[10:33] <CIA-17> nico * r4135 /packages/libs/libelf/ (. Config.in Makefile ipkg/): port libelf to buildroot-ng, update it to v0.8.8
[14:09] <CIA-17> nico * r4136 /packages/libs/ncurses/ (. Config.in Makefile ipkg/): port ncurses to buildroot-ng.
[14:20] <CIA-17> nico * r4137 /packages/libs/readline/ (. Config.in Makefile ipkg/): port readline to buildroot-ng, update it to v5.1
[14:47] <CIA-17> nico * r4138 /packages/libs/libtool/ (. Config.in Makefile ipkg/): port libtool to buildroot-ng
[15:01] <CIA-17> nico * r4139 /packages/libs/libnet/ (. Config.in Makefile ipkg/): port libnet to buildroot-ng
[15:22] <CIA-17> nico * r4140 /packages/libs/lzo/ (. Config.in Makefile ipkg/): port lzo to buildroot-ng
[15:39] <CIA-17> nico * r4141 /packages/libs/pcre/ (. Config.in Makefile ipkg/): port pcre to buildroot-ng
[16:37] <florian__> {Nico}: is on a comitting spree ;)
[16:42] <BrainSlayer> florian?
[16:56] <CIA-17> nico * r4142 /packages/libs/mysql/ (. Config.in Makefile ipkg/): port mysql (client library) to buildroot-ng
[17:02] <{Nico}> florian__: right, i'm off today and it's way too hot here to do anything else
[17:06] <{Nico}> is anyone else porting packages to buildroot-ng?
[17:07] <nbd> i'm working on package/kernel right now
[17:07] <nbd> (moving the kernel module stuff there so that it reuses the existing packaging code)
[17:27] <CIA-17> nico * r4143 /packages/libs/postgresql/ (. Config.in Makefile ipkg/): port postgresql (client library and cli) to buildroot-ng
[17:31] <CIA-17> nbd * r4144 /branches/buildroot-ng/openwrt/include/package.mk: check if the Package/$(NAME) template is defined before packaging or installing ipkg packages
[17:34] <{Nico}> nbd: i saw you changed ./scripts/rstrip.sh to also sstrip already stripped binaries
[17:34] <{Nico}> iirc, it make unusable binaries
[17:34] <nbd> yes
[17:34] <nbd> worked for me
[17:34] <nbd> didn't see any unusable binaries
[17:35] <nbd> do you have any unusable binaries in your build?
[17:36] <{Nico}> nbd: i have to check (and remember what my test case was)
[17:41] <{Nico}> nbd: how to you add compile-time dependencies to DEPENDS?
[17:42] <{Nico}> ie. my package needs openssl to build but the resulting binaries don't need it at runtime
[17:43] <nbd> PKG_BUILDDEP
[17:43] <nbd> but the dependency system needs to be changed
[17:43] <nbd> it's not flexible enough for hostapd
[17:43] <nbd> because there's hostapd and hostapd-mini
[17:43] <nbd> and mini has no build dependency on openssl
[17:45] <{Nico}> can't we move PKG_BUILDEP somewhere in 'define Package/foo' template?
[17:58] <nbd> yeah, i was planning on doing that
[17:59] <nbd> but the script that generates .pkgdep needs to be changed as well
[17:59] <nbd> so that the config is checked for the package dependencies
[18:09] <CIA-17> nico * r4145 /packages/libs/cyrus-sasl/ (. Config.in Makefile ipkg/): port cyrus-sasl to buildroot-ng
[18:17] <CIA-17> nbd * r4146 /branches/buildroot-ng/openwrt/ (5 files in 2 dirs): move some of the kernel module packaging to package/kernel/
[18:18] <nbd> {Nico}: if you have time. could you review package/kernel/Makefile and tell me if you think it's suitable for all modules?
[18:19] <CIA-17> nico * r4147 /packages/libs/openldap/ (. Config.in Makefile ipkg/): port openldap (client libraries and utilities) to buildroot-ng
[18:22] <CIA-17> nbd * r4148 /branches/buildroot-ng/openwrt/package/kernel/Makefile: allow multiple kernel config dependencies
[18:35] <CIA-17> nbd * r4149 /branches/buildroot-ng/openwrt/ (5 files in 3 dirs): move the cryptoapi kmod to package/kernel
[18:36] <{Nico}> nbd: ok
[18:43] <nbd> that should make dependency handling a lot easier
[18:44] <CIA-17> nico * r4150 /packages/libs/libevent/ (. Config.in Makefile ipkg/): port libevent to buildroot-ng
[18:49] <nbd> i'm backporting the jffs2 patches to wr right now
[18:52] <{Nico}> during the last meeting, we talk about adding missing nvram setting for wireless
[18:53] <[mbm]> for webif, yes
[18:53] <nbd> which ines?
[18:53] <nbd> s/ines/ones/
[18:53] <[mbm]> webif really should show every nvram variable openwrt uses
[18:54] <{Nico}> [mbm]: ok, i'll try to work on that
[18:54] <CIA-17> nbd * r4151 /branches/buildroot-ng/openwrt/target/image/brcm/Makefile: disable wgt634u images for linux 2.4 (only works with 2.6)
[18:55] <{Nico}> nbd: broadcom-wl in buildroot-ng is missing the libbcmcrypto.so (for the nas package)
[18:55] <nbd> is it missing in the tarball?
