**** BEGIN LOGGING AT Sun Sep 06 02:59:57 2009 Sep 06 03:36:54 I've made a new x86 target for a Kontron board. I've compiled in the SCSI + SATA drivers into the image.. Sep 06 03:37:27 is there a way of saying that the kernel provides kmod-scsi-core ? so that moduled that depend on it will still compile? Sep 06 03:37:52 (for eg usb-storage ) Sep 06 05:13:46 yet another openwrt/upstream package comparison, http://pastehtml.com/view/090906hjVWTPUJ.html Sep 06 07:50:50 good morning Sep 06 08:34:15 mornin' Hopscotch Sep 06 08:34:26 or afternoon. Sep 06 08:34:52 still morning here :) Sep 06 13:42:30 florian * r17524 /trunk/target/linux/uml/patches/004-fix_section_warning.patch: [uml] fix some section warnings Sep 06 13:42:52 florian * r17525 /trunk/target/linux/uml/patches/005-fix_bash4_builds.patch: [uml] fix build failures with bash4 Sep 06 13:44:16 <[florian]> swalker: how did you generate this list? I am really interested in this :) Sep 06 13:54:19 florian * r17526 /trunk/target/linux/brcm63xx/ (3 files in 3 dirs): [brcm63xx] add infrastructure to register gpio-input reset button Sep 06 14:38:05 Should a package have -L/usr/lib in its link command? Sep 06 14:38:48 net-snmp seems to be trying to link with the host system's /usr/lib/libm. Sep 06 14:42:40 that is wrong and should be fixed then Sep 06 14:42:53 How would I do that? Sep 06 14:43:03 I know next to nothing about the openwrt build process. Sep 06 14:43:11 by looking for what's injecting the -L/usr/lib Sep 06 14:43:20 it's most likely not comping from the openwrt build process Sep 06 14:43:32 it's probably hardcoded somewhere inside the package Sep 06 14:43:38 or in a file that is used by the package Sep 06 14:52:36 Bah. There's nowhere saying 'Fix this problem here' so I'm going to have to try and figure it out myself. Sep 06 14:52:41 This is going to be fun. Sep 06 17:02:16 thepeople * r17527 /trunk/package/base-files/files/ (3 files in 3 dirs): fix nonhotpluggable networks not being brought up. Closes #2781 Sep 06 17:26:39 nbd * r17528 /trunk/ (include/kernel.mk package/kernel/Makefile): now that $(BOARD) is part of the opkg ARCH, there's no reason to keep the board name around in kmod-* package versions anymore Sep 06 17:36:03 nbd * r17529 /trunk/package/ppp/files/ (ppp.sh pppoe.sh): move ppp ifname fixup from pppoe.sh to ppp.sh, so that it applies to all other ppp based protocols as well Sep 06 18:48:51 [florian]: Makefile grep'ping for the majority with manual fixing of packages that don't provide the needed or have out-of-date info, something uscan-like for the upstream checking would be preferable rather than doing it manually right now Sep 06 19:31:16 mb * r17530 /trunk/target/linux/brcm47xx/patches-2.6.28/ (3 files): bcm47xx: Clean up USB patches Sep 06 19:36:59 Am I right in think that the toolchain gets built using only the generic kernel patches, and not the target specific ones? Sep 06 19:41:54 why would the kernel patches matter for the toolchain? Sep 06 19:44:30 noz: yes Sep 06 19:49:09 mb * r17531 /trunk/target/linux/brcm47xx/patches-2.6.28/275-usb2-bcm5354-init.patch: bcm47xx: Apply bcm5354 USB fix Sep 06 19:56:44 nbd: Thanks. Would explain why the ppoll() patch wasn't being applied. Sep 06 19:56:56 nbd * r17532 /trunk/Config.in: do not forcibly disable stripping when selecting CONFIG_DEBUG, debugging symbols can be useful even if they are not put on the device (e.g. remote cross-gdb) Sep 06 19:58:30 the reason for this is that toolchains are potentially shared across different targets of the same