**** BEGIN LOGGING AT Thu Jul 01 02:59:57 2010 Jul 01 08:18:30 build #60 of orion is complete: Failure [failed compile_7] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/60 Jul 01 08:30:19 florian * r22010 /trunk/target/linux/au1000/patches-2.6.32/005-revert_mtx1_board_gpiolib.patch: [au1000] revert upstream commit that broke PCI until we have a better fix Jul 01 08:53:09 acoul * r22011 /packages/net/iputils/ (4 files in 2 dirs): net/iputils: fix ipv6 compile errors (thanks florian) Jul 01 09:24:45 Hello, I'm trying to create customized images using the SDK, but I fail to understand how the init-script for a package could be automatically enabled (ie. execute /etc/init.d/foo enable) Jul 01 09:26:08 the offline opkg install scans for init scripts provided by a package and enables them one by one Jul 01 09:27:01 for this to work you must ensure that the init script is runnable on the host, e.g. do not test for target-specific features outside of the start() or stop() procedure Jul 01 09:27:18 xMff: ah, i see, that's the problem Jul 01 09:28:41 xMff: so each init-script gets automatically enabled when building an image using the SDK? Jul 01 09:29:07 yes Jul 01 09:29:31 as long as the scripts cooperate Jul 01 09:31:29 xMff: so without further magic it's not possible to provide an init-script but not have it automatically enabled? Jul 01 09:32:35 it is. you can do something like: [ -n "${IPKG_INSTROOT}" ] && exit 0 Jul 01 09:33:19 xMff: hmm, nice idea. Thank you very much for this hints! Jul 01 09:33:54 yw Jul 01 10:40:26 markus * r22012 /trunk/tools/firmware-utils/src/ (addpattern.c trx.c): Jul 01 10:40:26 preset stable and try* flags for TRX2 headers Jul 01 10:40:26 * changed addpattern.c to preset the stable and try flags used in TRXv2 images Jul 01 10:40:26 to dupe CFE and pretend a stable image. Jul 01 10:40:26 * changed trx.c to calculate TRXv2 CRC with stable and try flags set to 0xFF Jul 01 10:40:26 like CFE does on startup Jul 01 10:40:27 * fixed compile warnings in trx.c by explicit casting Jul 01 10:48:17 markus * r22013 /trunk/target/linux/brcm47xx/image/Makefile: Jul 01 10:48:17 renamed wrt54g3gv2-vf images Jul 01 10:48:17 'factory' and 'sysupgrade' did not make much sense. A discussion Jul 01 10:48:17 with jow convinced me that .trx results in a helpdesk disaster. Jul 01 10:48:17 So I decided to use '.bin' for normal bin-headers and '.noheader.bin' Jul 01 10:48:17 for the trx-v2 image. Jul 01 10:48:17 I fixed the wiki accordingly. Jul 01 10:53:07 Hello all Jul 01 10:53:17 I'm trying to port openwrt to a new board Jul 01 10:53:21 a print server from edimax Jul 01 10:53:35 everything is described here Jul 01 10:53:38 https://forum.openwrt.org/viewtopic.php?pid=112016#p112016 Jul 01 10:54:28 to summarize, the cfe needs an extra header ( 12 bytes long starting with a 4 bytes signature and it's the only thing useful for the cfe) Jul 01 10:54:53 then in flash the firmware looks like PS16--------HDR0 Jul 01 10:55:58 I have already modified decompress.c in lzmaloader folder to correctly detect the trx header Jul 01 10:56:04 and also find_cfe_size Jul 01 10:56:18 but i'm a little stuck for find_root Jul 01 10:56:22 here is what i have done Jul 01 10:56:31 http://pastebin.com/U0hX9ZT8 Jul 01 10:57:08 i hjave some doubts about the off var Jul 01 10:57:11 whats wrong in find_root ? Jul 01 10:58:30 here is the boot Jul 01 10:58:32 http://pastebin.com/WnjVwAXC Jul 01 10:59:20 first error is that it can't erase region [0x20000, 0x20000] Jul 01 11:03:43 for the doesn't found squash fs i believe it's because i had my header after the trx creation, hence squashfs is 12 bytes later than normal Jul 01 11:06:50 hm Jul 01 11:07:01 then you need to keep the aligned offset Jul 01 11:07:40 yes i have think that i could create the trx using -f -a 0xfff4 squash.root Jul 01 11:08:27 line 10 and 11 are normal, even with aligned offset Jul 01 11:09:03 ok but the failed erasing ? Jul 01 11:09:37 erase_write used an unaligned address probalby Jul 01 11:10:32 but 0x20000 is algined no ? Jul 01 11:10:35 aligned Jul 01 11:10:47 are you sure its actually 0x20000 ? Jul 01 11:11:18 pretty sure yes Jul 01 11:11:34 anyway what the point of rewriting the trx ? Jul 01 11:12:37 jffs2 changes the on-flash data covered by the trx header which invalidates the crc Jul 01 11:13:01 therfore the trx is rewritten to not cover the jffs2 area so that crc32 stays the same even after jffs2 changes Jul 01 11:13:10 aaaah ok Jul 01 11:14:04 in general you should not mess with the offsets at all I think but only change the code that deals with the trx header to search for it in the off+0 - off+12 range Jul 01 11:14:54 