**** BEGIN LOGGING AT Sun Jan 31 02:59:57 2010 Jan 31 03:05:22 ping rtz Jan 31 03:08:14 cshore: pung Jan 31 03:08:16 o Jan 31 03:08:30 rtz: have you seen the hotplug patch I did Jan 31 03:09:14 rtz: if so, any comments? I know ede had some Jan 31 03:09:27 cshore: no Jan 31 03:09:29 rtz: trying to figure out how to proceed next Jan 31 03:09:34 cshore: where can I find it? Jan 31 03:09:46 rtz: are you on the -devel list? Jan 31 03:09:54 nope Jan 31 03:10:17 let me take a look at it Jan 31 03:11:07 http://openwrt.pastebin.ca/1772705 Jan 31 03:11:58 cshore: Hotplug automount fixes/enhancements? Jan 31 03:12:04 rtz: yes Jan 31 03:21:07 cshore: I admit I'm not very good at reviewing stuff I'm not directly involved with Jan 31 03:21:16 rtz: ok Jan 31 03:21:33 cshore: but besides the modularisation, I don't see anything too obvious Jan 31 03:21:46 rtz: good...I did do a fair bit of testing Jan 31 03:22:27 rtz: I'm trying to figure out how to modularize the swap bits...the problem is we want to try swap before a regular mount Jan 31 03:22:48 cshore: well, there is one thing Jan 31 03:22:53 rtz: oh? Jan 31 03:23:12 cshore: I don't completly understand the /etc/fstab versus /tmp/fstab stuff Jan 31 03:23:47 rtz: /etc/fstab is where mount looks, but in openwrt we use /tmp/fstab (because we don't want to write to flash), so it's a symlink Jan 31 03:24:23 cshore: and you update it for in case a device gets added Jan 31 03:24:29 rtz: yes Jan 31 03:25:04 rtz: particularly with the uuid stuff we can't know the device beforehand Jan 31 03:25:09 rtz: and label Jan 31 03:25:37 hmm, ok Jan 31 03:25:50 rtz: openwrt doesn't really use regular fstab...it's mostly for compatilbility Jan 31 03:26:02 I see Jan 31 03:28:40 rtz: for /etc/init.d/fstab and the automount stuff, I know how to do it without /etc/fstab at all Jan 31 03:29:53 rtz: but the modularization is an interesting problem Jan 31 03:30:07 rtz: particularly swap Jan 31 03:30:35 rtz: do you know if it's bad to try to mount a swap partition as a regular filesystem, or will it just fail with an error? Jan 31 03:31:17 cshore: well, it's like mounting random garbabge Jan 31 03:31:29 cshore: so I think, it will give an error Jan 31 03:31:54 you could try on your desktop system Jan 31 03:32:26 rtz: well I know it will fail with an error...I just want to know if it's actually bad Jan 31 03:33:52 rtz: I want to make the swap mount stuff separate from the mount stuff for modularization, but I can only do so if swap doesn't break mount Jan 31 03:34:15 rtz: if you follow what I mean Jan 31 03:35:20 cshore: well, I'm only guessing, but I don't see how this could be a problem Jan 31 03:35:56 cshore: i wouldn't run fsck on it, but otherwise, mount gives an error, you shelf it and carry on Jan 31 03:36:36 rtz: ok....actually I think i'll man swap 10-swap and mount 20-mount Jan 31 03:36:54 man=make Jan 31 03:40:33 rtz: if there are multiple files in hotplug for rules (e.g. 10-swap 20-fsck 30-mount) will the earlier one finish before the next one is started? Jan 31 03:43:38 cshore: look at /sbin/hotplug-call Jan 31 03:43:48 cshore: I think they wait for each other Jan 31 03:44:25 rtz: ok, yes it does Jan 31 03:46:31 rtz: I think I'm going to make a /lib/mount that I use include to source...that way I don't have to manually source a particular fs type or swap Jan 31 