[00:00] <wigyori> i'll play around with them a bit - i hope they won't complain about the '11111111' serial number ;p
[00:00] <nbd> yeah. please tell them about the wl-700gE and the wl-500gP as well
[00:00] <wigyori> okay. :)
[00:02] <tziOm> I notice that login is not enabled by default.. only passwd under "Login/password management utils" .. is this correct?
[00:02] <nbd> yes
[00:03] <nbd> we don't want any password protection on telnet
[00:03] <tziOm> and dropbear has its own method?
[00:03] <nbd> yes
[00:04] <tziOm> but is really the mini_fo enabled in the svn version
[00:16] <framer99> mbm: i took out this change in rc5: https://dev.openwrt.org/changeset/3491 and now mmc.o works for me in rc5
[00:17] <framer99> i can't see how that coudl be it with mmc.o writing gpio directly
[00:17] <framer99> right?
[00:19] <[mbm]> that's just fixing the permissions on /dev/gpio
[00:19] <framer99> roger that
[00:20] <tziOm> mbm - have you checked space-savings compiling everything into kernel and dropping module utils?
[00:20] <[mbm]> tziOm: would be too hard to find a configuration that worked for everyone
[00:21] <nbd> framer99: put the change back in and see if it still works
[00:21] <tziOm> mbm - one for wlhdd is atleast possible :)
[00:21] <framer99> nbd: ok
[00:21] <tziOm> its different when you have extention options ofcorse..
[00:21] <[mbm]> framer99: nothing in the mmc.c you pasted gives a damn what the file permissions on /dev/gpio are
[00:23] <framer99> i agree , just doublechecking that i knew what i was looking at
[00:23] <crazy_imp> tziOm: if i understand you right you only want to have a stripped down kernel for a special device?
[00:48] <framer99> nbd: of course, putting the gpio.c change back and mmc.o still works... but still inop with stock rc5 .trx. I just need to figure out what i did :)
[00:50] <framer99> it's not remaking gpio.o during build at all
[01:19] <[mbm]> hrm .. got the busybox plugins working for simple programs but there seems to be an issue with programs that use symbols from busybox
[01:22] <[mbm]> oh .. duh .. -rdynamic
[01:26] <[mbm]> ok, that works
[01:30] <[mbm]> nbd: http://pastebin.com/761047
[02:04] <nbd> [mbm]: cool
[02:04] <[mbm]> really small patch too
[02:07] <[mbm]> http://downloads.openwrt.org/people/mbm/busybox.diff
[02:15] <nbd> what's with /usr/lib/busybox/busybox.so?
[02:24] <[mbm]> fallback mechanism
[02:25] <[mbm]> otherwise there would be one .so for every external applet
[02:25] <[mbm]> this lets you group multiple as busybox.so
[02:25] <nbd> what happens if it can't find busybox.so?
[02:25] <[mbm]> nothing
[02:26] <nbd> so dlsym just returns null if the library pointer is null?
[02:26] <[mbm]> yeah
[02:26] <[mbm]> which allows me to be excessively lazy with exception handling
[02:27] <nbd> the name should probably have some length checking because of that static buffer that you're using
[02:28] <[mbm]> yeha, should be snprintf
[02:28] <nbd> otherwise it looks fine to me
[02:28] <[mbm]> but as a quick test it seems to work
[02:28] <nbd> yeah
[02:28] <nbd> i need some sleep now. i'll take care of ipkg tomorrow
[09:52] <wigyori> .
[10:40] <CIA-16> kaloz * r3907 /branches/buildroot-ng/openwrt/toolchain/gcc/ (13 files in 3 dirs): add GCC 4.1.1 and sync 4.1.0 with Kamikaze
[15:19] <frop> 88 lo
[15:19] <frop> definitely, kmod-ohci, kmod-usb2, kmod-storage...aren't added to the image
[15:19] <frop> also if i mark them with "*" in buildroot-ng
[18:38] <common> i offer to port 20 packages to buildroot-ng if you merge the wrv54 code from openixp
[22:11] <tziOm> [mbm], hey man.. Your mini_fo.patch for firstboot and mount_root .. where should it be placed in the svn version to be applied to the root?