the trx header itself contains the offsets to the actual partition start points so I think it should be enough to make the code find the shifted header Jul 01 11:17:31 ok but for the rewriting i should keep the edimax header, without that it will not boot anymore after reboot no ? Jul 01 11:17:40 yep Jul 01 11:18:13 use a separate variable, hdr0_off for example which contains the relative offset of the trx header Jul 01 11:18:30 which would be 0 for non-edimax devices and 12 otherwise Jul 01 11:18:47 this should work no ? : memcpy(block+EDIMAX_PS_HEADER_LEN, &trx, sizeof(trx)); (that's just to test) Jul 01 11:19:13 when rewriting the trx header, you need to copy the whole block, starting at the aligned boundary Jul 01 11:19:20 then modify the trx header in-place Jul 01 11:19:26 and finally write back the whole block Jul 01 11:19:39 mtd does not allow partial writes Jul 01 11:19:47 you can only "replace" the contents of a whole block Jul 01 11:20:13 ok I have understood that Jul 01 11:21:21 you should probably create a (static) function which takes a char * buffer + len, searches the trx header within it and returns a char * / void * pointer to the HDR0 data Jul 01 11:21:48 ok Jul 01 11:21:48 it would return either 0 or 12, possibly more in the future Jul 01 11:22:18 then use that function in all places that want a struct pointer to the trx Jul 01 11:22:22 ... header Jul 01 11:22:29 ok i see Jul 01 11:22:35 thanks for your explanations Jul 01 11:23:06 I have no idea whether this would work but this is the way I'd try it Jul 01 11:24:14 regarding the function I meant it would either return *(buffer+0) or *(buffer+12) of course Jul 01 11:25:41 k it should be easy to modify later Jul 01 12:02:49 I want to try uploading a kernel image to uboot (using loadb). Which image file do I use? -uImage or -zImage? Jul 01 12:15:22 openwrt is booting fine in ramdisk mode Jul 01 12:18:25 but why the nvram reported by openwrt is different than the cfe one ? Jul 01 13:32:24 build #74 of brcm47xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/74 Jul 01 13:44:03 acoul * r22014 /packages/net/mtr/ (3 files in 2 dirs): net/mtr: update to mtr-0.79 & fix compile issues Jul 01 13:50:26 anyone here used kermit to transfer a new kernel to u-boot? Jul 01 14:49:07 juhosg * r22015 /trunk/tools/firmware-utils/src/mktplinkfw.c: Jul 01 14:49:07 firmware-utils/mktplinkfw: add support for the TL-WA901ND Jul 01 14:49:07 Patch-by: Pieter "Fate" Hollants Jul 01 14:49:10 juhosg * r22016 /trunk/target/linux/ar71xx/files/drivers/leds/leds-rb750.c: ar71xx: fix leds-rb750 build failure on 2.6.34 Jul 01 15:10:22 marek * r22017 /packages/net/batman-advanced/ (4 files in 3 dirs): Jul 01 15:10:22 batman-adv: update to latest release (2010.0.0) Jul 01 15:10:22 * simplified Makefile (thanks Nico!) Jul 01 15:10:22 * moved the branches into an extra feed Jul 01 15:10:22 * updated the init script to match the latest features Jul 01 15:41:54 build #83 of atheros is complete: Failure [failed shell_6] Build details are at http://tksite.gotdns.org:8010/builders/atheros/builds/83 Jul 01 16:23:41 markus * r22018 /trunk/package/broadcom-diag/src/diag.c: Jul 01 16:23:41 fix GPIO pin for G3GV2-VF Jul 01 16:23:41 Signed-off-by: Niclas Koeser Jul 01 16:26:29 markus * r22019 /trunk/target/linux/brcm47xx/files/drivers/mtd/maps/bcm47xx-flash.c: Jul 01 16:26:29 brcm47xx: added dual image support to flashmap driver Jul 01 16:26:29 some CFE versions (i.e. WRT54G3Gv2-VF) expect two separate firmware Jul 01 16:26:29 images which each consume half of the available flash space. Jul 01 16:26:29 these changes check for the TRX header version and set the partition Jul 01 16:26:30 size correctly. Jul 01 16:26:31 Tested on G3Gv2-VF and Asus WL-500GP Jul 01 16:51:05 acoul * r22020 /packages/net/netperf/patches/ (. 001-dccp.patch): net/netperf: fix SOCK_DCCP compile issue (uClibc 0.9.31) Jul 01 17:18:09 build #65 of mpc52xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/65 Jul 01 18:07:50 juhosg * r22021 /branches/backfire/package/mac80211/ (3 files in 2 dirs): (log message trimmed) Jul 01 18:07:50 backfire: rt2x00: cleanup patches (backport of r21975) Jul 01 18:07:50 Remove 601-rt2x00-remove-mcu-requests-for-soc.patch as an equivalent patch Jul 01 18:07:50 is already included in upstream. Jul 01 18:07:50 Add 601-rt2x00-fix-hang-on-ifdown.patch which is not suitable for upstream Jul 01 18:07:50 inclusion but fixes a hang on ifdown under some circumstances and thus also Jul 01 18:07:51 fixes AP mode operation in OpenWRT. Jul 01 18:07:56 juhosg * r22022 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: use RTL8366{S,RB} prefixes for all