03:50:48 that is a directory /lib/mount Jan 31 03:51:08 or /lib/functions/mount Jan 31 03:51:59 ok, the tty code is really really odd, but I managed to make my stolen code work :) Jan 31 03:52:08 rtz: heh, ok Jan 31 03:52:22 cshore: it's really funny, you know Jan 31 03:52:23 rtz: you getting close? Jan 31 03:52:28 rtz: what's that? Jan 31 03:53:04 cshore: i looked at the tty setup code from busybox init, upstart and initng, and it looks exactly the same everywhere Jan 31 03:53:09 including comments Jan 31 03:53:13 rtz: heh Jan 31 03:54:00 but yes, I'm getting close Jan 31 03:54:09 rtz: so if your idea works, will init be eliminated? Jan 31 03:54:20 I hope so Jan 31 03:54:49 rtz: how do you deal with ram usage (i.e. making sure you don't start two too big tasks?) Jan 31 03:55:20 I would like to make an optional packe out of it first, then port the preinit code, then make it mandatory and then port the init services Jan 31 03:55:39 rtz: makes sense Jan 31 03:56:32 cshore: it's not that bad with normal services with regard to ram, but i'm not sure yet, how to deal with the big network and firewall init scripts Jan 31 03:56:57 probably split them a bit and check ram usage before a task is started Jan 31 03:58:05 rtz: make free ram a dependency? Jan 31 03:58:25 maybe Jan 31 03:58:42 I have to check more carefully how much ram is actually needed Jan 31 03:58:47 rtz: right Jan 31 03:59:16 rtz: maybe in the config, specify anticipated max ram Jan 31 03:59:39 rtz: and if the current max rams would be too big, make the process wait Jan 31 04:00:10 we will see Jan 31 04:00:15 rtz: otherwise you might have free ram now but while executing ram gets exhausted Jan 31 04:02:22 rtz: 16M = ~12MB free is pretty small Jan 31 04:22:01 build #2 of ifxmips is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ifxmips/builds/2 Jan 31 04:24:32 rtz: what do you think of making swap not a separate type of uci config, but chosen by specifiying fstype = swap? Jan 31 04:25:32 sounds actually better to me Jan 31 04:26:04 but i'm not sure what to do about backward compatibility Jan 31 04:26:11 rtz: right Jan 31 04:28:28 ok, I'm off to bed Jan 31 04:28:51 rtz: sleep tight Jan 31 08:26:01 anyone around with a WRT600N? Jan 31 08:37:16 swalker: do you have commit access...I have a bugfix patch for atheros for preinit Jan 31 08:39:07 cshore: yes, but only for packages afaik Jan 31 08:40:20 swalker: ah, ok, well I guess the patch will have to wait Jan 31 10:15:52 Hello everyone Jan 31 10:16:01 suro: hello Jan 31 10:16:19 Is this the right place to talk about adding support for new devices? Jan 31 10:16:48 suro: yes, but I'm not the one to talk with about that...not sure if anyone else is here right now Jan 31 10:17:08 unless it's a brcm63xx device Jan 31 10:17:39 looks like I'm lucky: BCM6358 ) Jan 31 10:17:44 heh Jan 31 10:17:50 what kind of router? Jan 31 10:18:36 It's a 3G/wifi router Huawei b560. sold by Orange Telecom under name "Flybox" Jan 31 10:19:09 do you have a firmware image? Jan 31 10:20:19 Yes, I was able to dump the image from it, but wasn't able to mount/read the sqfs Jan 31 10:20:57 hmmm....actually meant a firmware image file (.e.g for an upgrade) but that will work Jan 31 10:21:41 can you ftp it to ftp.cshore.com in the Incoming directory? Jan 31 10:21:46 sore Jan 31 10:21:49 sorry Jan 31 10:22:03 ftp