**** BEGIN LOGGING AT Mon Jun 13 02:59:57 2011 Jun 13 03:07:28 Does Automatic process group scheduling help to improve performance a bit for embeded device? Jun 13 03:30:10 OutBackDingo: https://dev.openwrt.org/wiki/people Jun 13 03:35:41 philipp64|laptop: ahhhhhhhh right Jun 13 04:16:17 what is the difference between supported and advertising in struct phy_device? Jun 13 05:35:31 build #41 of ar71xx is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/41 Jun 13 06:27:05 build #52 of at91 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/52 Jun 13 06:47:53 build #51 of ubicom32 is complete: Failure [failed shell_4 compile_3] Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/51 Jun 13 08:13:40 juhosg * r27165 /trunk/target/linux/ar71xx/files/arch/mips/ (ar71xx/early_printk.c include/asm/mach-ar71xx/ar933x_uart.h): ar71xx: update early_printk code Jun 13 08:13:41 juhosg * r27166 /trunk/target/linux/ar71xx/files/drivers/serial/ar933x_uart.c: ar71xx: use ar933x_uart.h in the AR933X serial driver Jun 13 08:13:43 juhosg * r27167 /trunk/tools/firmware-utils/src/ (buffalo-enc.c buffalo-lib.c): Jun 13 08:13:43 tools/firmware-utils: fix buffalo csum calculation Jun 13 08:13:43 Also print an error message if decryption fails. Jun 13 09:15:31 build #33 of gemini is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/33 Jun 13 10:28:52 Is it possible to setup ath9k drivers in 3 radio ar71xx boards in 2.4Ghz band .? to make hotspot coverage in 360 degree for longer range with external antenna Jun 13 11:00:59 Hauke: is a git patch with detected copies also okay? I finally found out I have to increase the number of candidates to get git to properly detecting copies (shrinking the patch to ~180k) Jun 13 11:07:57 build #54 of brcm63xx is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/54 Jun 13 11:08:56 KanjiMonster: I do not know how git submits these patches to svn Jun 13 11:12:16 Hauke: I'll test in my local svn Jun 13 11:34:08 Hauke: the git svn dcommit also takes the --find-copies-harder and -l arguments; using these it properly detects copy-and-modify (but since it's automated it doesn't take everything from -2.6.39; some patches are auto-based on .37 and .38 versions) Jun 13 11:35:06 Hauke: anyway, I have to regenerate the patch; I just saw my ipv6-bridge rfc patch crept into the patches Jun 13 11:46:12 build #46 of rb532 is complete: Failure [failed shell_4 compile_11] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/46 Jun 13 11:50:43 build #45 of x86 is complete: Failure [failed compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/45 Jun 13 12:01:37 build #44 of uml is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/44 Jun 13 12:46:17 florian * r27168 /trunk/target/linux/uml/patches-2.6.39/ (2 files): Jun 13 12:46:17 [uml] replace unfortify patch with a better one Jun 13 12:46:17 Thanks to Richard Weinberger for providing this patch Jun 13 13:07:27 build #44 of ppc44x is complete: Failure [failed shell_4 compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/44 Jun 13 13:17:12 mirko * r27169 /trunk/package/libbsd/ (. Makefile): Jun 13 13:17:12 [package] add library 'libbsd' Jun 13 13:17:12 This library contains core functionality usually (but obviously not always) Jun 13 13:17:12 provided by libc implementations which isneeded by OpenWrt core packages. Jun 13 13:17:12 For instance the functions 'strlcat', 'strlcpy' are not implemented in Jun 13 13:17:12 eglibc but needed by hotplug2 (udevtrigger) Jun 13 13:17:14 mirko * r27170 /trunk/package/hotplug2/Makefile: [package/hotplug] link hotplug2 against libbsd when eglibc is used Jun 13 13:45:15 philipp64|laptop: ping Jun 13 13:56:51 Is there a newer version of samba than 3.0.24 and I just dont see it, or is there some other reason there isn't a newer version? Jun 13 14:00:19 I realize the newer version will have a much larger binary Jun 13 14:08:12 netprince: we used some patches by avm to reduce the size of the samba binary and this is not ported to a newer version Jun 13 14:09:11 are the patches necessary? What is the size by default? Jun 13 14:14:56 the current version does not seem compatible with win7 without