**** BEGIN LOGGING AT Thu Aug 05 02:59:57 2010 Aug 05 03:00:02 then Aug 05 03:00:02 $(call Image/Build/Chk,$(1),U12H081T00_NETGEAR,2) Aug 05 03:00:56 make sense? Aug 05 03:01:19 oh and done forget the endef after the content of the define Aug 05 03:02:10 sorry trying to find out why buid failed Aug 05 03:02:16 but yeah that makes sense Aug 05 03:06:02 cshore, do u see any thing wrong with this code? Aug 05 03:06:12 mm ill pastebin Aug 05 03:06:50 cshore, http://paste.ubuntu.com/473364/ Aug 05 03:08:14 mmm i think i found it Aug 05 03:08:50 i was wrong it still errors hmm Aug 05 03:12:10 : error: too few arguments to function 'strncmp' Aug 05 03:12:12 :/ Aug 05 03:12:44 yeah, strncmp, means read to null (\0), or n characters...you've forgotten the n characters Aug 05 03:13:04 strncmp(boardnum, "8", 1) for example Aug 05 03:13:21 if a board num is only ever one chacter long Aug 05 03:13:38 ah ok so i just need ti add , 1 Aug 05 03:13:51 to* Aug 05 03:13:55 or whatever is appropriate Aug 05 03:14:34 ahhh i realize what i did wrong now Aug 05 03:14:39 and for the strcmp's those probably should be strncmp's as well, unless you *know* they will *always* end with a null Aug 05 03:15:03 strncmp needs to be strcmp Aug 05 03:15:25 depends, where do boardnum, and boardtype come from? Aug 05 03:15:30 nvram Aug 05 03:15:35 i.e. are they already sanitized Aug 05 03:15:41 mmm well actually Aug 05 03:15:46 this is code for diag.c Aug 05 03:15:47 so Aug 05 03:15:55 boardnum = getvar("boardnum"); Aug 05 03:15:55 boardtype = getvar("boardtype"); Aug 05 03:16:02 thats where they come from :/ Aug 05 03:16:09 ah, getvar should do the null I think Aug 05 03:16:18 so you should be able to strcmp Aug 05 03:16:23 :) Aug 05 03:17:24 kk time for a smoke, i hope it builds this time hehe Aug 05 03:23:29 cshore, thanks man! Aug 05 03:23:36 it worked Aug 05 03:23:44 i have a chk generated now :) Aug 05 03:24:21 but i havent added the code you told me just yet, bout to do that now Aug 05 03:26:57 cshore, im pretty sure that $(call Image/Build/Chk,$(1),U12H081T00_NETGEAR,2) goes under define Image/Build correct? Aug 05 03:27:58 * RealOpty tests Aug 05 03:29:19 cshore, wanna give me a hand on making patches so we can get them on svn :) ? Aug 05 03:36:03 * RealOpty reads http://wiki.openwrt.org/doc/devel/patches Aug 05 03:40:14 RealOpty: yes, and yes Aug 05 03:40:53 kk ive installed quilt and ive set the options it said on wiki into .quiltrc Aug 05 03:42:53 cshore, wanna give me some pointers? Aug 05 03:43:12 ive only changed the two makefiles we talked about and added one file into firmware-utils Aug 05 03:43:32 in an svn tree? Aug 05 03:43:42 do svn diff Aug 05 03:43:52 that should show you the combined diff of the changes Aug 05 03:45:04 kk :) Aug 05 03:45:25 now upload that to somewhere for you? Aug 05 03:46:02 make a ticket? Aug 05 03:46:17 kk :) Aug 05 03:46:39 I have a ticket for it. ill reopen it then later someone can close it again Aug 05 03:49:03 ahh i dont have to reopen it to attach file Aug 05 03:49:15 what ticket? Aug 05 03:50:56 cshore, https://dev.openwrt.org/attachment/ticket/7630/ Aug 05 03:51:23 theres the patch and the code for making the image Aug 05 03:51:34 that goes into /tools/firmware-utils/src Aug 05 03:52:21 instead do svn add mkchkimg.c (well actually the path to it from buildroot) Aug 05 03:52:33 and then svn diff Aug 05 03:52:39 then it's a single patch Aug 05 03:55:04 kk Aug 05 04:00:04 cshore, https://dev.openwrt.org/attachment/ticket/7630/allinone.patch Aug 05 04:01:33 sweet its nice to know how to make patches now :) Aug 05 04:02:34 you need to refresh your svn tree...the patch failed to apply Aug 05 04:02:53 due to changes in other things in the Makefile changing Aug 05 04:03:10 okie Aug 05 04:05:19 cshore, i did a svn up before making that last patch. Aug 05 04:05:25 also i think this might be related Aug 05 04:05:31 one sec Aug 05 04:05:50 Conflict discovered in 'target/linux/brcm47xx/files-2.6.34/drivers/mtd/maps/bcm47xx-flash.c' Aug 05 04:05:58 and i told it that mine conflicted Aug 05 04:06:35 (hauke gave me the patch and then later he applied it to svn) Aug 05 04:07:34 At revision 22498. Aug 05 04:09:17 hmmmm.....was that the only conflict? Aug 05 04:09:36 yeah just for kernel 2.6.34 AND 2.635 Aug 05 04:10:03 could you make the patch again Aug 05 04:10:14 indeed Aug 05 04:10:31 nvm, it was me Aug 05 04:10:35 oh Aug 05 04:10:36 kk Aug 05 04:10:43 nope, not me Aug 05 04:10:46 oh Aug 05 04:10:47 lol Aug 05 04:10:48 :) Aug 05 04:12:15 cshore, every time i make the patch it also wants to include haukes patch that he gave me (which is on svn now) Aug 05 04:12:36 doh! it was me, just not the problem I thought Aug 05 04:12:47 it applied now Aug 05 04:12:56 muhahah :) kk Aug 05 04:15:02 ok, building to make sure it doesn't have some silly FTB error, and then I'll apply it Aug 05 04:15:15 for sure :) Aug 05 04:15:27 time for me to flash