**** BEGIN LOGGING AT Tue Sep 27 02:59:57 2011
Sep 27 09:51:32 <CIA-73> loswillios * r28311 /packages/sound/pulseaudio/Makefile: [packages] pulseaudio: update to 1.0 final
Sep 27 12:38:47 <karlp> if gpiodev is deprecated, on trunk, shouldn't I have a /sys/classes/gpio by default? not a /sys/classes/gpiodev ?
Sep 27 12:39:13 <karlp> is there some kernel_menuconfig or regular menuconfig option I need to use to get the sysfs interface?
Sep 27 12:41:54 <mute> isnt it in /sys/bus or something
Sep 27 12:43:06 <karlp> I have /sys/bus/platform/devices/GPIODEV.0,
Sep 27 12:43:14 <karlp> anywhere else in bus I should look?
Sep 27 12:44:15 <mute> i dont recall and i messed up my routers IP 2 days ago so it's just a switch right now hehe
Sep 27 12:46:10 <karlp> pages like http://www.armbedded.eu/node/258 really seem to imply it should be in /sys/class/gpio
Sep 27 12:46:35 <karlp> I presume that something is missing in the board setup then, if it works for gpiodev, but not for sysfs
Sep 27 12:49:47 <mute> yea probably needs a gpio class driver
Sep 27 12:50:35 <larsc> karlp: you need GPIO_SYSFS
Sep 27 12:52:21 <karlp> where in the menu is that? that armbedded page refers to it to, but I couldn't find it
Sep 27 12:52:36 <karlp> is that in kernel_menuconfig->device drivers?
Sep 27 12:52:53 <larsc> yes
Sep 27 12:53:01 <larsc> device drivers->gpio
Sep 27 12:53:33 <larsc> your architecture needs to support gpiolib though
Sep 27 12:53:42 <karlp> right, I think that's what's missing.
Sep 27 12:53:55 <karlp> because I don't have an option for GPIO in device drivers for the kernel config
Sep 27 12:54:36 <karlp> ATHEROS_AR2315 says it does GENERIC_GPIO = y
Sep 27 12:54:42 <larsc> search for GPIOLIB symbol and check if it's requirements are set
Sep 27 12:56:00 <karlp> depends on ARCH_WANT_OPTIONAL_GPIOLIB?
Sep 27 12:56:16 <karlp> do I have to specifically add that?
Sep 27 12:57:08 <larsc> your arch is supposed to select that if it supports gpiolib
Sep 27 12:58:28 <karlp> the board and arch section just has CONFIG_GPIO_SUPPORT=t
Sep 27 12:58:50 <karlp> =y sorry, so I'm guessing when this board was setup, either sysfs was too new, and it ws never changed?
Sep 27 12:59:15 <larsc> possible
Sep 27 13:02:01 <larsc> it's worth a try to simply select ARCH_WANT_OPTIONAL_GPIOLIB?
Sep 27 13:02:08 <karlp> yeah, doing that now.
Sep 27 13:02:21 <karlp> let's see what happens.
Sep 27 14:10:31 <karlp> hmm, it enabled more config options ok, but I still don't have /sys/class/gpio.  What is the name of the kernel module I shoul dhave?
Sep 27 14:11:00 <karlp> on the bright side, that gave me a bunch of links in /sys/class/leds/gpioX
Sep 27 14:22:09 <larsc> hu?
Sep 27 14:22:33 <larsc> those should be in /sys/class/gpio/
Sep 27 14:23:56 <karlp> /sys/class/gpio doesn't exist. I have /sys/class/gpiodev, and /sys/class/leds
Sep 27 14:27:03 <karlp> bah, running make clean removed my changes in build_dir
Sep 27 14:27:13 <karlp> I guess I need to set up quilt again
Sep 27 14:27:43 <larsc> you have selected GPIOLIB and GPIO_SYSFS?
Sep 27 14:29:17 <karlp> I'm having a very hard time keeping track of menu options and KEY_NAMES, but I believe so.
Sep 27 14:29:38 <karlp> I added ARCH_WANT...
Sep 27 14:29:46 <karlp> and will thenr un scripts/diff_config and check
Sep 27 14:32:36 <karlp> cat build_dir/linux-atheros/linux-2.6.37.6/.config shows gpiolib=y and gpiosysfs=y
Sep 27 14:32:42 <karlp> this time I won't run clean first :)
Sep 27 14:55:32 <karlp> hmm, seems to not like that, I guess the existing gpio implementation is hardcoded to be old gpio_dev? http://pastebin.com/L16E4nq9
Sep 27 14:58:06 <[florian]> you probably need to to tweak a little the existing implementation yes
Sep 27 14:59:20 <karlp> I'm reading the docs at http://www.mjmwired.net/kernel/Documentation/gpio.txt
Sep 27 14:59:35 <karlp> is there anything else that's good to read? or any other "good" comparison implementations?
