**** BEGIN LOGGING AT Tue Nov 20 02:59:57 2012
Nov 20 03:23:34 <DonkeyHotei> mazilo: do you by any chance use future nine?
Nov 20 03:25:14 <mazilo> DonkeyHotei: I have anon-active F9 account which was given to me as a test account.
Nov 20 03:25:51 <DonkeyHotei> hmm
Nov 20 03:30:17 <mazilo> DonkeyHotei: R U planning to port your # to F9?
Nov 20 03:30:28 <DonkeyHotei> someone else's
Nov 20 03:31:39 <mazilo> DonkeyHotei: IC. I have never done any # porting. All I did was to let my number expire and moved on to another provider with a new number. This way, I won't waste $$$ to port my number.
Nov 20 03:33:32 <DonkeyHotei> the only way to keep a landline number is to port to cell or voip, and voip is cheaper
Nov 20 07:19:36 <OpenWrt-Commits> juhosg r34267 trunk/target/linux/ generic/config-3.7 realview/config-3.3 generic/config-3.6 ar71xx/config-3.3 * refresh all kernel configurations
Nov 20 07:19:38 <OpenWrt-Commits> juhosg r34268 trunk/package/mac80211/patches/622-rt2x00-fix-rt3352-ext-pa.patch * mac80211/rt2x00: support Rt3352 with external PA
Nov 20 07:19:39 <OpenWrt-Commits> juhosg r34269 trunk/ target/linux/generic/files/include/linux/rt2x00_platform.h package/mac80211/patches/623-rt2x00-rf_vals-rt3352-xtal20.patch * mac80211/rt2x00: add rf_vals for Rt3352 with Xtal=20MHz
Nov 20 07:19:40 <OpenWrt-Commits> juhosg r34270 trunk/target/linux/ ramips/files/arch/mips/ralink/rt305x/devices.c ramips/files/arch/mips/include/asm/mach-ralink/rt305x_regs.h * ramips: set clk_is_20mhz for rt2x00 on RT3352/RT5350
Nov 20 12:18:12 <OpenWrt-Commits> juhosg r34271 trunk/target/linux/ ar71xx/image/lzma-loader/src/loader2.lds ar71xx/image/lzma-loader/src/Makefile * ar71xx/image: relink loader.elf image to LOADADDR
Nov 20 12:18:13 <OpenWrt-Commits> juhosg r34272 trunk/target/linux/ar71xx/image/Makefile * ar71xx/image: provide lzma compressed elf kernel image
Nov 20 12:18:14 <OpenWrt-Commits> juhosg r34273 trunk/target/linux/ar71xx/base-files/sbin/wget2nand * ar71xx: install lzma compressed elf kernel image from wget2nand
Nov 20 12:18:15 <OpenWrt-Commits> juhosg r34274 trunk/target/linux/ar71xx/files/drivers/mtd/nand/rb4xx_nand.c * ar71xx: restore rb4xx kernel partition size to the original value
Nov 20 14:27:47 <OpenWrt-Commits> juhosg r34275 trunk/target/linux/ (6 files in 2 dirs) * ar71xx: fix GPIO function selection for AR934x
Nov 20 14:40:52 <OpenWrt-Commits> juhosg r34276 trunk/target/linux/ar71xx/files/drivers/mtd/nand/rb4xx_nand.c * ar71xx: remove buffer verification reference from rb4xx_nand.c
Nov 20 15:50:43 <jwendell> is it legal to get the port of linux made by a vendor and submit it to linux kernel?
Nov 20 15:51:19 <DonkeyHotei> define "made"
Nov 20 15:51:27 <jwendell> (I don't know the reason why that port is not mainstream)
Nov 20 15:51:37 <DonkeyHotei> what port?
Nov 20 15:51:40 <jwendell> made = build, create
Nov 20 15:52:39 <[florian]> jwendell: everything that gets built into the final vmlinux file, even if not GPL copyrighted (or similar) can be used
Nov 20 15:52:48 <[florian]> jwendell: for kernel modules you just need to be careful
Nov 20 15:53:47 <OpenWrt-Commits> juhosg r34277 trunk/target/linux/ar71xx/patches-3.3/a03-rb4xx_nand-add-NO_AUTOINCR-flag.patch * ar71xx: refresh rb4xx_nand patch
Nov 20 15:53:59 <jwendell> the vendor supplies at their website the [modified] kernel
Nov 20 15:54:11 <jwendell> I get that kernel, make a diff from a vanilla
Nov 20 15:54:18 <jwendell> and submit that patch upstream
Nov 20 15:54:29 <DonkeyHotei> what is it a port to?
