**** BEGIN LOGGING AT Fri Mar 15 02:59:58 2013 Mar 15 03:12:45 build #194 of rb532 is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/194 Mar 15 03:33:55 build #194 of ppc44x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/194 Mar 15 05:58:16 build #184 of sibyte is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/184 Mar 15 09:47:41 build #214 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/214 Mar 15 09:53:18 build #189 of xburst is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/189 Mar 15 10:54:53 i'm trying these values with mini_snmpd and an snmp browser : http://riceball.com/d/content/mini-howto-minisnmpd-small-snmpd-embedded-systems-openwrt Mar 15 10:55:03 some of these doesn't works Mar 15 12:21:15 build #166 of iop32x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/166 Mar 15 12:47:03 build #162 of gemini is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/162 Mar 15 13:31:59 nbd: ping? Mar 15 13:32:06 fullstop: pong Mar 15 13:32:31 I'm looking at the sierra directip stuff at the moment, and had a quick question regarding ar71xx/mips Mar 15 13:32:44 are their dma transfer requirements similar to ARM? Mar 15 13:32:55 That is, must they be aligned on 4 byte boundaries? Mar 15 13:33:31 i think that's a low-level detail that you don't have to deal with on most hardware Mar 15 13:33:42 You would think, yes. Mar 15 13:33:44 with most devices, you can provide unaligned DMA addresses Mar 15 13:33:49 there are some exceptions though Mar 15 13:34:13 e.g. the ethernet MAC on ar71xx (but not ar724x and newer) is incapable of doing unaligned DMA Mar 15 13:34:16 I'm having a problem which seems similar to one that I had on ARM with usbnet and alignment. Mar 15 13:34:35 This describes the problem pretty well: http://lists.infradead.org/pipermail/linux-arm-kernel/2010-August/022983.html Mar 15 13:34:58 I was wondering if there was a similar problem with directip and how the driver receives HIP messages from the modem. Mar 15 13:35:42 I had directip working with an ar71xx based device at some point (though it was one where the fw didn't support DHCP) Mar 15 13:35:49 Right. Same here. Mar 15 13:35:57 but not always Mar 15 13:36:31 and I think that might depend on how/when the initial skbs are allocated. Mar 15 13:37:27 I can dig further, but I figured that I would ask if mips had a similar dma requirement first before chasing my tail. Mar 15 13:38:59 it's not really about the architecture, but about what the HCD supports Mar 15 13:39:22 <_rmk_> fullstop: the email you refer to is not about the CPU architecture, it's about the restrictions imposed on what can be DMA'd by hardware Mar 15 13:39:27 Sorry, you are absolutely correct. Mar 15 13:40:06 i'm not aware of any DMA alignment issues in the ar71xx HCD Mar 15 13:40:22 <_rmk_> if you do set NET_IP_ALIGN to zero, the IP stack becomes less efficient on many CPUs Mar 15 13:40:23 so unless you have any specific evidence pointing at alignment issues, my guess is it's something else Mar 15 13:40:29 I've experienced the issue with both atmel and ti/davinci, so I mistakenly implied ARM Mar 15 13:40:42 okay, I'll look into it. Thanks. Mar 15 13:40:47 also, if it works sometimes it might just be a firmware issue Mar 15 13:40:54 the sierrawireless firmware is a bit fragile in my experience Mar 15 13:41:05 <_rmk_> lots of embedded DMA controllers seem to suffer from DMA restrictions :( Mar 15 13:41:14 if you do things in the wrong order, it may suddenly decide to just lock up Mar 15 13:41:25 _rmk_: i think it's mostly the DWC crap Mar 15 13:41:34 which unfortunately has spread to many designs Mar 15 13:41:35 The older cards had great firmware.. but I'm not so sure about this one. It's one of their first with LTE support. Mar 15 13:41:52 <_rmk_> ARM will work if you set NET_IP_ALIGN to zero, but then you