**** BEGIN LOGGING AT Thu Dec 31 02:59:58 2015 Dec 31 09:29:52 freemangordon: I fixed some problems in libcal & osso-product-info Dec 31 09:56:39 Pali: great Dec 31 09:56:49 changes are in git Dec 31 09:56:50 is there deb? Dec 31 09:56:58 but do now if it helps you Dec 31 09:57:00 ok, I'll build and test Dec 31 09:57:01 no deb yet Dec 31 09:59:26 ok, looks like I am running stock versions of both of those so I dont need to upgrade :) Dec 31 10:00:18 jonwil: grab my CAL , mount it and test Dec 31 10:00:22 :) Dec 31 10:01:23 soup slayer? Dec 31 10:01:28 uhm Dec 31 10:01:29 hi Dec 31 10:01:38 this better Dec 31 10:01:51 i'm trying an irc proxy Dec 31 10:02:04 kind of like screen but for irc Dec 31 10:03:25 yup Dec 31 10:03:30 pretty useful Dec 31 10:03:40 in case you dont have stable shell Dec 31 10:03:44 this should let me connect with n900's xchat and receive a backlog of last things spoken and all that Dec 31 10:04:03 i usually ssh this server and screen -rd an irssi Dec 31 10:04:14 but i like n900's xchat Dec 31 10:04:17 ceene: Thanks to Juhani Mäkelä we now have the "official" parse-certdata-txt.c file used to update the root certificates in Maemo Dec 31 10:04:26 jonwil: cool! Dec 31 10:04:28 so we dont need the python script that you edited last night Dec 31 10:04:32 lol Dec 31 10:04:37 all that work for nothing :P Dec 31 10:05:02 I sent a follow up question asking what to do with the output of parse-certdata-txt and how to get said output into the actual root CA sture Dec 31 10:05:03 but it's okay, it's better having the original! Dec 31 10:05:05 store Dec 31 10:05:17 seems look a cool guy, doesn't he? Dec 31 10:05:36 well it seems like he is willing to share info Dec 31 10:05:48 cool enough for us Dec 31 10:06:25 hopefully he responds with the rest of the steps to take :) Dec 31 10:07:16 then we can have something in SVN where you just download the latest certdata.txt from mozilla and run some things and get an up-to-date maemosec-certman-common-ca package :) Dec 31 10:11:06 and one part of the security system at least will be up-to-date Dec 31 10:11:53 that'd be a great step Dec 31 10:12:48 and alongside openssl1.0.2e we're cutting edge on a 6 years old device Dec 31 10:15:11 :) Dec 31 10:15:20 my laptop is from 2006 Dec 31 10:18:14 so Dec 31 10:18:16 hi Dec 31 10:18:21 uhm Dec 31 10:18:22 funny Dec 31 10:18:32 i can connect with two clients to the proxy Dec 31 10:18:35 at the same time Dec 31 10:18:37 cool :) Dec 31 10:18:52 ceene: you're a cool guy, man Dec 31 10:19:02 you're talking about a bouncer? Dec 31 10:19:10 and speaking my name triggers a notification on n900 Dec 31 10:19:21 apt-get install bip :) Dec 31 10:19:33 yes, something like that Dec 31 10:19:55 ceene: Updating OpenSSL is a great start but we also really need to update NSS inside microb Dec 31 10:20:07 how does one start doing that? Dec 31 10:20:13 to be honest, i don't use microb Dec 31 10:20:21 i use midori running on an easy debian image Dec 31 10:20:27 which i find to be blazing fast Dec 31 10:20:57 although i understand microb is used too by things like modest Dec 31 10:21:08 Yes microb is used for modest Dec 31 10:21:14 and also for conversations app Dec 31 10:21:17 and nokia maps Dec 31 10:21:49 Anyone here running a stock Maemo Fremantle PR1.3.1 install? (i.e. not running CSSU or updated versions of any system packages) Dec 31 10:22:52 jonwil: last time I checked, it seemed like NSS should keep some kind of backward compat with older apps Dec 31 10:23:16 it's called libnss3 both on maemo and debian sid Dec 31 10:23:21 (ABI-wise) Dec 31 10:23:23 so it's probably compatible Dec 31 10:23:32 and only repackaging up to date libnss3 is enough? Dec 31 10:24:06 so we might try building latest libnss3 Dec 31 10:24:08 ceene: maybe