**** BEGIN LOGGING AT Tue Jul 12 02:59:58 2016 Jul 12 03:38:07 build #246 of sunxi is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/sunxi/builds/246 Jul 12 06:15:19 nbd: my wife is complaining, and i see the device re-authorizing more than it should with less connection uptime. allegedly according to my assessment it's still working though so i guess that's progress Jul 12 11:47:17 build #301 of x86 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/301 Jul 12 12:28:38 build #272 of brcm47xx.mips74k is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx.mips74k/builds/272 Jul 12 14:04:17 nbd: mission failed. Jul 12 14:26:05 the tp-link hardware ID and model ID are used by the firmware check upgrade on factory images? dd if=/dev/mtd5 bs=4 count=1 skip=16 2>/dev/null | hexdump -v -n 4 -e '1/1 "%02x" Jul 12 14:29:04 hojuruku: also on sysupgrade images and by the bootloader on every boot IIRC Jul 12 14:35:09 KanjiMonster: great so if I don't have a firmware downloadable image for my router i can still work out the hwid - nice Jul 12 14:35:35 i think i'll see if i can upgrade my router throught he factory interface now.. i don't even have to patch any of the image building utils, as the hwid is passed over the command line right? Jul 12 14:36:13 probably, haven't tried adding support for any tp-link devices yet Jul 12 16:07:07 http://www.infowars.com/chinese-president-orders-peoples-liberation-army-to-prepare-for-combat/ WOOT. Lede could get done with dealing with the enemy like George Bushes's grandfather herbert walker bush. Only joking. Jul 12 16:38:40 * distemper reads "infowars" and thinks "what happened to this channel?" Jul 12 16:39:29 i sorry i should have used a drudge report link Jul 12 16:40:13 another day in ground zero just kidding, now back to pulling my tp-link hwid out of a uboot serial dump Jul 12 16:44:47 want another? just for laughs! http://www.infowars.com/former-prosecutor-sues-obama-al-sharpton-for-inciting-a-race-war/ You know what that proves? Charles Manson was a visionary! https://en.wikipedia.org/wiki/Helter_Skelter_(Manson_scenario) My first born child is half-african american, but never saw due to the actions of that XXX XXX XXX (3 banned adjectives) Christopher Kent Bowersox. Jul 12 16:48:20 is it possible to use USB storage to compensate for the lack of Flash? Jul 12 16:48:20 the "firmware" partition is everything but the uboot at the art ... ah ok. Jul 12 16:49:10 JyZyXEL: yes, put an ext4 partition on your usb, set up a mountpoint, set up extra path and LD_LIBRARY_PATH in /etc/profile.d - add an installation directory to /etc/opkg.conf Jul 12 16:49:36 JyZyXEL: and off you go, install heaps of stuff to opkg -d usb package Jul 12 16:50:07 hojuruku: first i probably need to add support for usb-storage and ext4 in make menuconfig Jul 12 16:51:37 or i probably should use something like f2fs instead? Jul 12 16:52:15 JyZyXEL: YOU DONT USE A MTD filesystem on a BLOCK device. You know what's on a USB stick? not just flash Jul 12 16:53:09 a little microprocessor - usb 3.0 devices are actually using low end ARMs - and software on it to do FTL to map the physical writes to the logical disk that you think you are using. It does the wear levelling. Jul 12 16:53:44 f2fs driver might be smaller than the ext4 Jul 12 16:54:03 why would you want to have a wear levelling filesystem, on a wear leavelling usb device. This reminds me of a guy who thought if I have 6 different virus scanning programs my computer will be more safe.... it didn't end well for his windows install. Jul 12 16:54:17 just simply because of the size Jul 12 16:54:47 JyZyXEL: i hate ext4, I prefer ext2, but for some reason openwrt doesn't package ext2 like they should. you can mount an ext2 filesystem without a journal with the ext4 fs driver. Jul 12 16:55:16 sure, but still the kernel size is what i care about the most Jul 12 16:55:54 bcachefs might be another tight alternative Jul 12 16:55:55 yeah F2FS is ok - samesung uses it on andriod and onenand a lot. Jul 12 16:56:05 Regardless of the sizes of various drivers, you shouldn't use a file system intended for MTDs on a regular disk device. Jul 12 16:56:14 i hear bcache has an awesome tight codebase Jul 12 16:57:17 f2fs is not a MTD filesystem it is a good choice for what you are trying to do. I RTFM. it's a filesystem optimized for the characteristics of NAND Jul 12 16:58:55 f2fs works fine everwhere, especially in low-memory conditions Jul 12 16:59:15 kind of like how the raspberry pi installer noobs does it's disk format :P i read something from linspire explaining in detail the best block sizes / alignments to use for SD cards etc. Jul 12 16:59:31 im just assuming that f2fs is smaller than ext4 Jul 12 17:01:42 https://developer.ridgerun.com/wiki/index.php/High_performance_SD_card_tuning_using_the_EXT4_file_system EXT4 on flash - how to tune Jul 12 17:04:15 ext4 kernel is 104439 bytes bigger than f2fs Jul 12 17:04:27 