**** BEGIN LOGGING AT Thu Oct 02 03:00:01 2014 Oct 02 05:55:04 rmilecki r42715 trunk/target/linux/bcm53xx/Makefile * bcm53xx: select LEDs support by default Oct 02 09:52:45 rmilecki r42716 trunk/target/linux/bcm53xx/patches-3.14/410-mtd-bcm47xxpart-alloc-memory-for-more-partitions.patch * bcm53xx: support devices with more (13+) partitions Oct 02 10:11:04 build #704 of ppc44x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/704 Oct 02 10:43:16 rmilecki r42717 trunk/target/linux/bcm53xx/patches-3.14/122-bcma-fill-core-details-for-every-device.patch * bcm53xx: backport bcma RFC patch that fixes LEDs Oct 02 11:07:16 zajec: ping Oct 02 11:08:07 nbd: Oct 02 11:08:45 does the device that you're working on have a recovery method that works without access to the serial console? Oct 02 11:10:26 don't know Oct 02 11:10:32 i do CTRL+C in CFE booting Oct 02 11:10:35 then "tftpd" Oct 02 11:10:49 and upload firmware using ( cd bin/bcm53xx/; tftp 192.168.1.1 -m binary -c put openwrt-bcm53xx-bcm4708-netgear-r6250-squashfs.chk ) Oct 02 11:11:40 we need to be careful about what kinds of recovery methods we put into the hands of users Oct 02 11:12:00 if the easiest way to recover is one where you have to erase all flash, that's a bad thing Oct 02 11:12:25 nbd: i'm afraid Broadcom always works this way Oct 02 11:12:45 we have a choice here Oct 02 11:13:03 nbd: I agree about smart upgrade OpenWrt -> OpenWrt Oct 02 11:13:11 but what about recovery Oct 02 11:13:15 or vendor -> OpenWrt Oct 02 11:13:24 we need an initramfs takeover image Oct 02 11:13:40 with a small web interface or shell for flashing the real image Oct 02 11:14:12 omg :| Oct 02 11:14:16 this can't bne Oct 02 11:16:01 so here's what i'd suggest: Oct 02 11:16:17 - we don't build images with generated ubi flash layout Oct 02 11:16:56 sysupgrade images have no ubi structure, only the block images suitable for pushing into ubi volumes Oct 02 11:17:39 vendor -> openwrt flashing is supported through an initramfs image that we build once (and rebuild whenever changes make it necessary) Oct 02 11:17:55 this initramfs takes regular sysupgrade images and does ubiformat only when necessary (or requested by the user) Oct 02 11:18:35 we could make the initramfs image flashable for devices that support flashing an image without direct access to the serial console Oct 02 11:24:44 zajec: what do you think? Oct 02 11:24:48 sec Oct 02 11:25:23 i don't get this ubiformat thing Oct 02 11:25:37 why some changes survive the fresh install? Oct 02 11:25:50 doesn't UBI use ubiformat on auto resize? Oct 02 11:25:54 maybe this needs some trick? Oct 02 11:26:23 you should just assume that ubi sprays its data all over the flash Oct 02 11:26:42 and trying to force it to discard data when new data is written by the boot loader is a bad idea Oct 02 11:26:50 because then it'll have to erase the flash all over again Oct 02 11:27:36 and what if you build some magic initramfs thing? Oct 02 11:27:41 you'll have to erase the flash anyway Oct 02 11:27:50 when flashing from vendorm firmware or recovering Oct 02 11:28:21 so, here's how it works: Oct 02 11:28:53 i'm afraid we'll make it overcompicated for users... and ourself as well (to maintain) Oct 02 11:28:53 since nand flash is big, we reserve enough room in the kernel partition to have an initramfs image fit in there, with a big safety margin Oct 02 11:28:57 * zajec listening to Oct 02 11:29:07 so when you flash the initramfs image, it won't touch the ubi area at all Oct 02 11:29:25 so for openwrt -> openwrt recovery, no ubiformat is necessary Oct 02 11:29:42 as for vendor -> openwrt - when the initramfs image detects that the first block in the ubi area is not ubi, it does ubiformat Oct 02 11:30:20 that way we minimize unnecessary flash wear Oct 02 11:30:27 as for overcomplicated: Oct 02 11:30:38 we'd intentionally sacrifice the easiest possible upgrade path Oct 02 11:30:47 but for a good and (imho) important cause Oct 02 11:30:52 and use the next best thing instead Oct 02 11:31:36 OpenWrt -> OpenWrt is totally different story, we have UBI, UBI