the device again, brb Aug 05 04:21:17 cshore, did u apply it yet? Aug 05 04:21:51 cshore, dont apply just yet. theres a typo i believe Aug 05 04:22:16 * RealOpty laughs then frowns Aug 05 04:22:20 too late, I'll redo it once you're ready Aug 05 04:22:35 yeah ill just make a patch to fix it Aug 05 04:22:44 its something to do with the board ID Aug 05 04:24:59 LOL thats waht i get for coping and pasting Aug 05 04:26:06 or maybe not hmm weird Aug 05 04:26:12 cshoe $(call Image/Build/Chk,$(1),U12H081T00_NETGEAR,2 Aug 05 04:26:24 what was $(1) again? Aug 05 04:28:05 cshore, i think that $(1) needs to be removed from the line i pasted above Aug 05 04:31:07 cshore, ahhh u didnt commit it yet, you just applied it to your tree :) Aug 05 04:31:31 anyways for some reason the board id isnt get transfer they way we have it coded Aug 05 04:31:46 that's right Aug 05 04:32:28 * RealOpty runs make V=99 Aug 05 04:32:35 or wait Aug 05 04:32:43 cshore, is there a way to just make the images? Aug 05 04:32:57 make target/linux/image Aug 05 04:33:08 sorry Aug 05 04:33:15 make target/linux/install Aug 05 04:33:42 hmmmmm Aug 05 04:33:50 no rule to make target Aug 05 04:35:28 LOL sorry Aug 05 04:35:32 im getting tired Aug 05 04:36:13 was typing on wrong console lol Aug 05 04:37:12 /home/will/OpenWRT/trunk_netgear/staging_dir/host/bin/mkchkimg -o /home/will/OpenWRT/trunk_netgear/bin/brcm47xx/openwrt-brcm47xx-squashfs.chk -k /home/will/OpenWRT/trunk_netgear/bin/brcm47xx/openwrt-brcm47xx-squashfs.trx -b U12H081T00_NETGEAR -r 2 Aug 05 04:37:24 weird Aug 05 04:37:27 its working.... Aug 05 04:37:30 so wtf Aug 05 04:38:03 what's not happening that should? Aug 05 04:38:30 the CFE didnt accept the image this time :/ Aug 05 04:39:08 http://paste.ubuntu.com/473382/ Aug 05 04:39:13 theres what the CFE says :/ Aug 05 04:41:27 brb gonna do more testing Aug 05 04:41:42 what does a hexeditor show for the .chk? Aug 05 04:43:01 it shows the header in there, do u want me to pastebin? Aug 05 04:43:10 brb im gonna try to flash again Aug 05 05:07:40 cshore, please forgive me for being an idiot. Aug 05 05:07:44 you can commit the patches.... Aug 05 05:07:52 it was a simple typo on my end. Aug 05 05:07:59 i tried to flash the ..trx Aug 05 05:08:45 pulling my hair out over nothing lol... Aug 05 05:17:02 cshore, ping Aug 05 05:31:07 cshore, if you around maybe you can help me figure out some LED stuff Aug 05 05:35:10 time to test the LEDs brb Aug 05 05:45:23 lol i guess thats progress. the power LED went off :/ Aug 05 07:11:56 lol i guess thats progress. the power LED went off :/ Aug 05 07:11:58 oops Aug 05 07:12:15 cshore, i got the LEDs working :D Aug 05 07:25:17 mb__, hello Aug 05 07:25:33 lol Aug 05 07:26:31 mb__, ping Aug 05 07:27:24 RealOpty: pong Aug 05 07:29:16 mb__, you do much with openwrt? Aug 05 07:30:32 Well, I'd like to do more. Aug 05 07:31:00 know anything about gpioctl? Aug 05 07:32:44 not really. Does it use the standard kernel GPIO framework? Aug 05 07:33:10 yup Aug 05 07:33:38 ok, well I know that Aug 05 07:34:00 whats a sortof generic way to test for a button then? Aug 05 07:36:22 Well in the kernel you register the button. That's all. Aug 05 07:36:45 After that you can access it with a number through the standard interface Aug 05 08:14:31 hello Aug 05 08:20:10 RealOpty: you mean how do you detect a button event once the buttons are defined in the board settings? Aug 05 08:26:38 cshore, yay your back :) Aug 05 08:26:49 nbd: Ok, I booted .35 on the AP and WDS still does not work Aug 05 08:26:54 cshore, basicly yes. Aug 05 08:27:34 will it say pin * is HIGH? Aug 05 08:27:59 for brcm47xx you edit diag.c and then make sure diag.c is loaded on boot Aug 05 08:28:04 oh I see Aug 05 08:28:30 cshore, i got gpioctl working now :) Aug 05 08:28:40 When the button is pressed it will say the opposite of what is says when the button is not pressed Aug 05 08:28:40 RealOpty: Whether it's high or low on pressed depends on how it is wired Aug 05 08:29:07 if it's active low, it will be high not pressed and low pressed Aug 05 08:29:07 cshore, also you can commit the patch :) every thing was working fine, i just had a typo Aug 05 08:29:26 if it's active high, it will be low not pressed and high pressed Aug 05 08:29:48 RealOpty: I'm getting a build error....let me see why Aug 05 08:30:07 weird Aug 05 08:31:08 cshore, did you do a make tools/firmware-utils/{clean,compile,install ? Aug 05 08:31:20 fresh svn checkout Aug 05 08:31:34 :\ Aug 05 08:31:38 * RealOpty will test also Aug 05 08:31:57 it looks like the gcc compile is broken Aug 05 08:32:01 or rather the download Aug 05 08:32:17 thats weird Aug 05 08:32:30 do u have the link to the file? ill test it too Aug 05 08:35:00 https://dev.openwrt.org/attachment/ticket/7630/allinone.patch this one? Aug 05 08:35:29 cshore, yes :) Aug 05 08:35:52 The docs on http://kamikaze.openwrt.org/docs/openwrt.html seem pretty much broken, too. Aug 05 