reducing security settings in win7 Jun 13 14:15:32 not compatible with ntlmv2 IIRC Jun 13 14:17:33 netprince: the unmodified size is ~3 times larger Jun 13 14:18:20 netprince: I started adding a new samba version, but nmbd is missing: https://dev.openwrt.org/ticket/8326 Jun 13 14:22:50 Many devices have >32M of storage, even more RAM, I wonder if it would make sense to have a version without the avm patches? Jun 13 14:23:57 netprince: yes it would make sense to have such a version for these users Jun 13 14:26:21 Hauke: how do I create a patch including the copy&modify info with svn? svn diff just has the changes after copying the files with svn cp Jun 13 14:26:55 KanjiMonster: How is 3.0-rc2 support going? Jun 13 14:27:09 hjwork: it's working for me ;) Jun 13 14:27:42 <^Willie^> ehh i have a simple question...in the luci webinterface at the "services" there is/was a "scripts init.d" tab that i can not find in the backfire releases but wat do i need to install to just get that single tab Jun 13 14:28:51 KanjiMonster: I do not know Jun 13 14:29:23 <^Willie^> unfortunately this example is in french but i talk about the left tab at the left side named "scripts d'initialisation" -> http://router-firmware-test.gamma.nu/LuCI/services/index.html. Jun 13 14:29:45 <^Willie^> wat is the package's name.. Jun 13 14:30:20 but when someone wants to review the patch it is much easier to review the differences from 2.6.39 that the hole patches. Jun 13 14:30:35 ^Willie^: luci-app-initmgr Jun 13 14:30:57 <^Willie^> xMff: hmm then there is something wrong with the backfire x86 releases :( Jun 13 14:31:29 <^Willie^> my diy builds it is there but the backfire precompiled set i do have luci-app-initmgr but not the rc.local field in the webui.. Jun 13 14:31:31 it has been merged into admin full in recent versions Jun 13 14:32:04 <^Willie^> since my diy builds fail from time to time :S Jun 13 14:32:21 <^Willie^> there do i wish to use a precompiled build and add wat is needed from remote host Jun 13 14:32:44 well what precompiled build are you talking about? Jun 13 14:33:02 10.03, 10.03.1-rc1, 10.03.1-rc2, 10.03.1-rc3, 10.03.1-rc4, 10.03.1-rc5 ? Jun 13 14:33:02 <^Willie^> but there was a ccs file missing or atleast the openwrt gui menu layout is incorrent and initmgr was there but no rc.local tab.. Jun 13 14:33:13 <^Willie^> http://downloads.openwrt.org/backfire/10.03.1-rc4/x86/ Jun 13 14:33:32 yeah, no rc.local tab implemented Jun 13 14:33:37 first appear in rc5 Jun 13 14:34:24 <^Willie^> argh since i do have it in R27156 (trunk build) Jun 13 14:35:36 <^Willie^> hmm guess i can better start over again from 0 .. Jun 13 14:35:54 <^Willie^> xMff: are the x86 .config files shared/stored somewhere in the svn ? Jun 13 14:37:06 <^Willie^> just that tab did make me redo a lot work last night :( Jun 13 14:39:25 build #40 of ixp4xx is complete: Failure [failed shell_4] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/40 Jun 13 14:39:45 <^Willie^> xMff: thanks ! Jun 13 14:41:48 Hauke: As long as it's about the patch, I can do that somewhat with git -> http://inuyasha.ath.cx/~jonas/0009-linux-generic-Add-support-for-3.0-rc2.patch (best I could do with options that are also supported by git svn dcommit; "--find-copies-harder -l1000") Jun 13 14:42:21 :P Jun 13 14:42:42 <^Willie^> lolz normal they pingtimeout with any text :p Jun 13 14:43:22 <^Willie^> do notice that in other channels then you highlite someone's name.. and they just timeout straight away ;-) Jun 13 14:45:08 ah well, then I'll bbl - off to play lego harry potter ;) Jun 13 14:45:44 <^Willie^> sounds easyer then wii harry potter :) Jun 13 14:46:14 <^Willie^> anyway that game is the same as asterix/spyro and a lot other games :( Jun 13 14:46:41 <^Willie^> one sdk, one gameplay, multiple theme's.. Jun 13 14:50:01 jow_laptop: Did you ever receive your buffalo power supply? Jun 13 14:51:59 fullstop: I think they shipped one to me. The tech support drone handed of my postal details to the rma department Jun 13 14:52:19 fullstop: that was friday, today is public holyday, so maybe during this week Jun 13 14:53:38 Do you have your wzr handy? Jun 13 14:53:43 