**** BEGIN LOGGING AT Sat Oct 08 02:59:57 2011
Oct 08 04:15:12 <builder_> build #73 of avr32 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/73
Oct 08 06:38:09 <Weedy> jow_laptop: is an empty QUERY_STRING a known thing with uhttpd or do i just suck?
Oct 08 11:40:23 <CIA-73> jogo * r28380 /trunk/target/linux/ar71xx/files/drivers/net/ag71xx/ag71xx_phy.c:
Oct 08 11:40:23 <CIA-73> ag71xx: close a race between the phy state machine and link state
Oct 08 11:40:23 <CIA-73> A fast stop/start cycle could leave the ag71xx interrupts and tx engine
Oct 08 11:40:23 <CIA-73> disabled when using a phy driver with a fixed link and the start/stop
Oct 08 11:40:23 <CIA-73> happens between two phy state machine polls.
Oct 08 11:40:23 <CIA-73> Prevent this by always forcing the link down on stop regardless of phy
Oct 08 11:40:24 <CIA-73> state and having a phy connected.
Oct 08 11:49:08 <KanjiMonster> "start/stop" <- meh, should have been "stop/start"
Oct 08 12:06:59 <jow_laptop> Weedy: context? empty query string?
Oct 08 12:08:02 <Weedy> ii figured out query_string != query_string_post
Oct 08 12:08:44 <jow_laptop> post data has nothing to do with query strings, you need to read and parse it from stdin
Oct 08 12:09:39 <Weedy> no one told me :/
Oct 08 12:12:43 <jow_laptop> its in the cgi spec
Oct 08 12:29:01 <CIA-73> jow * r28381 /packages/ipv6/radvd/Makefile: [packages] radvd: bump to v1.8.2 (#10196, CVE-2011-3601, CVE-2011-3602, CVE-2011-3603, CVE-2011-3604, CVE-2011-3605)
Oct 08 12:31:42 <CIA-73> jow * r28382 /branches/packages_10.03.1/ipv6/radvd/ (Makefile patches/): [packages_10.03.1] merge r28381
Oct 08 13:44:09 <CIA-73> nbd * r28383 /trunk/package/mac80211/patches/552-ath9k_fix_phyerror.patch: ath9k: extend the phy error disable patch
Oct 08 14:51:29 <CIA-73> nbd * r28384 /trunk/package/mac80211/patches/ (2 files): ath9k: add support for per-chain signal strength info
Oct 08 14:51:35 <CIA-73> nbd * r28385 /trunk/package/iw/ (7 files in 2 dirs): iw: update to v3.1, add support for per-chain signal strength info
Oct 08 15:23:41 <CIA-73> nbd * r28386 /trunk/package/mac80211/patches/140-mesh_pathtbl_backport.patch: mac80211: remove a WARN_ON that causes compile errors on 2.6.32
Oct 08 17:38:02 <CIA-73> nbd * r28387 /trunk/package/mac80211/files/lib/wifi/mac80211.sh: mac80211: set the country code early to allow crda to settle
Oct 08 18:11:41 <mazilo> Can anyone please fix asterisk18-chan-sccp-b package that causes the all the modules of asterisk18 to spill out of its sub-menu?
Oct 08 18:14:32 <CIA-73> nbd * r28388 /branches/backfire/package/mac80211/ (60 files in 3 dirs): mac80211: update to latest version from trunk
Oct 08 18:31:46 <philipp64|laptop> what's the clock rate or bogo-mips of the brcm and ar71xx platforms?
Oct 08 18:32:09 <EqUaTe> that's going to vary dramatically from device to device
Oct 08 18:32:27 <philipp64|laptop> envelope then.
Oct 08 18:32:48 <nbd> many ar71xx devices are have 400 mhz
Oct 08 18:32:51 <nbd> some have 680 mhz
Oct 08 18:33:35 <philipp64|laptop> super scalar?
Oct 08 18:34:05 <philipp64|laptop> the Geode will do about 1.6 instructions per tick...
Oct 08 18:35:22 <philipp64|laptop> but I don't know about RISC architectures like the MIPS.
Oct 08 18:35:34 <philipp64|laptop> it might be 1:1 ...
Oct 08 18:38:22 <nbd> i don't think it's superscalar
Oct 08 18:38:44 <KanjiMonster> mips 74k would be
Oct 08 18:38:51 <KanjiMonster> iirc
Oct 08 18:38:53 <philipp64|laptop> I didn't think so. pipelining takes a lot of expensive power and die surface area that could be used for registers and cache.
Oct 08 18:39:35 <philipp64|laptop> for the lower end processors, I mean.
Oct 08 18:39:42 <nbd> it does have a pipeline though
Oct 08 18:39:52 <philipp64|laptop> we used the MIPS 7K processors at Cisco and they were pipelined...
Oct 08 18:40:16 <philipp64|laptop> some of the higher end 4K processors were too.