to cshore.is-a-geek.com Jan 31 10:22:07 also in the docs there was a notice saying if I want the sources I can email them and if I agree to pay for writing and delivering a CD they can send it Jan 31 10:22:25 will do it Jan 31 10:22:37 i mean ftp.. Jan 31 10:25:54 suro: any difficulties? I know it's working for downloads as it's been used, but not sure about uploads Jan 31 10:27:50 suro: btw unless you extracted the squashfs from the mtd, you probably got the imagetag header in the image as well (which is in fact what I want) Jan 31 11:53:31 build #2 of iop32x is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/iop32x/builds/2 Jan 31 12:09:54 lars * r19437 /trunk/target/linux/xburst/files-2.6.32/drivers/rtc/rtc-jz4740.c: [xburst] rtc-jz4740: reset rtc clock if it has lost its state Jan 31 12:18:23 nbd * r19438 /trunk/package/mac80211/Makefile: mac80211: i maintain this package Jan 31 12:24:57 nbd * r19439 /trunk/package/ (hostapd/Makefile madwifi/Makefile wprobe/Makefile): add me as a maintainer for a few other packages Jan 31 12:48:39 juhosg * r19440 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: convert to use netdev_ops Jan 31 12:48:42 juhosg * r19441 /trunk/target/linux/ramips/Makefile: ramips: switch to 2.6.32 Jan 31 12:48:45 juhosg * r19442 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: pass 'raeth_priv' struct directly to dma specific functions Jan 31 12:48:47 juhosg * r19443 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: simplify tx descriptor initialization Jan 31 12:48:50 juhosg * r19444 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: simplify tx_next computation Jan 31 12:48:52 juhosg * r19445 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: remove unnecessary typecasts Jan 31 12:48:55 juhosg * r19446 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: add helper functions to {ed,dis}able interrupts Jan 31 13:46:19 kaloz * r19447 /trunk/target/linux/ixp4xx/patches-2.6.32/191-cambria_optional_uart.patch: refresh patch Jan 31 13:46:54 kaloz * r19448 /trunk/target/linux/ixp4xx/patches-2.6.32/ (190-cambria_support.patch 300-avila_fetch_mac.patch): [ixp4xx] fixup cambria/avila model detection Jan 31 13:47:16 kaloz * r19449 /trunk/target/linux/ixp4xx/Makefile: [ixp4xx] switch to 2.6.32 Jan 31 13:50:34 kaloz * r19450 /trunk/target/linux/ixp4xx/ (5 files in 3 dirs): [ixp4xx] remove support for older kernels Jan 31 14:48:48 swalker * r19451 /packages/ (libs/libgphoto2/Makefile multimedia/gphoto2/Makefile): [packages] libgphoto2, gphoto2: update to 2.4.8 Jan 31 14:59:38 hauke * r19452 /trunk/target/linux/amazon/ (config-2.6.21 patches/): [amazon] remove support for kernel 2.6.21 Jan 31 15:06:19 hauke * r19453 /trunk/target/linux/amazon/ (8 files in 7 dirs): [amazon] move files to correct position Jan 31 15:07:52 hauke * r19454 /trunk/target/linux/amazon/ (5 files in 3 dirs): [amazon] Apply patch 200-fix_deprecated_interrupt_definations.patch directly Jan 31 15:08:29 hauke * r19455 /trunk/target/linux/amazon/ (2 files in 2 dirs): [amazon] Apply patch 210-remove_unnedded_variables.patch directly Jan 31 15:09:17 hauke * r19456 /trunk/target/linux/amazon/ (3 files in 2 dirs): [amazon] Apply patch 220-fix_timer.patch directly Jan 31 15:09:50 hauke * r19457 /trunk/target/linux/amazon/ (2 files in 2 dirs): [amazon] Apply patch 240-irq_fix.patch