[18:55] <nbd> because the nas package references it
[18:56] <{Nico}> nbd: i guess it's missing in the archive
[18:57] <nbd> i just built it
[18:57] <nbd> it's there
[18:58] <CIA-17> nbd * r4152 /branches/buildroot-ng/openwrt/package/broadcom-wl/Makefile: change depends/default for nas
[18:58] <CIA-17> nico * r4153 /packages/libs/gmp/ (. Config.in Makefile ipkg/): port gmp to buildroot-ng
[18:59] Action: nbd tests the flash map and jffs2 backport
[19:08] <nbd> damn... where did i put my serial cable...
[19:09] <[mbm]> jffs2 patch didn't work? :P
[19:18] <nbd> [mbm]: i don't remember the ip of the device...
[19:26] <CIA-17> nico * r4154 /packages/libs/expat/ (. Config.in Makefile ipkg/): port expat to buildroot-ng
[19:41] <florian__> BrainSlayer: what ?
[19:44] <florian__> Kaloz: ping ?
[19:44] <{Nico}> florian__: i couldn't make it for rmll, how was it?
[19:44] <florian__> {Nico}: it was great
[19:45] <florian__> {Nico}: did you read the presentation ?
[19:45] <{Nico}> florian__: did anyone make a movie of your presentation?
[19:45] <florian__> {Nico}: unfortunately no
[19:45] <{Nico}> florian__: yep, good work
[19:46] <florian__> {Nico}: thanks, I was not sure how to make a good plan and I did not wanted to make it too technical
[19:46] <{Nico}> how many people attended?
[19:46] <florian__> between 40 and 50 it was a small room
[19:46] <florian__> brb
[19:50] <Kaloz> florian__: pong
[19:50] <nbd> finally found my cable
[19:51] <BrainSlayer> hey
[19:51] <BrainSlayer> @florian: do you know why the flash mapping for the brcm63xx port wont work?
[19:53] <florian__> BrainSlayer: sure, because the crappy flash map does only manage cfe, and is lost when the bootloader is redboot
[19:53] <BrainSlayer> i'm using still the cfe
[19:54] <BrainSlayer> if i would use redboot it would be easier
[19:54] <florian__> ah then I don't know, I did not tested it
[19:54] <florian__> I only managed to have a redboot device working
[19:54] <BrainSlayer> do you have a working redboot image for this these units?
[19:54] <florian__> yes
[19:54] <CIA-17> nbd * r4155 /branches/whiterussian/openwrt/ (4 files in 4 dirs): backport jffs2 changes and flash map from buildroot-ng
[19:54] <nbd> cool. worked right on first try :)
[19:54] <BrainSlayer> can you provide me this image?
[19:54] <[mbm]> you sound surprised
[19:55] <florian__> Kaloz: sent me his livebox bootloader which was not protected like mine
[19:55] <florian__> BrainSlayer: let me find it
[19:55] <BrainSlayer> okay
[19:55] <nbd> [mbm]: well, the flash map was a bit different on wr. i just manually edited the diff without rediffing afterwards
[19:55] <[mbm]> ah
[19:55] <florian__> Kaloz: do you remember what was the things that needed to be changed in the brcm63xx patches ?
[19:55] <BrainSlayer> pci-fixups
[19:55] <BrainSlayer> i already have done this
[19:55] <Kaloz> florian__: nope, but i will look at those as well just my last days were pretty fsck'ed up
[19:56] <Kaloz> florian__: it's too hot here, and i can hardly sleep..
[19:56] <nbd> yeah, i have trouble with the temperature as well
[19:56] <BrainSlayer> the same here. 36 degrees today
[19:57] <florian__> Kaloz: lol ok, I just wanted to commit the 2.6.17 patches and clean them up
[19:58] <BrainSlayer> @florian: i'm currently cheating with a vendor to get the broadcom adsl sources
[19:58] <florian__> someone here has already been working with cisco catalyst 3550 ? or similar ?
[19:58] <florian__> BrainSlayer: ok great
[19:58] <Kaloz> florian__: will look at those later
[19:58] <florian__> Kaloz: ok no pb
[19:59] <crazy_imp> florian__: openwrt on a cisco switch? cool :D
[20:00] <florian__> crazy_imp: no, I am having serious troubles with a catalyst 3550, it's the core layer3 switch of my school
[20:00] <florian__> it has been freezing with no reasons at all the last few days
[20:00] <florian__> althought the temperatur seems not to have changed in the room
[20:01] <nbd> so. what's missing for rc6?
[20:01] <[mbm]> I gave you a list
[20:02] <florian__> BrainSlayer: anyway the brcm63xx driver seems to be crappy as hell, so it would either be better to start reversing it
[20:02] <nbd> florian__: i don't think we can find enough people to do that
[20:03] <nbd> florian__: and reversing device drivers is not that easy...
[20:03] <florian__> nbd: I don't think so, I met several people at #openfreebox that are crazy enough to do things like that :)
[20:03] <nbd> oh, right. freebox is bcm63xx as well
[20:04] <nbd> well, then motivate these people :)
[20:04] <nbd> just make sure they don't work on another fork called 'openfreebox' :)
[20:04] <florian__> nbd: that's what I am doing
[20:04] <nbd> we have enough forks already :)
[20:04] <florian__> nbd: I know that
[20:04] <nbd> :)
[20:04] <[mbm]> we need to put an end to those forks
[20:05] <[mbm]> make everyone work on openwrt
[20:05] <[mbm]> :)
[20:05] <nbd> definitely
[20:05] <[mbm]> (after all, they all borrow heavily from it :P)
[20:05] <florian__> I know the people behind openfreebox and I can ensure you they are what we call in french "pipo"
[20:05] <nbd> what's pipo?