Nov 20 15:54:30 <jwendell> is that possible, once I didn't write anything?
Nov 20 15:54:51 <[florian]> jwendell: yes it is, by making this code available under a gpl or bsd license, you get the right to actually push it upstream
Nov 20 15:55:10 <[florian]> jwendell: since you probably will never get any Signed-off-by tags from the original author you can actually use your own Signed-off-by ta
Nov 20 15:55:13 <[florian]> *g
Nov 20 15:55:39 <[florian]> jwendell: what port are we talking about?
Nov 20 15:56:01 <jwendell> buffalo
Nov 20 15:56:13 <jwendell> it's an old model I'm studying
Nov 20 15:56:15 <[florian]> what is the target platform/SoC?
Nov 20 15:56:22 <jwendell> trendchip
Nov 20 15:56:31 <[florian]> be more specific :)
Nov 20 15:56:41 <jwendell> tc3162
Nov 20 15:57:08 <[florian]> sounds old
Nov 20 15:57:12 <[florian]> what kind of cpu does that use,
Nov 20 15:58:38 <DonkeyHotei> probably mips
Nov 20 15:58:39 <jwendell> I couldn't get into their console, and the default firmware hasn't ssh or telnet enabled
Nov 20 15:58:52 <jwendell> but they supply all the kernel
Nov 20 15:59:04 <jwendell> I'm not sure about the DSL modules...
Nov 20 15:59:22 <jwendell> yes, it's mips
Nov 20 16:00:24 <[florian]> I just found a tarball, let me look into this
Nov 20 16:01:13 <[florian]> it's even a mips without mmu
Nov 20 16:02:19 <jwendell> [florian], here is a complete tarball, from buffalo: http://buffalo.jp/php/los.php?to=gpl/wireless/gpl_wbmr-hp-gnv2_101.tar.bz2
Nov 20 16:02:34 <[florian]> thanks
Nov 20 16:04:00 <Chocks> kids
Nov 20 16:04:19 <[florian]> Chocks: hey, so did you find your issue?
Nov 20 16:04:39 <[florian]> jwendell: a 2.6.22, that's more like it could be doable then
Nov 20 16:04:42 <Chocks> not yet.  I started a built with eglibc 2.16 last night, but it's failed
Nov 20 16:05:07 <Chocks> some of my kernel patching gone awry.  which is a bit odd.
Nov 20 16:05:20 <KanjiMonster> [florian]: according to ralink it has a mmu
Nov 20 16:05:43 <[florian]> yep
Nov 20 16:05:51 <[florian]> seems like the old tarball I found was for some other chip
Nov 20 16:05:59 <[florian]> jwendell: the dsl driver is there, which is cool, but it's not GPL
Nov 20 16:06:05 <Chocks> if this fails, I'll add back in 2.13 and try again
Nov 20 16:06:20 <Chocks> it seems odd that it might be eglibc, but I have to rule it out
Nov 20 16:06:37 <DonkeyHotei> [florian]: is it linkable against multiple kernel versions?
Nov 20 16:07:02 <[florian]> DonkeyHotei: it's actually the dsl driver source code
Nov 20 16:07:59 <[florian]> ah that's just the command forwarder actually
Nov 20 16:08:24 <jwendell> [florian], so, there's the possibility of getting it ported to openwrt, right?
Nov 20 16:08:38 <[florian]> jwendell: certainly yes
Nov 20 16:08:43 <jwendell> good
Nov 20 16:09:02 <[florian]> the existing code is pretty ugly (as usual)
Nov 20 16:09:12 <jwendell> [florian], I'd like to start doing ports and not only doing trivial patches :)
Nov 20 16:09:27 <[florian]> jwendell: a rewrite from scratch is indicated
Nov 20 16:10:12 <jwendell> hmm
Nov 20 16:10:25 <[florian]> we'd never accept such an ugly code as-is
Nov 20 16:10:37 <jwendell> :(
Nov 20 16:11:27 <jwendell> I never did this before
Nov 20 16:11:31 <DonkeyHotei> [florian]: do you think it is applicable to ralink/mediatek's current dsl chips?