tools, we can do all the magic Oct 02 11:31:43 and don't lost wear info Oct 02 11:31:45 i mean openwrt -> openwrt *recovery* Oct 02 11:31:47 that is easy, cool Oct 02 11:31:52 i.e. original system doesn't boot Oct 02 11:31:59 oh Oct 02 11:32:00 sec Oct 02 11:33:09 btw. for the buffalo wzr-1750dhp the initramfs path is quite easy Oct 02 11:33:16 if you have a tftp server Oct 02 11:33:43 because you can just press the aoss button at boot, and it'll try to download a lzma compressed firmware.ram image from the tftp server Oct 02 11:33:51 and boot it directly (without flashing anything) Oct 02 11:34:01 that explains your ideas Oct 02 11:34:03 ;p Oct 02 11:34:20 you're focusing on Buffalo solutions I guess Oct 02 11:34:24 no Oct 02 11:34:30 because for normal devices I still can't see the point Oct 02 11:34:30 i'm focusing on preserving nand flash Oct 02 11:34:46 i'm pretty sure we will see devices with MLC flash Oct 02 11:34:56 give me a sec, i'll write my questio Oct 02 11:36:32 nbd, Hi. Did you get my ubus_connect_ctx() ubus patch? Oct 02 11:37:19 yep Oct 02 11:37:30 will take another look later, but on first glance it looks good Oct 02 11:37:41 nbd: i think we want to implement the same thing, but in other ways. I think you prefer some special image with initramfs that detects fresh installation and does ubiformat when necessary. I prefer simply doing ubiformat on the UBI auto resize, which also should happen after the installation only. Oct 02 11:38:25 right, but if i understand correctly, what you're describing will do the autoformat every time you reflash via boot loader Oct 02 11:38:34 and that's the part i don't like Oct 02 11:39:05 well, yes Oct 02 11:39:13 what about your solution? Oct 02 11:39:23 it will detect there was already an OpenWrt installed I guess? Oct 02 11:39:26 yes Oct 02 11:39:33 OK, then I've to adjust my solution Oct 02 11:39:46 i don't see a way to do it without initramfs Oct 02 11:39:50 I'd need to make it do the format only if there isn't UBI Oct 02 11:39:52 if you can come up with one, it would be nice Oct 02 11:39:57 i'll try Oct 02 11:40:37 because i think combining preserving ubi formatting and flashing pregenerated ubi data just doesn't work Oct 02 11:40:47 they don't mix well Oct 02 11:41:32 ndb, ack. thanks. Oct 02 12:06:44 rmilecki r42718 trunk/target/linux/ bcm53xx/patches-3.14/301-ARM-BCM5301X-Add-buttons-support-for-Netgear-R6250.patch bcm53xx/Makefile * bcm53xx: early buttons support (without interrupts) Oct 02 12:13:58 haha, bcm53xx is the no 1 in the log now ;) Oct 02 12:19:07 now what can we do about that... ;) Oct 02 12:19:15 nbd r42719 trunk/package/network/utils/uqmi/files/lib/netifd/proto/qmi.sh * uqmi: work out the ifname instead of relying on it being in uci Oct 02 12:19:29 nbd r42720 trunk/package/network/utils/uqmi/files/lib/netifd/proto/qmi.sh * uqmi: replace logger calls in netifd with echo Oct 02 12:19:49 nbd r42721 trunk/package/network/utils/uqmi/files/lib/netifd/proto/qmi.sh * uqmi: use the autoconnect feature Oct 02 12:19:58 nbd r42722 trunk/package/network/utils/uqmi/Makefile * uqmi: update to the latest version Oct 02 12:20:14 nbd r42723 trunk/package/network/utils/uqmi/files/lib/netifd/proto/qmi.sh * uqmi: do not wait for network connection before starting dhcp Oct 02 12:20:25 nbd r42724 trunk/package/network/utils/uqmi/files/lib/netifd/proto/qmi.sh * uqmi: set data format to 802.3 at startup Oct 02 12:23:06 :P Oct 02 12:24:51 jogo r42725 trunk/package/kernel/broadcom-wl/patches/910-fallback-sprom.patch * brcm-wl: use new brcm63xx fallback sprom Oct 02 12:25:00 jogo r42726 trunk/target/linux/brcm63xx/patches-3.14/361-MIPS-BCM63XX-add-raw-fallback-sproms-for-most-common.patch * brcm63xx: fix brcm63xx fallback typo Oct 02 12:25:16 jogo r42727 trunk/target/linux/ (36 files in 4 dirs) * brcm63xx: Add DT support for 963281TAN Oct 02 12:25:20 jogo r42728 trunk/target/linux/ brcm63xx/dts/bcm96345GW2.dts brcm63xx/profiles/01-generic.mk brcm63xx/image/Makefile * brcm63xx: Add DT support for 96345GW2 Oct 02 12:25:42 jogo r42729 trunk/target/linux/ brcm63xx/dts/bcm96348GW.dts