[00:08] <florian__> nbd: what else can maje a probe fail ?
[00:08] <nbd> what does the error look like now
[00:08] <nbd> ?
[00:13] <florian__> nbd: http://pastebin.ca/109171
[00:13] <florian__> with the correct size
[00:13] <florian__> but in every mtd map driver I see, there are rootfs things ? what for ?
[00:14] <florian__> and they use the kernel addr
[00:16] <nbd> florian__: try 1e400000 instead of be400000
[00:17] <nbd> because i think you need the physical address and not the virtual one
[00:18] <florian__> ok
[00:18] <florian__> same value for the size, right ?
[00:18] <nbd> yes
[00:19] <florian__> I am such a newbie, that's sad :)
[00:19] <nbd> it's not sad
[00:20] <florian__> ah ah !
[00:20] <nbd> you can't expect to be an expert right from the start
[00:20] <nbd> does it work now?
[00:21] <florian__> yep that's interesting
[00:21] <florian__> I will pastebin the result
[00:21] <florian__> http://pastebin.ca/109176
[00:22] <florian__> I think the redboot FIS directory must be wrong
[00:23] <nbd> the offset in parse_redboot_partitions needs fixing
[00:23] <nbd> i mean the one that you supply to it
[00:24] <florian__> yep, in the inventel driver it's 0, so I will try with 0
[00:24] <nbd> if it doesn't work, use something relative to the remapped virtual address
[00:24] <florian__> ok
[00:26] <florian__> how did you find out it was 1 instead of b ?
[00:26] <nbd> but it's great that we can use the generic cfi stuff instead of broadcom's crap
[00:26] <florian__> yes !
[00:26] <nbd> florian__: virtual address = physical address | 0x80000000
[00:27] <florian__> ah of course
[00:27] <florian__> nbd: don't be too happy, it is quite easy with RedBoot ;) it will need fixups for CFE
[00:28] <nbd> sure it will, but that won't affect the physical interface to the flash
[00:28] <florian__> right
[00:28] <florian__> well, looks like the remapped virtual address of the fis directory does not seem to be the right value
[00:29] <nbd> what offset did you add to it?
[00:29] <florian__> fis list gives me FIS directory at 0xBEBFF000 so I tried 0x1EBFF000
[00:30] <nbd> that's the physical address
[00:30] <nbd> you need to do it differently. let me check...
[00:34] <nbd> use (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x1000)
[00:34] <florian__> can I ask why :) ? why .
[00:35] <nbd> if you look at the offset that redboot tells you, you should notice that it start 0x1000 before the end of the flash
[00:35] <florian__> right
[00:35] <nbd> the flash size is stored in bcm963xx_mtd_info->size after the probe
[00:35] <nbd> and you need to supply a memory address
[00:36] <florian__> ok, I think I see
[00:36] <nbd> bcm963xx_map.virt is the address that's accessible from the kernel space, because it's the remapped one
[00:36] <florian__> yes
[00:36] <florian__> this does not seem to be the right address anyway
[00:38] <nbd> seems like that parameter is not supposed to carry an address
[00:38] <nbd> (looking at the redboot.c code right now)
[00:39] <nbd> use 0
[00:39] <florian__> I did actually
[00:39] <florian__> let me check ;)
[00:41] <florian__> it attemtps to detect 21 MTD partitions :)
[00:42] <nbd> i see the error
[00:42] <nbd> redboot.c is not prepared to look for the fis directory at such an 'odd' offset
[00:44] <florian__> so we know it is not : 0, physicall address, remapped address
[00:45] <nbd> http://downloads.openwrt.org/people/nbd/tmp/redboot.patch
[00:45] <nbd> use this
[00:45] <nbd> and then use (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x1000) as last parameter to the function
[00:45] <nbd> no
[00:45] <nbd> wait
[00:45] <nbd> bug in the patch
[00:46] <nbd> fixed now
[00:46] <florian__> ok
[00:50] <common> http://flickr.com/photos/24017027@N00/199751092/
[00:50] <florian__> so 0, and (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x1000) as last parameters
[00:50] <nbd> no, replace the 0
[00:50] <nbd> the patch doesn't add a new parameter
[00:51] <florian__> oops sorry :)
[00:53] <florian__> Searching for RedBoot partition table in bcm963xx at offset 0xbebff000
[00:53] <florian__> mtd: Giving out device 0 to bcm963xx
[00:54] <florian__> hu wait
[00:56] <florian__> no that's all
[00:56] <florian__> looks like it does not find the fis directory or cannot parse it
[00:59] <nbd> show me the full fis list
[01:00] <nbd> you told me the wrong fis directory offset
[01:00] <florian__> in query
[01:00] <florian__> ah, really ?