Nov 20 16:11:37 <jwendell> I mean, hack in the kernel
Nov 20 16:11:38 <[florian]> DonkeyHotei: I don't know
Nov 20 16:11:54 <DonkeyHotei> it would be good to find out
Nov 20 16:11:59 <[florian]> oh woah this even has dedicated cache routines
Nov 20 16:12:12 <[florian]> sounds like fun
Nov 20 16:12:20 <DonkeyHotei> also, i saw mentions somewhere that lantiq has a docsis chip
Nov 20 16:12:50 <DonkeyHotei> i thought all docsis was broadcom
Nov 20 16:17:25 <[florian]> jwendell: the ethernet driver is missing
Nov 20 16:19:32 <jwendell> really?
Nov 20 16:19:38 <jwendell> that basic thing...
Nov 20 16:20:14 <[florian]> or it's using an already supported ethernet chip, otherwise I did not see it in the tarball
Nov 20 16:27:37 <Chocks> on a different note, I patched lighttpd build to have websockets support.  Patches to mailing list soon
Nov 20 16:30:12 <OpenWrt-Commits> juhosg r34278 trunk/target/linux/ ar71xx/patches-3.6/a05-ar934x_nfc-add-buffer-verification.patch ar71xx/patches-3.3/a02-ar934x_nfc-add-NO_AUTOINCR-flag.patch ar71xx/patches-3.3/a05-ar934x_nfc-add-buffer-verification.patch ar71xx/files/drivers/mtd/nand/ar934x_nfc.c * ar71xx: make ar934x_nfc driver compatible with 3.7
Nov 20 16:30:14 <OpenWrt-Commits> juhosg r34279 trunk/target/linux/ ar71xx/patches-3.3/a04-rb750_nand-add-NO_AUTOINCR-flag.patch ar71xx/patches-3.6/a06-rb750_nand-add-buffer-verification.patch ar71xx/files/drivers/mtd/nand/rb750_nand.c ar71xx/patches-3.3/a06-rb750_nand-add-buffer-verification.patch * ar71xx: make rb750_nand driver compatible with 3.7
Nov 20 16:44:56 <Chocks> there. websockets. enjoy.
Nov 20 17:21:15 <jwendell> [florian], would you mentor me? I never ported or touched anything in the kernel level, but I'm willing too much to do that. so... I'd start with that router and use that tarball as a start
Nov 20 17:21:55 <[florian]> jwendell: I'd start making sure what's in the tarball is actually bootable
Nov 20 17:22:04 <jwendell> so, I could ask my doubts directly to you, of course without to bug you too much
Nov 20 17:22:05 <[florian]> then we can figure out what's missing, what can be reworked
Nov 20 17:22:09 <[florian]> no problem
Nov 20 17:23:02 <jwendell> [florian], good tip to start. tomorrow I'll solde the console pins
Nov 20 17:25:13 <Chocks> moo
Nov 20 17:27:28 <jwendell> [florian], so, in order to check if that kernel is bootable, I need to: 1) actually manage to build the kernel; 2) overwrite the flash kernel partition with my own kernel image
Nov 20 17:27:38 <jwendell> right? am i missing something?
Nov 20 17:28:19 <[florian]> I'd rather build a kernel that you load in ram
Nov 20 17:28:28 <[florian]> without actually flashing it
Nov 20 17:28:38 <[florian]> you should also explore what does the bootloader offer to you
Nov 20 17:29:42 <jwendell> ah great
Nov 20 17:43:37 <Chocks> the eglibc 2.16 build appears to have various problems
Nov 20 17:44:10 <Chocks> missing /usr/lib in library search path
Nov 20 17:45:26 <Chocks> yes, I see
Nov 20 17:45:55 <OpenWrt-Commits> juhosg r34280 trunk/target/linux/ (10 files in 2 dirs) * generic: convert yaffs to use kuid_t and kgid_t on 3.6+
Nov 20 17:47:40 <Chocks> 2.16 is missing  200-add-dl-search-paths.patch
Nov 20 17:48:21 * Chocks pokes eglibc guy, mirko
Nov 20 17:51:47 <Chocks> florian: that aside, 2.16 does not help my original problem.
Nov 20 18:16:04 <[florian]> Chocks: ok, I will add that patch
Nov 20 18:16:09 <Chocks> ta
Nov 20 18:16:29 <Chocks> building with 2.13 again now
Nov 20 18:16:36 <Chocks> but I suspect it's something else
Nov 20 18:24:10 <[florian]> Chocks: looks like we are missing quite some eglibc 2.16 from OE
Nov 20 18:24:21 <Chocks> hm
Nov 20 18:24:24 <Chocks> what else?
