**** BEGIN LOGGING AT Thu Aug 25 02:59:58 2011
Aug 25 09:39:19 <CIA-89> florian * r28081 /trunk/target/linux/ (8 files in 6 dirs):
Aug 25 09:39:19 <CIA-89> [x86] update alix2 platform driver
Aug 25 09:39:19 <CIA-89> Add support for Ed Wildgoose's alix platform driver plus my
Aug 25 09:39:19 <CIA-89> enhancement to add GPIO soft button support.
Aug 25 09:39:19 <CIA-89> Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
Aug 25 09:39:23 <CIA-89> florian * r28082 /trunk/target/linux/x86/Makefile:
Aug 25 09:39:23 <CIA-89> [x86] bump to 2.6.39.4
Aug 25 09:39:23 <CIA-89> Since linux gets developed on x86 more than all other hardware,
Aug 25 09:39:23 <CIA-89> hardly makes sense to have our default x86 version be trailing
Aug 25 09:39:24 <CIA-89> by over a year.
Aug 25 09:39:24 <CIA-89> Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
Aug 25 10:14:09 <CIA-89> florian * r28083 /packages/libs/mysql/Makefile: (log message trimmed)
Aug 25 10:14:09 <CIA-89> [package] Fix mysql dependencies
Aug 25 10:14:09 <CIA-89> Trying to fix some issues with php5 I discoverd that php5's mysql extension doesn't load:
Aug 25 10:14:09 <CIA-89> <snip>
Aug 25 10:14:09 <CIA-89> root@OpenWrt:/tmp# php-cli -m
Aug 25 10:14:09 <CIA-89> PHP Warning: PHP Startup: Unable to load dynamic library '/usr/lib/php/mysql.so' - File not found in Unknown on line 0
Aug 25 10:14:10 <CIA-89> [PHP Modules]
Aug 25 10:41:03 <CIA-89> hcg * r28084 /trunk/target/linux/omap35xx/patches-3.0/ (5 files): [omap35xx]: Re-order patches, add nand sub-page alignment
Aug 25 10:53:01 <CIA-89> florian * r28085 /trunk/package/udev/ (5 files in 2 dirs): (log message trimmed)
Aug 25 10:53:01 <CIA-89> [package] udev: bump to 173
Aug 25 10:53:01 <CIA-89> Yet another update, I inherited a wrong behaviour with older versions.
Aug 25 10:53:01 <CIA-89> It used to $(INSTALL_BIN) symlinks which results in copying the dereferenced
Aug 25 10:53:01 <CIA-89> file and not the symlink itself, now uses $(CP) and it's reported working (and running)
Aug 25 10:53:02 <CIA-89> by at least one person in IRC (using a glibc toolchain with mips). Works and runs on
Aug 25 10:53:03 <CIA-89> my toolchain (mipsel uClibc) too.
Aug 25 11:45:50 <WillieNL> new board in the collection asus wl520gu/gc
Aug 25 11:46:09 <WillieNL> will come without usb.. parts for the usb connection are on my old blownup wl500g :)
Aug 25 11:46:39 <WillieNL> any interest for an other howto ?
Aug 25 11:53:27 <CIA-89> hcg * r28086 /trunk/target/linux/omap35xx/patches-3.0/003-nand_subpage_align.patch: [omap35xx]: Remove subpage align patch
Aug 25 13:00:14 <WillieNL> anyone that still have a list with part nr's from the asus wl-520gC to gU mod (usb modification) ?
Aug 25 13:00:45 <WillieNL> i had 3 examples bookmarket and just can't find them anymore.. :S and yes can lookup part by part but .. i had a list :(
Aug 25 13:05:40 <WillieNL> hmm thanks found some rusian howto.. (yes saved the whole page now)
Aug 25 14:11:38 <WillieNL> fullstop: dunno if this is helpfull for you but they show some suppliers for in this example broadcom bcm4702kpb chips http://www.ic2ic.com/search.jsp?sSearchWord=BCM4702KPB&prefix=B
Aug 25 15:35:48 <fullstop> WillieNL: I'm looking into Atmel ARM9s now.