[01:00] <nbd> use (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x10000)
[01:01] <florian__> with you patch ?
[01:01] <nbd> but it will probably have the same effect as no patch and parameter 0
[01:01] <florian__> ok
[01:01] <nbd> but please try it anyway
[01:01] <nbd> yes, with my patch
[01:03] <florian__> Searching for RedBoot partition table in bcm963xx at offset 0xbebf0000
[01:03] <florian__> Creating 0 MTD partitions on "bcm963xx":
[01:04] <florian__> so it searches the partition at the right place
[01:04] <nbd> make a paste of your current source
[01:05] <florian__> http://pastebin.ca/109250
[01:07] <nbd> try to set the last parameter to 0 again
[01:08] <florian__> ok, with the redboot patch too
[01:08] <nbd> shouldn't make a difference
[01:10] <florian__> Searching for RedBoot partition table in bcm963xx at offset 0x7f0000
[01:10] <florian__> and it tries to detect 21 MTD partitions
[01:13] <florian__> if you have another idea, I take it, after I go sleeping
[01:13] <florian__> thank you very much for your help !
[01:15] <nbd> did your device come with redboot or did you reflash to redboot?
[01:15] <florian__> it came with redboot
[01:15] <florian__> and I reflashed with Kaloz's redboot coming from his livebox as well, so probably almost the same redboot
[01:15] <florian__> because mine was locked up
[01:15] <nbd> k
[01:17] <florian__> at the end of the kmsg I see that : mtd: Giving out device 14 to jffs_system
[01:17] <florian__> 0x003f0000-0x003ff000 : "FIS directory"
[01:17] <florian__> does it help ?
[01:20] <nbd> show me the full kmesg
[01:21] <florian__> http://pastebin.ca/109269
[01:24] <nbd> florian__: http://downloads.openwrt.org/people/nbd/tmp/redboot.patch
[01:34] <florian__> http://pastebin.ca/109280
[01:36] <florian__> I go sleeping, if you have a fix, I will test it tomorrow around 6am, and let you know :)
[01:37] <nbd> i think we almost have it
[01:38] <nbd> so if you want to stay up for a few minutes, i'll try to get it fixed :)
[01:38] <nbd> otherwise: good night :)
[01:45] <nbd> florian__: http://downloads.openwrt.org/people/nbd/tmp/redboot.c
[02:45] <nbd> hey common
[02:45] <common> friend of mine screw'd soldering a 2nd minipci slot on his wrv54g, the 2nd slot does not work, are you intrested in the device?, it works fine as long as you dont try to use the 2nd slot
[02:45] <nbd> sure
[02:45] <common> i'll ask him
[02:46] <common> he is pretty pissed off about it, as he solderd mine first, and my 2nd slot works ...
[02:50] <common> the bug i found in broadcoms upnp daemon will be disclosed without vendor patches, the assholes did not reply any mail
[02:51] <common> sometimes its really sick
[02:51] <common> you offer assistance fixing the shit they sell to make hard cash, and they just give a fuck about it
[07:05] <florian__> nbd: here is the output with your patch : http://pastebin.ca/109558
[08:49] <florian__> re
[09:00] <CIA-17> florian * r4359 /packages/net/click/ (. Config.in Makefile ipkg/): Port click to -ng
[09:51] <CIA-17> nico * r4360 /packages/utils/mpd/ (. Config.in Makefile ipkg/): port mpd to buildroot-ng
[09:54] <{Nico}> do we really need httpd address binding patch in bb?
[10:18] <florian__> {Nico}: it was a feature asked by a user, so I guess it's used at least byhim
[13:13] <nbd> florian_1: good morning
[13:13] <nbd> florian_1: i have another redboot.c for you
[13:13] <nbd> florian_1: same location: http://downloads.openwrt.org/people/nbd/tmp/redboot.c
[13:17] <florian_1> damn, this fscking school server had so many problems this morning, I could not do anything with it :)
[13:17] <nbd> no problem
[13:18] <florian_> I will have a try at you redboot tonight
[13:19] <florian_> anyway, looks quite strange to me that we have to modify redboot in order to get the partitions mounted, should not it rather be the flash map driver ?
[13:20] <nbd> the redboot you're using seems to do things slightly different
[13:20] <nbd> the changes that i'm doing might also work on other platforms
[13:21] <nbd> the only way to avoid having to change redboot.c would be duplicating it
[13:21] <florian_> ok, in fact, it is the stock redboot from linux-2.6.17
[13:24] <CIA-17> florian * r4361 /packages/libs/clinkc/ (. Config.in Makefile ipkg/): Port clinkc to -ng
[13:27] <florian_> nbd: is there a reference package for boolean options in Config.in ?
