**** BEGIN LOGGING AT Thu May 20 03:00:02 2010 **** BEGIN LOGGING AT Thu May 20 03:51:34 2010 May 20 04:32:38 florian, are you around? May 20 05:57:59 jow * r21511 /trunk/package/base-files/files/lib/functions/boot.sh: [package] base-files: fix a preinit hang in jffs2_ready() when no rootfs_data partition is found (#7362) May 20 06:03:41 jow * r21512 /packages/net/ddns-scripts/files/usr/lib/ddns/dynamic_dns_functions.sh: [packages] ddns-scripts: fix overriding ip_network (#7361) May 20 06:09:51 jow * r21513 /trunk/package/iproute2/ (7 files in 2 dirs): [package] iproute2: update to v2.6.31 (#7354) May 20 06:14:26 jow * r21514 /trunk/package/iproute2/Makefile: [package] iproute2: cleanup Makefile after r21513 May 20 06:15:03 <[florian]> Strykar: yes I am May 20 06:53:29 jow * r21515 /packages/net/haproxy/Makefile: [packages] haproxy: update to v1.4.6, tune optimization flags (#7350) May 20 06:55:56 bleh. I've had no time to work on openwrt .... 37 hours working so far in the past three days May 20 07:02:23 cshore: ;) May 20 07:04:26 I need to get rich so I can mess about with projects I really care about all day long May 20 07:04:33 jow * r21516 /trunk/package/uci/Makefile: [package] uci: revert pkg rev to 5 until updated sources are available (#7351) May 20 07:11:27 jow * r21517 /trunk/package/nvram/files/nvram.init: [package] nvram: fix nvram set invokation after r21497 May 20 07:17:38 cshore: ping ? May 20 07:17:49 pong May 20 07:17:54 Morning :) May 20 07:18:39 I think the extroot-sdcard script should be working now, and exit silently if the sdhc module does not exist May 20 07:18:47 However, someone posted an interesting remark on the forum May 20 07:19:07 "i see you duplicated some code form the sdcard init script. it would be cleaner to reuse the existing functions or create some library script under /lib/functions and reuse this in the init and the extroot script" May 20 07:19:27 right May 20 07:19:41 It sounds reasonable enough May 20 07:20:12 That library script should probably belong to broadcom-sdhc then May 20 07:20:33 yeah May 20 07:20:35 yes...unless it's common for other packages as well May 20 07:20:43 Hm not really May 20 07:20:53 sdhc specific stuff is involved in there May 20 07:21:14 Namely the GPIO settings (both for the module options and the gpiomasks) May 20 07:21:33 Which leads me to a second question May 20 07:22:03 I've read the "howto submit a patch" form the wiki May 20 07:22:14 yes? May 20 07:22:35 Is there any difference in the procedure when the patches are made against a package (like in this case) ? May 20 07:22:42 it'll be two because you'll be going from both packages and trunk May 20 07:22:44 (sorry sluggish typing, just woke up) May 20 07:22:48 Yes two indeed May 20 07:22:56 erm May 20 07:23:01 come again ? May 20 07:23:26 do you know how to use a local copy of the packages repository rather than just feeds.conf? May 20 07:23:48 Hm, I don't think I've ever done that May 20 07:24:07 There's a svn for packages too ? May 20 07:24:14 yes May 20 07:24:25 ok then May 20 07:24:32 1) checkout svn://svn.openwrt.org/openwrt/packages May 20 07:24:36 But then the patch will have nothing to do with trunk in fact May 20 07:24:43 2) copy feeds.conf.default to feeds.conf May 20 07:25:07 3) change the entry for packages from src-svn to src-link, point it to the absolute path of the packages checkout May 20 07:25:36 broadcom-sdhc will be moved to packages soon May 20 07:25:56 Oh, it's in trunk atm ? I hadn't noticed May 20 07:26:04 yeah in package/ May 20 07:26:13 Hm, I am confused now May 20 07:26:21 What is the difference between trunk/package and packages ? May 20 07:26:39 trunk/package is supposed to contain only essential stuff May 20 07:26:58 Ah I see. May 20 07:27:05 Okay so let's recap: May 20 07:27:22 1) get a fresh copy of the svn packages repo May 20 07:27:35 2) patch against packages for block-extroot May 20 07:27:49 3) patch against trunk/package for broadcom-sdhc (since it's there for the moment) May 20 07:27:53 4) mail the lot May 20 07:27:56 ? May 20 07:27:58 correct May 20 07:28:03 or do a ticket May 20 07:28:07 whatever you prefer May 20 07:28:12 Ok, thanks a lot May 20 07:28:35 I'll try the packages stuff right now May 20 07:28:57 one thing you need to keep in mind May 20 07:29:02 Yes ? May 20 07:29:18 ./scripts/feeds update will not do "svn up" on the checkout anymore if you use the src-link method May 20 07:29:27 it will just re-read the directory May 20 07:29:37 Yeah, I'll need to svn up manually from ../packages May 20 07:29:39 so you have to do the svn up manually there in the future May 20 07:29:43 correct May 20 07:29:44 Dully noted May 20 07:29:47 duly* haha May 20 07:30:00 Morning typos. May 20 07:30:22 you should also uninstall all packages before switching the source May 20 07:30:35 erm no, scratch that May 20 07:30:40 feeds should handle that