[20:05] <florian__> they are talking about vaporware, just like multidesk OS
[20:06] <nbd> oh :)
[20:06] <florian__> I think pipo is a real concept
[20:06] <CIA-17> nico * r4156 /packages/libs/curl/ (. Config.in Makefile ipkg/): port curl to buildroot-ng
[20:06] <florian__> {Nico}: how can we translate the french expression "c'est pipo" ?
[20:06] <florian__> what's the status of the adm5120 port ?
[20:06] <florian__> what about rtl8181 too ?
[20:07] <florian__> I mean do we have contacts with those projects ?
[20:07] <[mbm]> I don't think anyone is maintaining those
[20:07] <florian__> ok, what about taking them ?
[20:08] <{Nico}> florian__: c'est pipo > it's bullshit
[20:08] <florian__> {Nico}: ok
[20:08] <Kaloz> florian__: 8186 and 8651 (8181 isn' interesting)
[20:08] <{Nico}> i have to go now, cu later
[20:09] <florian__> Kaloz: ok, I started building a kernel for 8651 with nonmips toolchain, have not tested it yet
[20:09] <florian__> see ya {Nico}
[20:12] <nbd> [mbm]: btw. i've thought about the diag stuff for looking for the reset button on all gpios - it might cause problems
[20:12] <nbd> [mbm]: because the way i see it, on adm6996 it might trigger the switch reset if we set the 'rc' pin to low
[20:12] <nbd> [mbm]: which would happen if we enable input on all gpio pins
[20:13] <[mbm]> who said anything about screwing up the reset button support?
[20:13] <nbd> you wanted to do it for unknown hardware
[20:13] <nbd> for failsafe
[20:14] <[mbm]> I never said "let's treat every gpio like a reset"
[20:14] <nbd> i think it was something like looking for gpio changes on any pin if we don't know the model
[20:14] <nbd> so that we catch reset that way
[20:14] <[mbm]> no.
[20:15] <nbd> then what was it?
[20:15] <nbd> maybe i misunderstood what you said
[20:15] <[mbm]> I said "can we use the gpio variables like gipio_reset nvram variables for boards that we don't know the gpio?"
[20:16] <nbd> but there was another suggestion, iirc
[20:16] <nbd> let me check the logs
[20:16] <[mbm]> might have been a mention of interrupt support
[20:17] <nbd> 22:46 < [mbm]> or some way of watching all the gpios for a change and assuming that the change means the reset was pressed
[20:18] <[mbm]> huh? that has to be out of context
[20:18] <BrainSlayer> @florian: did you saw the bogomips problem in the brcm63xx too?
[20:19] <nbd> [mbm]: you were discussing ideas on how diag should deal with the situation of not knowing the reset button
[20:19] <[mbm]> nbd: yeah, but I don't remember ever saying that or thinking that was a good idea
[20:19] <nbd> first you mentioned looking at the _reset variables
[20:19] <nbd> and then this as an alternative
[20:20] <[mbm]> right, that part I remember
[20:20] <nbd> i have it in my logs and i do remember it
[20:20] <[mbm]> *shrug* well, we both agree it's a bad idea
[20:20] <nbd> yeah
[20:21] <nbd> i'll add reset_gpio as fallback
[20:21] <[mbm]> if I remember correctly there were a few reports of platforms we needed to add to diag
[20:22] <nbd> hmm... just tested the asus wl-500g deluxe and reset_gpio is wrong there
[20:23] <nbd> reset_gpio is 7
[20:23] <nbd> the reset button is on 6
[20:23] <[mbm]> well, it should be a fallback
[20:23] <nbd> and what if that fallback is completely useless?
[20:24] <[mbm]> and what if we don't know and don't even attempt to guess the reset?
[20:24] <[mbm]> (is that somehow better?)
[20:24] <nbd> i'm still thinking whether there's a better way to do it
[20:25] <[mbm]> well, we should also consider the hotplug stuff we talked about
[21:36] <nbd> [mbm]: btw. i'm currently making the whiterussian sdk buildroot-ng compatible
[21:36] <[mbm]> ?
[21:36] <nbd> i want it to be able to build both the old and the new packages at the same time
[21:37] <[mbm]> ok, are you also making buildroot-ng backwards compatible?
[21:37] <nbd> so that we can build buildroot-ng packages for 'legacy' systems
[21:37] <nbd> not yet - might do that later
[21:37] <nbd> i think i'll do that for the sdk only
[21:38] <nbd> allowing people to integrate old packages into the new buildroot is probably not a good idea
[21:38] <[mbm]> true
[22:01] <CIA-17> nbd * r4157 /branches/whiterussian/openwrt/package/ (28 files in 6 dirs): add buildroot-ng support to the whiterussian sdk
[22:01] <nbd> :)
[22:09] <nshm> btw, do you know what's the status of midge merge into openwrt?
[22:11] <[mbm]> midge?
[22:14] <Kaloz> adm5120
[22:14] <Kaloz> brb
[22:14] <nshm> it's the port of openwrt on adm5120
[22:15] <nshm> latest state -> http://midge.vlad.org.ua/svn/trunk/openwrt-midge/
[22:15] <[mbm]> ah, interesting butI think there are more important items on the todo list
[22:16] <nshm> I see, I was just wondering if there are some troubles or it's just due to the lack of time
[22:16] <nshm> thanks [mbm]
[22:17] <CIA-17> nbd * r4158 /branches/buildroot-ng/openwrt/ (7 files in 4 dirs): make the buildroot-ng sdk compatible with the old packages
[22:17] <[mbm]> nshm: are you the midge maintainer?