08:36:22 It says to enable WDS do "option mode     wds", but that does not work at all. It results in invalid mode messages Aug 05 08:40:07 SWEET Aug 05 08:40:14 i found the reset button gpio Aug 05 08:40:50 cshore, i have another patch for you. its diag.c with all the LEDs and button. Aug 05 08:42:58 cshore, ok so the button was active high Aug 05 08:43:05 does that mean anything for the polarity? Aug 05 08:43:32 it means you don't want to set active low in the diag.c Aug 05 08:43:56 or rather .active_low = 0 Aug 05 08:44:06 { .name = "reset", .gpio = 1 << 6 }, Aug 05 08:44:11 that should be good yeah? Aug 05 08:44:41 what button is it? Aug 05 08:44:48 reset button Aug 05 08:44:52 on gpio 6 Aug 05 08:45:20 wouldn't it just be .gpio = 6? At least that's the way brcm63xx works Aug 05 08:45:40 what do other buttons do ? Aug 05 08:45:51 theres no other buttons. Aug 05 08:45:59 I mean for other devices Aug 05 08:46:06 oh Aug 05 08:46:16 the code looks just like that for all the other ones Aug 05 08:46:26 then I'd do that Aug 05 08:47:49 kk Aug 05 08:48:15 you should attach both patches to open tickets (could both be a 'add support for xxxx' ticket Aug 05 08:48:33 on a single I mean Aug 05 08:48:45 thats the title of this ticket? Aug 05 08:49:12 I can't test the patches until the build failure is fixed (or I debug it myself) Aug 05 08:49:51 add button, led, and image creation support for netgear xxxxx Aug 05 08:49:56 cshore, where is the location of diag.c? Aug 05 08:50:01 i think i edited the wrong one Aug 05 08:50:08 i changed the one in my build_dir Aug 05 08:50:16 yeah that's the wrong one Aug 05 08:50:40 /target/linux/brcm47xx/files/something I think Aug 05 08:50:48 cshore, so open new ticket, yeah? Aug 05 08:51:53 yeah Aug 05 08:52:33 cshore, do u think that trunk/package/broadcom-diag/src is the correct one to edit? Aug 05 08:52:59 that sounds right Aug 05 08:53:50 cshore, would it be safe to copy and paste my diag.c over the other one? Aug 05 08:53:56 i think it would Aug 05 08:54:58 nbd: If I add an AP interface (which I do not want, actually), it gets a step further. But still errors: http://pastebin.org/448500 Aug 05 09:06:54 ought to be, but why not just cp ? Aug 05 09:07:08 less chance of random whitespace changes Aug 05 09:07:25 yeah thats what i ment :) Aug 05 09:07:43 about to upload the all in one super duper add support patch Aug 05 09:09:17 cshore, hmm i thought that hauke closed this ticket, but he didnt. it only says its closed in the svn Aug 05 09:14:06 cshore, https://dev.openwrt.org/ticket/7702 Aug 05 09:15:18 so there it is. my days work lol Aug 05 09:15:50 the image generation was a major part, as it will allow the user to flash from the factory firmware. Aug 05 09:22:08 I am so pissed by this. Proprietary wl also does not work. It does not load at all. No error message whatsoever. Aug 05 09:23:08 mb__, me too! Aug 05 09:23:17 mb__: huh? you should only set up a regular AP on one side, and a regular STA on the other side and then add 'option wds 1' on both sides Aug 05 09:23:18 i have the same prob Aug 05 09:23:42 no idea why you're trying to add a second virtual interface Aug 05 09:24:05 nbd, morning :) Aug 05 09:24:15 morning Aug 05 09:24:23 nbd: I am trying to add a second, because all other things do not work Aug 05 09:24:54 nbd: I have an AP with the wds_sta=1 set. Then I have an openwrt STA with wds 1 set. It does not work. Aug 05 09:25:02 I cannot ping the AP Aug 05 09:25:02 details? Aug 05 09:25:08 does it authenticate? Aug 05 09:25:14 The wireless works Aug 05 09:25:15 is there a wlan0.sta* interface on the AP? Aug 05 09:25:35 The AP is not openwrt Aug 05 09:25:50 i know Aug 05 09:26:13 do you have hostapd 0.7.2 (or newer) on the AP? Aug 05 09:26:20 latest GIT Aug 05 09:26:23 good Aug 05 09:26:31 so... did it create the interface? Aug 05 09:26:37 no Aug 05 09:26:44 odd Aug 05 09:26:54 check the hostapd log Aug 05 09:27:05 It is a pretty standard WPA-PSK setup. I just added the wds option. Aug 05 09:27:08 to see if it tried Aug 05 09:27:21 it needs to create a STA VLAN for the incoming wds link Aug 05 09:28:57 There are no special messages Aug 05 09:29:10 Does hostapd automatically create those interfaces? Aug 05 09:29:15 yes Aug 05 09:29:28 it should detect incoming 4-addr frames Aug 05 09:29:34 then create an interface and move the sta into it Aug 05 09:29:41 that's necessary for being able to bridge stuff properly Aug 05 09:31:04 So I just tried backfire-2.4-SVN. It does not succeed loading wl. (No message). In backfire release it loads. Aug 05 09:32:41 nbd, do u know of a device that has N protocol for wireless and works with openwrt? Aug 05 09:33:55 brb Aug 05 09:39:52 nbd: Ok, I flashed backfire-2.4 release and it finally works (with bridged STA) Aug 05 09:54:14 nbd, what does wl: module license 'unspecified' taints kernel. mean? Aug 05 09:54:51 RealOpty: it means that wl is proprietory and not