directly Jan 31 15:13:15 hauke * r19458 /trunk/target/linux/amazon/ (12 files in 5 dirs): [amazon] Add kernel 2.6.32 support Jan 31 15:13:53 hauke * r19459 /trunk/target/linux/amazon/files/ (3 files in 2 dirs): i[amazon] use SYS_HAS_EARLY_PRINTK instead of prom_printf Jan 31 15:14:16 hauke * r19460 /trunk/target/linux/amazon/files/ (4 files in 4 dirs): [amazon] export amazon_get_cpu_hz Jan 31 16:10:44 jow * r19461 /trunk/target/linux/atheros/base-files/lib/preinit/15_preinit_iface_atheros: [atheros] fix syntax error in preinit hook (#6610) Jan 31 17:25:17 nbd: Thanks for merging my patch and removing the rought edges! Also, kudos for your progress with running android on top of OpenWrt. Jan 31 17:28:11 i wish i had the time to complete it Jan 31 17:28:28 but right now wifi driver development is eating up almost all of my hacking time ;) Jan 31 17:54:29 Actually, we might have a (paid) project in the not-too-distant future where being able to run Android on top of OpenWrt on some existing phone hardware would be extremely useful. Would you be available for doing OpenWrt development on subcontracting basis? Jan 31 17:55:56 i don't know if i'll have the time for it Jan 31 17:56:32 juhosg * r19462 /trunk/target/linux/ramips/files/ (3 files in 3 dirs): ramips_eth: pass sys_clk via platform_data Jan 31 17:56:35 juhosg * r19463 /trunk/target/linux/ramips/files/drivers/net/ramips_eth.h: ramips_eth: update register offsets Jan 31 17:56:37 juhosg * r19464 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips_eth: cleanup mac_address changing Jan 31 17:56:40 juhosg * r19465 /trunk/target/linux/ramips/rt288x/config-2.6.32: ramips: enable RTL8366SR driver for the RT288x devices Jan 31 17:56:42 juhosg * r19466 /trunk/target/linux/ramips/ (patches-2.6.30/ rt288x/config-2.6.30 rt305x/config-2.6.30): ramips: nuke 2.6.30 support Jan 31 17:56:46 juhosg * r19467 /trunk/target/linux/ar71xx/ (64 files in 3 dirs): ar71xx: nuke 2.6.3[01] stuff Jan 31 17:56:53 I'll drop you an email, in any case. Jan 31 17:57:11 When it materializes, that is. Jan 31 17:59:17 sounds good Jan 31 17:59:18 juhosg * r19468 /trunk/target/linux/ar71xx/ (patches-2.6.30/ patches-2.6.31/): ar71xx: nuke remaining patches Jan 31 18:46:21 ping nbd Jan 31 18:50:21 pong Jan 31 18:50:51 Did you see my automount post to the devel list? Jan 31 18:51:33 yes, but i didn't have time to read it yet Jan 31 18:51:37 nbd: and if so, other than modularization (which I'm now working on), what did you think Jan 31 18:51:40 ah, ok Jan 31 18:51:49 i'll see if i can find the time to review it later Jan 31 18:52:03 thanks Jan 31 18:52:16 btw, I noticed you were talking about devtmpfs yesterday Jan 31 18:52:35 devtmpfs is automatically loaded by the kernel if there is a rootfs= line to the kernel Jan 31 18:53:30 loaded = mounted Jan 31 18:54:42 but since we don't want it (at least past preinit), I say preinit should unmount it, once pivot is done Jan 31 18:55:05 preinit can umount it before the pivot Jan 31 18:55:13 actually just before hotplug2 does the coldplugging Jan 31 18:55:36 it's only useful so that preinit can get an open reference to /dev/console, imho Jan 31 18:58:22 yeah, the fact that it creates device nodes with wrong permissions and possibly names is kind of pointless Jan 31 18:59:18 you still need some kind of hotplug thing to get it right (or udev) Jan 31 19:12:23 BTW, anyone else seen the strange issue with the missing /dev/null node? Jan 31 19:12:55 I was a bit surprised to see a fast-growing regular file in its place. Jan 31 19:15:52 what platform? Jan 31 19:46:16 yes I see it too Jan 31 19:50:24 It appears hotplug isn't creating /dev/null during coldplug Jan 31 20:02:07 nope, that's not the problem Jan 31 20:02:25 it's a silly thing I did in preinit...I used /dev/null before it was created Jan 31 20:05:08 http://openwrt.pastebin.ca/1773430 solves it Jan 31 20:23:40 <[florian]> cshore: I am testing 2.6.32 on brcm63xx riht now Jan 31 20:23:56 [florian]: cool Jan 31 20:24:29 <[florian]> cshore: I did more splitting of the patches so that it should be easier to separate things and get them mainline later Jan 31 20:24:46 [florian]: sounds good Jan 31 20:24:57 [florian]: did you see the note about /dev/null? Jan 31 20:25:37 [florian]: how much of 63xx is in mainline now? Jan 31 20:26:12 <[florian]> cshore: as of 2.6.33 everything we have in openwrt, except the recent leds/button stuff is mainline, as well as spi Jan 31 20:26:25 <[florian]> cshore: but since spi is far from fully funtionnal, it does not really matter Jan 31 20:26:37 [florian]: very cool Jan 31 20:27:41 [florian]: have you thought about the led/reset button module any more, yet? Jan 31 20:31:53 <[florian]> cshore: saw the /dev/null now Jan 31 20:34:35 [florian]: /dev/null isn't being created, so there's a (growing) regular file in it's place Jan 31 20:35:10 [florian]: because I foolishly sent output to /dev/null before the node was created in preinit Jan 31 20:46:40 <[florian]> cshore: I will commit that afterwards Jan 31 20:49:26 [florian]: nice Jan 31 20:49:34 [florian]: ok, thanks Jan 31 20:50:22 danage: [florian]'s been really good at getting the kernel stuff in shape for mainline Jan 31 20:51:14 [florian]: did you find time to review my rdc patches? Jan 31 20:52:02 <[florian]> rtz: yup, have not fully tested it yet Jan 31 20:52:39 [florian]: any changes necessary for the pci driver? Jan 31 20:54:24 <[florian]> rtz: not really Jan 31 20:55:32 [florian]: ok Jan 31 20:56:10 rtz: hey, how goes it? Jan 31 20:56:22 builder: dane Jan 31 20:56:26 builder: dance Jan 31 20:56:27 0-< Jan 31 20:56:29 0-/ Jan 31 20:56:30 0-\ Jan 31 20:56:41 [florian]: could you remove the comment for the the wdt timeout at the top of the file before commiting? Jan 31 20:56:44 it's wrong Jan 31 20:56:55 [florian]: or should I post a new patch Jan 31 20:57:18 <[florian]> rtz: it will be fine Jan 31 20:57:26 cshore: it goes Jan 31 20:57:44 cshore: but I can't test easily :/ Jan 31 20:57:59 nbd * r19469 /trunk/package/mac80211/patches/540-rx_mgmt_handling.patch: mac80211: fix aggregation setup on wds ap interfaces Jan 31 20:58:04 nbd * r19470 /trunk/package/mac80211/patches/550-radiotap_ps_buffer.patch: mac80211: fix rekeying in ap mode with connected powersave stations Jan 31 20:58:26 rtz: why, not enough hardware? Jan 31 20:59:34 cshore: well, the difference between the Visual Studio debugger I used for the core and printf is really big Jan 31 20:59:59 rtz: heh Jan 31 21:00:48 [florian]: did you have time to look at the phy stuff? Jan 31 21:00:53 florian * r19471 /trunk/target/linux/brcm63xx/ (153 files in 37 dirs): [brcm63xx] move files to files-2.6.30, to ease newer kernel integration