[22:18] <nshm> no
[22:18] <nshm> sadly no :)
[22:34] <florian__> re
[22:35] <florian__> nshm: well actually as the
[22:35] <florian__> midge structure is the same as openwrt it is potentially easy to backport things from one repository to another
[22:37] <[mbm]> probably need someone to update midge to the latest openwrt release and then generate the patches
[22:37] <nbd> and then port everything to buildroot-ng
[22:38] <Kaloz> [mbm]: midge -> vladwrt
[22:38] <[mbm]> ah
[22:38] <[mbm]> him
[22:38] <Kaloz> nshm: the author was gone with "he will update his patches for the latest kamikaze" for months now
[22:38] <[mbm]> gone is the key word
[22:39] <florian__> lol
[22:39] <florian__> maybe we could integrate his work as well
[22:39] <florian__> I am really impressed with the increasing number of targets being supported
[22:39] <florian__> even experimentally
[22:39] <nbd> yeah, it's way cool
[22:39] <malbon> I'm looking at adm5120 stuff, just need to fix my jtag and I'll be working on it right away.
[22:40] <Kaloz> i would do 2.6, but i don't have any adm5120 board
[22:40] <florian__> the rtl8181 and adm5120 are similar according to linux-mips, in which things ? both are mips clones ?
[22:40] <malbon> they are, in the same way that bcm63xx is.
[22:41] <florian__> ok, but bcm63xx is able to run mips patented instructions, I am not sure for rtl8181
[22:41] <nbd> 8181 is non-mips
[22:41] <[mbm]> florian__: yeah there are a large number of new platforms, but I'm not sure that adding them is a good idea if nobody is going to be supporting them
[22:41] <malbon> adm5120 can too.
[22:42] <florian__> ok, adm5120 uses a lexra core right ?
[22:43] <malbon> hmm no I didn't think so.
[22:44] <florian__> ok I thought it was one
[22:44] <malbon> nope, mips 4Kc core
[22:45] <florian__> the adm5120 patches contains a lot of "Unpublished" references, is this GPL ?
[22:45] <malbon> florian__: did you get your bcm63xx flash driver working with redboot?
[22:45] <nshm> florian__: that's probably is the main trouble
[22:45] <malbon> I was looking at the one Kaloz pointed me at, the 2.6.12 patch.
[22:46] <nshm> florian__: although I wonder how kernel patches can have another license
[22:46] <florian__> malbon: I did not do any work since I sent you the fis list
[22:47] <malbon> florian__: ah, ok, well I didn't see the list, it must have got lost somewhere in my client.
[22:47] <malbon> florian__: I'm not used to irssi yet.
[22:47] <malbon> florian__: can you send it again please?
[22:47] <florian__> malbon: may I query it ?
[22:48] <florian__> malbon: by the way I am not (still) registered
[22:48] <malbon> florian__: give it a go, see if it comes though this time
[22:49] <[mbm]> florian__: if it's compiled into the kernel, that should make the file gpl
[22:52] <florian__> [mbm]: even with the Unpublished thing ?
[22:52] <florian__> that's good to know
[22:52] <florian__> everything that statically goes into the kernel is gpl right ?
[22:52] <[mbm]> should be, but I don't think anyone has ever tested that
[22:53] <florian__> seems quite hard
[22:54] <florian__> malbon: is is ok for the fis list output ??
[22:54] <malbon> florian__: no :( I didn't recieve it.
[22:55] <florian__> ah ok
[22:55] <florian__> let me pastebin it
[22:55] <malbon> ok,
[22:55] <malbon> why don't you register?
[22:56] <florian__> hum because I don't take the time to find a channel operator and explain him I want this pseudo :)
[22:57] <[mbm]> /m nickserv help register
[22:57] <[mbm]> :P
[22:57] <florian__> [mbm]: you know me :) I have to be repeated 6 or 7 times before really understanding :)
[22:57] <[mbm]> what's with the __'s now anyways?
[22:58] <florian__> florian is already taken
[22:59] <florian__> malbon: http://pastebin.ca/91716
[23:00] <malbon> florian__: bec00000 is the end of flash?
[23:02] <[mbm]> should be
[23:03] <[mbm]> 8M flash starting at be400000 would end at bec00000
[23:03] <[mbm]> which is what the pastebin seems to show
[23:03] <malbon> yeah, just confirming more than anything else.
[23:03] <malbon> that's the normal position for the fis directory then.
[23:04] <malbon> end - 1 blocks
[23:04] <malbon> I'll take a look at the code.
[23:13] <malbon> florian__: if you look at another mtd map, then you'll see they all call:
[23:14] <malbon> parse_redboot_partitions(struct mtd_info *master, struct mtd_partition **pparts);
[23:15] <malbon> that's not called anywhere in your init_brcm_physmap()
[23:25] <nbd> [mbm]: btw. currently verbose.mk accepts V= only from the command line. i'd like to change that so that it also works from the environment, ok?
[23:25] <[mbm]> reason?
[23:26] <nbd> i like to do export V=99 so that i always get verbose output in that shell
[23:26] <[mbm]> ok
[23:26] <CIA-17> nbd * r4159 /branches/buildroot-ng/openwrt/include/verbose.mk: accept V= from the environment as well
[23:27] <[mbm]> btw the naming of the lzma-loader variables sucks
[23:27] <[mbm]> KERNEL_ENTRY is not the kernel entry point
[23:29] <CIA-17> nbd * r4160 /branches/buildroot-ng/openwrt/include/verbose.mk: fix KBUILD_VERBOSE if V is unset
[00:00] --- Wed Jul 19 2006
[14:09] <CIA-17> nico * r4136 /packages/libs/ncurses/ (. Config.in Makefile ipkg/): port ncurses to buildroot-ng.