GPL Aug 05 09:54:58 o ok Aug 05 09:55:23 since wifi detect dosnt detect my card Aug 05 09:55:39 how can i go about trying to get it to find it? Aug 05 09:55:47 not make it work but just at least find it lol Aug 05 10:18:36 mm is there suport for a bcm4321 wifi card? Aug 05 10:23:29 i know that its only supported with the broadcom binary Aug 05 10:23:30 RealOpty: there should be Aug 05 10:23:55 jow_laptop, have any tips on trying to get it detected in openwrt? Aug 05 10:24:33 dd-wrt can make it work, openwrt can too!!!! :) Aug 05 10:24:53 what do you run? Aug 05 10:25:01 wifi detect Aug 05 10:25:05 ... Aug 05 10:25:07 cat /proc/net/dev Aug 05 10:25:09 I mean revision Aug 05 10:25:15 arch Aug 05 10:25:19 o Aug 05 10:25:20 hardware Aug 05 10:25:23 brcm47xx Aug 05 10:25:44 with b43 ? Aug 05 10:25:48 nope Aug 05 10:25:57 post the full dmesg Aug 05 10:26:04 kk. Aug 05 10:26:48 http://paste.ubuntu.com/473455/ Aug 05 10:26:53 jow_laptop, ^ Aug 05 10:28:29 nvram show too please Aug 05 10:29:20 jow_laptop, do u want all the pci/1/1 stuff and board data? Aug 05 10:29:26 yes Aug 05 10:29:29 the nvram is larger then what i can copy Aug 05 10:29:40 mmm sec i can fix the terminal Aug 05 10:29:47 make a file Aug 05 10:30:36 jow_laptop, http://paste.ubuntu.com/473459/ Aug 05 10:31:21 jow_laptop, i changed two settings in nvram. ill grab them for ya Aug 05 10:33:01 kk heres the settings i changed (cause dd-wrt changed them so i thought i would try) Aug 05 10:33:03 pci/1/1/txpid2ga0=55 Aug 05 10:33:03 pci/1/1/txpid2ga1=78 Aug 05 10:33:12 they were both set at 91 Aug 05 10:35:13 I am not sure whether the driver supports the pci/x/x/ notation Aug 05 10:36:22 ok it does Aug 05 10:36:44 here is what dd-wrt does to the nvram. http://svn.dd-wrt.com:8000/dd-wrt/browser/src/router/services/sysinit/sysinit-broadcom.c Aug 05 10:37:02 search for "struct nvram_tuple wnr834bv2_pci_1_1_params" Aug 05 10:37:14 (if that helps) Aug 05 10:37:26 idk why they change the values :\ Aug 05 10:39:10 the nvram you pasted is complete enough to kick the driver into operation Aug 05 10:40:48 can you add some debug printks to initvars_flash() in broadcom-wl-*/driver/bcmsrom.c ? Aug 05 10:42:53 in particular I want to know what the drivers idea of "devpath" is Aug 05 10:43:02 and whether it manges to find the vars Aug 05 10:44:48 sure ive been wanting a good reason to reflash Aug 05 10:46:42 the pci/1/1/ variables are the srom of the radios, the rest is not that interesting Aug 05 10:46:57 jow_laptop, i need the file path plz Aug 05 10:47:13 i dont think its trunk/packages/broadcom-wl Aug 05 10:47:31 build_dir/linux-brcm47xx/broadcom-wl-*/driver/ Aug 05 10:48:06 I suggest to follow http://wiki.openwrt.org/doc/devel/patches Aug 05 10:48:23 and on your box, comment out the wl in /etc/modules.d/ Aug 05 10:48:29 reboot Aug 05 10:48:39 jow_laptop, svn diff isnt good enough? Aug 05 10:48:57 then just scp the wl.ko file from your buildroot to /tmp on the router and insmod /tmp/wl.ko for testing Aug 05 10:49:04 its easier than reflashing for every test Aug 05 10:49:13 svn diff won't work, you need to patch the source Aug 05 11:01:11 jow_laptop, what part of http://wiki.openwrt.org/doc/devel/patches do i follow ? Aug 05 11:01:24 and Aug 05 11:01:25 initvars_flash(si_t *sih, osl_t *osh, char **base, uint len) Aug 05 11:01:40 thats the part you want the kernel messages into yeah? Aug 05 11:02:00 yes Aug 05 11:03:35 jow_laptop, suggestions on what i should have it print? Aug 05 11:03:53 or rather Aug 05 11:04:22 how to printk devpath the correct way? Aug 05 11:05:16 printk("devpath=%s\n", devpath); Aug 05 11:05:59 i should put that under si_devpath yeah? Aug 05 11:06:12 * RealOpty is no C coder ;p Aug 05 11:06:30 yes, before the "grab" comment Aug 05 11:07:14 within the for loop, add printk("s=%s\n", s); before the if(...) continue statement Aug 05 11:07:40 all that should end up in dmesg if you did it right Aug 05 11:08:07 jow_laptop, kk Aug 05 11:08:23 now command to rebuild the package? Aug 05 11:08:32 you added the patch? Aug 05 11:08:56 I edited the file directly :\ Aug 05 11:10:06 this will not work Aug 05 11:10:23 ok Aug 05 11:10:28 jow_laptop, do i do this Aug 05 11:10:30 http://wiki.openwrt.org/doc/devel/patches#adding.a.new.patch ? Aug 05 11:10:38 yes Aug 05 11:10:42 (after i undo the changes) lol Aug 05 11:10:53 make package/broadcom-wl/{clean,prepare} V=99 QUILT=1 Aug 05 11:11:09 cd build_dir/linux-brcm47xx/broadcom-wl-* Aug 05 11:11:12 quilt push -a Aug 05 11:11:18 quilt new 900-debug.patch Aug 05 11:11:31 quilt edit driver/bcmsrom.c Aug 05 11:11:35 quilt refresh Aug 05 11:11:41 cd ../../../ Aug 05 11:11:52 make package/broadcom-wl/{update,clean,compile} V=99 Aug 05 11:12:07 :) Aug 05 11:12:18 scp build_dir/linux-brcm47xx/broadcom-wl-*/driver/wl.ko root@router:/tmp Aug 05 11:12:21 ssh root@router Aug 05 11:12:27 insmod /tmp/wl.ko Aug 05 11:12:29 dmesg Aug 05 11:12:49 and as I