yes Jun 13 14:54:11 does your SSID (mac address) on the label begin with 02 or 4C? Jun 13 14:54:25 00:24 Jun 13 14:55:02 okay. Have you used it much, at all? Jun 13 14:55:25 briefly, maybe three days in total Jun 13 14:56:08 I'm pretty sure that 00:24 has the rtl8366s and the 4C:E6 has the rtl8366rb, but that's based on examining ~40 units. Jun 13 14:56:41 let me open it Jun 13 14:57:36 openwrt doesn't probe the ethernet yet, so I'm looking for a way to make a universal image for both the 300 and 301, along with the 300 with the "rb" chipset., Jun 13 14:58:03 ah, damn ... its under the heatshield Jun 13 14:58:13 or rf shield Jun 13 14:59:20 maybe I should solder serial whily I'm at it Jun 13 14:59:40 *while Jun 13 15:01:37 I looked into the buffalo uboot code, and they are probing the MII bus. Jun 13 15:07:41 xMff: I have a target-specific base-files/etc/config/network file, but it's not appearing on the target image when I boot. any ideas why? Jun 13 15:13:31 mirko: pong Jun 13 15:23:07 fullstop: it should be fairly easy; the registers of both devices are accessed the same way, the ID registers are just at different addresses; so to make a "universal" rtl8366 driver you just need to probe for both and then use the corresponding driver/ops Jun 13 15:26:25 KanjiMonster: I'll look into that some more. Jun 13 15:26:29 KanjiMonster: Thanks. Jun 13 15:26:45 What does this mean for /etc/config/network, where the device name is specified in the config? Jun 13 15:27:05 Is w25q128 fully supported by openwrt? Jun 13 15:28:46 philipp64|laptop: you're using eglibc with OpenWrt I guess? Jun 13 15:29:01 (in case you're the philipp with whom i talked on -devel) Jun 13 15:29:31 fullstop: they probably should then use a common name (rtl8366 or rtl8366x or something like that) Jun 13 15:29:43 fullstop: compare https://dev.openwrt.org/browser/trunk/target/linux/generic/files/drivers/net/phy/rtl8366s.c#L1006 vs https://dev.openwrt.org/browser/trunk/target/linux/generic/files/drivers/net/phy/rtl8366rb.c#L1177 ;) Jun 13 15:31:04 fullstop: I already thought about letting swconfig make switches also available as switchX (the same as ethX or wlanX) Jun 13 15:37:36 <^Willie^> mirko: jup he is/was using eglibc Jun 13 15:42:24 mirko: I was trying to, yes. Jun 13 15:44:08 KanjiMonster: I like the idea of merging the two drivers together; they are incredibly similar. Jun 13 15:44:23 However, that still doesn't help me with the switch config. Jun 13 15:44:43 can anyone tell me why target/linux/x86/geos/base-files/etc/config/network wouldn't show up on my built system? Jun 13 15:45:01 (assuming I'm building x86-geos as the target, that is...) Jun 13 15:45:37 <^Willie^> philipp64|laptop: hmm for me while building glibc targets i notices some files where on my image but in the wrong subdirs. Jun 13 15:45:43 KanjiMonster: afaik they have completely different vlan representations, the s has 4096 table entries while the other one can only have 16(?) but with vids up to 4096 Jun 13 15:46:03 <^Willie^> for example some lua/luci stuff was not working untill i moved some libs from /???/lib to /lib.. Jun 13 15:46:15 this was a separate build... with uclibc. Jun 13 15:46:48 sounds like your --prefix= or --libdir= wasn't being set correctly. Jun 13 15:47:12 <^Willie^> philipp64|laptop: jup i think so but for glibc based builds try seeking that on the forums :) Jun 13 15:47:37 <^Willie^> also in the end i have managed getting everything working in the "default" uclibc based builds anyway Jun 13 15:47:52 philipp64|laptop: ^Willie^: i'm gonna use the eglibc in a project now - experienced some issues but got most sorted out.. Jun 13 15:48:10 just noticed /usr/lib is not in the default ld search path *grr* Jun 13 15:48:20 <^Willie^> indeed /lib is Jun 13 15:48:32 <^Willie^> but i dunno if luci or any other part expect it in /lib Jun 13 15:48:42 <^Willie^> like my iptables where even seeking in /lib :( Jun 13 15:49:28 <^Willie^> but full luci stuff included still +- 9.5M combined image size Jun 13 15:49:41 <^Willie^> for a glibc based build Jun 13 15:50:35 nbd: ping Jun 13 15:54:43 well, I guess the mystery of why /etc/config/network