defines (backport of 21976) Jul 01 18:08:05 juhosg * r22023 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.h rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: add VLAN handling functions to rtl8366_smi_ops (backport of r21977) Jul 01 18:08:13 juhosg * r22024 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.h rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: add cpu_port, num_ports and num_vlan_mc to struct rtl8366_smi (backport of r21978) Jul 01 18:08:20 juhosg * r22025 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: move common VLAN handling functions to rtl8366_smi.c (backport of 21979) Jul 01 18:08:27 juhosg * r22026 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.h rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: move buf to struct rtl8366_smi (backport of 21980) Jul 01 18:08:34 juhosg * r22027 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: move common debugfs code to rtl8366_smi.c (backport of 21981) Jul 01 18:08:41 juhosg * r22028 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.h rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: use common rtl8366_mib_counter structure (backport of 21982) Jul 01 18:09:01 juhosg * r22029 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.h rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: standardize read_debugfs_mibs functions (backport of r21983) Jul 01 18:09:01 juhosg * r22030 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366_smi.c rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: add common read_debugfs_mibs function (backport of 21984) Jul 01 18:09:02 juhosg * r22031 /branches/backfire/target/linux/generic-2.6/files/drivers/net/phy/ (rtl8366rb.c rtl8366s.c): backfire: generic: rtl8366: minor cleanup (backport of r21985) Jul 01 18:09:04 juhosg * r22032 /branches/backfire/target/linux/ar71xx/base-files/etc/defconfig/dir-825-b1/network: backfire: [ar71xx] fix dir-825-b1 default network config (backport of r22005) Jul 01 19:07:48 florian * r22033 /trunk/target/linux/au1000/ (3 files in 2 dirs): [au1000] fix PCI, thanks br1, put back 2.6.32.14 support by default Jul 01 19:41:53 florian * r22034 /branches/backfire/target/linux/au1000/patches-2.6.32/005-mtx1_fix_pci.patch: [backfire] backport r22033 Jul 01 19:43:05 nbd * r22035 /trunk/package/mac80211/patches/530-ath9k_nf_sanitize.patch: ath9k: properly sanitize calibrated noise floor values on all hardware Jul 01 20:51:32 nbd * r22036 /trunk/package/mac80211/patches/530-ath9k_nf_sanitize.patch: ath9k: add a missing null pointer check and some further cleanup Jul 01 20:58:49 <[Fate]> "openwrt coding style" == kernel coding style? Jul 01 20:59:53 for c stuff: pretty much, yes Jul 01 21:06:11 <[Fate]> I seem to be blind, where does http://pastebin.ca/1893041 not match the coding style? Jul 01 21:08:55 dunno, looks ok to me on a first glance, though i usually put the else on the same line as the }, but that's just personal taste Jul 01 21:09:18 tbh, i've never actually read all of the kernel coding style guidelines Jul 01 21:09:24 i just mirror what i see in other code Jul 01 21:09:25 ;) Jul 01 21:09:27 lulz Jul 01 21:15:08 <[Fate]> ok, and should I put "#define helper_func()" inside of a func just before it's first use or in front of the func? (it's not used in any other func) Jul 01 21:18:39 does it have to be a macro? or could it be an inline function instead? Jul 01 21:18:46 inline functions are usually preferred Jul 01 21:19:31 <[Fate]> good point indeed Jul 01 21:24:47 build #60 of ixp4xx is complete: Failure [failed shell_6] Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/60 Jul 01 21:26:12 build #69 of ep93xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/69 Jul 01 21:41:37 nbd * r22037 /trunk/package/mac80211/patches/531-ath9k_nf_cleanup.patch: ath9k: clean up some code duplication related to noise floor handling Jul 01 21:41:43 nbd * r22038 /trunk/package/mac80211/patches/540-ath9k_hang_check.patch: ath9k: fix false positives in the baseband hang check by repeating the test a few times before pronouncing the hardware dead and resetting it Jul 01 22:22:41 nbd * r22039 /branches/backfire/package/mac80211/ (31 files in 2 dirs): merge latest mac80211 from trunk (as of r22038) to backfire Jul 01 22:28:07 florian * r22040 /trunk/target/linux/au1000/patches-2.6.32/005-mtx1_fix_pci.patch: [au1000] fix line wrapped patch Jul 01 22:28:17 florian * r22041 /trunk/target/linux/au1000/ (8 files in 4 dirs): [au1000] add 2.6.34 support Jul 01 22:28:25 florian * r22042 /trunk/target/linux/generic/patches-2.6.34/ (3 files): [kernel] refresh 2.6.34 patches **** ENDING LOGGING AT Fri Jul 02 02:59:57 2010