said, prevent the the one currently there from autoloading Aug 05 11:12:59 kk already did that Aug 05 11:13:32 if you need to edit the patch Aug 05 11:13:46 follow the same procedure as above but leave out the "quilt new" step Aug 05 11:15:45 there's also a #define at the top of bcmsrom.c Aug 05 11:15:49 #define BS_ERROR(args) Aug 05 11:15:52 change it to Aug 05 11:15:56 #define BS_ERROR(args) printk args Aug 05 11:16:08 this will enable some debugging messages Aug 05 11:16:10 :) i was gonna do that earlier Aug 05 11:16:17 cause thats what i thought u was talking about Aug 05 11:16:21 there are similar #defines in other files too Aug 05 11:16:21 then i looked in the code Aug 05 11:16:35 and there wasnt anything there for it the section we edited Aug 05 11:18:44 you can also edit package/broadcom-wl/Makefile and try adding "-DBCMDBG" to the CFLAGS Aug 05 11:19:09 jow_laptop, printk or printf? Aug 05 11:19:16 printk Aug 05 11:19:26 printf would probably work too Aug 05 11:19:32 but I always used printf Aug 05 11:21:58 jow_laptop, if i edit pack/broad-wl/makef then i need to reflash yeah? Aug 05 11:23:14 oh nvm Aug 05 11:23:19 im too tired lol Aug 05 11:23:26 * RealOpty edits the file Aug 05 11:25:06 * RealOpty compiles Aug 05 11:26:36 jow_laptop, just the wl.ko file needs copying yeah? Aug 05 11:27:57 brb rebooting Aug 05 11:31:34 jow_laptop, i did everything to the book you wrote. Aug 05 11:31:41 nothings in dmesg lol Aug 05 11:31:57 and i confirmed that it didnt load the wl.ko thats on the router Aug 05 11:32:21 after that i downloaded wl.ko Aug 05 11:32:36 then i did insmod /tmp/wl.kl Aug 05 11:32:39 and its now loaded Aug 05 11:33:10 all it says is Aug 05 11:33:11 wl: module license 'unspecified' taints kernel. Aug 05 11:33:11 Disabling lock debugging due to kernel taint Aug 05 11:35:34 Applying ./patches/900-debug.patch using plaintext: Aug 05 11:35:34 patching file driver/bcmsrom.c Aug 05 11:35:34 Hunk #1 succeeded at 1819 (offset -4 lines). Aug 05 11:35:53 ^== what make said Aug 05 11:37:39 jow_laptop, wanna see the .patch? Aug 05 11:43:18 odd Aug 05 11:44:28 jow_laptop, http://paste.ubuntu.com/473477/ Aug 05 11:44:32 there is the patch. Aug 05 11:45:53 jow_laptop, why do i get this message File driver/bcmsrom.c is already in patch 910-fallback-sprom.patch Aug 05 11:45:59 when trying to make a change to the file? Aug 05 11:46:07 this is normal Aug 05 11:46:09 ignore it Aug 05 11:46:19 did you add "-DBCMDBG" to Makefile? Aug 05 11:46:24 yes Aug 05 11:46:32 sec ill show u where i added it Aug 05 11:46:44 CFLAGS="$(TARGET_CFLAGS) -DBCMDBG -I. -I$(PKG_BUILD_DIR)/driver/include" \ Aug 05 11:48:10 ah that was wrong :) Aug 05 11:48:12 MODFLAGS Aug 05 11:48:15 add it there Aug 05 11:48:31 where -DMODULE already is Aug 05 11:49:44 kk. Aug 05 11:49:53 it dont matter for mini atm correct ;p Aug 05 11:49:59 yeah Aug 05 11:50:16 kk Aug 05 11:50:24 another thing ive noticed Aug 05 11:50:29 if u look at the patch i pasted Aug 05 11:50:48 i dont see the changes to BS_ERROR in there Aug 05 11:50:58 but when i go back in to edit i do see the changes???? Aug 05 11:52:38 * RealOpty runs make and goes for a smoke Aug 05 11:53:02 good idea :) Aug 05 11:58:24 hmm Aug 05 11:58:37 jow_laptop, ok so i rmmod wl then uploaded new one Aug 05 11:58:41 and i get this.... Aug 05 11:58:42 insmod: can't insert '/tmp/wl.ko': unknown symbol in module, or unknown parameter Aug 05 12:00:06 dmesg Aug 05 12:00:25 reboot between reloads Aug 05 12:00:35 rmmod usually does not work cleany Aug 05 12:00:37 cleanly Aug 05 12:00:40 kk Aug 05 12:01:09 heres what dmesg said Aug 05 12:01:09 wl: Unknown symbol wlc_iovar_dump Aug 05 12:01:10 wl: Unknown symbol wl_apsta_dbg Aug 05 12:01:10 wl: Unknown symbol wlc_iovar_dump Aug 05 12:01:10 wl: Unknown symbol wl_apsta_dbg Aug 05 12:01:12 rebooting now Aug 05 12:01:36 ah hm Aug 05 12:01:52 let me check Aug 05 12:03:47 insmod /tmp/wl.ko Aug 05 12:03:48 Killed Aug 05 12:06:25 jow_laptop, http://paste.ubuntu.com/473485/ Aug 05 12:06:42 for some reason it made a kernel oops with dnsmasq also :\ Aug 05 12:06:52 try adding this to the MODFLAGS too: -D"wlc_iovar_dump Aug 05 12:07:10 try adding this to the MODFLAGS too: -D"wlc_iovar_dump(...)" -D"wl_apsta_dbg(...)" Aug 05 12:07:26 ... = ? Aug 05 12:07:36 space ;p ? Aug 05 12:08:20 ... = ... Aug 05 12:08:29 write it this way Aug 05 12:08:35 pl Aug 05 12:08:38 ok* Aug 05 12:09:10 ah and you might need to quote the " Aug 05 12:09:13 or use ' Aug 05 12:09:20 i was about to say lol Aug 05 12:09:46 then remake yeah? Aug 05 12:09:48 yes Aug 05 12:09:55 without update Aug 05 12:10:01 just clean,compile Aug 05 12:10:56 ok i started it with update then killed it and started over without Aug 05 12:11:01 hope that dont break it ;p Aug 05 12:12:33 jow_laptop, compile error http://paste.ubuntu.com/473487/ Aug 05 12:14:29 its the quotes. Aug 05 12:14:36 what should they be :\ Aug 05 12:15:08 ah, yeah Aug 05 