**** BEGIN LOGGING AT Fri Aug 16 02:59:58 2013
Aug 16 03:47:44 <Devastator> ok, wireless mac fixed..
Aug 16 04:30:04 <Devastator> blogic when you have the time, can you tell me what is the available values for ralink,portmap and ralink,led-polarity? no rush :)
Aug 16 05:20:21 <builder> build #237 of octeon is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/237
Aug 16 08:21:17 <[florian]> zajec: ping?
Aug 16 09:17:42 <zajec> [florian]: pong
Aug 16 09:22:06 <[florian]> zajec: from the last email you sent, according to perf you are spending only about 5% of the time idle, so you are CPU bound
Aug 16 09:22:18 <[florian]> zajec: how fast is the MIPS core on your router?
Aug 16 09:23:43 <zajec> BCM4706
Aug 16 09:23:44 <zajec> sec
Aug 16 09:24:07 <zajec> 400MHz or 600MHz...
Aug 16 09:25:56 <zajec> can I paste 5 lines?
Aug 16 09:26:01 <zajec> # cat /proc/cpuinfo
Aug 16 09:26:03 <zajec> system type             : Broadcom BCM5300 (Unknown Board)
Aug 16 09:26:04 <zajec> processor               : 0
Aug 16 09:26:06 <zajec> cpu model               : MIPS 74Kc V4.9
Aug 16 09:26:07 <zajec> BogoMIPS                : 299.26
Aug 16 09:27:53 <zajec> [florian]: so it's 600Mhz... that sounds like pretty fast for SoCs
Aug 16 09:28:41 <zajec> [florian]: and yeah, I believe i'm CPU bound, I see 99/100% usage
Aug 16 09:29:06 <zajec> [florian]: but why? that "perf top" output doesn't show anything special taking - let's say) - 50% of the CPU
Aug 16 09:29:11 <Hauke> [florian]: and zajec gets ~800 MBit/s with the vendor driver
Aug 16 09:29:18 <zajec> [florian]: looks like a normal, balanced case
Aug 16 09:29:52 <OpenWrt-Commits> jow r37798 trunk/package/ utils/e2fsprogs/Makefile utils/e2fsprogs/patches/001-link-against-libuuid.patch utils/e2fsprogs/patches/002-no_malloc_h.patch utils/e2fsprogs/patches/003-com_err_version.patch * Revert "[package] e2fsprogs: update to 1.42.8 and add e4defrag"
Aug 16 09:29:58 <OpenWrt-Commits> jow r37799 trunk/package/system/procd/Makefile * procd: update to git head
Aug 16 09:31:09 <zajec> 100Mb/s I get now isn't that bad... it's pretty common top speed offered by ISPs
Aug 16 09:31:16 <zajec> but I still would like to get better
Aug 16 09:41:41 <MartijnVdS> My ISP is going to offer 500/500 soon
Aug 16 09:43:23 <[florian]> zajec: that's just bad considering how much you could get
Aug 16 09:46:45 <zajec> [florian]: yeah, and i don't believe it's a matter of modified network stack
Aug 16 09:46:49 <zajec> they could hack it a bit
Aug 16 09:46:59 <zajec> but not so much to have to much better performance
Aug 16 09:47:13 <[florian]> well there could be quite a lot happening in the hardware
Aug 16 09:47:29 <[florian]> some recent broadcom switches include features like layer 3 routing in switch
Aug 16 09:49:14 <zajec> is there some way of verifying clock speed?
Aug 16 09:49:19 <zajec> maybe it's underclocked...?
Aug 16 09:51:59 <[florian]> there might be some cp0 registers controlling speed, but I assume CFE should set the CPU to full-speed
Aug 16 09:54:31 <zajec> [florian]: do you think it's a matter of "layer 3 routing"?
Aug 16 09:54:44 <zajec> [florian]: i'm not sure if this is supported by this hardware at all
Aug 16 09:55:11 <zajec> AFAIK Broadcom's firmware uses standard et0 driver + switch driver
Aug 16 09:57:41 <[florian]> the switch might be doing the routing
Aug 16 09:58:04 <[florian]> if you have access to the original firmware, you might just see if packets are making it to the CPU interface when routing is enabled
Aug 16 09:58:58 <zajec> [florian]: how could it work? i mean switch could so simple routing like "pass packets from VLAN#3 to VLAN#2", but what about firewall then?