Nov 20 18:24:37 <[florian]> http://sourceware.org/ml/libc-ports/2007-12/msg00000.html
Nov 20 18:24:46 <Chocks> someone must have just added it for kicks and not actually tested it
Nov 20 18:25:53 <Chocks> however, apart from this particular issue, which might have nothing to do with eglibc, I'm using 2.15 on ramips and an ARM9 platform no problem
Nov 20 18:26:11 <Chocks> will update for ar91xx later too
Nov 20 18:27:13 <[florian]> what kind of arm9 platform do you use?
Nov 20 18:28:02 <Chocks> atmel91SAM9645
Nov 20 18:28:18 <Chocks> not strictly supported by OpenWrt, but pretty generic comprehensive support for that family
Nov 20 18:28:39 <[florian]> ok
Nov 20 18:30:03 <Chocks> however, the customer thinks they want Angstrom, so I don't know which way that might go yet.
Nov 20 18:55:02 <Chocks> so, still waiting for my 2.13 build
Nov 20 18:55:08 <Chocks> but the problem would appear to be netlink related
Nov 20 18:55:20 <Chocks> judging by:   http://kernel.org/doc/pending/hotplug.txt
Nov 20 19:00:08 <jwendell> [florian], I'm trying to build that kernel with a toolchain I already have built here (openwrt, mips for lantiq). can I do that?
Nov 20 19:00:36 <jwendell> (I'm getting lots of warnings  warning: "__MIPSEL__" is not defined [-Wundef])
Nov 20 19:01:03 <Chocks> looks like someone made some comedy changes to that kerenl
Nov 20 19:01:06 <jwendell> from what I see in config files, that device is big endian (just like lantiq)
Nov 20 19:01:27 <Chocks> you might have to define it manually
Nov 20 19:02:21 <jwendell> Chocks, actually it's NOT mipsel, it's just mips (big endian)
Nov 20 19:02:43 <jwendell> so, I don't understand the warning here
Nov 20 19:02:44 <Chocks> -D__MIPSEL__=0
Nov 20 19:03:05 <Chocks> someone without good coding style made some hacky changes
Nov 20 19:04:12 <jwendell> looking at arch/mips/Makefile, I saw this: cflags-$(CONFIG_CPU_BIG_ENDIAN)		+= $(shell $(CC) -dumpmachine |grep -q 'mips.*el-.*' && echo -EB $(undef-all) $(predef-be))
Nov 20 19:04:32 <jwendell> it calls gcc -dumpmachine
Nov 20 19:04:49 <jwendell> and expects an output
Nov 20 19:04:54 <Chocks> hm
Nov 20 19:05:21 <jwendell> my output is: mips-openwrt-linux-uclibc
Nov 20 19:05:49 <Chocks> right.  They originally were using some other comedy toolchain that did something special for them
Nov 20 19:06:43 <jwendell> this is in the vanilla kernel
Nov 20 19:06:45 <jwendell> also
Nov 20 19:07:00 <jwendell> so, it's not their trick :)
Nov 20 19:08:17 <Chocks> aww
Nov 20 19:44:04 <mazilo> Is anyone here working to port OpenWRT to this PogoPlug P21 (http://archlinuxarm.org/platforms/armv6/pogoplug-provideov3) which is based on a PLX/Oxford Semiconductor NAS782 SoC?
Nov 20 19:44:39 <Chocks> have you checked the mailing list?
Nov 20 19:45:29 <DonkeyHotei> doesn't sound too difficult, but armv6 sounds a bit old
Nov 20 19:45:43 <Chocks> not that old
Nov 20 19:47:29 <mazilo> No and U haven
Nov 20 19:47:36 <mazilo> t checked the mailing list yet.
Nov 20 19:47:47 <Chocks> lots of armv5 stuff still about
Nov 20 19:48:09 <mazilo> The reason I asked is this device can be had for less than US$20 these days.
Nov 20 19:48:20 <Chocks> one should always check the mailing list first
Nov 20 19:48:54 <mazilo> Yes and it is based on ARMv6. So, does a RaspberryPi.
Nov 20 19:49:03 <mazilo> OK.
Nov 20 19:49:08 <Chocks> https://wiki.xkyle.com/Install_Openwrt_10.03_on_a_Seagate_Dockstar
Nov 20 19:49:37 <mazilo> I have some Seagate DockStar running a self-built OpenWRT firmware.