[13:27] <nbd> what do you mean?
[13:27] <florian_> https://dev.openwrt.org/browser/trunk/openwrt/package/collectd/Config.in
[13:31] <{Nico}> i think we should drop boolean option in packages
[13:31] <florian_> ok, how do you want it to be done ?
[13:32] <{Nico}> just enable "reasonable defaults" from the boolean options
[13:32] <{Nico}> look at mpd, it used to have boolean options for codec support, i enabled all codecs
[13:33] <florian_> {Nico}: ok, thanks, I will have a look
[13:34] <{Nico}> we might add that later back and build -custom packages where every options would be configurable
[13:34] <florian_> {Nico}: I don't think these are booleans like the Config.in I pointed, it's a bit different
[13:34] <florian_> here you do "configure time" plugin detection
[13:37] <{Nico}> with this, we could have packages with the same name built with different options and i find it bad
[13:38] <{Nico}> just do as if all options were selected
[13:38] <florian_> {Nico}: and how do I disable it through menuconfig ?
[13:38] <florian_> by the way, this will be required for OpenVPN too
[13:41] <{Nico}> florian_: forget menuconfig options
[13:41] <{Nico}> just do enable-{cpu,load..etc} in ./configure
[13:46] <florian_> oki
[13:47] <florian_> anyway, you are right, if well done, the configure script should enable on the fly the options
[13:47] <florian_> if it relies on system library, but it is not always the case
[13:48] <{Nico}> nbd: what do you think?
[13:49] <nbd> no idea how to get this right
[13:53] <CIA-17> florian * r4362 /packages/utils/collectd/ (. Config.in Makefile Makefile files/ files/ ipkg/): Port collectd to -ng
[13:53] <{Nico}> what do you think of -custom packages idea (like in pmacct and snort) that could be enabled only when CONFIG_DEVEL=y
[13:53] <florian_> I think it's a good idea
[13:55] <{Nico}> florian_: you should use "install -m0755" for init-scripts to always have correct permissions
[13:55] <florian_> ok
[13:55] <florian_> I fix it
[13:56] <CIA-17> florian * r4363 /packages/utils/collectd/Makefile: Fix init script install permission
[13:59] <CIA-17> florian * r4364 /packages/net/ctorrent/ (. Config.in Makefile ipkg/): Port ctorrent to -ng
[14:01] <florian_> is not PKG_BUILDDEP and DEPENDS:= redundant ?
[14:03] <{Nico}> PKG_BUILDEP is for compile-time dependencies, while DEPENDS will end up in the ipkg Depends: control field
[14:04] <florian_> at the beginning I did not fill PKG_BUILDDEP and compile time dependencies were solved too
[14:04] <{Nico}> DEPENDS will also feed compile-time dependencies
[14:05] <florian_> ok
[14:05] <florian_> it is just in the case the package name is different from the library name like mysql and libmysqlclient for instance ?
[14:06] <{Nico}> PKG_BUILDEPS should be used when you need another package to be compiled before
[14:06] <{Nico}> but you don't need any of its libs at runtime
[14:06] <{Nico}> for your package
[14:06] <florian_> ok
[14:07] <florian_> I have to write documentation about that
[14:07] <florian_> {Nico}: by the way did you test the au1000 port ?
[14:07] <{Nico}> not yet
[14:08] <florian_> maybe you did not had time, and I understand, I just want if you take 5 min to flash :) to tell me if you get atheros working "out of the box"
[14:09] <{Nico}> florian_: making a new -ng test branch now, i'll try it later this afternoon, when it's done ;)
[14:10] <florian_> {Nico}: no pb !
[14:30] <{Nico}> nbd: what's the script to update the config when upgrading busybox?