isn't being installed from base-files will have to wait another day. Jun 13 15:56:39 <^Willie^> hmm i'm still at the point that i that grepping a lot files seeking for the reason of a error when i have no clue wich file is the source of my problem ;) Jun 13 15:56:47 <^Willie^> in case i can't find it at the mailing lists.. Jun 13 15:57:17 philipp64|laptop: paste the output of "make package/base-files/{clean,compile} V=99" Jun 13 15:59:17 KanjiMonster: regardless of probing, I have the rtl8366rb driver up and running, and swconfig dev rtl8366rb show returns what appears to be valid information. Jun 13 15:59:37 build #52 of s3c24xx is complete: Failure [failed shell_4 compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/52 Jun 13 15:59:43 However, despite the byte counters increasing nothing appears to be passing through. Jun 13 16:04:28 WAN port works. Jun 13 16:04:37 but none of the lan appear to be Jun 13 16:11:15 build #40 of xburst is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/40 Jun 13 16:29:07 xMff: http://fpaste.org/QXU8/ Jun 13 16:35:55 philipp64|laptop: thats target/linux/x86/base-files.mk Jun 13 16:35:57 dunno why Jun 13 16:37:18 ah, was introduced for https://dev.openwrt.org/changeset/25103 Jun 13 16:39:57 xMff: you lost me. Jun 13 16:40:19 generic x86 has a base-files hook which removes etc/config/network Jun 13 16:40:31 this was added to make soekris defconfig stuff work Jun 13 16:41:22 that's broken. why would a profile have implications across all targets of a given processor family? Jun 13 16:41:37 what needs to be done to fix it? Jun 13 16:42:30 r25103 Jun 13 16:46:49 somehow the soekris stuff must be changed Jun 13 16:46:57 to cope with an existing genric config Jun 13 16:49:58 <^Willie^> or make a profile for soekris stuff .. Jun 13 16:50:16 <^Willie^> that overrule the generic config when select soekris ;) Jun 13 16:50:24 build #33 of mpc52xx is complete: Failure [failed shell_4 compile_3] Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/33 Jun 13 16:50:56 <^Willie^> anyway diner time here.. bbl Jun 13 16:52:06 can that go into the x86/profiles/*.mk files? or... how do I detect if I'm building a specific profile? Jun 13 16:59:50 philipp64|laptop: the .config should hold a CONFIG_* fr the profile Jun 13 17:00:45 CONFIG_TARGET_lantiq_xway_GIGASX76X would be the GIGASX76X profile of the xway subtarget of the lantiq target Jun 13 17:17:36 well there's the thing Jun 13 17:17:53 profiles should be nothing more than package selections Jun 13 17:18:15 if something needs different base-files, kernel stuff whatever it should be a subtarget Jun 13 17:19:13 profiles where once used for stuff like "broadcom box with broadcom wifi driver", "broadcom box with madwifi driver", "broadcom box with usb modules" Jun 13 17:19:50 especially since some soekris versions are already sub-targets (?) I think they all should be Jun 13 17:19:55 not some sub-targets and some profiles Jun 13 17:23:09 Let's try that again ... Hauke: As long as it's about the patch, I can do that somewhat with git -> http://inuyasha.ath.cx/~jonas/0009-linux-generic-Add-support-for-3.0-rc2.patch (best I could do with options that are also supported by git svn dcommit; "--find-copies-harder -l1000") Jun 13 17:29:49 xMff: are you sure about that? both the rb and the s register with dev->vlans = *_NUM_VIDS; (= 4096); and both have a 4k mode. Perhaps you are thinking about the rtl8306? that one has only 16 vlans. Jun 13 17:30:16 KanjiMonster: ah so both can take 4096 vlan entries? Jun 13 17:30:25 afair it used to be not the case in the past Jun 13 17:30:39 xMff: according to the code, yes Jun 13 17:31:05 I really with one could do the same for the other drivers Jun 13 17:31:16 to end the vlan vs vid confusion Jun 13 17:32:22 I think if you ignore all rtl8366s vs rb instances the code seems to have a similarity of 80-90% (guessed ;) Jun 13 17:52:35 regarding the rtl8366x.. I configured my device so that it will use the rtl8366rb driver, using the latest backfire branch svn.. Jun 13 17:53:11 The WAN port works great! However, the lan port.. not so much. Any ideas on how to debug this? Jun 13 17:56:30 