[22:11] <[mbm]> it's already in there
[22:13] <tziOm> hmm.. looking at build_mipsel/root/sbin/mount_root after enabling mini_fo in menuconfig .. no change there.. still symlinking. is this the wrong place to look?
[22:15] <tziOm> [mbm], ?
[22:15] <[mbm]> wait, are you looking at kamikaze or whiterussian?
[22:15] <tziOm> its the svn version.. thats pretty much what i know atm.. howto figure out?
[22:16] <[mbm]> "svn version" is useless
[22:16] <[mbm]> there's multiple versions in svn
[22:17] <tziOm> aha..
[22:17] <tziOm> what is the correct command to co the kamikaze then? if thats the edge.
[22:18] Action: [mbm] points at dev.openwrt.org
[22:18] <[mbm]> and while I'd love to hold you hand, I have work to do
[22:18] <tziOm> thanks.
[22:37] <[mbm]> nbd: hmm .. thinking about troubleshooting from the end user perspective .. what if when booted into failsafe mode there's a command to kick off a normal boot, displaying the startup in a telnet session .. would have to use ip aliases or something so the network didn't drop but it would give them a look at what would otherwise require a serial
[22:59] <crazy_imp> can i kill the make in the buildroot with ctrl-c and it will continue where i killed it if i start make again? (does it have something like keepwork?)
[23:01] <milan_h> maybe it will restart the current package, but usually it continues where you killed it.. at least in svn
[23:01] <[mbm]> as long as you don't modify the files it should resume somewhat intelligently (although don't hold me to that)
[23:02] <crazy_imp> fine, thx :)
[23:03] <crazy_imp> btw, i would be nice if the sources for the packages are downloaded while compiling the allready downloaded package, so it would be fast on isdn
[23:04] <[mbm]> think buildroot-ng should do that
[23:08] <tziOm> [mbm], what is your success with the bcm43xx descape git?
[23:14] <tziOm> [mbm], using kamikaze for sure now, and still the same!
[23:35] <tziOm> no?
[23:38] <milan_h> is there any reason why, in the brcm-2.6 kernel config in svn, CONFIG_SLAB is not enabled?
[23:38] <Kaloz> because it uses slob?
[23:39] <milan_h> hehe yes
[23:39] <milan_h> I mean why does it not use slab, but slob?
[23:40] <milan_h> with slob I get "only" 550 KByte/s out of my WAN connection (PPPoE), which is capable of 680 KByte/s, and I'm now testing slab...
[23:40] <Kaloz> iirc this was reported to be fixed in 2.6.16.19.. when .17 is out, i will upgrade it
[23:43] <milan_h> so slob is currently worse that it should be?
[23:43] <milan_h> well anyway, I'll try out slab now, will be back in a minute or so....
[23:43] <Kaloz> maybe it has a bug. but Kamikaze isn't the stable branch :P
[23:46] <blux> where are the JTAG people? hands up plz
[23:47] <blux> still working on my Flash thing. . writing work, CFE is in the Flash, nvram is in, but: the ert won't boot
[23:48] <milan_h> ok, back
[23:49] <blux> has someone an idea?
[23:49] <milan_h> with slab instead of slob, I get the full speed of my WAN connection again, an I can also use ssh again while having much traffic
[23:50] <milan_h> but if it's fixed with slob in 2.6.17, that's good too, ofc
[23:50] <Kaloz> milan_h: test 2.6.16.19 or 20 with slob
[23:53] <milan_h> I suppose that the easiest way to do that would be adding the incremental patches to target/linux/generic-2.6/patches...?
[23:53] <Kaloz> nah, simply modify the version number in the makefile
[23:54] <milan_h> then it'll download 40+ MByte again.. but ok who cares
[23:57] <blux> I'm ignored?