brcm63xx/profiles/01-generic.mk brcm63xx/image/Makefile * brcm63xx: Add DT support for 96348GW Oct 02 12:25:47 jogo r42730 trunk/target/linux/ brcm63xx/dts/bcm96348GW-10.dts brcm63xx/profiles/01-generic.mk brcm63xx/image/Makefile * brcm63xx: Add DT support for 96348GW-10 Oct 02 12:25:55 jogo r42731 trunk/target/linux/ brcm63xx/dts/bcm96348GW-11.dts brcm63xx/profiles/01-generic.mk brcm63xx/image/Makefile * brcm63xx: Add DT support for 96348GW-11 Oct 02 12:25:59 jogo r42732 trunk/target/linux/ brcm63xx/dts/bcm96348R.dts brcm63xx/profiles/01-generic.mk brcm63xx/image/Makefile * brcm63xx: Add DT support for 96348R Oct 02 12:26:09 jogo r42733 trunk/target/linux/ brcm63xx/profiles/01-generic.mk brcm63xx/image/Makefile brcm63xx/dts/bcm96338GW.dts * brcm63xx: Add DT support for 96338GW Oct 02 12:26:13 jogo r42734 trunk/target/linux/ brcm63xx/patches-3.14/370-bcm63xx-add-of_board_ids-for-all-supported-boards.patch brcm63xx/dts/bcm96338W.dts brcm63xx/profiles/01-generic.mk brcm63xx/image/Makefile * brcm63xx: Add DT support for 96338W Oct 02 12:26:24 jogo r42735 trunk/target/linux/ brcm63xx/profiles/01-generic.mk brcm63xx/dts/bcm96358VW.dts brcm63xx/image/Makefile * brcm63xx: Add DT support for 96358VW Oct 02 12:26:27 jogo r42736 trunk/target/linux/ brcm63xx/image/Makefile brcm63xx/profiles/01-generic.mk brcm63xx/dts/bcm96358VW2.dts * brcm63xx: Add DT support for 96358VW2 Oct 02 12:26:28 posers ;) Oct 02 12:26:36 :P Oct 02 12:26:43 jogo r42737 trunk/target/linux/ (32 files in 4 dirs) * brcm63xx: Add DT support for 96368MVNgr Oct 02 12:26:53 jogo r42738 trunk/target/linux/ (33 files in 4 dirs) * brcm63xx: Add DT support for 96368MVWG Oct 02 12:27:06 jogo r42739 trunk/target/linux/brcm63xx/patches-3.14/ (36 files) * brcm63xx: reorder some patches according to their upstream state Oct 02 12:27:22 jogo r42740 trunk/target/linux/brcm63xx/patches-3.14/ (45 files) * brcm63xx: update spw500v sprom to values from gpl tarball Oct 02 12:27:29 zajec: better? ;p Oct 02 12:27:36 i knew it was too quitet recently :p Oct 02 12:27:40 *quiet Oct 02 12:28:17 nbd: do you know anything about Buffalo firmwares? I've downloaded one from http://www.buffalo-technology.com/en/products/network-devices/dual-band-11ac/routers/wzr-1750dhp-buffalo-airstation/ Oct 02 12:28:28 it contains some binary file wzr_1750dhp_eu_222 Oct 02 12:28:44 hexdump -C wzr_1750dhp_eu_222 | head Oct 02 12:28:46 00000000 62 67 6e 00 00 00 00 00 00 00 00 b1 01 9f c1 30 |bgn............0| Oct 02 12:28:48 not a TRX file Oct 02 12:28:58 do they use encrypted firmwares only for all the time? Oct 02 12:29:19 for web flashing they use encrypted firmware Oct 02 12:29:26 i haven't looked into that part yet Oct 02 12:29:49 does bootloader allow flashing TRX files? Oct 02 12:30:36 no idea Oct 02 12:30:43 i use the initramfs method i described earlier Oct 02 12:31:08 could you try flashing trx one day? Oct 02 12:31:28 has your device not arrived yet? Oct 02 12:31:57 no Oct 02 12:55:30 build #750 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/750 Oct 02 13:04:49 owah, it seems it compiled this time Oct 02 13:06:30 I have to diff these .config to understand why it's so different and what do I have to enable or disable in order to not change kernel's md5sum Oct 02 13:09:07 another question, with who I have to talk about creating profiles for octeon target? currently there is only generic Oct 02 13:09:40 make a proposal on the list Oct 02 13:13:37 nbd ok Oct 02 13:15:00 nbd also, about the kernel md5sum that comes with the package, like kernel_3.10.49-, it only changes if I add or remove kernel related things like modules etc, right? Oct 02 13:16:17 yes Oct 02 13:16:37 it's the md5sum of the kernel config Oct 02 13:22:02 basically the sections called "Kernel modules" and "Kernel Build options" in .config Oct 02 13:22:09 I guess Oct 02 13:34:44 the generated kernel config, which is influenced by the above Oct 02 13:43:33 build #673 of sibyte is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/673 Oct 02 13:43:56 nbd any other section besides the above can affect kernel