[14:33] <nbd> gen_busybox_menuconfig
[14:33] <nbd> but don't use the result of it directly
[14:33] <nbd> diff against the old one to see if some defaults are off
[14:33] <nbd> because it's not always doing the right thing
[14:35] <{Nico}> it seems there was some changes, i have "wget: unrecognized option `--passive-ftp'" when running ipkg update
[14:48] <{Nico}> nbd: i don't understand how to run the script
[14:55] <CIA-17> nico * r4365 /packages/lang/perl/ (. Config.in Makefile ipkg/): port (micro)perl to buildroot-ng
[15:33] <CIA-17> florian * r4366 /packages/net/cups/ (. Config.in Makefile ipkg/): Port cups to -ng
[15:40] <CIA-17> florian * r4367 /packages/net/cutter/ (. Config.in Makefile ipkg/): Port cutter to -ng
[15:45] <CIA-17> florian * r4368 /packages/utils/deco/ (. Config.in Makefile ipkg/): Port deco to -ng
[16:32] <CIA-17> florian * r4369 /packages/net/dhcp6/ (. Config.in Makefile ipkg/): Port dhcp6 to -ng, now builds two packages : dhcp6-server and dhcp6-client
[16:54] <CIA-17> florian * r4370 /packages/net/dhcp/ (. Config.in Makefile ipkg/): Port dhcp to -ng
[18:40] <CIA-17> florian * r4371 /packages/net/dhcp-forwarder/ (. Config.in Makefile ipkg/): Port dhcp-forwarder to -ng
[18:50] <CIA-17> florian * r4372 /packages/net/dhcp-forwarder/Makefile: Add missing conffiles section
[19:05] <CIA-17> florian * r4373 /packages/libs/libnids/Makefile: Add libnet compile time dependency
[19:05] <CIA-17> florian * r4374 /packages/net/dsniff/ (. Config.in Makefile ipkg/): Port dsniff to -ng
[19:12] <CIA-17> florian * r4375 /packages/net/elhttp/ (. Config.in Makefile ipkg/): Port elhttp to -ng
[19:38] <CIA-17> florian * r4376 /packages/libs/bzip2/ (. Makefile): Add bzip2 library
[19:40] <nbd> florian_: can you test bcm63xx again?
[19:52] <florian_> nbd: not yest
[19:52] <florian_> not yet sorry
[19:54] <florian_> but be sure that tonight it will be ok :)
[19:58] <nbd> ok
[20:06] <CIA-17> florian * r4377 /packages/net/elinks/ (. Config.in Makefile ipkg/): Port elinks to -ng
[20:23] <florian_> should I put ethtool in utils or net ?
[20:24] <florian_> I think utils is more relevant
[21:41] <florian_> {Nico}: what do you think about ethtool ?
[21:42] <florian_> ok it is net, like its previous section in control file
[21:44] <CIA-17> florian * r4378 /packages/net/ethtool/ (. Config.in Makefile ipkg/): Port ethtool to -ng
[21:46] <florian_> nbd: is your redboot.c still ok for testing ?
[21:46] <nbd> yes
[21:47] <nbd> brb
[21:47] <florian_> ok
[21:47] <florian_> same here brb
[21:58] <nbd> re
[21:59] <Bartman007> what target are you work with redboot on?
[21:59] <nbd> broadcom 63xx
[21:59] <Bartman007> ah. I shoulda thought of that.
[22:02] <florian_> re
[22:05] <florian_> nbd: Searching for RedBoot partition table in bcm963xx at offset 0x7f0000
[22:05] <florian_> No RedBoot partition table detected in bcm963xx
[22:05] <florian_> Creating 0 MTD partitions on "bcm963xx":
[22:05] <florian_> with the latest patch
[22:10] <nbd> oh, found the bug
[22:10] <nbd> got a new one up...
[22:11] <florian_> oki
[22:13] <florian_> nbd: please make parse_redboot_partitions not static
[22:13] <nbd> done
[22:13] <florian_> thanks
[22:13] <florian_> it's just for next time
[22:15] <florian_> http://pastebin.ca/110612
[22:16] <nbd> strange
[22:16] <nbd> oh, wait
[22:17] <nbd> got another one
[22:17] <nbd> brb
[22:19] <CIA-17> florian * r4379 /packages/utils/fetchmail/ (. Config.in Makefile ipkg/): Port fetchmail to -ng
[22:22] <florian_> nbd: http://pastebin.ca/110624
[22:22] <florian_> it is stopping, but seems to detect all partitions
[22:28] <CIA-17> florian * r4380 /packages/net/fping/ (. Config.in Makefile ipkg/): Port fping to -ng
[22:29] <florian_> {Nico}: ping ?
[22:40] <CIA-17> florian * r4381 /packages/net/fprobe/ (. Config.in Makefile ipkg/): Port fprobe to -ng
[23:09] <florian_> nbd: back ?
[23:09] <nbd> yes
[23:09] <florian_> ok, did you see the pastebin ?
[23:09] <florian_> between commits :)
[23:11] <CIA-17> florian * r4382 /packages/net/fprobe/Makefile: Oops, forgot to add libpthread dependency
[23:13] <CIA-17> florian * r4383 /packages/net/fprobe-ulog/ (. Config.in Makefile ipkg/): Port fprobe-ulog to -ng
[00:00] --- Wed Aug 2 2006
[00:08] <nbd> what does the error look like now
[00:08] <nbd> ?