Dec 31 10:24:36 if it's not, well... we need to replace the libnss3 source in the microb-engine tree and see what happens Dec 31 10:24:52 (and/or diff source, bleh, the usual stuff) Dec 31 10:34:31 hmmm, I still cant find out how libcurl on a stock N900 gets its root certificates Dec 31 10:42:04 Pali: https://github.com/community-ssu/osso-product-info/blob/master/osso-product-info.c#L87 Dec 31 10:42:06 :) Dec 31 10:42:57 *p = 0; if(osso_set_product_info(code, p) < 0) Dec 31 10:43:00 here it is Dec 31 10:43:06 yes Dec 31 10:43:30 but first fix why '-q' param crash! Dec 31 10:43:39 also, are you sure "RX-51_2009SE_21.2011.38-1_PR_MR0" should be in quotes Dec 31 10:43:52 Pali: it no longer crashes, after your fixes Dec 31 10:43:56 quotes are removed by bash Dec 31 10:44:06 sure? Dec 31 10:44:08 freemangordon: and what is output now? Dec 31 10:44:35 ~ # osso-product-info -q OSSO_VERSION Dec 31 10:44:35 ERROR: Could not find value for key /device/sw-release-ver Dec 31 10:44:35 Error getting variable OSSO_VERSION value Dec 31 10:44:50 shell remove quotes everytime, so if you want to pass ' or " into shell you need to escape it Dec 31 10:44:57 ok Dec 31 10:45:04 freemangordon: great, query it fixed Dec 31 10:45:34 that *p = 0; should be p ++; IMO Dec 31 10:45:37 correct? Dec 31 10:46:03 or remove it alltogether and pass p + 1 to the function Dec 31 10:46:11 *p=0; p++; Dec 31 10:46:22 you need to remove char '=' Dec 31 10:46:34 so optarg is key Dec 31 10:46:37 and p is value Dec 31 10:46:45 format is "key=value" Dec 31 10:48:13 Pali: but oh, yes Dec 31 10:50:27 ++p ? Dec 31 10:51:41 KotCzarny: in C without checking lvalue is ++p and p++ same, gcc optimize p++ code Dec 31 10:52:24 ahm Dec 31 10:52:39 didnt know gcc does it anyway Dec 31 10:55:46 Pali: hmm, another bug https://github.com/community-ssu/osso-product-info/blob/master/libossoproductinfo.c#L76 Dec 31 10:56:13 hm? Dec 31 10:56:24 we search for /device/sw-release-ver, but set sw-release-ver Dec 31 10:56:34 its ok Dec 31 10:56:42 sw-release-ver is CAL key Dec 31 10:56:58 /device/sw-release-ver is sysinfod DBUS key Dec 31 10:57:39 "sw-release-ver" is correct CAL key (section) name Dec 31 10:57:49 yes, but see what get_product_info does Dec 31 10:59:13 maybe replace product_info_keys[i].name with key at line 370? Dec 31 10:59:23 but both in this case should be valid Dec 31 11:00:25 get_info call sysinfo functions Dec 31 11:00:59 and? Dec 31 11:01:00 oh, maybe I should rebott Dec 31 11:01:14 yes, sysinfo read CAL suff when starting Dec 31 11:01:16 for the value in CAL to take effect, lemme try Dec 31 11:01:19 ok Dec 31 11:01:21 makes sense Dec 31 11:01:23 either restart sysinfo Dec 31 11:01:26 or reboot device Dec 31 11:01:38 rebooting Dec 31 11:01:40 stop sysinfod && start sysinfod Dec 31 11:02:03 too late, also I wont to boot into 4.4 Dec 31 11:02:07 *want Dec 31 11:07:21 Pali: it is ok after reboot Dec 31 11:07:48 great Dec 31 11:07:58 push fixes to git Dec 31 11:08:06 will do Dec 31 11:08:07 and ideally also new packages into cssu-devel Dec 31 11:08:33 no, as my SB target is for linux, don;t want to switch it now Dec 31 11:08:53 maybe later Dec 31 11:09:55 oh, i see libnss3' source is microb-engine Dec 31 11:11:34 Pali: http://talk.maemo.org/showthread.php?p=1493038#post1493038 Dec 31 11:11:36 :) Dec 31 11:12:05 :) Dec 31 11:12:11 great :-) Dec 31 11:12:20 freemangordon: cool! Dec 31 11:13:01 Pali: BTW osso-product-info is missing CFLAGS for optimization and debug symbols Dec 31 11:13:02 freemangordon: can you test HW if working? Dec 31 11:13:09 what HW? Dec 31 11:13:13 wifi, ... Dec 31 11:13:17 touchscreen, .. Dec 31 11:13:19 etc Dec 31 11:13:37 basicly what should be supported according to elinux.org/N900 Dec 31 11:13:45 test for regressions Dec 31 