[14:20] <CIA-17> nico * r4137 /packages/libs/readline/ (. Config.in Makefile ipkg/): port readline to buildroot-ng, update it to v5.1
[14:47] <CIA-17> nico * r4138 /packages/libs/libtool/ (. Config.in Makefile ipkg/): port libtool to buildroot-ng
[15:01] <CIA-17> nico * r4139 /packages/libs/libnet/ (. Config.in Makefile ipkg/): port libnet to buildroot-ng
[15:22] <CIA-17> nico * r4140 /packages/libs/lzo/ (. Config.in Makefile ipkg/): port lzo to buildroot-ng
[15:39] <CIA-17> nico * r4141 /packages/libs/pcre/ (. Config.in Makefile ipkg/): port pcre to buildroot-ng
[16:37] <florian__> {Nico}: is on a comitting spree ;)
[16:42] <BrainSlayer> florian?
[16:56] <CIA-17> nico * r4142 /packages/libs/mysql/ (. Config.in Makefile ipkg/): port mysql (client library) to buildroot-ng
[17:02] <{Nico}> florian__: right, i'm off today and it's way too hot here to do anything else
[17:06] <{Nico}> is anyone else porting packages to buildroot-ng?
[17:07] <nbd> i'm working on package/kernel right now
[17:07] <nbd> (moving the kernel module stuff there so that it reuses the existing packaging code)
[17:27] <CIA-17> nico * r4143 /packages/libs/postgresql/ (. Config.in Makefile ipkg/): port postgresql (client library and cli) to buildroot-ng
[17:31] <CIA-17> nbd * r4144 /branches/buildroot-ng/openwrt/include/package.mk: check if the Package/$(NAME) template is defined before packaging or installing ipkg packages
[17:34] <{Nico}> nbd: i saw you changed ./scripts/rstrip.sh to also sstrip already stripped binaries
[17:34] <{Nico}> iirc, it make unusable binaries
[17:34] <nbd> yes
[17:34] <nbd> worked for me
[17:34] <nbd> didn't see any unusable binaries
[17:35] <nbd> do you have any unusable binaries in your build?
[17:36] <{Nico}> nbd: i have to check (and remember what my test case was)
[17:41] <{Nico}> nbd: how to you add compile-time dependencies to DEPENDS?
[17:42] <{Nico}> ie. my package needs openssl to build but the resulting binaries don't need it at runtime
[17:43] <nbd> PKG_BUILDDEP
[17:43] <nbd> but the dependency system needs to be changed
[17:43] <nbd> it's not flexible enough for hostapd
[17:43] <nbd> because there's hostapd and hostapd-mini
[17:43] <nbd> and mini has no build dependency on openssl
[17:45] <{Nico}> can't we move PKG_BUILDEP somewhere in 'define Package/foo' template?
[17:58] <nbd> yeah, i was planning on doing that
[17:59] <nbd> but the script that generates .pkgdep needs to be changed as well
[17:59] <nbd> so that the config is checked for the package dependencies
[18:09] <CIA-17> nico * r4145 /packages/libs/cyrus-sasl/ (. Config.in Makefile ipkg/): port cyrus-sasl to buildroot-ng
[18:17] <CIA-17> nbd * r4146 /branches/buildroot-ng/openwrt/ (5 files in 2 dirs): move some of the kernel module packaging to package/kernel/
[18:18] <nbd> {Nico}: if you have time. could you review package/kernel/Makefile and tell me if you think it's suitable for all modules?
[18:19] <CIA-17> nico * r4147 /packages/libs/openldap/ (. Config.in Makefile ipkg/): port openldap (client libraries and utilities) to buildroot-ng
[18:22] <CIA-17> nbd * r4148 /branches/buildroot-ng/openwrt/package/kernel/Makefile: allow multiple kernel config dependencies
[18:35] <CIA-17> nbd * r4149 /branches/buildroot-ng/openwrt/ (5 files in 3 dirs): move the cryptoapi kmod to package/kernel
[18:36] <{Nico}> nbd: ok
[18:43] <nbd> that should make dependency handling a lot easier
[18:44] <CIA-17> nico * r4150 /packages/libs/libevent/ (. Config.in Makefile ipkg/): port libevent to buildroot-ng
[18:49] <nbd> i'm backporting the jffs2 patches to wr right now
[18:52] <{Nico}> during the last meeting, we talk about adding missing nvram setting for wireless
[18:53] <[mbm]> for webif, yes
[18:53] <nbd> which ines?
[18:53] <nbd> s/ines/ones/
[18:53] <[mbm]> webif really should show every nvram variable openwrt uses
[18:54] <{Nico}> [mbm]: ok, i'll try to work on that
[18:54] <CIA-17> nbd * r4151 /branches/buildroot-ng/openwrt/target/image/brcm/Makefile: disable wgt634u images for linux 2.4 (only works with 2.6)
[18:55] <{Nico}> nbd: broadcom-wl in buildroot-ng is missing the libbcmcrypto.so (for the nas package)
[18:55] <nbd> is it missing in the tarball?