May 20 07:30:44 heh ok :) May 20 07:31:11 checking out May 20 07:33:06 anyone seen epointsystems lately. I want to talk to him about the block-mount of /dev/sda vs /dev/sda1 that he needs (i.e. whole device filesystems) May 20 07:33:58 cshore: search the openwrt-devel list archive to find his mail addr May 20 07:34:26 Updating feed 'packages' from '/home/smokey/openwrt/packages' ... May 20 07:34:30 Ok that worked :) May 20 07:36:07 xMff: so broadcom-sdhc will move to packages, but extroot stays in trunk ? May 20 07:36:27 Mr_Smoke: yes May 20 07:36:46 Ok so I won't patch anything against packages/ actually May 20 07:36:48 Mr_Smoke: why do you need to patch extroot btw May 20 07:37:17 Mr_Smoke: can't sdhc just provide a lib file that is placed in extroots lib dir? May 20 07:37:21 xMff: because if extroot lives on the sdhc, the the module needs to be loaded before the extroot overlay can be mounted (even detected) May 20 07:37:27 hm May 20 07:37:49 otherwise all extroot users will have sdhc glue around thats usually not needed May 20 07:37:58 no big deal, maybe it needs some stuff May 20 07:38:13 but if it is just adding a new file then there should be no need to patch extroot May 20 07:38:29 only if you need to alter standard stuff (special cases and such) May 20 07:38:31 You're right, but I think we had that discussion before and the outcome was to patch extroot, can't remember why exactly May 20 07:39:17 The only need is that the sdhc preinit stuff should happen before extroot attempts to mount the extroot overlay, so I guess broadcom-sdhc can handle that by just putting the lib/preinit file in the right place indeed May 20 07:39:29 I think I numbered it properly already May 20 07:39:52 yes, since both packages must be built-in anyway, the install order should be no issue May 20 07:40:07 Yeah, it makes sense. Will do that: create a generic functions script which the initscript and preinitscript will refer to, and place it all in broadcom-sdhc May 20 07:41:03 The only thing is that if we do this, then the sdhc will be mounted during preinit, always May 20 07:41:11 So perhaps I should add a flag in the conf file May 20 07:41:21 Just in case the user doesn't want that behaviour May 20 07:41:44 maybe May 20 07:41:50 I see no adverse side-effects, but maybe it's worth it anyways May 20 07:41:53 how do you handle the device specific gpio stuff? May 20 07:42:15 /etc/sdhc.conf I think May 20 07:42:20 is the user required to edit the sources prior to building? May 20 07:42:35 Mr_Smoke: you could have a function defined in a preinit file with a number that is greater than the function you are altering, and in that file add a new function of the same name that does the sdhc specific stuff. That will override the standard behaviour with sdhc behaviour. And that doesn't require patching extroot May 20 07:42:38 Well I haven't gone through that testing already: May 20 07:42:43 ok May 20 07:43:07 I think if the user edits the config file, and the preinit script is run in its proper turn, then it should be able to read the jffs2 data May 20 07:43:17 So that the user won't need to edit the source before compiling May 20 07:43:17 maybe you should create a Config.in file in your package to provide a menuconfig menu to specify the gpios, so that they're correctly prepopulated May 20 07:43:21 aah ok May 20 07:43:21 But that needs double checking May 20 07:43:53 hm yeah, right. sdhc can be initialized after jffs2 May 20 07:43:57 cshore: that's also very true, but I don't need to replace anything in fact, it's all new stuff May 20 07:44:08 (from the preinit point of view) May 20 07:44:27 It's just a matter of making /dev/sdcard available when determint_usb_root (or sthg) gets called May 20 07:44:43 xMff: yes, that's what I need to check May 20 07:44:47 Mr_Smoke: if you use the built-in flash the same way extroot does you can do what you want May 20 07:45:13 cshore: what do you mean ? May 20 07:45:47 extroot uses the flash to define the device (partition) used for the root filesystem May 20 07:46:01 and uses UCI to read the config info from fstab May 20 07:46:41 Oh yeah, you mean it reads /e/c/fstab from jffs2 May 20 07:46:55 and functions.sh for the processing of it ? May 20 07:47:03 (in layman's terms) May 20 07:47:22 sthg like that May 20 07:47:29 yeah I get the gist May 20 07:47:52 package/block-extroot/files/50_determine_usb_root May 20 07:48:15 and package/block-extroot/files/extmount.sh May 20 07:48:24 set_jffs_mp May 20 07:48:59 Yeah, I'll work out something similar then May 20 10:04:57 build #43 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/43 May 20 11:32:10 nbd: ping May 20 12:12:31 build #45 of s3c24xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx/builds/45 May 20 12:13:30 xMff: quick check here, just making sure I get the preinit sequence stuff : before any decision is made on where to