Aug 16 09:59:27 <zajec> [florian]: i think i could try compiling originalfirmware with my changes
Aug 16 09:59:33 <[florian]> oh right, forgot that that the wan interface is separated
Aug 16 09:59:51 <zajec> [florian]: i'm just not sure if I can add for example tcpdump package to it
Aug 16 10:00:10 <[florian]> you might just want to watch the packet count
Aug 16 10:00:18 <zajec> [florian]: it's connected to the same switch, we just give it a separated VLAN id
Aug 16 10:00:18 <[florian]> that will tell you if packets flow through the interface
Aug 16 10:00:25 <zajec> hm, ok
Aug 16 10:00:27 <[florian]> zajec: oh then that is very well possible
Aug 16 10:00:29 <zajec> that should be simple enough
Aug 16 10:00:45 <[florian]> the switch can be programmed to do routing between 1 port and multiple
Aug 16 10:01:06 <[florian]> just found a netgear r6200 in the office, will give it a try
Aug 16 10:01:20 <zajec> [florian]: wow, that would be great
Aug 16 10:01:28 <zajec> [florian]: i really appeciate any extra help :)
Aug 16 10:01:38 <zajec> [florian]: it may always appear i'm doing sth stupid ;)
Aug 16 10:01:44 <merbanan> broadcom hardware can have routing accelerators
Aug 16 10:01:47 <[florian]> don't seem like it
Aug 16 10:01:53 <zajec> so i'd love to get an opinion of someone with his hardware
Aug 16 10:02:15 <[florian]> merbanan: DSL platforms have FAPs which is a different architecture than the one used in bcm47xx/53xx
Aug 16 10:02:54 <merbanan> so 63xx are the only ones with FAP ?
Aug 16 10:03:14 <[florian]> 33xx (cable) also have them
Aug 16 10:03:33 <[florian]> but 47xx/53xx don't have these afair
Aug 16 10:06:28 <merbanan> we get 200mbit routing on dual code 400mhz mips
Aug 16 10:06:40 <merbanan> gig when we enable the FAP
Aug 16 10:08:00 <[florian]> that's not really surprising, the FAP really helps
Aug 16 11:19:01 <OpenWrt-Commits> jow r37800 trunk/package/system/ubox/Makefile * ubox: update to git head
Aug 16 13:14:43 <OpenWrt-Commits> jow r37801 trunk/package/system/ubox/Makefile * ubox: update to git head
Aug 16 14:54:43 <russell--> latest stable kernel, 3.10.7, not 3.10.4 ;-)
Aug 16 14:58:53 <dape> works by editing Makefile
Aug 16 15:32:24 <russell--> weird, make isn't building a -squashfs-sysupgrade.bin, just -factory.bin and -tftp.bin (for wzr600dhp)
Aug 16 15:46:24 <russell--> was working as of r37763
Aug 16 15:46:48 * russell-- sees nothing in diffs since then that readily explains
Aug 16 16:05:06 <DonkeyHotei> can anyone help with http://pastebin.com/Kz9gLhC6 ?
Aug 16 16:08:49 <russell--> looks like the rootfs was too big?
Aug 16 16:08:51 <russell--> weird
Aug 16 16:09:35 <DonkeyHotei> russell--: how so?
Aug 16 16:15:01 <russell--> -rw-r--r-- 1 russell users 201694649 Aug 16 09:08 ./usr/lib/libcrypto.so.1.0.0
Aug 16 16:15:09 <russell--> that seems a little large
Aug 16 16:26:57 <russell--> nbd: my libopenssh ipk is 51M ???
Aug 16 16:27:13 <russell--> sorry, libopenssl*
Aug 16 16:27:32 <jow_laptop> debug options enabled in menuconfig?
Aug 16 16:28:06 <russell--> quite possibly, looking
Aug 16 16:32:13 <russell--> lol, yes.  thank you, jow_laptop
Aug 16 16:32:20 <russell--> much better now ;-)
Aug 16 16:37:41 <heffer> hmm why is dev.openwrt.org so unbelievably unreliable? :D
Aug 16 16:41:05 <Devastator> maybe because git.openwrt.org is reliable?! heheh, might be unrelated but..
Aug 16 17:29:04 <builder> build #320 of rb532 is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/320
Aug 16 17:31:11 <builder> build #320 of ppc44x is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/320
Aug 16 18:10:22 <nbd> russell--: did you disable stripping?
Aug 16 18:10:24 <nbd> :)
Aug 16 18:18:28 <builder> build #333 of uml is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/333
Aug 16 18:56:44 <builder> build #303 of sibyte is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/303
Aug 16 20:56:21 <builder> build #313 of avr32 is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/313
Aug 16 22:33:51 <russell--> jow_laptop, blogic: recent changes fixed my remote syslogging problem, apparently
Aug 16 22:59:01 <russell--> at least on my buffalo wzr600dhp, will try others
**** ENDING LOGGING AT Sat Aug 17 02:59:58 2013