[18:55] <nbd> because the nas package references it
[18:56] <{Nico}> nbd: i guess it's missing in the archive
[18:57] <nbd> i just built it
[18:57] <nbd> it's there
[18:58] <CIA-17> nbd * r4152 /branches/buildroot-ng/openwrt/package/broadcom-wl/Makefile: change depends/default for nas
[18:58] <CIA-17> nico * r4153 /packages/libs/gmp/ (. Config.in Makefile ipkg/): port gmp to buildroot-ng
[18:59] Action: nbd tests the flash map and jffs2 backport
[19:08] <nbd> damn... where did i put my serial cable...
[19:09] <[mbm]> jffs2 patch didn't work? :P
[19:18] <nbd> [mbm]: i don't remember the ip of the device...
[19:26] <CIA-17> nico * r4154 /packages/libs/expat/ (. Config.in Makefile ipkg/): port expat to buildroot-ng
[19:41] <florian__> BrainSlayer: what ?
[19:44] <florian__> Kaloz: ping ?
[19:44] <{Nico}> florian__: i couldn't make it for rmll, how was it?
[19:44] <florian__> {Nico}: it was great
[19:45] <florian__> {Nico}: did you read the presentation ?
[19:45] <{Nico}> florian__: did anyone make a movie of your presentation?
[19:45] <florian__> {Nico}: unfortunately no
[19:45] <{Nico}> florian__: yep, good work
[19:46] <florian__> {Nico}: thanks, I was not sure how to make a good plan and I did not wanted to make it too technical
[19:46] <{Nico}> how many people attended?
[19:46] <florian__> between 40 and 50 it was a small room
[19:46] <florian__> brb
[19:50] <Kaloz> florian__: pong
[19:50] <nbd> finally found my cable
[19:51] <BrainSlayer> hey
[19:51] <BrainSlayer> @florian: do you know why the flash mapping for the brcm63xx port wont work?
[19:53] <florian__> BrainSlayer: sure, because the crappy flash map does only manage cfe, and is lost when the bootloader is redboot
[19:53] <BrainSlayer> i'm using still the cfe
[19:54] <BrainSlayer> if i would use redboot it would be easier
[19:54] <florian__> ah then I don't know, I did not tested it
[19:54] <florian__> I only managed to have a redboot device working
[19:54] <BrainSlayer> do you have a working redboot image for this these units?
[19:54] <florian__> yes
[19:54] <CIA-17> nbd * r4155 /branches/whiterussian/openwrt/ (4 files in 4 dirs): backport jffs2 changes and flash map from buildroot-ng
[19:54] <nbd> cool. worked right on first try :)
[19:54] <BrainSlayer> can you provide me this image?
[19:54] <[mbm]> you sound surprised
[19:55] <florian__> Kaloz: sent me his livebox bootloader which was not protected like mine
[19:55] <florian__> BrainSlayer: let me find it
[19:55] <BrainSlayer> okay
[19:55] <nbd> [mbm]: well, the flash map was a bit different on wr. i just manually edited the diff without rediffing afterwards
[19:55] <[mbm]> ah
[19:55] <florian__> Kaloz: do you remember what was the things that needed to be changed in the brcm63xx patches ?
[19:55] <BrainSlayer> pci-fixups
[19:55] <BrainSlayer> i already have done this
[19:55] <Kaloz> florian__: nope, but i will look at those as well just my last days were pretty fsck'ed up
[19:56] <Kaloz> florian__: it's too hot here, and i can hardly sleep..
[19:56] <nbd> yeah, i have trouble with the temperature as well
[19:56] <BrainSlayer> the same here. 36 degrees today
[19:57] <florian__> Kaloz: lol ok, I just wanted to commit the 2.6.17 patches and clean them up
[19:58] <BrainSlayer> @florian: i'm currently cheating with a vendor to get the broadcom adsl sources
[19:58] <florian__> someone here has already been working with cisco catalyst 3550 ? or similar ?
[19:58] <florian__> BrainSlayer: ok great
[19:58] <Kaloz> florian__: will look at those later
[19:58] <florian__> Kaloz: ok no pb
[19:59] <crazy_imp> florian__: openwrt on a cisco switch? cool :D
[20:00] <florian__> crazy_imp: no, I am having serious troubles with a catalyst 3550, it's the core layer3 switch of my school
[20:00] <florian__> it has been freezing with no reasons at all the last few days
[20:00] <florian__> althought the temperatur seems not to have changed in the room
[20:01] <nbd> so. what's missing for rc6?
[20:01] <[mbm]> I gave you a list
[20:02] <florian__> BrainSlayer: anyway the brcm63xx driver seems to be crappy as hell, so it would either be better to start reversing it
[20:02] <nbd> florian__: i don't think we can find enough people to do that
[20:03] <nbd> florian__: and reversing device drivers is not that easy...
[20:03] <florian__> nbd: I don't think so, I met several people at #openfreebox that are crazy enough to do things like that :)
[20:03] <nbd> oh, right. freebox is bcm63xx as well
[20:04] <nbd> well, then motivate these people :)
[20:04] <nbd> just make sure they don't work on another fork called 'openfreebox' :)
[20:04] <florian__> nbd: that's what I am doing
[20:04] <nbd> we have enough forks already :)
[20:04] <florian__> nbd: I know that
[20:04] <nbd> :)
[20:04] <[mbm]> we need to put an end to those forks
[20:05] <[mbm]> make everyone work on openwrt
[20:05] <[mbm]> :)
[20:05] <nbd> definitely
[20:05] <[mbm]> (after all, they all borrow heavily from it :P)
[20:05] <florian__> I know the people behind openfreebox and I can ensure you they are what we call in french "pipo"
[20:05] <nbd> what's pipo?