12:15:11 wait Aug 05 12:17:18 need to prepare a build, might take a while Aug 05 12:19:20 np Aug 05 12:19:32 i appreciate your help on this :) Aug 05 12:19:49 once wifi is working this router will be fully functional with openwrt :D Aug 05 12:20:21 hi all Aug 05 12:20:46 does anyone know the proper status of the dwg_otg driver for ramips platform? Aug 05 12:21:20 so far hauke made the patch for skipping over the board data, now ive added the LEDs/button to diag.c, and ive found opensource code for generating the image header that the router needs. Aug 05 12:22:24 jow_laptop, is there a reason that /etc/diag.sh dosnt call the LED 'connected' ? Aug 05 12:32:14 RealOpty: probably not Aug 05 12:32:31 RealOpty: what should "connected" indicate anyway? Aug 05 12:33:09 well on this router, it turns the "internet light" from amber to green Aug 05 12:33:43 and i think there is another router that uses it too Aug 05 12:34:52 well 4 total routers Aug 05 12:35:08 RealOpty: ah, well then you might display the "wan" status ther Aug 05 12:35:09 e Aug 05 12:36:08 on this device if the cable is disconnected the light turns off Aug 05 13:06:07 jow_laptop, any progress bro? Aug 05 13:08:21 compiling... Aug 05 13:11:15 jow_laptop, whats the command to erase nvram? Aug 05 13:11:22 mtd erase nvram Aug 05 13:11:24 ty Aug 05 13:11:52 be back in a bit need to do some testing Aug 05 13:15:03 impressive. Aug 05 13:15:20 erasing nvram didnt break the cfe Aug 05 13:16:29 time for more testing. Aug 05 13:46:39 RealOpty: http://luci.subsignal.org/~jow/990-fix-debug.patch Aug 05 13:46:59 RealOpty: put that into package/broadcom-wl/patches/ and retry building with -DBCMDBG Aug 05 13:48:35 ok Aug 05 13:49:44 and remove the other -D...(...) flags Aug 05 13:57:11 jow_laptop, http://paste.ubuntu.com/473527/ Aug 05 13:58:12 you're running out of ram Aug 05 13:58:22 probably because the debug module is so big Aug 05 13:58:46 you can try to nuke some processes like dnsmasq, httpd etc. Aug 05 13:59:15 I assume there is no external storage on the router? Aug 05 14:00:11 nope Aug 05 14:00:18 sec ill kill dnsmas Aug 05 14:02:04 jow_laptop, can i kill hotplug? Aug 05 14:02:45 yes Aug 05 14:03:02 how big is the .ko file? Aug 05 14:03:19 if everything fails, you might need to include it into squashfs Aug 05 14:04:08 its 2027kb Aug 05 14:04:45 Mem: 13588 10560 3028 0 368 Aug 05 14:05:00 hmm Aug 05 14:05:12 can i install that ummm Aug 05 14:05:27 whats it called... some kind of swap thing that starts with 'c' Aug 05 14:05:39 compcache? Aug 05 14:05:42 yes Aug 05 14:05:50 never used it, don't know Aug 05 14:05:59 a network fs mount would probably work too Aug 05 14:06:08 its the .ko file in tmpfs that eats the whole ram Aug 05 14:06:15 and then a big chunk of ram if it gets loaded Aug 05 14:06:24 both at the same time is probably too much Aug 05 14:07:59 hmm compcache didnt wanna work this round.. Aug 05 14:08:23 jow_laptop, can i mount cifs with busybox? Aug 05 14:08:43 * RealOpty makes new image Aug 05 14:10:16 jow_laptop, we can break this thing if we need too... i have jtag :) Aug 05 14:12:42 kk flashing brb Aug 05 14:18:17 jow_laptop, theres nothing new in dmesg Aug 05 14:19:50 Linux version 2.6.34.1 (will@bedroom) (gcc version 4.3.3 (GCC) ) #6 Thu Aug 5 07:12:17 MST 2010 <- just compiled that kernel Aug 05 14:22:21 you sure wl.ko got replaced? Aug 05 14:23:22 nope. Aug 05 14:23:27 how should i check? Aug 05 14:23:57 when i compiled all i told the compiler was 'make' Aug 05 14:25:07 jow_laptop, where is the wl.ko located on the flash? Aug 05 14:25:14 ill do a md5sum Aug 05 14:25:17 and compair. Aug 05 14:29:09 jow_laptop, md5sums dont match Aug 05 14:32:34 odd... Aug 05 14:32:49 the patch is applied to the build dir. Aug 05 14:37:51 both patches Aug 05 14:38:08 jow_laptop, PM me any new u have. i need to get some rest Aug 05 16:10:52 nbd before i go i'll post this here again (could be important?) because it isn't so crowded here *hrhr* Aug 05 16:10:53 nbd i get the following message on my ath9k client "ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020" - don't know when it Aug 05 16:10:53 started ... my logread is full of it - after it the association was successfull (that was the last message in logread) Aug 05 16:12:48 l8er Aug 05 17:17:32 Memphiz: https://dev.openwrt.org/ticket/7700 <- no patch, no fix!!!11one. Also, why not the newer rc? Aug 05 17:43:13 KanjiMonster: the newest rc works for me (i've forgotten to submit my additional post *hrhr*) ... so it works with the newest rc Aug 05 17:43:48 KanjiMonster: i'll attach a fix right now (was at work when i opened the ticket) Aug 05 17:44:02 err a patch (its no fix ...) Aug 05 17:46:23 :D Aug 05 17:47:19 KanjiMonster: is a svn diff in trunk/feeds/packages sufficient? Aug 05 17:50:49 did it ... Aug 05 17:53:26 nbd i know you are full of work ... but i only want to be sure ... did you note the "switch of gpio mode" for