end up with _lots_ of alignment faults from the IP stack Mar 15 13:42:17 <_rmk_> and on fast IP links, that can be quite a problem for CPU bandwidth Mar 15 13:42:49 Yes. This router is only used with the data card, so I'm already limited by the cellular network. Mar 15 13:44:25 nbd r36034 trunk/package/network/services/uhttpd/Makefile * uhttpd: update to latest version, fixes an occasional crash on CGI scripts Mar 15 13:52:02 [florian]: ping Mar 15 13:53:29 <[florian]> wbx: pong Mar 15 13:55:17 * dlooh note la presence d'un florian Mar 15 13:59:16 nbd r36035 trunk/package/mac80211/patches/300-pending_work.patch * ath9k: fix issues with disabling rx/tx antennas Mar 15 14:04:28 [florian]: did you find the openbios-ppc file in your distribution? Mar 15 14:04:52 <[florian]> wbx: yep, thanks Mar 15 14:09:44 [florian]: what is the problem with the virtex-ml507 target? Mar 15 14:12:13 Unable to get size of device tree file 'ppc.dtb Mar 15 14:12:17 this one? Mar 15 14:13:35 <[florian]> it failed for me somewhere in dtb parsing Mar 15 14:13:42 <[florian]> even though I used a qemu HEAD Mar 15 14:13:57 i get the above message... Mar 15 14:36:17 juhosg r36036 trunk/target/linux/ar71xx/files/arch/mips/ath79/mach-tl-wr841n-v8.c * ar71xx: Error fix in Kernel for MR3420v2 Ethernet Switch configuration Mar 15 14:36:19 juhosg r36037 trunk/target/linux/ar71xx/files/arch/mips/ath79/mach-rw2458n.c * ar71xx: rw2458n - Use eeprom (art) from wifi card Mar 15 14:36:21 juhosg r36038 trunk/target/linux/ar71xx/image/Makefile * ar71xx: build images for the TL-MR11U v2 Mar 15 14:40:20 juhosg r36039 trunk/ (72 files in 28 dirs) * kernel: update linux 3.8 to 3.8.3 Mar 15 15:22:35 nbd r36040 trunk/package/mac80211/patches/300-pending_work.patch * ath9k: add a fix to reduce the number of spurious tx hang detections Mar 15 15:28:20 zioproto r36041 packages/net/tinc/Makefile * [packages] tinc: Version bump 1.0.20 Mar 15 15:50:24 juhosg r36042 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: remove superfluous pdata checks Mar 15 15:50:26 juhosg r36043 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: simplify and rename ar8327_config_port Mar 15 15:50:29 juhosg r36044 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: move the ar8327_get_port_init_status function Mar 15 15:50:31 juhosg r36045 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: don't use platform_data in ar8327_init_port Mar 15 15:50:34 juhosg r36046 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: separate platform_data based AR8327 code Mar 15 15:50:37 juhosg r36047 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: add OF configuration support for AR8327 Mar 15 15:53:35 juhosg r36048 trunk/target/linux/mpc85xx/patches-3.8/140-powerpc-85xx-tl-wdr4900-v1-support.patch * mpc85xx: add AR8327 initvals to the TL-WDR4900 devicetree Mar 15 16:06:54 juhosg r36049 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: add ar8xxx_start helper Mar 15 16:06:57 juhosg r36050 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: start AR8327 switch from the probe routine Mar 15 16:12:14 juhosg r36051 trunk/target/linux/ mpc85xx/image/Makefile mpc85xx/profiles/tp-link.mk * mpc85xx: add profile and build image for the TL-WDR4900 v1 board Mar 15 19:35:11 build #150 of octeon is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/150 Mar 15 22:25:05 nbd r36052 branches/attitude_adjustment/package/mac80211/patches/300-pending_work.patch * ath9k: add a fix to reduce the number of spurious tx hang detections Mar 15 22:29:57 nbd r36053 trunk/package/mac80211/patches/300-pending_work.patch * mac80211: merge a fix for a race condition on station removal Mar 16 01:26:12 nbd r36054 trunk/package/busybox/Makefile * busybox: fix disabling /sbin/init and logread if procd is used as init **** ENDING LOGGING AT Sat Mar 16 02:59:58 2013