**** BEGIN LOGGING AT Tue Apr 17 02:59:56 2007 Apr 17 09:21:54 kaloz * r6981 /trunk/ (6 files in 3 dirs): add support for the Compex WP18 and NP18A Apr 17 09:25:06 kaloz * r6982 /trunk/target/linux/ixp4xx-2.6/Makefile: disable jffs2 support for ixp4xx Apr 17 11:59:52 <[florian]> sn9: ok, will have a test Apr 17 12:46:53 florian * r6983 /trunk/target/linux/adm5120-2.6/files/arch/mips/pci/pci-adm5120.c: Re-enable PCI subsystem :) Apr 17 12:47:40 florian * r6984 /trunk/target/linux/brcm63xx-2.6/files/arch/mips/pci/pci-bcm96348.c: Re-enable PCI subsystem here too :)) Apr 17 12:49:08 [florian]: any improvement from the new patch? Apr 17 12:53:41 <[florian]> sn9: not yet tried, was doing some adm5120 hacking, I had pending patches for it for some days Apr 17 12:53:49 <[florian]> sn9: will do right now Apr 17 12:55:16 test both jffs2 and squashfs images with the patch, but commit the change if squash still works, regardless of jffs2 Apr 17 12:56:37 ram gets wasted without it Apr 17 13:03:13 if you're waiting on a build, i have ready images to test Apr 17 13:24:06 <[florian]> have to go doing something, we will be back later Apr 17 15:32:04 <[florian]> sn9: I am back Apr 17 15:32:21 ok Apr 17 15:35:06 <[florian]> sn9: which erase size do you use ? 64k still does not work :/ Apr 17 15:35:32 define "use" Apr 17 15:35:56 <[florian]> which image have you tested ? Apr 17 15:37:28 i have not tested yet, as i still have not satisfactorily repaired the serial console on my device, and i was asleep by the time Sargun got back to irc, so he didn't have the opportunity to test the new image, either Apr 17 15:37:57 <[florian]> ok, because it shows the exact same bootlog as before Apr 17 15:38:08 but my device has a 64k size Apr 17 15:38:57 if the squashfs image still works, commit the change (you may leave out the addition of jffs2 until it works) Apr 17 15:39:22 <[florian]> humm, no, I won't commit the change, we should have device specific base-files instead of copying them that way Apr 17 15:39:50 it's just a symlink Apr 17 15:41:15 one of the copyings is thereby removed Apr 17 15:45:02 <[florian]> I will have a test at rt61 Apr 17 15:46:03 back when i got 2.4 to finally boot, ethernet did not work properly. does it now? Apr 17 15:46:20 <[florian]> yes it does Apr 17 15:46:41 <[florian]> the ethernet driver also implements NAPI, which is pretty good tto reduce the cpu work Apr 17 15:47:31 the makefile without the latest big.patch causes a bad waste of ram Apr 17 15:47:52 because all the tmpfs things get mounted twice Apr 17 15:48:26 <[florian]> there is a common problem with every target which do not have base-files, is that hotplug litterally takes all the ram Apr 17 15:48:29 <[florian]> and goes crazy Apr 17 15:48:52 the big.patch eliminates the extra hotplug call Apr 17 15:49:50 <[florian]> sn9: about the gpio line, did you manage finding it ? no hurry Apr 17 15:50:28 i could go get the device and follow the traces again, with the SoC datasheet as reference Apr 17 15:51:03 if you have a good eye and a light at the correct angle, you can do the same Apr 17 15:54:23 i'd rather not harp on this, but i consider replacing that line in the makefile to be an important bugfix Apr 17 16:00:25 florian * r6985 /trunk/target/linux/brcm63xx-2.6/ (8 files in 4 dirs): brcm63xx cleanups; add a watchdog driver, indent serial console driver, add support for another sst flash Apr 17 16:05:18 <[florian]> the lzma stuff is a good thing, now you can always reflash using the internal tftpd server Apr 17 16:09:14 yes Apr 17 16:09:21 <[florian]> nice, got wifi working as well :) Apr 17 16:10:21 so then that line change in the makefile would make squashfs perfect Apr 17 16:11:35 florian * r6986 /trunk/package/rt61/Makefile: rt61 is not broken on 2.6 kernels, use a daily snapshot instead Apr 17 16:13:48 <[florian]> will do it, thinking of an alternative way of doing it Apr 17 16:15:17 it'll be easier to think of an alternative way when introducing gpio-related stuff to the userspace Apr 17 16:15:30 all in due time Apr 17 16:18:23 i now suspect that the jffs2 problem might have something to do with the root image being so much smaller than the partition, since eventually, the missing symbol problem on 2.4 was releagted to manifesting only in the jffs2 images Apr 17 16:44:06 [florian]: do you think that's possible? Apr 17 16:45:50 florian * r6987 /trunk/package/ppp/patches/205-pppoe_iface_name.patch: Fix the device names too long (#1591) Apr 17 16:49:20 florian * r6988 /trunk/target/linux/rdc-2.6/files/drivers/net/r6040.c: Suppress the useless ioctl function Apr 17 17:29:02 [florian]: aww, c'mon -- makefile update? Apr 17 18:13:19 <[florian]> sn9: wait Apr 17 18:13:27 ok... Apr 17 18:14:43 <[florian]> do you have any rdc device with usb ? Apr 17 18:15:27 i'm not aware of the existence of such a device, but they are possible according to the datasheet Apr 17 18:19:35 what did you have in mind? Apr 17 18:22:22 <[florian]> enable usb support, and see whether it uses an existing hcd driver or something custom Apr 17 18:22:42 the datasheet should say Apr 17 18:23:37 i would not expect it to be custom Apr 17 18:23:55 it is x86, after all Apr 17 18:32:20 * sn9 wonders what the "wait" is about... Apr 17 18:38:44 florian * r6989 /trunk/target/linux/rdc-2.6/ (Makefile config/default): More stuff enabled in config, fix squashfs image, thanks sn9 Apr 17 18:39:25 thank you Apr 17 18:42:19 does the "reboot" cmd at the shell work? it hung on 2.4 Apr 17 19:05:22 <[florian]> it works, yes, as well as halt Apr 17 20:19:44 anyone familiar with __strtol_internal? I'm runing into a bit of a problem with it and python Apr 17 22:14:34 florian * r6990 /trunk/package/rt61/patches/: Patches for rt61 are no longer required Apr 17 22:15:51 florian * r6991 /trunk/target/linux/adm5120-2.6/ (4 files in 3 dirs): Move NAND driver from rb1xx to adm5120, rbmipsnand fixed by David Goodenough, thanks ! Apr 17 22:16:29 florian * r6992 /trunk/target/linux/rb1xx-2.6/: Remove this target since rb-2.6 will build both rb532 and adm5120 later on Apr 18 00:44:49 nico * r6993 /packages/net/lighttpd/Makefile: update lighttpd to 1.4.15 Apr 18 00:46:40 nico * r6994 /packages/lang/lua/Makefile: update lua to 5.1.2 Apr 18 00:53:24 nico * r6995 /packages/lang/ (11 files in 5 dirs): replace sablevm and sablevm-classpath with sablevm-sdk, add a sablevm-mini package containing a minimal set of classes and a working example Apr 18 01:06:08 nico * r6996 /packages/net/freeradius/ (14 files in 3 dirs): update freeradius to 1.0.6 (closes: #1331), rename patches to match our naming convention, available dictionnaries are now handled from the Makefile (no need for a dict patch anymore) **** ENDING LOGGING AT Wed Apr 18 02:59:56 2007