[00:13] <florian__> nbd: http://pastebin.ca/109171
[00:13] <florian__> with the correct size
[00:13] <florian__> but in every mtd map driver I see, there are rootfs things ? what for ?
[00:14] <florian__> and they use the kernel addr
[00:16] <nbd> florian__: try 1e400000 instead of be400000
[00:17] <nbd> because i think you need the physical address and not the virtual one
[00:18] <florian__> ok
[00:18] <florian__> same value for the size, right ?
[00:18] <nbd> yes
[00:19] <florian__> I am such a newbie, that's sad :)
[00:19] <nbd> it's not sad
[00:20] <florian__> ah ah !
[00:20] <nbd> you can't expect to be an expert right from the start
[00:20] <nbd> does it work now?
[00:21] <florian__> yep that's interesting
[00:21] <florian__> I will pastebin the result
[00:21] <florian__> http://pastebin.ca/109176
[00:22] <florian__> I think the redboot FIS directory must be wrong
[00:23] <nbd> the offset in parse_redboot_partitions needs fixing
[00:23] <nbd> i mean the one that you supply to it
[00:24] <florian__> yep, in the inventel driver it's 0, so I will try with 0
[00:24] <nbd> if it doesn't work, use something relative to the remapped virtual address
[00:24] <florian__> ok
[00:26] <florian__> how did you find out it was 1 instead of b ?
[00:26] <nbd> but it's great that we can use the generic cfi stuff instead of broadcom's crap
[00:26] <florian__> yes !
[00:26] <nbd> florian__: virtual address = physical address | 0x80000000
[00:27] <florian__> ah of course
[00:27] <florian__> nbd: don't be too happy, it is quite easy with RedBoot ;) it will need fixups for CFE
[00:28] <nbd> sure it will, but that won't affect the physical interface to the flash
[00:28] <florian__> right
[00:28] <florian__> well, looks like the remapped virtual address of the fis directory does not seem to be the right value
[00:29] <nbd> what offset did you add to it?
[00:29] <florian__> fis list gives me FIS directory at 0xBEBFF000 so I tried 0x1EBFF000
[00:30] <nbd> that's the physical address
[00:30] <nbd> you need to do it differently. let me check...
[00:34] <nbd> use (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x1000)
[00:34] <florian__> can I ask why :) ? why .
[00:35] <nbd> if you look at the offset that redboot tells you, you should notice that it start 0x1000 before the end of the flash
[00:35] <florian__> right
[00:35] <nbd> the flash size is stored in bcm963xx_mtd_info->size after the probe
[00:35] <nbd> and you need to supply a memory address
[00:36] <florian__> ok, I think I see
[00:36] <nbd> bcm963xx_map.virt is the address that's accessible from the kernel space, because it's the remapped one
[00:36] <florian__> yes
[00:36] <florian__> this does not seem to be the right address anyway
[00:38] <nbd> seems like that parameter is not supposed to carry an address
[00:38] <nbd> (looking at the redboot.c code right now)
[00:39] <nbd> use 0
[00:39] <florian__> I did actually
[00:39] <florian__> let me check ;)
[00:41] <florian__> it attemtps to detect 21 MTD partitions :)
[00:42] <nbd> i see the error
[00:42] <nbd> redboot.c is not prepared to look for the fis directory at such an 'odd' offset
[00:44] <florian__> so we know it is not : 0, physicall address, remapped address
[00:45] <nbd> http://downloads.openwrt.org/people/nbd/tmp/redboot.patch
[00:45] <nbd> use this
[00:45] <nbd> and then use (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x1000) as last parameter to the function
[00:45] <nbd> no
[00:45] <nbd> wait
[00:45] <nbd> bug in the patch
[00:46] <nbd> fixed now
[00:46] <florian__> ok
[00:50] <common> http://flickr.com/photos/24017027@N00/199751092/
[00:50] <florian__> so 0, and (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x1000) as last parameters
[00:50] <nbd> no, replace the 0
[00:50] <nbd> the patch doesn't add a new parameter
[00:51] <florian__> oops sorry :)
[00:53] <florian__> Searching for RedBoot partition table in bcm963xx at offset 0xbebff000
[00:53] <florian__> mtd: Giving out device 0 to bcm963xx
[00:54] <florian__> hu wait
[00:56] <florian__> no that's all
[00:56] <florian__> looks like it does not find the fis directory or cannot parse it
[00:59] <nbd> show me the full fis list
[01:00] <nbd> you told me the wrong fis directory offset
[01:00] <florian__> in query
[01:00] <florian__> ah, really ?