Nov 20 19:49:50 <Chocks> it really seems like you haven't done any kind of googling
Nov 20 19:50:17 <russell--> i think a dockstar isn't an oxnas chip
Nov 20 19:50:40 <russell--> yeah, it's a kirkwood thing
Nov 20 19:50:51 <Chocks> like sheeva
Nov 20 19:51:11 <DonkeyHotei> btw, using () makes a link not clickable in some irc clients (like mine) unlike <>
Nov 20 19:51:19 <mazilo> When I did some googling, all I got was and impression that the NAS782 SoC code is proprietary.
Nov 20 19:53:16 <[florian]> jwendell: are you sure you are setting CROSS_COMPILE to point to your OpenWrt toolchain?
Nov 20 19:53:44 <jwendell> [florian], yep. I have fixed those warnings
Nov 20 19:53:58 <jwendell> the issue now is an compile error...
Nov 20 19:54:15 <[florian]> you might want to use an older toolchain
Nov 20 19:54:25 <[florian]> newer ones tend to unveil a bunch of warnings
Nov 20 19:54:47 <jwendell> I see
Nov 20 19:54:55 <jwendell> the error is: arch/mips/kernel/time.c:349:3: error: impossible constraint in 'asm'
Nov 20 19:54:57 <DonkeyHotei> "The SoC used in these models is not supported in the mainline Linux kernel, and all sources are covered under an NDA with PLX. We support these devices with a custom kernel that is flashed alongside the original kernel in flash memory. This is a safe and non-destructive process, and allows the devices to use all of the software in our repositories." right on the page mazilo linked
Nov 20 19:55:18 <jwendell> weird thing is that time.c line 349 has nothing to do with that error message
Nov 20 19:55:44 <Chocks> donkey: fail
Nov 20 19:56:20 <KanjiMonster> DonkeyHotei: despite their claims, there are kernel sources publically available, and it even seems the drivers carry a correct GPL license
Nov 20 19:56:38 <russell--> the archlinux dude did some work on porting the vendor oxnas patches to an earlier kernel to something more recent, not sure how far that got
Nov 20 19:56:40 <[florian]> jwendell: that might be because of some inline function pulled in, you might want to have a look at the pre-processed file (.i)
Nov 20 19:56:57 <russell--> i have a B01 or something
Nov 20 19:57:26 <russell--> there was also some craziness due to the stock u-boot, iirc
Nov 20 19:57:33 <KanjiMonster> russell--: well, their build system pulls from https://github.com/WarheadsSE/OX820-2.6-linux/ which is still 2.6.31
Nov 20 19:57:50 <russell--> yeah, there was a 3.x version on github
Nov 20 19:57:54 <russell--> i think
Nov 20 19:58:06 <DonkeyHotei> Chocks: ?
Nov 20 19:59:08 <russell--> https://github.com/WarheadsSE/OX820-3.1-Linux ... last touched 9 months ago
Nov 20 20:35:17 <Chocks> florian: no win on 2.13
Nov 20 20:35:29 <Chocks> what the hell else might it be?
Nov 20 20:37:48 <mazilo> russell--: Thanks and I will check out the link.
Nov 20 20:40:32 <OpenWrt-Commits> juhosg r34281 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: add missing locking calls
Nov 20 20:40:33 <OpenWrt-Commits> juhosg r34282 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: remove an unblanced mutex_unlock call
Nov 20 20:40:34 <OpenWrt-Commits> juhosg r34283 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * generic: ar8216: fix port number check in 'ar8xxx_mib_work_func'
Nov 20 21:19:07 <Chocks> damn
Nov 20 21:32:20 <Chocks> definitely not the newly built kernel either
Nov 20 21:35:14 <OpenWrt-Commits> juhosg r34284 trunk/target/linux/generic/config-3.7 * generic: add some missing symbols for 3.7
Nov 20 21:38:10 <Chocks> florian: I found it
Nov 20 21:38:23 <OpenWrt-Commits> juhosg r34285 trunk/target/linux/ (20 files in 2 dirs) * cns21xx: add support for 3.6
Nov 20 21:38:27 <Chocks> the new build has a problem with its button_hotplug module
Nov 20 21:56:38 <Chocks> why?  I have no idea. it's very strange
Nov 20 21:57:59 * Chocks wiggles florian
Nov 20 22:02:51 <written_tabletpc> .
Nov 20 22:03:22 <Chocks> your . is stuck
Nov 21 00:03:48 <Chocks> moo all
Nov 21 00:24:51 <written_tabletpc> re :-)
Nov 21 00:25:17 <Chocks> yes
**** ENDING LOGGING AT Wed Nov 21 02:59:58 2012