md5sum? I'm gonna test anyway, but it took me 12 hours to build :( Oct 02 13:44:39 what I'm trying to achieve is basically a official image with a bigger rootfs Oct 02 13:46:07 This is the config that took me 12 hours to build https://downloads.openwrt.org/barrier_breaker/14.07/octeon/generic/config.octeon_generic Oct 02 13:47:39 the .config generated when I only select target octeon and change rootfs size, after a distclean, is much smaller, but then, kernel md5sum is different and I'm not able to use official repos Oct 02 13:49:09 I want to write a new kernel into nor-flash using u-boot. In dmesg I see 0x000003c00000-0x000003f00000 : "kernel" Oct 02 13:49:09 So far I red correctly: erase 3c00000 3f00000 ; tftp 400000 kernel ; cp.b 400000 3c00000 ; imi 3c00000 Oct 02 13:49:09 Is this the correct when using tftp? Oct 02 13:52:20 I know "tftp" and "imi", but I've never used "erase" and "cp.b" before :-) Oct 02 14:38:03 build #685 of avr32 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/685 Oct 02 14:47:08 yep, there is an ongoing discussion in the mailing list.. after jow's explanations, from my tests, hash is definitely not consistent Oct 02 15:25:44 cyrusff: btw, net/mtr has been moved to the new packages feed Oct 02 15:33:12 nbd are you still here? I've been talking to jow about creating some kind of netifd syslog error when adding a vlan interface and 8021q is not loaded/installed, can you point me where I should start? thanks! Oct 02 15:46:10 Devastator: it should probably go in the vlan add path in system-linux.c Oct 02 16:57:28 oh cool Oct 02 16:57:36 any plans for a BB release party ? :P Oct 02 17:36:05 nbd what is the level of support and feature for Marvell http://www.marvell.com/wireless/assets/Marvell_Avastar_88W8864_SoC.pdf Oct 02 17:36:09 in OpenWRT Oct 02 17:37:06 nonexistant Oct 02 17:37:35 thanks :) Oct 02 17:37:41 you made me a happy camper :) Oct 02 17:53:01 build #776 of at91 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/776 Oct 02 19:16:26 cyrus r42741 trunk/package/network/ipv6/map/ files/map.sh Makefile src/mapcalc.c * map: fix portsets starting with 0 and use regular NAT for 1:1 MAP Oct 02 19:17:03 cyrus r42742 branches/barrier_breaker/package/ network/ipv6/map/Makefile network/ipv6/map/src/mapcalc.c network/ipv6/map/files/map.sh * map: fix portsets starting with 0 and use regular NAT for 1:1 MAP Oct 02 19:30:54 rmilecki r42743 trunk/target/linux/ bcm53xx/patches-3.14/302-ARM-BCM5301X-Add-DT-for-Netgear-R6300-V2.patch bcm53xx/image/Makefile * bcm53xx: early support for Netgear R6300 V2 Oct 02 19:38:04 cyrus r42744 trunk/package/network/config/netifd/Makefile * netifd: add IPIP tunnel support (thx Hans Dedecker) Oct 02 19:38:13 cyrus r42745 trunk/package/network/config/netifd/files/etc/init.d/network * netifd: fix an error message during network shutdown Oct 02 19:38:24 cyrus r42746 trunk/package/network/ config/ipip config/ipip/Makefile config/ipip/files config/ipip/files/ipip.sh * IPIP: IP in IP package support Oct 02 19:38:45 cyrus r42747 trunk/package/network/ipv6/6rd/files/6rd.sh * 6rd: Tos support Oct 02 19:38:51 cyrus r42748 trunk/package/network/ipv6/6to4/files/6to4.sh * 6to4: Tos support Oct 02 19:39:03 cyrus r42749 trunk/package/network/ipv6/6in4/files/6in4.sh * 6in4: Tos support Oct 02 19:48:14 cyrus r42750 trunk/package/network/services/authsae/Makefile * authsae: update to latest version Oct 02 19:48:20 cyrus r42751 trunk/package/utils/busybox/files/cron * busybox: remove validation message from bootup Oct 02 19:48:37 cyrus r42752 trunk/package/libs/toolchain/Makefile * toolchain: Fix libthreaddb not being installed Oct 02 20:01:28 nbd r42753 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c Oct 02 20:01:28 ar8216: set the pvid based on runtime state instead of port number to fix learning with enable_vlan = 0 Oct 02 23:43:02 build #187 of omap is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/187 Oct 02 23:57:36 build #92 of brcm63xx.smp is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx.smp/builds/92 **** ENDING LOGGING AT Fri Oct 03 03:00:00 2014