[01:00] <nbd> use (bcm963xx_map.virt + bcm963xx_mtd_info->size - 0x10000)
[01:01] <florian__> with you patch ?
[01:01] <nbd> but it will probably have the same effect as no patch and parameter 0
[01:01] <florian__> ok
[01:01] <nbd> but please try it anyway
[01:01] <nbd> yes, with my patch
[01:03] <florian__> Searching for RedBoot partition table in bcm963xx at offset 0xbebf0000
[01:03] <florian__> Creating 0 MTD partitions on "bcm963xx":
[01:04] <florian__> so it searches the partition at the right place
[01:04] <nbd> make a paste of your current source
[01:05] <florian__> http://pastebin.ca/109250
[01:07] <nbd> try to set the last parameter to 0 again
[01:08] <florian__> ok, with the redboot patch too
[01:08] <nbd> shouldn't make a difference
[01:10] <florian__> Searching for RedBoot partition table in bcm963xx at offset 0x7f0000
[01:10] <florian__> and it tries to detect 21 MTD partitions
[01:13] <florian__> if you have another idea, I take it, after I go sleeping
[01:13] <florian__> thank you very much for your help !
[01:15] <nbd> did your device come with redboot or did you reflash to redboot?
[01:15] <florian__> it came with redboot
[01:15] <florian__> and I reflashed with Kaloz's redboot coming from his livebox as well, so probably almost the same redboot
[01:15] <florian__> because mine was locked up
[01:15] <nbd> k
[01:17] <florian__> at the end of the kmsg I see that : mtd: Giving out device 14 to jffs_system
[01:17] <florian__> 0x003f0000-0x003ff000 : "FIS directory"
[01:17] <florian__> does it help ?
[01:20] <nbd> show me the full kmesg
[01:21] <florian__> http://pastebin.ca/109269
[01:24] <nbd> florian__: http://downloads.openwrt.org/people/nbd/tmp/redboot.patch
[01:34] <florian__> http://pastebin.ca/109280
[01:36] <florian__> I go sleeping, if you have a fix, I will test it tomorrow around 6am, and let you know :)
[01:37] <nbd> i think we almost have it
[01:38] <nbd> so if you want to stay up for a few minutes, i'll try to get it fixed :)
[01:38] <nbd> otherwise: good night :)
[01:45] <nbd> florian__: http://downloads.openwrt.org/people/nbd/tmp/redboot.c
[02:45] <nbd> hey common
[02:45] <common> friend of mine screw'd soldering a 2nd minipci slot on his wrv54g, the 2nd slot does not work, are you intrested in the device?, it works fine as long as you dont try to use the 2nd slot
[02:45] <nbd> sure
[02:45] <common> i'll ask him
[02:46] <common> he is pretty pissed off about it, as he solderd mine first, and my 2nd slot works ...
[02:50] <common> the bug i found in broadcoms upnp daemon will be disclosed without vendor patches, the assholes did not reply any mail
[02:51] <common> sometimes its really sick
[02:51] <common> you offer assistance fixing the shit they sell to make hard cash, and they just give a fuck about it
[07:05] <florian__> nbd: here is the output with your patch : http://pastebin.ca/109558
[08:49] <florian__> re
[09:00] <CIA-17> florian * r4359 /packages/net/click/ (. Config.in Makefile ipkg/): Port click to -ng
[09:51] <CIA-17> nico * r4360 /packages/utils/mpd/ (. Config.in Makefile ipkg/): port mpd to buildroot-ng
[09:54] <{Nico}> do we really need httpd address binding patch in bb?
[10:18] <florian__> {Nico}: it was a feature asked by a user, so I guess it's used at least byhim
[13:13] <nbd> florian_1: good morning
[13:13] <nbd> florian_1: i have another redboot.c for you
[13:13] <nbd> florian_1: same location: http://downloads.openwrt.org/people/nbd/tmp/redboot.c
[13:17] <florian_1> damn, this fscking school server had so many problems this morning, I could not do anything with it :)
[13:17] <nbd> no problem
[13:18] <florian_> I will have a try at you redboot tonight
[13:19] <florian_> anyway, looks quite strange to me that we have to modify redboot in order to get the partitions mounted, should not it rather be the flash map driver ?
[13:20] <nbd> the redboot you're using seems to do things slightly different
[13:20] <nbd> the changes that i'm doing might also work on other platforms
[13:21] <nbd> the only way to avoid having to change redboot.c would be duplicating it
[13:21] <florian_> ok, in fact, it is the stock redboot from linux-2.6.17
[13:24] <CIA-17> florian * r4361 /packages/libs/clinkc/ (. Config.in Makefile ipkg/): Port clinkc to -ng
[13:27] <florian_> nbd: is there a reference package for boolean options in Config.in ?