the ar71xx flash on your TODO list? (for fixing the kexec issue with mac adr and ath9k eeprom) ... Aug 05 17:53:34 "switch off" i mean Aug 05 18:03:53 does anyone know if the new minstrel_ht rate control for ath9k brings any benefits - is it wise to activate it? Aug 05 18:05:24 it's already activated per default Aug 05 18:08:15 since when Aug 05 18:08:35 a few month? Aug 05 18:17:22 loswillios: mhh strange ... its not selected here on ar71xx Aug 05 18:30:45 Memphiz: a diff should be enough. I am thinking about whether selecting the rc should depend on the kirkwood target or whether it's okay for all targets since it's trunk anyway, so breakage should be expected ;) Aug 05 18:31:56 Memphiz: http://nbd.name/gitweb.cgi?p=openwrt.git;a=commit;h=a1b55c5c7b849f7d57df504b5707fbb4e33095f1 Aug 05 18:35:18 loswillios: thx ... somehow it isn't reflected in may menuconfig Aug 05 18:35:27 anyways ... i have to go off for now ... bye Aug 05 20:08:04 build #80 of ep93xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/80 Aug 05 20:24:51 Hauke, ping Aug 05 20:25:05 RealOpty: pong Aug 05 20:25:36 Hauke, i found some opensource program that generates the .chk file :) Aug 05 20:26:06 RealOpty: that's nice Aug 05 20:26:19 so now its possible to flash openwrt from the factory firmware. and the cfe will accept this too! Aug 05 20:26:21 indeed Aug 05 20:26:45 Also ive mapped out the LEDs and reset button for diag.c Aug 05 20:28:19 Hauke, here is the patch https://dev.openwrt.org/ticket/7702 check it out :) Aug 05 20:28:38 jow_laptop, ping Aug 05 20:40:55 Hauke, any tips on trying to make wifi work? Aug 05 20:46:53 RealOpty: no not realy Aug 05 20:47:17 RealOpty: what wireless chip do you have? Aug 05 22:37:11 RealOpty: pong Aug 05 22:37:14 build #78 of cobalt is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/78 Aug 05 22:58:15 wtf damn bugs in ubuntu screensaver.... Aug 05 22:58:30 xMff, im back now Aug 05 22:58:35 what should we do next Aug 05 22:59:41 RealOpty: you got the debug driver loaded? Aug 05 23:00:10 im not sure really. Aug 05 23:00:28 is there some command i have to tell make to incluce it in the image? or rebuild it or something? Aug 05 23:02:03 how did you rebuild? Aug 05 23:02:10 'make' Aug 05 23:02:11 anything in dmesg? Aug 05 23:02:19 nope just the kernel taint Aug 05 23:02:24 did you clean the boradcom-wl package before? Aug 05 23:02:36 every time i rebuilt it, yes Aug 05 23:02:51 whats the size of wl.ko in /lib/modules/*/ ? Aug 05 23:03:35 1962536 Aug 5 14:10 wl.ko Aug 05 23:03:55 run "strings" on it Aug 05 23:06:10 xMff, and grep dbg ? Aug 05 23:06:20 no just paste it somewhere Aug 05 23:11:42 xMff, http://paste.ubuntu.com/473747/ Aug 05 23:12:35 RealOpty: hm, it does not seem to contain debug messages Aug 05 23:13:24 what commands should i run to make sure the patches are applied? Aug 05 23:13:44 it should happen automatically Aug 05 23:13:52 thats what i thought :/ Aug 05 23:14:06 can you run strings on build_dir/linux-brcm47xx/broadcom-wl-*/driver/wl.ko ? Aug 05 23:16:06 xMff, http://paste.ubuntu.com/473751/ Aug 05 23:17:22 RealOpty: oops sorry. it has debug. Aug 05 23:17:35 which one? Aug 05 23:17:38 both Aug 05 23:17:47 weird Aug 05 23:17:55 wanna see the dmesg? Aug 05 23:17:59 yes Aug 05 23:18:01 it dont say anything Aug 05 23:18:03 but np Aug 05 23:18:43 xMff, http://paste.ubuntu.com/473752/ Aug 05 23:20:17 RealOpty: did you ever try b43 with this board? Aug 05 23:20:44 yes Aug 05 23:20:52 it didnt detect wifi also Aug 05 23:20:56 ah I see Aug 05 23:21:05 its a brcm4321 with N protocol. Aug 05 23:21:13 on my wrt54gs, my SSB reports something like this: Aug 05 23:21:13 ssb: Core 1 found: IEEE 802.11 (cc 0x812, rev 0x07, vendor 0x4243) Aug 05 23:21:22 your ssb does not find a 802.11 core Aug 05 23:21:27 ah. Aug 05 23:21:42 So how do i tell kernel to find the wifi? Aug 05 23:21:42 maybe the radio is wired differently Aug 05 23:21:51 mmhmm Aug 05 23:21:53 sdio or something Aug 05 23:22:43 do you have a bootlog of the vendor firmware? Aug 05 23:22:55 yes Aug 05 23:23:56 and can you install lspci ? Aug 05 23:24:08 and run lspci -vv Aug 05 23:24:38 http://paste.ubuntu.com/473754/ (sorry its kinda messy, i was poking around with the firmware) Aug 05 23:25:36 xMff, i take it i have to install lspci from makemenu? Aug 05 23:25:45 you can try opkg too Aug 05 23:26:56 is the package name lspci ? Aug 05 23:27:07 pciutils iirc Aug 05 23:27:37 yep thats the one Aug 05 23:28:48 xMff, http://paste.ubuntu.com/473756/ Aug 05 23:29:05 heh. ok Aug 05 23:29:14 thats what i said lol Aug 05 23:29:52 anyway, its a BCM4329 and google indicates that it is SDIO Aug 05 23:30:12 xMff, would having the source of the factory firmware help any? Aug 05 23:30:18 no Aug 05 23:30:19 and would installing dd-wrt help any? Aug 05 23:31:46 though its not that easy getting things to work in dd-wrt lol Aug 05 23:33:02 well, I need to read into the sdio stuff