[20:05] <florian__> they are talking about vaporware, just like multidesk OS
[20:06] <nbd> oh :)
[20:06] <florian__> I think pipo is a real concept
[20:06] <CIA-17> nico * r4156 /packages/libs/curl/ (. Config.in Makefile ipkg/): port curl to buildroot-ng
[20:06] <florian__> {Nico}: how can we translate the french expression "c'est pipo" ?
[20:06] <florian__> what's the status of the adm5120 port ?
[20:06] <florian__> what about rtl8181 too ?
[20:07] <florian__> I mean do we have contacts with those projects ?
[20:07] <[mbm]> I don't think anyone is maintaining those
[20:07] <florian__> ok, what about taking them ?
[20:08] <{Nico}> florian__: c'est pipo > it's bullshit
[20:08] <florian__> {Nico}: ok
[20:08] <Kaloz> florian__: 8186 and 8651 (8181 isn' interesting)
[20:08] <{Nico}> i have to go now, cu later
[20:09] <florian__> Kaloz: ok, I started building a kernel for 8651 with nonmips toolchain, have not tested it yet
[20:09] <florian__> see ya {Nico}
[20:12] <nbd> [mbm]: btw. i've thought about the diag stuff for looking for the reset button on all gpios - it might cause problems
[20:12] <nbd> [mbm]: because the way i see it, on adm6996 it might trigger the switch reset if we set the 'rc' pin to low
[20:12] <nbd> [mbm]: which would happen if we enable input on all gpio pins
[20:13] <[mbm]> who said anything about screwing up the reset button support?
[20:13] <nbd> you wanted to do it for unknown hardware
[20:13] <nbd> for failsafe
[20:14] <[mbm]> I never said "let's treat every gpio like a reset"
[20:14] <nbd> i think it was something like looking for gpio changes on any pin if we don't know the model
[20:14] <nbd> so that we catch reset that way
[20:14] <[mbm]> no.
[20:15] <nbd> then what was it?
[20:15] <nbd> maybe i misunderstood what you said
[20:15] <[mbm]> I said "can we use the gpio variables like gipio_reset nvram variables for boards that we don't know the gpio?"
[20:16] <nbd> but there was another suggestion, iirc
[20:16] <nbd> let me check the logs
[20:16] <[mbm]> might have been a mention of interrupt support
[20:17] <nbd> 22:46 < [mbm]> or some way of watching all the gpios for a change and assuming that the change means the reset was pressed
[20:18] <[mbm]> huh? that has to be out of context
[20:18] <BrainSlayer> @florian: did you saw the bogomips problem in the brcm63xx too?
[20:19] <nbd> [mbm]: you were discussing ideas on how diag should deal with the situation of not knowing the reset button
[20:19] <[mbm]> nbd: yeah, but I don't remember ever saying that or thinking that was a good idea
[20:19] <nbd> first you mentioned looking at the _reset variables
[20:19] <nbd> and then this as an alternative
[20:20] <[mbm]> right, that part I remember
[20:20] <nbd> i have it in my logs and i do remember it
[20:20] <[mbm]> *shrug* well, we both agree it's a bad idea
[20:20] <nbd> yeah
[20:21] <nbd> i'll add reset_gpio as fallback
[20:21] <[mbm]> if I remember correctly there were a few reports of platforms we needed to add to diag
[20:22] <nbd> hmm... just tested the asus wl-500g deluxe and reset_gpio is wrong there
[20:23] <nbd> reset_gpio is 7
[20:23] <nbd> the reset button is on 6
[20:23] <[mbm]> well, it should be a fallback
[20:23] <nbd> and what if that fallback is completely useless?
[20:24] <[mbm]> and what if we don't know and don't even attempt to guess the reset?
[20:24] <[mbm]> (is that somehow better?)
[20:24] <nbd> i'm still thinking whether there's a better way to do it
[20:25] <[mbm]> well, we should also consider the hotplug stuff we talked about
[21:36] <nbd> [mbm]: btw. i'm currently making the whiterussian sdk buildroot-ng compatible
[21:36] <[mbm]> ?
[21:36] <nbd> i want it to be able to build both the old and the new packages at the same time
[21:37] <[mbm]> ok, are you also making buildroot-ng backwards compatible?
[21:37] <nbd> so that we can build buildroot-ng packages for 'legacy' systems
[21:37] <nbd> not yet - might do that later
[21:37] <nbd> i think i'll do that for the sdk only
[21:38] <nbd> allowing people to integrate old packages into the new buildroot is probably not a good idea
[21:38] <[mbm]> true
[22:01] <CIA-17> nbd * r4157 /branches/whiterussian/openwrt/package/ (28 files in 6 dirs): add buildroot-ng support to the whiterussian sdk
[22:01] <nbd> :)
[22:09] <nshm> btw, do you know what's the status of midge merge into openwrt?
[22:11] <[mbm]> midge?
[22:14] <Kaloz> adm5120
[22:14] <Kaloz> brb
[22:14] <nshm> it's the port of openwrt on adm5120
[22:15] <nshm> latest state -> http://midge.vlad.org.ua/svn/trunk/openwrt-midge/
[22:15] <[mbm]> ah, interesting butI think there are more important items on the todo list
[22:16] <nshm> I see, I was just wondering if there are some troubles or it's just due to the lack of time
[22:16] <nshm> thanks [mbm]
[22:17] <CIA-17> nbd * r4158 /branches/buildroot-ng/openwrt/ (7 files in 4 dirs): make the buildroot-ng sdk compatible with the old packages
[22:17] <[mbm]> nshm: are you the midge maintainer?