[23:57] <Kaloz> blux: no, just noone was here who could help you
[23:57] <blux> ok
[00:00] --- Wed Jun 7 2006
[00:00] <nbd> yeah. please tell them about the wl-700gE and the wl-500gP as well
[00:00] <wigyori> okay. :)
[00:02] <tziOm> I notice that login is not enabled by default.. only passwd under "Login/password management utils" .. is this correct?
[00:02] <nbd> yes
[00:03] <nbd> we don't want any password protection on telnet
[00:03] <tziOm> and dropbear has its own method?
[00:03] <nbd> yes
[00:04] <tziOm> but is really the mini_fo enabled in the svn version
[00:16] <framer99> mbm: i took out this change in rc5: https://dev.openwrt.org/changeset/3491 and now mmc.o works for me in rc5
[00:17] <framer99> i can't see how that coudl be it with mmc.o writing gpio directly
[00:17] <framer99> right?
[00:19] <[mbm]> that's just fixing the permissions on /dev/gpio
[00:19] <framer99> roger that
[00:20] <tziOm> mbm - have you checked space-savings compiling everything into kernel and dropping module utils?
[00:20] <[mbm]> tziOm: would be too hard to find a configuration that worked for everyone
[00:21] <nbd> framer99: put the change back in and see if it still works
[00:21] <tziOm> mbm - one for wlhdd is atleast possible :)
[00:21] <framer99> nbd: ok
[00:21] <tziOm> its different when you have extention options ofcorse..
[00:21] <[mbm]> framer99: nothing in the mmc.c you pasted gives a damn what the file permissions on /dev/gpio are
[00:23] <framer99> i agree , just doublechecking that i knew what i was looking at
[00:23] <crazy_imp> tziOm: if i understand you right you only want to have a stripped down kernel for a special device?
[00:48] <framer99> nbd: of course, putting the gpio.c change back and mmc.o still works... but still inop with stock rc5 .trx. I just need to figure out what i did :)
[00:50] <framer99> it's not remaking gpio.o during build at all
[01:19] <[mbm]> hrm .. got the busybox plugins working for simple programs but there seems to be an issue with programs that use symbols from busybox
[01:22] <[mbm]> oh .. duh .. -rdynamic
[01:26] <[mbm]> ok, that works
[01:30] <[mbm]> nbd: http://pastebin.com/761047
[02:04] <nbd> [mbm]: cool
[02:04] <[mbm]> really small patch too
[02:07] <[mbm]> http://downloads.openwrt.org/people/mbm/busybox.diff
[02:15] <nbd> what's with /usr/lib/busybox/busybox.so?
[02:24] <[mbm]> fallback mechanism
[02:25] <[mbm]> otherwise there would be one .so for every external applet
[02:25] <[mbm]> this lets you group multiple as busybox.so
[02:25] <nbd> what happens if it can't find busybox.so?
[02:25] <[mbm]> nothing
[02:26] <nbd> so dlsym just returns null if the library pointer is null?
[02:26] <[mbm]> yeah
[02:26] <[mbm]> which allows me to be excessively lazy with exception handling
[02:27] <nbd> the name should probably have some length checking because of that static buffer that you're using
[02:28] <[mbm]> yeha, should be snprintf
[02:28] <nbd> otherwise it looks fine to me
[02:28] <[mbm]> but as a quick test it seems to work
[02:28] <nbd> yeah
[02:28] <nbd> i need some sleep now. i'll take care of ipkg tomorrow
[09:52] <wigyori> .
[10:40] <CIA-16> kaloz * r3907 /branches/buildroot-ng/openwrt/toolchain/gcc/ (13 files in 3 dirs): add GCC 4.1.1 and sync 4.1.0 with Kamikaze
[15:19] <frop> 88 lo
[15:19] <frop> definitely, kmod-ohci, kmod-usb2, kmod-storage...aren't added to the image
[15:19] <frop> also if i mark them with "*" in buildroot-ng
[18:38] <common> i offer to port 20 packages to buildroot-ng if you merge the wrv54 code from openixp
[22:11] <tziOm> [mbm], hey man.. Your mini_fo.patch for firstboot and mount_root .. where should it be placed in the svn version to be applied to the root?