Sep 27 15:02:01 <[florian]> ar71xx, ar7, bcm63xx all have a fairly straight forward implementation of gpiolib
Sep 27 15:05:18 <karlp> would you recommend adding OPTIONAL_GPIOLIB and GENERIC_GPIO, or just all the way to REQUIRE_GPIOLIB?
Sep 27 15:08:44 <[florian]> all the way to REQUIRE_GPIOLIB should be fine
Sep 27 15:39:16 <mazilo> Does anyone here know how to use mod_dingaling to call someone on GTalk (not Google Voice)?
Sep 27 15:53:52 <blogic> what a bizarre name for a plugin
Sep 27 15:55:53 <Chocks> what a bizarre nick
Sep 27 15:58:12 <blogic> ?
Sep 27 16:19:12 <karlp> should I be removing this call then: platform_device_register_simple("GPIODEV", 0, &res, 1); to move from gpiodev to sysfs? (it's in __init ar231x_register_devices()
Sep 27 16:19:43 <[florian]> karlp: you don't need to remove that line, both can co-exist
Sep 27 16:22:04 <karlp> so I should be able to keep the ar23xx_gpio_get/set/direction things too, and it's mostly going to be removing the extra wrappers?
Sep 27 16:22:15 <karlp> I don't need to rewrite them to use a struct gpio_chip instead?
Sep 27 16:24:17 <[florian]> no you should not because gpiodev uses GENERIC_GPIO indiferently
Sep 27 17:53:08 <CIA-73> tripolar * r28312 /packages/net/openssh/Makefile: [packages]: upate openssh to 5.9p1
Sep 27 20:39:57 <Delboy_> blogic: ping
Sep 27 20:54:37 <blogic> yes
Sep 27 20:54:40 <blogic> what ticket
Sep 27 20:54:49 <blogic> and dont slp me :) i am not a slapper
Sep 27 20:54:51 <blogic> :)
Sep 27 20:55:10 <blogic> anyhow i am fixing lantiq tickets as is, so tell me which one to look at
Sep 27 20:55:50 <Delboy_> hmm, what ticket ?
Sep 27 20:55:51 <Delboy_> :D
Sep 27 20:56:12 <Delboy_> i never talked about any ticket
Sep 27 20:56:14 <Delboy_> anyway
Sep 27 20:56:42 <blogic> ok, someone pinged me about a ticket
Sep 27 20:56:47 <cshore> ping jow
Sep 27 20:56:49 <blogic> ok, was not u then
Sep 27 20:56:58 <blogic> what can i do for you  :)
Sep 27 20:57:16 <Delboy_> wanted to ask you about patches for 2.6.39 and 3.0 inside lantiq, there are some changes inside 2.6.39 that are not inside 3.0
Sep 27 20:57:24 <blogic> yes
Sep 27 20:57:30 <blogic> i am rebasing 3.0
Sep 27 20:57:42 <blogic> i will stay at .39 until 3.2
Sep 27 20:57:49 <cshore> hey blogic, do you know what the current preferred method for checking wireless txpower is?
Sep 27 20:57:54 <blogic> 3.2 will get a lot of fixes merged
Sep 27 20:57:57 <Delboy_> aa that was next question
Sep 27 20:58:12 <Delboy_> cuz i wanted to create a patch for those things
Sep 27 20:58:14 <blogic> the patches-3.0 is more a test
Sep 27 20:58:20 <blogic> ok
Sep 27 20:58:29 <Delboy_> but i didn't know should i wait for 3.2
Sep 27 20:58:40 <Delboy_> or do it now for 3.0
Sep 27 20:58:40 <blogic> just pmed you a link
Sep 27 20:59:17 <Delboy_> ok thx
Sep 27 21:00:51 <blogic> cshore: on ath9k ?
Sep 27 21:00:57 <cshore> yes
Sep 27 21:01:36 <cshore> iwlist ath0 txpower is returning 'unknown transmit-power information'
Sep 27 21:03:55 <KanjiMonster> ath0 is madwifi, isn't it?
Sep 27 21:03:58 <EqUaTe> cshore: i think jow's on holiday someplace.. he said he'd be back in a week or something like that yesterday.. or possibly the day before, i lose track.