[22:18] <nshm> no
[22:18] <nshm> sadly no :)
[22:34] <florian__> re
[22:35] <florian__> nshm: well actually as the
[22:35] <florian__> midge structure is the same as openwrt it is potentially easy to backport things from one repository to another
[22:37] <[mbm]> probably need someone to update midge to the latest openwrt release and then generate the patches
[22:37] <nbd> and then port everything to buildroot-ng
[22:38] <Kaloz> [mbm]: midge -> vladwrt
[22:38] <[mbm]> ah
[22:38] <[mbm]> him
[22:38] <Kaloz> nshm: the author was gone with "he will update his patches for the latest kamikaze" for months now
[22:38] <[mbm]> gone is the key word
[22:39] <florian__> lol
[22:39] <florian__> maybe we could integrate his work as well
[22:39] <florian__> I am really impressed with the increasing number of targets being supported
[22:39] <florian__> even experimentally
[22:39] <nbd> yeah, it's way cool
[22:39] <malbon> I'm looking at adm5120 stuff, just need to fix my jtag and I'll be working on it right away.
[22:40] <Kaloz> i would do 2.6, but i don't have any adm5120 board
[22:40] <florian__> the rtl8181 and adm5120 are similar according to linux-mips, in which things ? both are mips clones ?
[22:40] <malbon> they are, in the same way that bcm63xx is.
[22:41] <florian__> ok, but bcm63xx is able to run mips patented instructions, I am not sure for rtl8181
[22:41] <nbd> 8181 is non-mips
[22:41] <[mbm]> florian__: yeah there are a large number of new platforms, but I'm not sure that adding them is a good idea if nobody is going to be supporting them
[22:41] <malbon> adm5120 can too.
[22:42] <florian__> ok, adm5120 uses a lexra core right ?
[22:43] <malbon> hmm no I didn't think so.
[22:44] <florian__> ok I thought it was one
[22:44] <malbon> nope, mips 4Kc core
[22:45] <florian__> the adm5120 patches contains a lot of "Unpublished" references, is this GPL ?
[22:45] <malbon> florian__: did you get your bcm63xx flash driver working with redboot?
[22:45] <nshm> florian__: that's probably is the main trouble
[22:45] <malbon> I was looking at the one Kaloz pointed me at, the 2.6.12 patch.
[22:46] <nshm> florian__: although I wonder how kernel patches can have another license
[22:46] <florian__> malbon: I did not do any work since I sent you the fis list
[22:47] <malbon> florian__: ah, ok, well I didn't see the list, it must have got lost somewhere in my client.
[22:47] <malbon> florian__: I'm not used to irssi yet.
[22:47] <malbon> florian__: can you send it again please?
[22:47] <florian__> malbon: may I query it ?
[22:48] <florian__> malbon: by the way I am not (still) registered
[22:48] <malbon> florian__: give it a go, see if it comes though this time
[22:49] <[mbm]> florian__: if it's compiled into the kernel, that should make the file gpl
[22:52] <florian__> [mbm]: even with the Unpublished thing ?
[22:52] <florian__> that's good to know
[22:52] <florian__> everything that statically goes into the kernel is gpl right ?
[22:52] <[mbm]> should be, but I don't think anyone has ever tested that
[22:53] <florian__> seems quite hard
[22:54] <florian__> malbon: is is ok for the fis list output ??
[22:54] <malbon> florian__: no :( I didn't recieve it.
[22:55] <florian__> ah ok
[22:55] <florian__> let me pastebin it
[22:55] <malbon> ok,
[22:55] <malbon> why don't you register?
[22:56] <florian__> hum because I don't take the time to find a channel operator and explain him I want this pseudo :)
[22:57] <[mbm]> /m nickserv help register
[22:57] <[mbm]> :P
[22:57] <florian__> [mbm]: you know me :) I have to be repeated 6 or 7 times before really understanding :)
[22:57] <[mbm]> what's with the __'s now anyways?
[22:58] <florian__> florian is already taken
[22:59] <florian__> malbon: http://pastebin.ca/91716
[23:00] <malbon> florian__: bec00000 is the end of flash?
[23:02] <[mbm]> should be
[23:03] <[mbm]> 8M flash starting at be400000 would end at bec00000
[23:03] <[mbm]> which is what the pastebin seems to show
[23:03] <malbon> yeah, just confirming more than anything else.
[23:03] <malbon> that's the normal position for the fis directory then.
[23:04] <malbon> end - 1 blocks
[23:04] <malbon> I'll take a look at the code.
[23:13] <malbon> florian__: if you look at another mtd map, then you'll see they all call:
[23:14] <malbon> parse_redboot_partitions(struct mtd_info *master, struct mtd_partition **pparts);
[23:15] <malbon> that's not called anywhere in your init_brcm_physmap()
[23:25] <nbd> [mbm]: btw. currently verbose.mk accepts V= only from the command line. i'd like to change that so that it also works from the environment, ok?
[23:25] <[mbm]> reason?
[23:26] <nbd> i like to do export V=99 so that i always get verbose output in that shell
[23:26] <[mbm]> ok
[23:26] <CIA-17> nbd * r4159 /branches/buildroot-ng/openwrt/include/verbose.mk: accept V= from the environment as well
[23:27] <[mbm]> btw the naming of the lzma-loader variables sucks
[23:27] <[mbm]> KERNEL_ENTRY is not the kernel entry point
[23:29] <CIA-17> nbd * r4160 /branches/buildroot-ng/openwrt/include/verbose.mk: fix KBUILD_VERBOSE if V is unset
[00:00] --- Wed Jul 19 2006