Jan 31 21:00:56 <[florian]> rtz: not yet, the patch is much bigger Jan 31 21:01:21 [florian]: actually, it's only a bunch of files :/ Jan 31 21:01:23 florian * r19472 /trunk/package/base-files/files/lib/preinit/20_device_fs_mount: Jan 31 21:01:23 [package] /dev/null was used before it was created, so an Jan 31 21:01:23 empty file was created. This prevented the actual device node from Jan 31 21:01:23 being created. The attached patch fixes that, patch from cshore Jan 31 21:02:01 <[florian]> rtz: I know, it's not that "direct" for me to review it though ;) Jan 31 21:02:18 [florian]: no problem Jan 31 21:02:28 [florian]: it's very incomplete Jan 31 21:02:52 [florian]: I'm only interested, if I should go ahead with it or shelf it Jan 31 21:02:58 rtz: printf debugging isn't so bad once you're used to it, but I like to have a macro, so that I can turn debugging on or off and leave the printf's in the code rather than constantly removing/adding Jan 31 21:03:48 rtz: actually I define debug levels...so some messages only get emitted at higher levels Jan 31 21:03:53 cshore: yes, but I have to complile, copy it to the router, run it, and find out where to place the next printf statements Jan 31 21:04:33 rtz: ah, it's the build cycle more than printf's that are the problem Jan 31 21:04:45 cshore: yes Jan 31 21:05:01 cshore: and you can't easily decide to take a look at another variable Jan 31 21:05:45 rtz: that's why at high debug levels I just spit out everything I think I might want Jan 31 21:07:10 cshore, [florian]: about that faulty grep line... how about using grep -q? Jan 31 21:07:29 nbd: doh, of course Jan 31 21:07:41 [florian]: can you add that? Jan 31 21:07:44 cshore: WRT160NL Jan 31 21:07:58 Guest90121: heh, we've found the problem Jan 31 21:08:06 <[florian]> nbd: allright Jan 31 21:09:27 cshore: when? Jan 31 21:09:55 florian * r19473 /trunk/package/base-files/files/lib/preinit/20_device_fs_mount: [package] quieten the check for devs, thanks nbd Jan 31 21:10:03 Guest90121: well a patch has a just been committed Jan 31 21:10:45 Guest90121: your query led me to look, but it was an easy bug to find Jan 31 21:11:06 but=and Jan 31 21:12:01 Ah, okay. I can see the patch. That was quick, indeed. Jan 31 21:12:10 Congrats. Jan 31 21:18:28 I have to go for an hour Jan 31 21:18:28 bye Jan 31 21:27:52 nbd * r19474 /trunk/package/mac80211/patches/560-ath9k_fix_eap_handling.patch: mac80211: fix an eap frame handling bug in ath9k - fixes client mode rekeying issues Jan 31 21:37:26 florian * r19475 /trunk/target/linux/brcm63xx/files-2.6.30/drivers/mtd/maps/bcm963xx-flash.c: [brcm63xx] fix printing of mtd offsets, remove whitespaces. Jan 31 21:39:37 [florian]: how long do you think for .32? Jan 31 21:40:06 ping cshore Jan 31 21:40:13 ping jow_laptop Jan 31 21:40:17 pong Jan 31 21:40:27 just read your mail regarding /dev/null Jan 31 21:40:41 you can use the -q flag of grep to have the same effect Jan 31 21:40:46 jow_laptop: a fix with grep -q has already been applied Jan 31 21:40:50 I see Jan 31 21:40:58 jow_laptop: things are happening fast today Jan 31 21:41:31 jow_laptop: nbd pointed out -q and florian applied the change Jan 31 21:43:34 good Jan 31 22:00:01 btw: as modularization of my hotplug patch is resulting in significant refactoring, I'd suggest not bother to review it for a while Jan 31 