11:14:12 freemangordon: modem is working or not? Dec 31 11:14:28 and SGX? Dec 31 11:14:36 Pali: comeon, this is xterm Dec 31 11:14:55 see wifi icon in status menu :) Dec 31 11:15:44 Pali: going to install syslog and reboot, will pastebin boot log in a minute Dec 31 11:18:03 hmm, where is syslog? in SDK repo? Dec 31 11:19:24 syslog? Dec 31 11:19:27 in nokia ssu repo Dec 31 11:19:31 you want apt-get sysklogd Dec 31 11:19:31 ksyslogd Dec 31 11:20:23 http://maemo.org/packages/view/sysklogd/ Dec 31 11:20:31 its in device repo Dec 31 11:21:19 "E: Couldn't find package sysklogd" :) Dec 31 11:21:40 it is in SDK repo Dec 31 11:22:19 you need also klogd Dec 31 11:22:31 yeah Dec 31 11:22:47 http://repository.maemo.org/pool/maemo5.0/free/s/sysklogd/sysklogd_1.5-5maemo7+0m5_armel.deb Dec 31 11:22:58 http://repository.maemo.org/pool/maemo5.0/free/s/sysklogd/klogd_1.5-5maemo7+0m5_armel.deb Dec 31 11:23:18 I installed the repo Dec 31 11:24:35 Pali: leds work Dec 31 11:24:42 RGB led that is Dec 31 11:25:11 seems like mce does not disable backlight Dec 31 11:26:06 and kbd backlight as well Dec 31 11:27:54 Pali: http://pastebin.com/haRz2j2Y Dec 31 11:28:04 also, charging doesn't seem to work Dec 31 11:29:28 sr_init: No PMIC hook to init smartreflex Dec 31 11:29:35 yeah Dec 31 11:29:55 it has always been like that Dec 31 11:30:07 omap-sham 480c3000.sham: initialization failed. Dec 31 11:30:24 omap-sham worked, I fixed in older version Dec 31 11:30:47 isp1704_charger isp1704: failed to register isp1704 with error -19 Dec 31 11:30:48 could be Dec 31 11:30:54 --> isp is broken Dec 31 11:31:01 needed for charger autodetection Dec 31 11:31:07 it worked in older version too! Dec 31 11:31:33 but -> "omap_ssi 48058000.ssi-controller: ssi controller 0 initialized (1 ports)!" Dec 31 11:31:55 and omap_ssi_port 4805a000.ssi-port: ssi port 0 successfully initialized (cawake=151) Dec 31 11:32:19 BT seems to work as well Dec 31 11:32:58 BT is working?? Dec 31 11:33:09 wait, maybe this is older version Dec 31 11:33:27 modprobe bq27x00_battery failed Dec 31 11:35:24 bq27x00_battery was renamed to bq27xxx_battery Dec 31 11:35:37 it is working, see: "bq27xxx-battery 2-0055: battery is not calibrated! ignoring capacity values" Dec 31 11:36:14 ok Dec 31 11:36:35 cellular: ssc[958]: modem connection cannot be established, retry (9 of 10) Dec 31 11:36:37 :( Dec 31 11:36:44 same problem as previous Dec 31 11:36:49 it is there for a long time Dec 31 11:36:56 but I think you once fixed it Dec 31 11:37:14 yes, and I was thinking the fix is uppstreamed Dec 31 11:37:32 looks like it is not in upstream... Dec 31 11:38:13 yeah Dec 31 11:39:34 libcmtspeechdata: hw6x_backend: unable to open device /dev/cmt_speech ('No such file or directory'). Dec 31 11:40:04 yep Dec 31 11:40:11 do you have drivers loaded? Dec 31 11:40:18 nokia_modem and others? Dec 31 11:40:48 dunno, I rebooted in KP to charge the battery as I got battery low warning Dec 31 11:40:56 will check in a while Dec 31 11:41:54 hmm, my patch seems to be upstreamed Dec 31 11:42:17 freemangordon: edit /sbin/preinit and add "nokia-modem" to line "load_modules rx51_battery" Dec 31 11:42:40 this is early init stage modules loading Dec 31 11:42:46 rx51_battery is needed here for BSI Dec 31 11:43:10 and nokia-modem is needed for gpio stuff in /sbin/preinit Dec 31 11:43:22 ok Dec 31 11:43:40 but now check if nokia-modem is loaded or not Dec 31 11:43:51 ok Dec 31 11:44:37 * jonwil wonders if this kernel work will help with Neo900 Dec 31 11:45:20 yes Dec 31 11:45:29 most of the stuff will be the same Dec 31 11:46:47 My current work is to do whatever I can in support of improving the security of Maemo Dec 31 11:47:09 freemangordon, post output from $ lshal Dec 31 11:47:25 ke-recv needs fixing too... Dec 31 