Aug 25 15:36:00 <Chocks> moo
Aug 25 15:36:04 <fullstop> I'd really like a Cortex-A5, but nobody makes them.
Aug 25 15:42:39 <WillieNL> fullstop: hmm oxe800dse :D
Aug 25 15:42:48 <WillieNL> or was that armv5 ?
Aug 25 15:44:17 <WillieNL> fullstop: based on a 200mhz arm926ej http://kompl.yc.cz/download/DS-0025_OXES800DSE_datasheet.pdf
Aug 25 15:44:34 <WillieNL> including sata/usb/ethernet/gpio/almosth everything ?
Aug 25 15:44:57 <WillieNL> hmm sata is external via here.
Aug 25 15:46:22 <WillieNL> lets say chip + 10 resistors 5 capacitors and power regulation and your have a 6 by 10 cm SBC
Aug 25 15:46:41 <WillieNL> argh sata/ethernet is external
Aug 25 15:47:20 <WillieNL> think that marvells have a better solution with more cpu performance on the soc.
Aug 25 15:52:09 <fullstop> WillieNL: Ideally, it would have usb, ethernet, spi, i2c, can and a ddr2 interface.
Aug 25 15:52:27 <fullstop> the A8 and OMAP3 look nice, but they are kind of expensive.
Aug 25 15:52:56 <fullstop> ~ $30 USD without any external flash and ram
Aug 25 15:53:16 <WillieNL> hmm also oxe800's are discontinued i think/guess
Aug 25 15:53:34 <WillieNL> but stick a simple heatsink on them fixed the factory design problem "heat" :)
Aug 25 15:57:47 <WillieNL> fullstop: hmm i have seen from chinese supplyers they where able to ship a complete nas board for +- 75 euro based on a Marvell 88F6281 1.2 G (ARM926EJ-S)
Aug 25 15:57:54 <WillieNL> with ddr2 800 ram on them
Aug 25 15:58:06 <WillieNL> i can pm you a link if you like
Aug 25 16:02:58 <CIA-89> florian * r28087 /trunk/toolchain/uClibc/patches-0.9.32/160-mips_signalfd_sfd_nonblock.patch: [toolchain] uclibc: backport upstream signalfd patch
Aug 25 16:14:27 * Chocks pokes full stop
Aug 25 16:15:00 <Chocks> fullstop too
Aug 25 16:18:04 <Chocks> fullstop: turns out that the lot of g300NHs we bought is a mix of v1/v2
Aug 25 17:19:13 <fullstop> Chocks: Yeah?  I heard from someone else who called Buffalo, and their support stated that Amazon must have accidentally sent the Australian model because the G300NH2 is not in the US market.
Aug 25 17:19:27 <Chocks> weird.
Aug 25 17:19:28 <fullstop> I think the FCC would be quite upset about that.
Aug 25 17:19:38 <Chocks> as it happens, these are going to australia
Aug 25 17:19:56 <fullstop> What is the regulatory domain?  Do you have the extra channels?
Aug 25 17:20:07 <Chocks> I don't have one to hand this second; will check
Aug 25 17:23:18 <Chocks> ok, so tell me (or wiki me!) on the differences between the two versions
Aug 25 17:28:49 <fullstop> http://wiki.openwrt.org/toh/buffalo/wzr-hp-g300nh2
Aug 25 17:29:13 <fullstop> http://wiki.openwrt.org/toh/buffalo/wzr-hp-g300h
Aug 25 17:29:48 <fullstop> Different flash, switch and cpu, although they share the same mips core.
Aug 25 17:30:01 <fullstop> wlan is pci instead of integrated.
Aug 25 17:39:56 <Chocks> so, no chance of using same kernel?
Aug 25 17:41:49 <Chocks> I'd settle for an initramfs only base image which could then choose a complete version for its hardware
Aug 25 17:42:39 <fullstop> I suppose that you could attempt to detect the platform and use different init code...
