:01 did not try Dec 22 20:13:02 forgot to clean after it was applied first Dec 22 20:19:12 heh local-build.inc example is wrong Dec 22 20:30:21 mrmoku: there? Dec 22 20:30:35 I need a guru. The only one I may know is you :-) Dec 22 20:31:00 I have the phone of Rael if you need Dec 22 20:31:15 :-) Dec 22 20:31:24 I'm playing with policy based routing Dec 22 20:31:38 huh Dec 22 20:31:42 I have two gateways and I'm trying to use one for http using marking and ip route Dec 22 20:31:48 But it won't work properly I don't know why Dec 22 20:31:55 Ainulindale: policy based routing? no guru for that Dec 22 20:32:00 Packet is marked properly I can see it in the iptable chain Dec 22 20:33:16 Ainulindale: I need a guru too... an e guru ;) Dec 22 20:33:28 That's not me Dec 22 20:33:33 If it's graphical in any way Dec 22 20:33:34 That's not me Dec 22 20:33:39 yeah, like me is no routing guru :P Dec 22 20:33:48 but maybe PaulFertser is Dec 22 20:34:21 PaulFertser is just a guru no matter which topic Dec 22 20:34:28 So yeah he might know Dec 22 20:34:42 But it's annoying Dec 22 20:35:25 the main goal being Dec 22 20:35:32 I have gateway1 and gateway2 on the same subnet Dec 22 20:35:37 And I want to redirect http to gateway1 Dec 22 20:35:42 and everything else to gateway2 Dec 22 20:35:49 s/redirect/route/ Dec 22 20:36:59 If you have any kind of sensible solution I'm in Dec 22 20:40:04 Ainulindale: hey :) Dec 22 20:40:17 Ainulindale: but how exactly do you setup your routes? Dec 22 20:42:03 arda ~ # ip ru Dec 22 20:42:04 0: from all lookup local Dec 22 20:42:04 2500: from all fwmark 0x2 lookup nc Dec 22 20:42:04 32765: from all fwmark 0x1 lookup free Dec 22 20:42:16 table nc has a default route for gw 1 Dec 22 20:42:20 and free gw 2 Dec 22 20:42:30 as simple as it gets Dec 22 20:42:37 I must have forgotten a stupid thing Dec 22 20:43:34 If I don't have any default roule in the default rt_table Dec 22 20:43:39 It fails Dec 22 20:43:48 If I do, it'll keep using it Dec 22 20:45:44 Ainulindale: so it looks like it never matches fwmark, right? Dec 22 20:45:59 it does match the chains in iptables Dec 22 20:46:19 27640 6935K restore tcp -- eth2 any anywhere anywhere state RELATED,ESTABLISHED Dec 22 20:46:22 23 1252 nc tcp -- any any anywhere anywhere state NEW tcp dpt:http Dec 22 20:46:25 3685 396K free all -- any any anywhere anywhere state NEW Dec 22 20:46:28 (mangle, prerouting) Dec 22 20:46:40 (eth2 being the iface on which the two gateways are located) Dec 22 20:47:20 Ainulindale: what about the RELATED,ESTABLISHED packets that match dpt:http? Dec 22 20:47:41 they're still marked Dec 22 20:47:53 Chain restore (1 references) pkts bytes target prot opt in out source destination Dec 22 20:47:56 27640 6935K CONNMARK all -- any any anywhere anywhere CONNMARK restore Dec 22 20:48:13 (the free and nc chains are using mark and connmark save) Dec 22 20:53:19 Ainulindale: where do the packets go (and from which address) if you tcpdump sniff on the outbound interface? Dec 22 20:53:54 Without the rule to lookup the default rt_table. Dec 22 20:54:37 well if I remove the rule then I lose connection so... :-) Dec 22 20:55:06 Ainulindale: are you testing from the host itself? Dec 22 20:55:29