Sep 27 21:04:06 <EqUaTe> was last time i checked
Sep 27 21:04:33 <cshore> KanjiMonster: hmmm.....maybe that's the problem....I thought I was suing ath9k, but maybe the config got borked
Sep 27 21:04:45 <blogic> cshore: athstats
Sep 27 21:04:55 <blogic> ath0 is madwifi
Sep 27 21:05:06 <cshore> odd, I have ath9k modules
Sep 27 21:05:12 <_trine> ath9k should be wlan0 should it not
Sep 27 21:06:53 <cshore> ok, sounds like something fishy with the config
Sep 27 21:06:57 <KanjiMonster> _trine: right
Sep 27 21:07:31 <builder_> build #85 of at91 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/85
Sep 27 21:07:39 <cshore> how would I be able to tell if I have both ath9k and madwifi?
Sep 27 21:08:13 <blogic> lsmod ?
Sep 27 21:08:15 <EqUaTe> opkg can probably tell you
Sep 27 21:08:21 <EqUaTe> lsmod would only show them if both are loaded
Sep 27 21:08:23 <EqUaTe> which is unlikely
Sep 27 21:08:30 <cshore> blogic: sorry which modules are madwifi
Sep 27 21:08:33 <cshore> well ath9k is loaded
Sep 27 21:08:34 <EqUaTe> imo ath9k probably failed to load as madwifi loaded first
Sep 27 21:08:36 <EqUaTe> hrm
Sep 27 21:08:36 <blogic> ath_pci
Sep 27 21:08:37 <EqUaTe> interesting
Sep 27 21:08:43 <EqUaTe> lsmod | grep ath
Sep 27 21:08:54 <_trine> iifconfig will tell you
Sep 27 21:09:19 <_trine> if you have wlan0 its ath9k if you have ath0 its madwifi
Sep 27 21:09:44 <blogic> cat /proc/bus/pci/devices
Sep 27 21:09:52 <cshore> but ath9k drivers are loaded not madwifi
Sep 27 21:10:05 <blogic> wrong shell ?
Sep 27 21:10:19 <cshore> it's ath9k
Sep 27 21:10:21 <EqUaTe> do you definitely have an ath0 device?
Sep 27 21:10:30 <EqUaTe> you may have named it ath0 in the config somehow..
Sep 27 21:10:40 <cshore> I'm thinking that the config I inherited has been renamed
Sep 27 21:11:00 <cshore> rather has wlan0 renamed as ath0
Sep 27 21:11:07 <_trine> just do wifi detect
Sep 27 21:11:57 <_trine> to a file and look
Sep 27 21:12:26 <cshore> hmmm....wifi isn't being detected by wifi detect
Sep 27 21:12:43 <cshore> but iwconfig works
Sep 27 21:12:45 <_trine> wifi detect > filetolookat
Sep 27 21:13:18 <cshore> _trine: yeah did that...it's empty
Sep 27 21:13:29 <KanjiMonster> cshore: wifi detect won't output anything if there's a non-empty /etc/config/wireless
Sep 27 21:13:40 <cshore> KanjiMonster: ah, doh!
Sep 27 21:13:54 <blogic> cshore: cat /proc/bus/pci/devices    and pastebin the result
Sep 27 21:14:04 <EqUaTe> so mv /etc/config/wireless /etc/config/wireless-temp; wifi detect > test; mv /etc/config/wireless-temp /etc/config/wireless
Sep 27 21:14:32 <cshore> well it detects a mac80211 driver, so I'm thinking it's ath9k
Sep 27 21:15:31 <cshore> blogic: http://pastebin.com/cA10G61r
Sep 27 21:16:14 <cshore> is iwlist deprecated, or should it work?
Sep 27 21:16:15 <blogic> ath9k ;)
Sep 27 21:16:57 <cshore> I think I'm going to have to dig through my inherited config some more though
Sep 27 21:17:51 <KanjiMonster> cshore: iwlist is deprecated, but it I think it should work. If it's recent enough you could try iw (dunno if it also allows *getting* the txpower; I only see setting it on the linux-wireless wiki)
Sep 27 21:18:19 <cshore> KanjiMonster: not iw is only setting
Sep 27 21:18:25 <cshore> *no,
Sep 27 21:18:43 <cshore> KanjiMonster: but iwinfo does work
Sep 27 21:19:00 <cshore> KanjiMonster: it's just not friendly for screen-scraping
Sep 27 21:24:45 <builder_> build #83 of ubicom32 is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/83
Sep 27 21:36:44 <cshore> apparently the interface name is a configuration option in /etc/config/wireless
Sep 27 23:45:21 <CIA-73> jmccrohan * r28313 /packages/utils/lcd4linux/ (Makefile patches/110-uclibc-logarithm-fix.patch): [Package] lcd4linux: update to r1159. Fixes #8897
**** ENDING LOGGING AT Wed Sep 28 02:59:56 2011