**** BEGIN LOGGING AT Thu Jul 15 02:59:57 2010 Jul 15 06:49:44 good morning Jul 15 07:36:55 juhosg * r22187 /trunk/target/linux/ar71xx/files/ (3 files in 2 dirs): Jul 15 07:36:55 ar71xx: in-flash macs and eeprom for mach-eap7660d Jul 15 07:36:55 This adds in-flash mac-addresses and calibration data for the Senao Jul 15 07:36:55 EAP7660D board. Jul 15 07:36:55 Signed-off-by: Daniel Golle Jul 15 07:36:57 juhosg * r22188 /trunk/package/mac80211/ (3 files in 2 dirs): Jul 15 07:36:57 ath5k: in-flash macs and eeprom for mach-eap7660d Jul 15 07:36:58 This adds ath5k_platform.h to mac80211 and patches ath5k to respect Jul 15 07:36:58 MAC addresses and eeprom content from platform_data, if supplied. Jul 15 07:36:58 Signed-off-by: Daniel Golle Jul 15 07:37:00 juhosg * r22189 /trunk/target/linux/ (2 files in 2 dirs): generic: move ath5k_platform.h from ar71xx to generic files Jul 15 07:37:03 juhosg * r22190 /trunk/package/madwifi/ (2 files in 2 dirs): Jul 15 07:37:03 madwifi: in-flash macs for mach-eap7660d Jul 15 07:37:03 This patch makes madwifi respect (at least) the platform_data Jul 15 07:37:03 supplied MAC address. Jul 15 07:37:04 Signed-off-by: Daniel Golle Jul 15 13:04:59 juhosg * r22191 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366rb.c rtl8366s.c): generic: rtl8366: use SWITCH_TYPE_NOVAL for reset_mib attributes Jul 15 13:05:05 juhosg * r22192 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366rb.c rtl8366s.c): Jul 15 13:05:05 generic: rtl8366: get rid of private vlan_{mc,4k} structs Jul 15 13:05:05 The usage of bitfields causes problems in Little-Endian mode. Remove Jul 15 13:05:05 them and manipulate the bits directly in the appropriate functions. This Jul 15 13:05:05 change also reduces the size of the binary a bit. Jul 15 13:05:08 juhosg * r22193 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: move switch device to the rtl8366_smi struct Jul 15 13:05:12 juhosg * r22194 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366rb.c rtl8366s.c): generic: rtl8366: get rid of private rtl8366{s,rb} structs Jul 15 13:05:18 juhosg * r22195 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: introduce rtl8366_smi_alloc Jul 15 13:05:22 juhosg * r22196 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: add setup to rtl8366_smi_ops struct Jul 15 13:05:29 juhosg * r22197 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: add common rtl8366_sw_{get,set}_port_pvid functions Jul 15 13:05:31 juhosg * r22198 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h): generic: rtl8366: make rtl8366_{g,s}et_pvid static Jul 15 13:05:34 juhosg * r22199 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366rb.c rtl8366s.c): generic: rtl8366: generealize rtl8366{s,rb}_sw_get_port_mib functions Jul 15 13:05:38 juhosg * r22200 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: add common rtl8366_sw_get_port_mib function Jul 15 13:05:41 juhosg * r22201 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: add is_vlan_valid function to smi_ops Jul 15 13:05:45 juhosg * r22202 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366rb.c rtl8366s.c): generic: rtl8366: use smi->num_ports and smi->ops->get_vlan_4k Jul 15 13:05:48 juhosg * r22203 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: add common rtl8366_sw_get_vlan_info Jul 15 13:05:52 juhosg * r22204 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h rtl8366rb.c rtl8366s.c): generic: rtl8366: add common rtl8366_sw_{get,set}_vlan_ports functions Jul 15 13:05:55 juhosg * r22205 /trunk/target/linux/generic/files/drivers/net/phy/ (rtl8366_smi.c rtl8366_smi.h): generic: rtl8366: make rtl8366_set_vlan static Jul 15 13:05:57 juhosg * r22206 /trunk/target/linux/generic/files/drivers/net/phy/rtl8366_smi.c: generic: rtl8366: add debugfs file to show PVIDs Jul 15 13:38:27 Hello, could someone please elaborate on the apparent obsolescence of brcm-2.4? Does this mean that the open-source broadcom wifi driver is production quality now? Jul 15 13:40:59 the binary driver is available for 2.6 now so there's no need to keep 2.4 around Jul 15 13:42:12 epointsystem: did you get my