Aug 25 17:43:18 <Chocks> are the code differences in the kernel entirely related to choose peripherals?
Aug 25 17:43:34 <Chocks> or are there different compile options?
Aug 25 17:44:59 <fullstop> The most  important piece is to set up flash so that you can actually boot.
Aug 25 17:45:22 <Chocks> of course
Aug 25 17:45:43 <fullstop> they share the same mips core -- you can try to boot the G300NH image on the G300NH2, but it won't find the root file system.
Aug 25 17:46:13 <fullstop> You can select both profiles in the build and end up with an image for the G300NH and an image for the G300NH2.
Aug 25 18:13:29 <Chocks> sure.  but I'd prefer to have one image where possible.  that solves a lot of hassle
Aug 25 18:14:54 <Chocks> I'll have a poke in that kernel code, and change everything you did ;-)
Aug 25 18:25:52 <fullstop> Chocks: have fun!  :D  You'll have to find a way to probe the hardware.
Aug 25 18:26:23 <Chocks> I'd expect to be able to get at the CPU ID/rev
Aug 25 18:27:13 <Chocks> I wonder if you could make a combined one with the a300g too.
Aug 25 18:30:04 <fullstop> the ag300 would be hard to separate from the g300nh2
Aug 25 18:30:22 <fullstop> same cpu
Aug 25 18:30:23 <Chocks> oh yes?
Aug 25 18:30:49 <fullstop> possibly different revisions, but that is certainly not a guarantee.
Aug 25 18:31:47 <Chocks> might be able to do something with flash detection, etc
Aug 25 18:31:53 <fullstop> same flash chips
Aug 25 18:32:07 <Chocks> the 300/301 was combined, I presume that does some network hardware detection
Aug 25 18:32:21 <fullstop> 300/301 are the same except for the realtek switch
Aug 25 18:32:27 <fullstop> and the realtek driver probes that now
Aug 25 18:32:32 <Chocks> right
Aug 25 18:32:42 <fullstop> that is, the rtl8366rb and rtl8366s drivers were merged
Aug 25 18:32:55 <fullstop> making the two platforms the same.
Aug 25 18:32:56 <Chocks> the a300g has more RAM, so that's something else
Aug 25 18:33:38 <fullstop> I don't think that most people want to combine images, though.
Aug 25 18:34:00 <fullstop> The heuristics of detection would become a nightmare.  :D
Aug 25 18:38:03 <Chocks> well yes.  I don't think it's too bad; but of course you need to add something every time they change hardware.  as usual.
Aug 25 18:44:35 <fullstop> I wouldn't be happy if the G300NH2 added 64MB of ram and it was suddenly detected as an AG300
Aug 25 18:45:11 <Chocks> haha
Aug 25 18:45:16 <Chocks> that's why we have testing
Aug 25 18:45:45 <Chocks> ok, so I want to add an init script, but I don't want the server to start by default
Aug 25 18:46:33 <Chocks> service
Aug 25 19:43:32 <luka12345|wiik> anyone know what is environment variable to use inside Makefile for this kind of string 'mips-openwrt-linux-gnu'
Aug 25 19:44:21 <luka12345|wiik> it is also passed when running configure script
Aug 25 19:44:21 <luka12345|wiik> checking host system type... mips-openwrt-linux-gnu
Aug 25 19:44:21 <luka12345|wiik> checking target system type... mips-openwrt-linux-gnu
Aug 25 19:59:32 <blogic> hi luka12345|wiik
Aug 25 20:04:56 <luka12345|wiik> blogic: hi
Aug 25 20:05:13 <luka12345|wiik> im updating Makefile for pjsip :)
Aug 25 20:06:20 <luka12345|wiik> i found out that $(GNU_TARGET_NAME)=mips-openwrt-linux
Aug 25 20:06:41 <luka12345|wiik> but pjsip adds -gnu internaly...