Oct 08 18:40:39 <philipp64|laptop> but you paid for it in doubling your power consumption.
Oct 08 18:40:39 <nbd> i think all mips processors are pipelined
Oct 08 18:41:07 <philipp64|laptop> even the licensed ones? like the Sibyte?
Oct 08 18:41:14 <philipp64|laptop> and the blackfin?
Oct 08 18:41:30 <philipp64|laptop> could be... been a long time since I worked with MIPS.
Oct 08 18:41:38 <nbd> not sure about blackfin
Oct 08 18:42:20 <KanjiMonster> *checks* 74k is dual issue superscalar; that means the newer broadcoms, ar933x/ar934x or ralink ra3662/ra3883 would be
Oct 08 18:42:41 <philipp64|laptop> ok, so 2:1. cool.
Oct 08 18:43:28 <philipp64|laptop> was talking to Daniel about the price of the Alix vs. an off-the-shelf BRCM box.
Oct 08 18:43:43 <philipp64|laptop> trying to figure out what the 500% price increase gets you. :-)
Oct 08 18:43:58 <philipp64|laptop> beside the sheet metal case, I mean.
Oct 08 18:44:15 <karlp> what's the alix?
Oct 08 18:44:20 <philipp64|laptop> and the ability to run glibc-compiled binary codec's.
Oct 08 18:44:35 <philipp64|laptop> karlp: http://www.pcengines.ch/
Oct 08 18:44:46 <philipp64|laptop> it's a Geode (AMD x86) based SBC.
Oct 08 18:45:29 <philipp64|laptop> 500-800Mhz, 256 or 512MB, CF, 2-3 LAN, mini-PCI or mini-PCIe (usb only, 0 lanes)
Oct 08 18:45:57 <philipp64|laptop> they're handy when you need to work in an x86 environment.
Oct 08 18:46:11 <philipp64|laptop> for ABI reasons.
Oct 08 18:47:03 <philipp64|laptop> the Geode also has a nice AES crypto-accelerator onchip...
Oct 08 18:47:05 <KanjiMonster> philipp64|laptop: well, not quite, afaict it can only exectute a load/store in parallel to a non-load/store, so YMWV (but iirc it supports out-of-order execution)
Oct 08 18:47:56 <philipp64|laptop> I don't know how much of a hit that would be, since a load usually gets followed by an add, subtract, shift, or, and, compare, etc.
Oct 08 18:48:15 <philipp64|laptop> unless you need both operands to be fetched, of course...
Oct 08 18:48:26 <philipp64|laptop> but with that much register space...
Oct 08 18:50:11 <philipp64|laptop> I see the Fish River Island is finally turning up on google...
Oct 08 18:51:45 <philipp64|laptop> the "yocto" project (whatever that is) is using it. doesn't give the part #, though.
Oct 08 18:52:51 <philipp64|laptop> oh, here it is... Z530, eMenlow.
Oct 08 18:53:00 <philipp64|laptop> (atom)
Oct 08 19:21:09 <karlp> fish river island?
Oct 08 19:40:02 <crow> alix is nice toy have an 3d3
Oct 08 19:42:32 <philipp64|laptop> yeah, I have a 2d13 and a 6f2... the 6f2 will take a Wifi mini-PCI card and a 3G or Wimax mini-PCIe card.
Oct 08 19:43:31 <philipp64|laptop> karlp: fish river island was the Intel internal name for a nano-ITX (or was it XTX?) board with a Z530 and eg30T chips.
Oct 08 19:44:38 <philipp64|laptop> about the size of a deck of playing cards, NAND, SIM reader, Ethernet, 1.5 mini-PCIe slots, 1.2GHz processor... CF, <5W power consumption.
Oct 08 19:44:52 <philipp64|laptop> you could get it with HDMI/LVDS/mini-DVI output.
Oct 08 19:45:28 <philipp64|laptop> it's a prototyping/SDK package, though. not intended for production.
Oct 08 19:45:53 <philipp64|laptop> there's stuff out there on it on the Wind River web site, and Kontron, etc.
Oct 08 19:46:36 <crow> philipp64|laptop that with 512MB of memory would be nice, and maybe something more then current 500MHz ..
Oct 08 19:54:01 <philipp64|laptop> as I remember it came with 2G of DRAM.
Oct 08 19:54:24 <philipp64|laptop> it also has decent MPEG-4 decoding capability.
Oct 08 19:54:42 <philipp64|laptop> (it's the same IP as goes into the STB architectures...)
Oct 08 19:59:06 <philipp64|laptop> I believe it will do PiP at 1080p.
Oct 08 21:32:43 <roh> hi here
Oct 08 21:33:06 <jow_laptop> hi
Oct 08 23:19:20 <mwalle> hi, are there any guidelines for new board support patches?
Oct 08 23:19:58 <mwalle> sth like, it has to be in the mainline kernel before we allow inclusion
**** ENDING LOGGING AT Sun Oct 09 02:59:57 2011