Aug 05 23:33:05 it might take a while Aug 05 23:33:17 for sure. thanks though :) Aug 05 23:34:07 xMff, i have a dmesg on a ticket thats from dd-wrt. not sure if it says anything interesting. Aug 05 23:34:25 https://dev.openwrt.org/attachment/ticket/7630/netgear-dd-wrt-dmesg Aug 05 23:34:26 let me try something Aug 05 23:35:14 kk, brb Aug 05 23:35:17 nope, not interesting Aug 05 23:40:36 it could be that the driver needs to be recompiled for sdio busses Aug 05 23:40:42 need to talk to nbd Aug 05 23:40:51 probably tomorrow or so Aug 05 23:41:43 xMff, here is the first dmesg from openwrt before i had to recover the device with jtag Aug 05 23:41:44 https://dev.openwrt.org/attachment/ticket/7630/firstdmesg.txt Aug 05 23:42:34 xMff, ah ok :) Aug 05 23:42:49 im just happy openwrt is running on the thing :) Aug 05 23:43:00 using another device for wifi atm Aug 05 23:43:05 for b43/ssb there was some sdio support patch Aug 05 23:43:15 but I have no idea whether it is in mainline already Aug 05 23:43:32 http://lwn.net/Articles/351493/ Aug 05 23:44:39 you could check kernel_menuconfig and see whether ssb offers some sdio related options Aug 05 23:44:47 for sure Aug 05 23:46:14 xMff, yes Aug 05 23:46:16 it does Aug 05 23:46:18 check this Aug 05 23:46:34 xMff, http://paste.ubuntu.com/473763/ Aug 05 23:47:06 ok, if you have some time, give that one a try together with b43 and see what happens Aug 05 23:47:27 xMff, there is also this one Aug 05 23:47:29 Prompt: Intel Wireless Multicom WiMAX Connection 3200 over SDIO Aug 05 23:47:39 no, thats unrelated Aug 05 23:47:46 if i remember correctly, the factory firmware said something about intel Aug 05 23:47:46 ok Aug 05 23:48:14 maybe your device has some wimax chip too but I kinda doubt that Aug 05 23:59:38 xMff, i dont see this package in the location that it says its in. Aug 06 00:00:25 you might need to edit target/linux/brcm47xx/config-2.6.* Aug 06 00:00:32 and add the symbol there Aug 06 00:00:38 CONFIG_xxx=y Aug 06 00:01:28 so add CONFIG_B43_SDIO=y Aug 06 00:01:48 yes Aug 06 00:01:58 hmm Aug 06 00:02:15 if its only related to B43 it should go into package/mac80211/ maybe Aug 06 00:02:24 xMff, maybe a SSB_SDIOHOST_POSSIBLE=y too? Aug 06 00:03:43 no, that sounds like some arch specific flag Aug 06 00:04:01 the POSSIBLE should go into the kenrel config I think Aug 06 00:12:19 xMff, http://paste.ubuntu.com/473774/ Aug 06 00:12:34 that ones required yeh? Aug 06 00:12:45 yes Aug 06 00:13:04 i dont see it in the k_makemenu Aug 06 00:13:52 i added config_b43_sdio=y to the targ/lin/br47/conf.2.6.24 Aug 06 00:15:27 2.6.34* Aug 06 00:17:28 xMff, Aug 06 00:17:47 does this mean that b43 has to mean not selected Aug 06 00:17:48 Depends on: NETDEVICES [=y] && WLAN [=y] && B43 [=n] Aug 06 00:19:36 jow * r22499 /trunk/tools/sed/patches/ (. 010-amd64_freebsd_build.patch): [tools] sed: fix compilation on FreeBSD/AMD64 Aug 06 00:19:58 no it means that it is currently not selected Aug 06 00:20:26 oh weird. Aug 06 00:20:32 normal Aug 06 00:20:36 oh Aug 06 00:20:40 b43 is built outside of the kernel Aug 06 00:20:46 by package/mac80211 Aug 06 00:20:51 ok well ill be back soon gotta run Aug 06 00:56:20 jow * r22500 /trunk/package/ipset/patches/ (. 001-2.6.35.patch): [package] ipset: fix ipset for kernel 2.6.35 Aug 06 00:58:52 jow * r22501 /trunk/package/kernel/modules/netsupport.mk: [package] kernel: package L2TP v3 ("Pseudowire") modules, restrict them to 2.6.35 Aug 06 01:05:54 jow * r22502 /packages/net/l2tpv3tun/ (. Makefile): [packages] add l2tpv3tun - control utility for pseudowire tunnels, patch by Thomas Heil Aug 06 01:33:25 jow * r22503 /trunk/target/linux/ (generic/config-2.6.35 x86/config-default): [kernel] add missing symbols for 2.6.35 Aug 06 01:36:16 jow * r22504 /trunk/target/linux/generic/config-2.6.35: [kernel] fix order fail in previous commit Aug 06 01:39:42 ok im back Aug 06 01:40:26 xMff, im not having success getting this to sdio thing to work. Aug 06 01:40:32 ok Aug 06 01:42:32 xMff, do i need to add the experimental=y option to config file? Aug 06 01:42:41 maybe, not sure Aug 06 01:58:58 xMff, should i enable SSB debugging? Aug 06 02:01:38 jow * r22505 /trunk/target/linux/generic/config-2.6.34: [kernel] add missing symbols for 2.6.34, explicitely enable bootmem - required by crashlog (#7695) Aug 06 02:02:23 jow * r22506 /trunk/target/linux/generic/config-2.6.35: [kernel] explicitely enable bootmem on 2.6.35 - required by crashlog (#7695) Aug 06 02:05:11 jow * r22507 /trunk/package/kernel/modules/netsupport.mk: [package] kernel: fix missing symbol introduced by l2tp Aug 06 02:13:20 xMff, https://dev.openwrt.org/attachment/ticket/7702/WNR834BV2_Support.patch commit that (please) :) Aug 06 02:38:39 xMff, what kernel version does Kamikaze use? Aug 06 02:38:44 8.09 branch Aug 06 02:47:35 anyone know if the package kmod-brcm-wl-mimo (in 8.09) is for 802.11n ? **** ENDING LOGGING AT Fri Aug 06 02:59:58 2010