pivot root, jffs2 (whether it be all_jffs2 or just the rw part of squashfs/jffs2) is mounted to /tmp/overlay, right ? May 20 12:14:56 I got that from /lib/preinit/40_mount_jffs2 May 20 12:59:09 jow_laptop: there ? May 20 13:07:59 <[florian]> rtz2: ping? May 20 13:08:51 oh hi there [florian] May 20 13:09:07 Can I ask you a quick question about patch submission ? May 20 13:10:05 <[florian]> sure May 20 13:11:03 [florian]: pong May 20 13:11:23 Ok, I'm working on a patch for broadcom-sdhc May 20 13:11:25 [florian]: sorry, I wasn't able to test the patch yet May 20 13:11:43 I'm not sure whether I should come up with patches that live in $package/patches May 20 13:11:53 [florian]: the device is in use right now and I can't take it down easiyl May 20 13:12:03 Or whether to send them to (whoever) so that the tarball is changed 'upstream' May 20 13:12:12 What's the best practice here ? May 20 13:12:14 [florian]: I plan to do it sometime this night, so I can tell you tomorrow May 20 13:12:38 <[florian]> rtz2: cool, I am fairly confident it should work fine, r6040 has a working mdio bus, unlike ar7 for instance ;) May 20 13:12:39 I assume it's the former, but this package is really openwrt-specific, so it might be different May 20 13:12:58 <[florian]> Mr_Smoke: well you can come up with patches for $package/patches and send them to openwrt-devel May 20 13:13:08 <[florian]> Mr_Smoke: but even better, you can probably send such patches upstream May 20 13:14:03 [florian]: any opinion about that: https://lists.openwrt.org/pipermail/openwrt-devel/2010-May/007133.html ? May 20 13:14:52 [florian]: that's the deal, I'm not sure exactly who upstream is for broadcom-sdhc ... even less so as it is apparently about to change maintainers :) May 20 13:15:48 <[florian]> Mr_Smoke: ah about broadcom-sdhc, well just try to find the most active guy on the package May 20 13:16:12 [florian]: it's pretty much ignored in openwrt May 20 13:16:37 <[florian]> rtz2: the package depends pretty closely on the kernel so I believe it should remain in trunk/package/ May 20 13:17:42 Ok then May 20 13:17:48 Cheers :) May 20 13:18:09 <{Nico}> [florian]: that's not a requirement anymore, we have many kernel modules in packages/ (drivers, xtables-addons...) May 20 13:18:28 <[florian]> {Nico}: oh yes, that's right May 20 13:18:32 [florian]: in this case, somebody should feel responsible for reviewing patches for it May 20 13:18:36 <[florian]> Mr_Smoke: was not buddhay or someone like this pretty active? May 20 13:18:51 the current situation is not nice May 20 13:19:12 [florian]: he doesn't have write access and his patch is still pending May 20 13:19:40 [florian]: or ignored May 20 13:19:48 depending on how you see it May 20 13:21:21 <[florian]> rtz2: who are we talking about Mr_Smoke or buddhay? May 20 13:21:54 I haven't actually submitted any real patch so far May 20 13:22:15 [florian]: buddhay May 20 13:22:17 I think May 20 13:23:04 buddhay May 20 13:23:13 This might seem like a silly question but ... apart from preinit integration, is sdhc likely to evolve that much ? May 20 13:23:53 Mr_Smoke: does it support other devices besides broadcom? May 20 13:24:38 rtz2: the source says it should, but I have clearly no idea May 20 13:25:10 Mr_Smoke: well, this for example would be a nice future May 20 13:25:30 True May 20 13:25:56 Who's the current maintainer ? May 20 13:26:10 <[florian]> I think nobody is May 20 13:26:14 Ah May 20 13:26:16 I see May 20 13:26:22 <[florian]> because we did not grant commit access to buddhay so that he can maintain it May 20 13:26:26 I'll get in touch with bud then May 20 13:26:41 <[florian]> moving broadcom-shdc apart would be a good occasion to give him commit access as well May 20 13:26:52 Sounds good to me :) May 20 13:27:00 thus my proposal to move to move it to packages May 20 13:27:25 All I know is I've got sdhc/preinit working and I just wanted to submit stuff so that others could use it too :) May 20 13:27:41 Got to go to work now, see you later. May 20 13:29:34 hi :) May 20 13:30:20 I'm toying around with backfire and openwrt backfire. System works so far but I can't see any network device. May 20 13:30:26 (neither on 2.6 nor on 2.4) May 20 14:18:06 <[florian]> Gottox: mentionning your device would certainly help ;) May 20 14:28:42 [florian]: wl500w May 20 14:28:59 I fixed it myself by installing -mimo :) May 20 14:29:50 <[florian]> ah ah **** BEGIN LOGGING AT Thu May 20 16:18:43 2010 May 20 18:19:15 jow_laptop: what do you think about this: https://lists.openwrt.org/pipermail/openwrt-devel/2010-May/007133.html ? May 20 20:45:25 rtz2: I spoke with him already, he sent a request for maintainership and I'll move it to the package feed once he got access May 20 20:46:27 jow_laptop: ok May 20 23:14:10 nbd: you still here? May 20 23:14:21 nbd: got disconnected ... **** ENDING LOGGING AT Fri May 21 02:59:57 2010