[13:27] <nbd> what do you mean?
[13:27] <florian_> https://dev.openwrt.org/browser/trunk/openwrt/package/collectd/Config.in
[13:31] <{Nico}> i think we should drop boolean option in packages
[13:31] <florian_> ok, how do you want it to be done ?
[13:32] <{Nico}> just enable "reasonable defaults" from the boolean options
[13:32] <{Nico}> look at mpd, it used to have boolean options for codec support, i enabled all codecs
[13:33] <florian_> {Nico}: ok, thanks, I will have a look
[13:34] <{Nico}> we might add that later back and build -custom packages where every options would be configurable
[13:34] <florian_> {Nico}: I don't think these are booleans like the Config.in I pointed, it's a bit different
[13:34] <florian_> here you do "configure time" plugin detection
[13:37] <{Nico}> with this, we could have packages with the same name built with different options and i find it bad
[13:38] <{Nico}> just do as if all options were selected
[13:38] <florian_> {Nico}: and how do I disable it through menuconfig ?
[13:38] <florian_> by the way, this will be required for OpenVPN too
[13:41] <{Nico}> florian_: forget menuconfig options
[13:41] <{Nico}> just do enable-{cpu,load..etc} in ./configure
[13:46] <florian_> oki
[13:47] <florian_> anyway, you are right, if well done, the configure script should enable on the fly the options
[13:47] <florian_> if it relies on system library, but it is not always the case
[13:48] <{Nico}> nbd: what do you think?
[13:49] <nbd> no idea how to get this right
[13:53] <CIA-17> florian * r4362 /packages/utils/collectd/ (. Config.in Makefile Makefile files/ files/ ipkg/): Port collectd to -ng
[13:53] <{Nico}> what do you think of -custom packages idea (like in pmacct and snort) that could be enabled only when CONFIG_DEVEL=y
[13:53] <florian_> I think it's a good idea
[13:55] <{Nico}> florian_: you should use "install -m0755" for init-scripts to always have correct permissions
[13:55] <florian_> ok
[13:55] <florian_> I fix it
[13:56] <CIA-17> florian * r4363 /packages/utils/collectd/Makefile: Fix init script install permission
[13:59] <CIA-17> florian * r4364 /packages/net/ctorrent/ (. Config.in Makefile ipkg/): Port ctorrent to -ng
[14:01] <florian_> is not PKG_BUILDDEP and DEPENDS:= redundant ?
[14:03] <{Nico}> PKG_BUILDEP is for compile-time dependencies, while DEPENDS will end up in the ipkg Depends: control field
[14:04] <florian_> at the beginning I did not fill PKG_BUILDDEP and compile time dependencies were solved too
[14:04] <{Nico}> DEPENDS will also feed compile-time dependencies
[14:05] <florian_> ok
[14:05] <florian_> it is just in the case the package name is different from the library name like mysql and libmysqlclient for instance ?
[14:06] <{Nico}> PKG_BUILDEPS should be used when you need another package to be compiled before
[14:06] <{Nico}> but you don't need any of its libs at runtime
[14:06] <{Nico}> for your package
[14:06] <florian_> ok
[14:07] <florian_> I have to write documentation about that
[14:07] <florian_> {Nico}: by the way did you test the au1000 port ?
[14:07] <{Nico}> not yet
[14:08] <florian_> maybe you did not had time, and I understand, I just want if you take 5 min to flash :) to tell me if you get atheros working "out of the box"
[14:09] <{Nico}> florian_: making a new -ng test branch now, i'll try it later this afternoon, when it's done ;)
[14:10] <florian_> {Nico}: no pb !
[14:30] <{Nico}> nbd: what's the script to update the config when upgrading busybox?