arch Sep 06 19:58:49 nbd: Indeed. I guess at some point soon, we need to decide which of the two patches under consideration is going to go in. Sep 06 19:59:15 I'll trawl the kernel git to see if I can guess whcih will make it upstream Sep 06 20:00:29 ok, sounds good Sep 06 20:06:24 The good news is that one of them has made it into linux-next Sep 06 20:07:59 nbd: a freindof mine created an updated click patch for usermode Sep 06 20:08:04 interested in it ? Sep 06 20:08:06 sure Sep 06 20:08:19 nbd: email it where Sep 06 20:08:47 openwrt-devel@lists.openwrt.org Sep 06 20:17:08 nbd: sent Sep 06 20:17:15 thx Sep 06 20:18:46 nbd: u know any iissue with wireless ath9k on dlink dir615 rev c1now working ? Sep 06 20:19:18 seems someones tried the 9k build and /etc/config/wireless is empty 0 byte Sep 06 20:25:12 mb * r17533 /trunk/target/linux/brcm47xx/patches-2.6.30/ (4 files): bcm47xx: Port usb fixes to .30 Sep 06 20:29:28 OutBackDingo: haven't played with ath9k yet Sep 06 20:30:07 hahahha whos doing the ath9k work . Sep 06 20:31:05 dunno Sep 06 20:43:46 mb * r17534 /trunk/target/linux/brcm47xx/patches-2.6.30/270-ehci-ssb.patch: Fix merge error Sep 06 20:45:33 nbd: still had no time for ath5k yet huh Sep 06 20:45:46 80% done, actually Sep 06 20:45:54 nice Sep 06 20:46:01 it's not working yet, because of something that seems like an rf buffer problem Sep 06 20:46:16 i think i ported almost all of the required changes Sep 06 20:46:22 just haven't tracked down that issue yet Sep 06 20:46:36 ported from ? Sep 06 20:47:15 the madwifi hal and driver Sep 06 20:47:19 but rewritten of course Sep 06 20:47:53 i've sent preliminary patches to nick, but he also didn't find any issues yet Sep 06 20:48:28 you know FreeBSD states they have TDMA for atheros in 8x Sep 06 20:48:51 too bad it doesnt run on Ubiquiti :) Sep 06 20:48:59 that tdma implementation has its own set of issues Sep 06 20:49:15 for instance it uses no ack, so there's no software or hardware level retransmission for pakets Sep 06 20:49:25 which can be a problem if you have a fluctuating link Sep 06 20:50:12 and im pretty sure Ubiquitis M AirMAX is based on the SoftTDMA work Sep 06 20:50:36 from what i can see in the binaries, it's similar in principle, but a different implementation Sep 06 20:51:10 i mean the idea behind it is quite simple Sep 06 20:51:35 in a point to point setup you split the beacon interval in two parts Sep 06 20:51:39 each side gets its slot Sep 06 20:51:52 sends a burst of packages with no-ack along with the beacon Sep 06 20:51:58 limited to the available airtime in the slot Sep 06 20:52:07 without carrier sensing Sep 06 20:52:19 each side synchronizes to the other side's beacon Sep 06 20:52:23 so links should be solid or there could be issues Sep 06 20:52:26 to keep the timeslot divsion intact Sep 06 20:52:27 yes Sep 06 20:52:50 too bad they dont do adhoc Sep 06 20:53:09 or ahdemo Sep 06 20:53:18 a tdma-like protocol for adhoc is a completely different order of complexity Sep 06 20:53:55 not sure anybody really had TDMA adhoc Sep 06 20:54:17 though skypiolt claims something similiar Sep 06 20:54:23 err skypilot Sep 06 20:55:03 mb * r17535 /trunk/target/linux/brcm47xx/patches-2.6.30/280-activate_ssb_support_in_usb.patch: bcm47xx: Fix EHCI/OHCI dependencies Sep 06 21:15:15 mb * r17536 /trunk/target/linux/brcm47xx/patches-2.6.28/280-activate_ssb_support_in_usb.patch: bcm47xx: Fix EHCI/OHCI dependencies Sep 07 01:03:35 any ideas http://pastebin.com/m73943274 **** ENDING LOGGING AT Mon Sep 07 02:59:57 2009