email about the block-mount on sd cards? Jul 15 13:55:06 cshore: no, I do not remember reading it. Jul 15 13:55:23 xMff: thanks for the clarification. Jul 15 13:56:50 epointsystem: b43 remains default but for unsupported devices (N-phy etc.) the bibary driver can be used Jul 15 13:57:02 binary Jul 15 13:58:28 So what is the recommended setup for the venerable WRT54GL and devices like Asus WL-520gU? Jul 15 13:59:09 as rule of thumb; if the wifi core is old, use b43 Jul 15 13:59:37 if you need multi-ssid, 11n, ap2ap wds or client brides, use wl Jul 15 14:02:35 What is the name of the package with the wl driver? Jul 15 14:02:49 if you absolutely need brcm-2.4 for legacy reasons, use backfire Jul 15 14:02:58 kmod-brcm-wl Jul 15 14:03:09 in package/broadcom-wl Jul 15 14:03:17 Ah, okay. Jul 15 14:04:30 cshore: did you try your image yet btw? Jul 15 14:05:10 It is not going to be merged into the backfire branch, right? Jul 15 14:05:33 i.e. it's only in the trunk and whatever next release will be based on it. Jul 15 14:07:35 yes Jul 15 14:12:51 And will the b43 driver be updated in the backfire branch? Jul 15 14:13:07 yes, along with compat wireless Jul 15 14:32:10 Thank you for the information! Bye. Jul 15 15:34:33 nico * r22207 /branches/backfire/package/mac80211/Makefile: [backfire] package/mac80211: don't build on 2.4 Jul 15 15:38:15 build #82 of brcm63xx is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/brcm63xx/builds/82 Jul 15 15:38:17 build #79 of brcm47xx is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/79 Jul 15 15:38:19 build #77 of sibyte is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/sibyte/builds/77 Jul 15 15:38:22 build #74 of octeon is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/74 Jul 15 16:06:00 ping nbd Jul 15 16:06:10 pong Jul 15 16:06:13 hi Jul 15 16:06:25 I have no good news Jul 15 16:06:37 the issue is still present Jul 15 16:06:47 ok, so we have a leak somewhere else Jul 15 16:08:57 but observing the issue I noticed that during the problem I have 2000 interrupts per second while in normal condition I have about 1000 interrupts per second Jul 15 16:16:13 can you see if you can find something that makes the problem happen more frequently? Jul 15 16:16:23 i.e. frequently disconnecting the client under load Jul 15 16:16:25 stuff like that Jul 15 16:17:34 i'll send you an updated version of my simple buffer leak tracing hack later Jul 15 16:17:46 maybe that'll give us an indication of where buffers are leaked Jul 15 16:17:52 ok Jul 15 16:20:02 I have not observed that kind of stuff i.e. frequently disconnecting Jul 15 16:20:35 the system works fine till the queue is full Jul 15 16:21:00 no, what i meant is: does frequently disconnecting the client speed up the buffer leak? Jul 15 16:21:53 no Jul 15 16:22:01 ok Jul 15 16:22:19 is it possible the system is in some kind of livelock? Jul 15 16:22:30 or interrupt storm? Jul 15 16:23:19 if it was an interrupt storm, you'd have trouble keeping an interactive shell Jul 15 16:23:40 does the problem also appear if you don't have any 11n clients connected and use legacy only? Jul 15 16:24:23 I have just 11n clients Jul 15 16:24:43 try forcing the ap into legacy mode Jul 15 16:24:48 and see if you can still reproduce the issue Jul 15 16:26:36 but it is weird that in normal condition, also with more traffic, I have around 1000 interrupts per second while in this condition the interrupts are twice Jul 15 16:27:03 check the files in debugfs Jul 15 16:27:16 they show you, what kind of interrupts are coming in Jul 15 16:31:23 I have not compiled debugfs Jul 15 16:36:13 what hardware are you testing on, btw Jul 15 16:37:06 Ubiquiti rocket m Jul 15 17:13:44 nico * r22208 /packages/net/tinc/Makefile: [packages] tinc: fix build on powerpc (closes: #7072), reformat Jul 15 19:08:49 jow * r22209 /branches/backfire/package/6in4/ (. Makefile files/6in4.hotplug files/6in4.sh): [backfire] merge r21589, r21612 and r22102 Jul 15 19:34:10 nico * r22210 /packages/libs/libdlna/Makefile: [packages] libdlna: fix build failure (closes: #7445) Jul 15 19:45:42 juhosg * r22211 /trunk/target/linux/ramips/files/ (3 files in 2 dirs): ramips: add mdio_cfg configuration for the ethernet driver Jul 15 19:45:45 juhosg * r22212 /trunk/target/linux/ramips/files/arch/mips/ralink/rt288x/ (devices.c devices.h): ramips: add generic ethernet device for the RT288x Jul 15 19:45:48 juhosg * r22213 /trunk/target/linux/ramips/files/arch/mips/ralink/rt288x/mach-rt-n15.c: ramips: register ethernet device on the RT-N15 Jul 15 19:45:51 juhosg * r22214 /trunk/target/linux/ramips/rt288x/ (config-2.6.32 config-2.6.34): ramips: enable ethernet driver for rt288x Jul 15 22:01:52 jow * r22215 /trunk/package/firewall/ (5 files in 2 dirs): Jul 15 22:01:52 [package] firewall: Jul 15 22:01:52 - notrack support was broken in multiple ways, fix it Jul 15 22:01:52 - also consider a zone conntracked if any redirect references it (#7196) Jul 15 22:08:03 jow * r22216 /branches/backfire/package/firewall/ (Makefile files/uci_firewall.sh): [backfire] firewall: consider zones referenced by redirects as conntracked (#7196) Jul 15 22:42:50 jow * r22217 /packages/net/bind/ (Makefile files/named.init): [packages] bind: fix pid file path (#7218) Jul 15 23:24:03 jow * r22218 /trunk/package/firewall/files/lib/core_init.sh: [package] firewall: fix another notrack related bug Jul 15 23:54:17 nico * r22219 /packages/net/iputils/Makefile: Jul 15 23:54:17 [packages] iputils: don't build on 2.4 Jul 15 23:54:17 * uses setsockopt(sock, SOL_SOCKET, SO_MARK, ...) only available on 2.6 Jul 15 23:57:45 nico * r22220 /packages/multimedia/ (4 files in 4 dirs): Jul 15 23:57:45 [packages] gstreamer & gst-plugins-*: don't build on 2.4 Jul 15 23:57:45 * uses prctl(PR_SET_NAME...) only available on 2.6 Jul 16 00:03:57 nico * r22221 /packages/multimedia/gstreamer/Makefile: [packages] gstreamer: fix libgstreamer dependencies not inherited from default (missed from [22220]) Jul 16 00:07:01 jow * r22222 /trunk/package/dnsmasq/ (Makefile files/dhcp.conf files/dnsmasq.init): [package] dnsmasq: add and enable DNS rebind protections Jul 16 00:07:46 yay I got 22222 Jul 16 00:55:08 xMff: didn't you steal 20000 also? Jul 16 00:55:44 no that was thepeople Jul 16 01:00:18 jow * r22223 /branches/backfire/package/dnsmasq/ (Makefile files/dhcp.conf files/dnsmasq.init): [backfire] merge r21523, r21871 & r22222 Jul 16 01:05:07 xMff, Jul 16 01:05:11 is CIA-38 a bot ? Jul 16 01:05:16 yes Jul 16 01:05:25 that says when svn is updated :o Jul 16 01:05:29 thats badass! Jul 16 01:06:06 xMff, are u a developer ? Jul 16 01:06:30 * realopty edits the wiki Jul 16 01:11:01 I suppose I am Jul 16 01:23:44 xMff, how do i go about getting an ;official; download for targeted hardware? Jul 16 01:24:27 ? Jul 16 01:27:02 xMff, mmmm the regular downloads dont have kmod-ide-** or usb/2, usb-storage, and some other things i cant remember Jul 16 01:27:33 id like to have an official download for some hardware :\ Jul 16 01:28:52 xMff, http://downloads.openwrt.org/snapshots/trunk/brcm47xx/ Jul 16 01:29:16 basicly i wanna have an image other people can use out of the box Jul 16 01:32:13 well step one is submitting an appropriate profile for this particular hw Jul 16 01:32:26 into a ticket? Jul 16 01:32:50 should i just upload my .config ?? Jul 16 01:33:09 step two is covincing the buildbot maintainer to compile subtargets which in turn requires processing power ... Jul 16 01:33:27 a real profile as patch in a ticket Jul 16 01:33:47 see target/linux/brcm47xx/profiles/ Jul 16 01:34:32 who is the buildbot maintainer? Jul 16 01:34:36 nbd ? Jul 16 01:35:08 thepeople Jul 16 01:35:39 kk Jul 16 02:02:15 cshore, ping Jul 16 02:10:17 acoul, ping Jul 16 02:28:37 nico * r22224 /packages/utils/cwiid/Makefile: Jul 16 02:28:37 [packages] cwiid: don't build on 2.4, reformat, cleanup Jul 16 02:28:37 * use the uinput subsystem, only available on 2.6 Jul 16 02:33:32 nico * r22225 /packages/multimedia/fswebcam/Makefile: Jul 16 02:33:32 [packages] fswebcam: don't build on 2.4 Jul 16 02:33:32 * uses Video 4 Linux 2, only available on 2.6 Jul 16 02:36:40 * realopty created http://wiki.openwrt.org/doc/uci/ushare **** ENDING LOGGING AT Fri Jul 16 02:59:56 2010