[14:33] <nbd> gen_busybox_menuconfig
[14:33] <nbd> but don't use the result of it directly
[14:33] <nbd> diff against the old one to see if some defaults are off
[14:33] <nbd> because it's not always doing the right thing
[14:35] <{Nico}> it seems there was some changes, i have "wget: unrecognized option `--passive-ftp'" when running ipkg update
[14:48] <{Nico}> nbd: i don't understand how to run the script
[14:55] <CIA-17> nico * r4365 /packages/lang/perl/ (. Config.in Makefile ipkg/): port (micro)perl to buildroot-ng
[15:33] <CIA-17> florian * r4366 /packages/net/cups/ (. Config.in Makefile ipkg/): Port cups to -ng
[15:40] <CIA-17> florian * r4367 /packages/net/cutter/ (. Config.in Makefile ipkg/): Port cutter to -ng
[15:45] <CIA-17> florian * r4368 /packages/utils/deco/ (. Config.in Makefile ipkg/): Port deco to -ng
[16:32] <CIA-17> florian * r4369 /packages/net/dhcp6/ (. Config.in Makefile ipkg/): Port dhcp6 to -ng, now builds two packages : dhcp6-server and dhcp6-client
[16:54] <CIA-17> florian * r4370 /packages/net/dhcp/ (. Config.in Makefile ipkg/): Port dhcp to -ng
[18:40] <CIA-17> florian * r4371 /packages/net/dhcp-forwarder/ (. Config.in Makefile ipkg/): Port dhcp-forwarder to -ng
[18:50] <CIA-17> florian * r4372 /packages/net/dhcp-forwarder/Makefile: Add missing conffiles section
[19:05] <CIA-17> florian * r4373 /packages/libs/libnids/Makefile: Add libnet compile time dependency
[19:05] <CIA-17> florian * r4374 /packages/net/dsniff/ (. Config.in Makefile ipkg/): Port dsniff to -ng
[19:12] <CIA-17> florian * r4375 /packages/net/elhttp/ (. Config.in Makefile ipkg/): Port elhttp to -ng
[19:38] <CIA-17> florian * r4376 /packages/libs/bzip2/ (. Makefile): Add bzip2 library
[19:40] <nbd> florian_: can you test bcm63xx again?
[19:52] <florian_> nbd: not yest
[19:52] <florian_> not yet sorry
[19:54] <florian_> but be sure that tonight it will be ok :)
[19:58] <nbd> ok
[20:06] <CIA-17> florian * r4377 /packages/net/elinks/ (. Config.in Makefile ipkg/): Port elinks to -ng
[20:23] <florian_> should I put ethtool in utils or net ?
[20:24] <florian_> I think utils is more relevant
[21:41] <florian_> {Nico}: what do you think about ethtool ?
[21:42] <florian_> ok it is net, like its previous section in control file
[21:44] <CIA-17> florian * r4378 /packages/net/ethtool/ (. Config.in Makefile ipkg/): Port ethtool to -ng
[21:46] <florian_> nbd: is your redboot.c still ok for testing ?
[21:46] <nbd> yes
[21:47] <nbd> brb
[21:47] <florian_> ok
[21:47] <florian_> same here brb
[21:58] <nbd> re
[21:59] <Bartman007> what target are you work with redboot on?
[21:59] <nbd> broadcom 63xx
[21:59] <Bartman007> ah. I shoulda thought of that.
[22:02] <florian_> re
[22:05] <florian_> nbd: Searching for RedBoot partition table in bcm963xx at offset 0x7f0000
[22:05] <florian_> No RedBoot partition table detected in bcm963xx
[22:05] <florian_> Creating 0 MTD partitions on "bcm963xx":
[22:05] <florian_> with the latest patch
[22:10] <nbd> oh, found the bug
[22:10] <nbd> got a new one up...
[22:11] <florian_> oki
[22:13] <florian_> nbd: please make parse_redboot_partitions not static
[22:13] <nbd> done
[22:13] <florian_> thanks
[22:13] <florian_> it's just for next time
[22:15] <florian_> http://pastebin.ca/110612
[22:16] <nbd> strange
[22:16] <nbd> oh, wait
[22:17] <nbd> got another one
[22:17] <nbd> brb
[22:19] <CIA-17> florian * r4379 /packages/utils/fetchmail/ (. Config.in Makefile ipkg/): Port fetchmail to -ng
[22:22] <florian_> nbd: http://pastebin.ca/110624
[22:22] <florian_> it is stopping, but seems to detect all partitions
[22:28] <CIA-17> florian * r4380 /packages/net/fping/ (. Config.in Makefile ipkg/): Port fping to -ng
[22:29] <florian_> {Nico}: ping ?
[22:40] <CIA-17> florian * r4381 /packages/net/fprobe/ (. Config.in Makefile ipkg/): Port fprobe to -ng
[23:09] <florian_> nbd: back ?
[23:09] <nbd> yes
[23:09] <florian_> ok, did you see the pastebin ?
[23:09] <florian_> between commits :)
[23:11] <CIA-17> florian * r4382 /packages/net/fprobe/Makefile: Oops, forgot to add libpthread dependency
[23:13] <CIA-17> florian * r4383 /packages/net/fprobe-ulog/ (. Config.in Makefile ipkg/): Port fprobe-ulog to -ng
[00:00] --- Wed Aug 2 2006