Aug 25 21:15:52 <blogic> i will commit my fixes tomorrow
Aug 25 21:16:13 <blogic> and mail you the owsip makefile
Aug 25 21:44:21 <luka12345|wiik> blogic: ok, great
Aug 25 21:45:01 <luka12345|wiik> i have compiled pjsua from pjsip package
Aug 25 21:45:38 <luka12345|wiik> but when i try to run it i get an error: /bin/ash: pjsua: not found
Aug 25 21:46:34 <luka12345|wiik> but the file is there: 1.3M Aug 25  2011 /usr/bin/pjsua
Aug 25 21:46:59 <russell--> weird, getting parallel build errors on x86 build, but -j1 seems to progress just fine.  hard to figure out where the problem is.
Aug 25 21:47:19 <luka12345|wiik> strace is also showing that; could it be because i dont have enough free ram
Aug 25 21:48:00 <luka12345|wiik> russell--: what happens when you run parallel build again?
Aug 25 21:48:08 <russell--> it dies again
Aug 25 21:48:24 <russell--> (i tried that first! ;-)
Aug 25 21:49:10 <luka12345|wiik> hmm; my parallel build issues were to wrong dependencies...
Aug 25 21:49:28 <luka12345|wiik> but running it again would fix it
Aug 25 21:50:46 <russell--> yeah, i've had that work before too
Aug 25 21:51:21 <russell--> i'm doing an 'select all packages' build, so there are lots of candidates
Aug 25 21:51:51 <russell--> i have a little one-liner that greps through compile.txt files looking for 'error' strings that usually helps
Aug 25 21:52:18 <russell--> but this time, nothing jumped out as something that would block progress
Aug 25 21:53:11 <russell--> clearly, i haven't exhaustive explored the boundaries of the problem, just noticed it
Aug 25 21:53:24 <luka12345|wiik> well, you could try to compile one package at the time with -j<n>
Aug 25 21:53:44 <luka12345|wiik> and see wich one will fail
Aug 25 21:58:53 <russell--> i think i'll just be patient and wait to see if my -j1 will complete, then my immediate problem is solved.  i suspect some dependency error, but it's a heisenbuggy proposition to try to track down
Aug 25 21:59:50 <russell--> i can still parallel in an afternoon nap
Aug 25 22:00:10 <russell--> sleep while building++
Aug 25 22:20:26 <WillieNL> russell--: parallel building on x86 resulted here in a messed up build order in the toolchain stages.
Aug 25 22:20:47 <WillieNL> without -jX everything works fine.
Aug 25 23:00:22 <CIA-89> jmccrohan * r28088 /packages/utils/lcd4linux/Makefile: [package] lcd4linux: bump to r1158
Aug 25 23:20:31 <russell--> WillieNL: I was past the toolchain stuff
Aug 25 23:21:43 <russell--> interesting
Aug 25 23:21:46 <russell--> checking whether we are cross compiling... configure: error: in `/aux/src/openwrt/build_dir/host/grub-0.97':
Aug 25 23:21:49 <russell--> configure: error: cannot run C compiled programs.
Aug 25 23:21:52 <russell--> If you meant to cross compile, use `--host'.
Aug 25 23:22:21 <WillieNL> russell--: hmm then i dunno but i have seen a lot weird issues that fade away by stop using -jX on the openwrt trunk :)
Aug 25 23:22:53 <russell--> r28078
Aug 25 23:29:03 <russell--> something about grub isn't being liked, i also just sync'd up
Aug 25 23:29:07 <russell--> trying again
Aug 25 23:31:34 <WillieNL> stage1/2 mismatch ?
Aug 26 02:38:44 <russell--> hmm.  make package/grub/clean doesn't remove build_dir/host/grub-0.97 like maybe you'd expect it to
Aug 26 02:40:28 <russell--> also, not finding asm/errno.h
Aug 26 02:44:31 <russell--> this seems to be for the host portion
**** ENDING LOGGING AT Fri Aug 26 02:59:57 2011