[22:11] <[mbm]> it's already in there
[22:13] <tziOm> hmm.. looking at build_mipsel/root/sbin/mount_root after enabling mini_fo in menuconfig .. no change there.. still symlinking. is this the wrong place to look?
[22:15] <tziOm> [mbm], ?
[22:15] <[mbm]> wait, are you looking at kamikaze or whiterussian?
[22:15] <tziOm> its the svn version.. thats pretty much what i know atm.. howto figure out?
[22:16] <[mbm]> "svn version" is useless
[22:16] <[mbm]> there's multiple versions in svn
[22:17] <tziOm> aha..
[22:17] <tziOm> what is the correct command to co the kamikaze then? if thats the edge.
[22:18] Action: [mbm] points at dev.openwrt.org
[22:18] <[mbm]> and while I'd love to hold you hand, I have work to do
[22:18] <tziOm> thanks.
[22:37] <[mbm]> nbd: hmm .. thinking about troubleshooting from the end user perspective .. what if when booted into failsafe mode there's a command to kick off a normal boot, displaying the startup in a telnet session .. would have to use ip aliases or something so the network didn't drop but it would give them a look at what would otherwise require a serial
[22:59] <crazy_imp> can i kill the make in the buildroot with ctrl-c and it will continue where i killed it if i start make again? (does it have something like keepwork?)
[23:01] <milan_h> maybe it will restart the current package, but usually it continues where you killed it.. at least in svn
[23:01] <[mbm]> as long as you don't modify the files it should resume somewhat intelligently (although don't hold me to that)
[23:02] <crazy_imp> fine, thx :)
[23:03] <crazy_imp> btw, i would be nice if the sources for the packages are downloaded while compiling the allready downloaded package, so it would be fast on isdn
[23:04] <[mbm]> think buildroot-ng should do that
[23:08] <tziOm> [mbm], what is your success with the bcm43xx descape git?
[23:14] <tziOm> [mbm], using kamikaze for sure now, and still the same!
[23:35] <tziOm> no?
[23:38] <milan_h> is there any reason why, in the brcm-2.6 kernel config in svn, CONFIG_SLAB is not enabled?
[23:38] <Kaloz> because it uses slob?
[23:39] <milan_h> hehe yes
[23:39] <milan_h> I mean why does it not use slab, but slob?
[23:40] <milan_h> with slob I get "only" 550 KByte/s out of my WAN connection (PPPoE), which is capable of 680 KByte/s, and I'm now testing slab...
[23:40] <Kaloz> iirc this was reported to be fixed in 2.6.16.19.. when .17 is out, i will upgrade it
[23:43] <milan_h> so slob is currently worse that it should be?
[23:43] <milan_h> well anyway, I'll try out slab now, will be back in a minute or so....
[23:43] <Kaloz> maybe it has a bug. but Kamikaze isn't the stable branch :P
[23:46] <blux> where are the JTAG people? hands up plz
[23:47] <blux> still working on my Flash thing. . writing work, CFE is in the Flash, nvram is in, but: the ert won't boot
[23:48] <milan_h> ok, back
[23:49] <blux> has someone an idea?
[23:49] <milan_h> with slab instead of slob, I get the full speed of my WAN connection again, an I can also use ssh again while having much traffic
[23:50] <milan_h> but if it's fixed with slob in 2.6.17, that's good too, ofc
[23:50] <Kaloz> milan_h: test 2.6.16.19 or 20 with slob
[23:53] <milan_h> I suppose that the easiest way to do that would be adding the incremental patches to target/linux/generic-2.6/patches...?
[23:53] <Kaloz> nah, simply modify the version number in the makefile
[23:54] <milan_h> then it'll download 40+ MByte again.. but ok who cares
[23:57] <blux> I'm ignored?
[23:57] <Kaloz> blux: no, just noone was here who could help you
[23:57] <blux> ok
[00:00] --- Wed Jun 7 2006