http://3gfp.com/wp/2014/07/formatting-sd-cards-for-speed-and-lifetime/ this is interesting too. The microprocessors are tune to 4MB block sizes, and it's the microprocessor bundled with the flash that affects flash performance more than flash itself. wow. Jul 12 17:05:18 even sd cards have ram and a write cache :P Jul 12 17:05:53 so im guessing that pretty much everything that is not in the "Kernel modules --->" menu, can be later installed via opkg on to the USB storage, correct? Jul 12 17:14:19 kmods too, as long as they're matching your running version Jul 12 17:15:55 i wonder if i should tag everything i want to the USB storage as: in menuconfig Jul 12 17:20:12 then maybe i could somehow conveniently install all that stuff on to the usb storage Jul 12 17:49:11 are the archer c5's based on arm working yet? Jul 12 17:53:55 what's a good router with hardware crypto? Jul 12 17:55:13 x86_64 Jul 12 18:11:28 haha Jul 12 18:11:50 what's the status of the tp-link archers with arm? Jul 12 19:40:30 https://forum.openwrt.org/viewtopic.php?pid=331331#p331331 cooking up openwrt on virgin hardware :P notice the hardware ID Jul 12 19:45:20 build #246 of mpc85xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/mpc85xx/builds/246 Jul 12 19:45:47 [Tue 2016-07-12 09:54:48 AM PDT] JyZyXEL: i hate ext4, I prefer ext2, but for some reason openwrt doesn't package ext2 like they should. you can mount an ext2 filesystem without a journal with the ext4 fs driver. <------- and you can also format ext4 without a journal Jul 12 19:46:14 DonkeyHotei: i know but busybox can't , hence why you format ext2 to get the same job done :) Jul 12 19:47:09 DonkeyHotei: refering to my patch for procd to use lz4 ext4 fs without journal formatted with busybox mkfs.ext2 - much lighter on the roms than the old way. Jul 12 19:48:38 i hear good things about bcachefs, it might be a nice alternative choice Jul 12 20:21:04 hummm.... Jul 12 20:21:33 i'm not going to make a lede version for this new vietnamese tp-link router with different hardware ID and 8mb of flash, openwrt proper should work... and have it set up with vietnamese luci :P Jul 12 20:21:42 revenge is sweet :P Jul 12 20:22:11 or they can unban me Jul 12 20:22:31 that's rather petty Jul 12 20:22:55 yeah but it's not hard for them to add 6 lines to a makefile to make it work. Jul 12 20:26:05 build #253 of adm8668 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/adm8668/builds/253 Jul 12 20:28:54 the vietnamese userbase hasn't heard of lede, lets keep it that way :P Jul 12 20:35:29 hojuruku: that's a small revenge and it's quite pity - lede can still pick your patches and include in their tree if they wish Jul 12 20:36:44 yeah i know. openwrt does linux hardware config different i know. Jul 12 21:21:37 build #356 of realview is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/realview/builds/356 Jul 12 21:24:45 how does /proc/cpuinfo detect hardware? Jul 12 21:44:02 in a lot of cases, it's just from kernel command parameters Jul 12 21:44:35 i know if you take a ubiquity picostation, flash it with a nanostation build, it'll work, and report itself as a nanostation in cpuinfo Jul 12 21:44:38 and vice versa Jul 12 21:44:49 linux/ar71xx/files/drivers/mtd/tplinkpart.c parses the tp-link kernel header Jul 12 21:45:02 now where is the machine type chosen that goes into cpuinfo Jul 12 21:54:48 look in the mach-**** file Jul 12 21:54:53 for stuff like this Jul 12 21:55:04 MIPS_MACHINE(ATH79_MACH_TL_WR2543N, "TL-WR2543N", "TP-LINK TL-WR2543N/ND", Jul 12 21:55:04 tl_wr2543n_setup); Jul 12 21:55:31 i believe the unravelling of all that stuff begins there Jul 12 21:55:52 that snip is from mach-tl-wr2543n.c Jul 12 21:57:24 yeah i've got the same Jul 12 21:57:41 i want to know where the kernel looks up the tp-link hardware Id from the MTD partitions Jul 12 21:58:13 I believe it's handled differently for different boards Jul 12 21:58:30 and some of them rely entirely on command line parameters to sort it out Jul 12 22:05:28 i have a tp-841nd-v9 with a different hardware ID and a different flash size Jul 12 22:10:53 __setup("board=", mips_machtype_setup); oh i got it. when openwrt makes the firmware it appends kernel arguments Jul 12 22:11:15 e.g. on my existing router board=TL-WR941ND console=ttyS0,115200 rootfstype=squashfs,jffs2 noinitr Jul 12 22:11:49 we'll that's easy Jul 12 22:37:10 build #370 of brcm47xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/370 Jul 12 23:30:10 does openwrt's ssl use http://cryptodev-linux.org/ or only OCF? Jul 12 23:30:31 yah, sorry, thought I was clear on that, it's decided in command line, and command line comes from that stuff I was pasting Jul 12 23:30:45 * groz been away for a bit Jul 13 00:36:59 build #369 of brcm47xx.legacy is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx.legacy/builds/369 Jul 13 02:34:37 build #364 of bcm53xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/bcm53xx/builds/364 **** ENDING LOGGING AT Wed Jul 13 02:59:58 2016