fullstop: check your dmesg, do you have somthing like "RTL5937 ver. chip found"? Jun 13 17:57:15 KanjiMonster: Yes. Jun 13 17:57:50 KanjiMonster: the wan port works... so I am guessing that it is a vlan thing. Jun 13 17:58:27 KanjiMonster: swconfig output: http://pastebin.com/Kj1Uj0An Jun 13 17:58:28 no the wan port is a dedicated phy Jun 13 17:58:35 a separate nic so to say Jun 13 17:59:00 xMff: interesting. I did not know that. Jun 13 18:00:36 fullstop: might be something other; I just remembered that the rb and s require different PLL values on ar71xx; that might be the problem Jun 13 18:01:14 KanjiMonster: I took care of that. Jun 13 18:01:45 fullstop: what does your /e/c/network say? Jun 13 18:04:02 KanjiMonster: http://pastebin.com/QtmjKCrQ Jun 13 18:04:28 fullstop: change lan's ifname from eth0 to eth0.1 Jun 13 18:04:32 It is essentially the same as the rtl8366s with a different device name Jun 13 18:04:37 KanjiMonster: I tried that as well.. Jun 13 18:04:58 fullstop: try rebooting after changing it Jun 13 18:05:07 One step ahead of you.. ;-) Jun 13 18:05:18 That didn't make a difference. Jun 13 18:05:54 fullstop: alternativly change the vlan 1's ports to "0 1 2 3 5" instead of "0 1 2 3 5t" Jun 13 18:10:39 okay, I just changed it to 0 1 2 3 5 with eth0 Jun 13 18:10:42 Booting now Jun 13 18:16:01 KanjiMonster: nothing different.. Jun 13 18:16:05 let me disable vlan Jun 13 18:19:49 KanjiMonster: nope.. :( Jun 13 18:20:13 fullstop: hrm. I'm currently out of ideas. Jun 13 18:20:51 The byte counters are going up.. both in ifconfig and swconfig Jun 13 18:30:55 KanjiMonster: it looks good Jun 13 18:35:21 Hauke: should I resend the patches with the linux patches refreshed? (the one I linked you already has two issues I found fixed; I removed the ipv6 bridge patch again and I added an atm patch that recently got added to .39 - it was hidden in a lantiq commit ;) Jun 13 18:39:55 KanjiMonster: I might have figured it out. There was a typo on the eth0 pll when I initialized the rtl8366rb. eth1 was fine, and worked. Jun 13 18:40:09 I'm rebuilding the image and I'll let you know if it worked or not. Jun 13 18:47:38 KanjiMonster: don't you have commit access? Jun 13 20:07:51 in mach-wzr-hp-g300nh.c, what is the reasoning behind explicitly setting the port speed on eth0 but not on eth1? Jun 13 20:14:46 Hauke: not yet Jun 13 20:15:42 fullstop: yes, since eth0 is directly wired to the switch chip, it's always connected Jun 13 20:18:57 KanjiMonster: should I commit it or would you like to wait till you have access Jun 13 20:23:32 Hauke: feel free to commit - btw, I only added a subset of your compat(-wireless) patches; all that enable features (or the bcma ones) I didn't include because because I assume the next compat-wireless we'll use will include them anyway Jun 13 20:24:22 KanjiMonster: yes I saw it Jun 13 20:31:24 KanjiMonster: Okay, that makes sense. Jun 13 20:31:49 KanjiMonster: I still don't have a working eth0.. :) I did notice something interesting, though.. Jun 13 20:32:56 I can see that it is using the rtl8366rb driver and, yet, swconfig's output states "Switch 1: rtl8366rb(RTL8366S), ports 6 (cpu @ 5), vlans: 16 Jun 13 20:33:20 Where does RTL8366S come from? A quick look at the source suggests that it is the driver name. Jun 13 20:33:57 dmesg shows rtl8366rb rtl8366rb: RTL5937 ver. 3 chip found Jun 13 20:53:10 ugh. it might have been a build problem. :( Jun 13 21:52:28 mirko * r27171 /packages/Xorg/lib/qt4/Makefile: [packages/qt4] only build certain components if selected Jun 13 22:38:30 blogic * r27172 /trunk/target/linux/lantiq/ (2 files in 2 dirs): Jun 13 22:38:30 [lantiq] Jun 13 22:38:30 Fix for bug where the iniatialisation of the MAC data private strucure Jun 13 22:38:30 is copied into the wrong location. Jun 13 22:38:30 Ticket #9547 Jun 13 22:38:30 Signed-off-by: Conor O'Gorman Jun 13 22:50:03 build #50 of atheros is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/50 Jun 14 02:37:00 blogic: ping Jun 14 02:40:29 build #42 of sibyte is complete: Failure [failed shell_4] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/42 **** ENDING LOGGING AT Tue Jun 14 02:59:57 2011