22:00:12 i.e. until I do a new one Jan 31 22:17:57 nbd * r19476 /trunk/package/mac80211/patches/570-ap_vlan_injection.patch: mac80211: fix lookup for vlan'd stations for injected frames - fixes wpa rekeying with wds ap Jan 31 22:23:49 nbd * r19477 /trunk/package/mac80211/patches/570-ap_vlan_injection.patch: mac80211: make the vlan injection patch more specific Jan 31 23:24:55 argh Jan 31 23:25:06 I can't seem to get anything done today :/ Jan 31 23:25:15 rtz: been there, done that Jan 31 23:27:32 happens Jan 31 23:52:02 hello everybody, just a question, what's the difference between hostapd and wpad? Jan 31 23:52:20 wpad is a multicall binary of hostapd and wpa_supplicant Jan 31 23:52:46 saves space Jan 31 23:53:23 useful for example if one will connect as a client and operate as master? Jan 31 23:53:42 yeah, or if you just want to have the functionality of both available in the firmware Jan 31 23:53:56 ohh, good to know, thx! Jan 31 23:57:26 one more thing, hostapd-mini can bring me HT modes in ath9k when operating without encryption? or do I need hostapd (full) Jan 31 23:59:09 -mini is enough Jan 31 23:59:22 the full version adds extra eap authentication methods Jan 31 23:59:55 ok thx again nbd! Feb 01 00:16:02 hi Feb 01 00:16:29 cshore: your dev/null patch is not enough dev/null is still a normal file Feb 01 00:16:48 but i think i found the failure Feb 01 00:16:55 oh? Feb 01 00:17:26 /lib/preinit/30_device_fs_daemons Feb 01 00:17:33 also has a > /dev/null entry Feb 01 00:17:42 this creates a file befor the device is created Feb 01 00:17:47 could you make a patch Feb 01 00:17:50 yeah Feb 01 00:17:57 i test then send it to the mailing list Feb 01 00:18:07 florian should test what he commits ;) Feb 01 00:18:30 don't forget to use grep -q to replace the >/dev/null Feb 01 00:18:34 if it's a grep Feb 01 00:18:45 actually it was my bad Feb 01 00:18:57 but that fix was needed too Feb 01 00:22:14 cshore i know Feb 01 00:22:31 but it seesm to be ony 50% of the needed work ;) Feb 01 00:22:58 I was testing to see if anyone was paying attention Feb 01 00:23:05 :) Feb 01 00:23:15 do you believe me? Feb 01 00:25:06 :) Feb 01 00:27:19 no but it doesnt matter ;) Feb 01 00:27:25 its fixecd now Feb 01 00:27:39 fixed Feb 01 00:28:00 http://openwrt.pastebin.com/m7cd5d71 Feb 01 00:28:22 does somebody have a build failure with 'make V=99 tools/firmware-utils compile' ? Feb 01 00:28:54 jow_laptop that all is not needed Feb 01 00:29:01 i send a patch to the mailing list Feb 01 00:29:08 this is all that is needed Feb 01 00:29:15 puchu: but it won't hurt Feb 01 00:29:29 thats true Feb 01 00:29:56 okay Feb 01 00:29:58 ill go now Feb 01 00:30:05 have a nice day/night @all Feb 01 00:30:06 puchu: I know its not all needed but the redirection is ineffiecient too, clutters the code and might lead to further problems Feb 01 00:30:23 thats true Feb 01 00:30:36 jow_laptop: yeah, I forgot about -q to grep Feb 01 00:30:56 no problem Feb 01 00:31:03 I'll commit that now Feb 01 00:31:06 bye Feb 01 00:32:47 jow * r19478 /trunk/package/base-files/files/lib/ (5 files in 2 dirs): [package] base-files: preinit: get rid of /dev/null redirections and use grep -q where applicable Feb 01 00:33:36 jow * r19479 /trunk/target/linux/x86/ (2 files in 2 dirs): [x86] preinit: use grep -q where applicable Feb 01 00:34:13 jow * r19480 /trunk/target/linux/brcm-2.4/base-files/lib/preinit/ (05_set_failsafe_switch_brcm 15_set_preinit_interface_brcm): [brcm-2.4] preinit: use grep -q where applicable Feb 01 00:34:36 jow * r19481 /trunk/target/linux/ifxmips/nfs/base-files/lib/preinit/01_init_nfs_ifxmips: [ifxmips] preinit: use grep -q where applicable Feb 01 00:34:53 jow * r19482 /trunk/target/linux/atheros/base-files/lib/preinit/15_preinit_iface_atheros: [atheros] preinit: use grep -q where applicable Feb 01 00:35:01 nbd * r19483 /trunk/package/wprobe/src/user/wprobe-lib.c: wprobe: fix missing return code check Feb 01 00:35:09 jow * r19484 /trunk/target/linux/pxcab/base-files/lib/preinit/45_failsafe_pxcab: [pxcab] preinit: use grep -q where applicable Feb 01 00:35:31 done :) Feb 01 00:36:12 jow_laptop: now that your comiting spree is done can you please check tools/firmware-utils ? it doesn't compile ;) Feb 01 00:36:40 solca: no issues here, you need to provide details Feb 01 00:38:49 jow_laptop: http://openwrt.pastebin.com/d2e84c039 Feb 01 00:39:17 you need make V=99 Feb 01 00:39:54 solca: we need make V=99 to do debugging Feb 01 00:40:06 solca: .../compile Feb 01 00:40:11 not separated by space Feb 01 00:45:08 jow_laptop: oops sorry, it's this: http://openwrt.pastebin.com/d6d777d4d Feb 01 00:46:26 jow_laptop: target is ar71xx profile WRT160NL Feb 01 01:02:16 build #2 of uml is complete: Failure [failed compile_2] Build details are at http://tksite.gotdns.org:8010/builders/uml/builds/2 Feb 01 01:07:48 solca: http://openwrt.pastebin.com/m22d5b79b Feb 01 01:10:12 jow_laptop: it works! Feb 01 01:10:21 I know Feb 01 01:10:31 :) Feb 01 01:17:45 jow_laptop: can you check and commit this: http://home.galileo.edu/~solca/openwrt/wrt160nl-nvram.diff Feb 01 01:17:48 jow * r19485 /trunk/tools/firmware-utils/ (Makefile src/bcm_tag.h): [tools] firmware-utils: copy bcm_tag.h to src/ and get rid of cc3 macro, fixes imagetag compilation Feb 01 01:19:49 it fixes nvram on wrt160nl by workarounding mmap implementation, it protects the mmap mem so it's safer too Feb 01 01:20:04 solca: so the implementation works just fine there? Feb 01 01:20:41 jow_laptop: I debug it why nvram was working some times and mostly not Feb 01 01:20:59 my mmap() manpage does not know MAP_LOCKED Feb 01 01:21:22 culprit was mmap unless with MAP_LOCKED Feb 01 01:22:14 I see, but do you know whether MAP_LOCKED is supported on Linux 2.4.30 ? Feb 01 01:24:26 maybe it is better to call mlock() instead, this should work on Linux 2.4.x too if I read it correctly Feb 01 01:25:29 maybe, the patch works fine for brcm-2.4 Feb 01 01:25:41 (ie wrt54gl) Feb 01 01:26:15 then I have no objections, apparently it is just ignored then. I was concerned that the symbol might not be defined Feb 01 01:31:26 jow_laptop: yea it works fine in 2.4 Feb 01 01:32:48 jow * r19486 /trunk/package/nvram/ (Makefile src/nvram.c): [package] nvram: request memory locked pages in mmap() and make it available for ar71xx Feb 01 01:34:20 jow_laptop: thx! Feb 01 01:35:11 np Feb 01 01:56:03 jow * r19487 /trunk/tools/firmware-utils/Makefile: [tools] firmware-utils: unify cc and cc2 macros, simplifies Makefile Feb 01 02:00:59 jow * r19488 /trunk/package/base-files/files/lib/network/config.sh: [package] base-file: revert unrelated change accidentally committed in r19478 **** ENDING LOGGING AT Mon Feb 01 02:59:56 2010