11:48:21 although I will also be devoting a chunk of time to the PC upgrade I am about to undertake :) Dec 31 11:48:40 And to playing Fallout 4 once said PC upgrade is complete :) Dec 31 11:48:58 Pali: ok, just a second Dec 31 11:50:33 [ 278.728698] nokia-modem n900-modem: Could not get gpio 0 Dec 31 11:51:19 Pali: http://pastebin.com/Y9z2uJWY Dec 31 11:51:58 gpio 0 ? Dec 31 11:52:06 looks like bug in nokia-modem or in DTS Dec 31 11:52:08 Pali: kbd open sensor doesn't seem to work either Dec 31 11:52:15 Pali: yeah Dec 31 11:52:21 yes, it is GPIO Dec 31 11:52:36 and mce search for /sys/ ... gpio-swicth Dec 31 11:52:57 this needs to be fixed in mce to use /sys/class/gpio/ Dec 31 11:53:03 yep Dec 31 11:53:09 now kernel export just static data for gpio-switch Dec 31 11:54:02 anyway in upstream kernel keyboard open sensors is input device in /dev/input/event... Dec 31 11:54:06 so needs to be fixed Dec 31 11:54:55 I will fix mce, probably later today Dec 31 11:55:56 Pali: seems devm_gpiod_get_index fails Dec 31 11:56:28 something in DTS Dec 31 11:57:46 Pali: sorry, still can;t grok that DTS stuff :( Dec 31 12:04:30 Pali: gpio 0 is index, not the gpio istelf Dec 31 12:05:18 gpio index 0 is 70 Dec 31 12:07:13 Pali: and it is used in rx51_gpio_switches in pdata-quirks.c, could it be that gpio switch driver is using it? Dec 31 12:16:23 freemangordon: maybe, check if .failed is set to true Dec 31 12:16:34 if is failed = true, then it is *not* used Dec 31 12:18:53 it is not set Dec 31 12:20:08 RX51_GPIO_CMT_APESLPX is gpio 70 Dec 31 12:20:24 so? Dec 31 12:20:26 are you sure that isp1704/7 needs cmt_apeslpx? Dec 31 12:20:41 I am talking about nokia-modem :) Dec 31 12:21:01 ah, right! Dec 31 12:21:27 but sscd try to use gpio_switch and enabled/disable modem itself Dec 31 12:21:32 not via nokia_modem.ko Dec 31 12:22:03 freemangordon: try to boot with nokia-modem.pm=1 Dec 31 12:22:16 power management should be handed by userland Dec 31 12:22:20 ok Dec 31 12:23:21 freemangordon: no Dec 31 12:23:33 kernel ignores gpio only if pm is set to zero Dec 31 12:23:48 so boot with nokia-modem.pm=0 Dec 31 12:24:02 nokia-modem.pm=0 is also in CONFIG_CMDLINE Dec 31 12:24:03 this is what is in the config file Dec 31 12:24:20 but why then you have line "nokia-modem n900-modem: Could not get gpio 0" in dmesg? Dec 31 12:24:51 "Enable power management (0=disabled, 1=userland based [default])"); Dec 31 12:24:51 line "Could not get gpio" is only in function nokia_modem_gpio_probe() Dec 31 12:25:01 this should be 1, not zero Dec 31 12:25:06 and function nokia_modem_gpio_probe() is called only if pm != 0 Dec 31 12:25:37 no, it must be zero, see nokia_modem_gpio_probe() function Dec 31 12:25:55 it just export gpio numbers via /sys/class/gpio nothing more Dec 31 12:26:05 same code which is in gpio-switch Dec 31 12:26:28 and we need to export those gpios by gpio-switch, so sscd could handle it by correct path Dec 31 12:26:47 hmm, correct Dec 31 12:27:14 rmmod nokia_modem Dec 31 12:27:15 oh, wait, I did modprobe by hand Dec 31 12:27:21 * Wizzup is excited since he ran into these modem issues some time ago and got stuck Dec 31 12:27:21 modprobe nokia_modem pm=0 Dec 31 12:27:23 lemme try again Dec 31 12:27:25 * Wizzup is silent again Dec 31 12:27:27 :nod: Dec 31 12:29:05 Pali: but then I'd rather fix pminit Dec 31 12:29:15 *preinit Dec 31 12:29:39 simple modprobe nokia_modem should work (without args) Dec 31 12:30:03 s/_/-/ Dec 31 12:30:03 Pali meant: simple modprobe nokia-modem should work (without args) Dec 31 12:30:08 if you have set nokia-modem.pm=0 in cmdline Dec 31 12:30:42 put it does not Dec 31 12:30:57 * freemangordon is afk for a while Dec 31 12:40:03 Pali: for some reason cmdline does not work Dec 31 12:41:06 bah :-( Dec 31 12:41:23 ~ # cat /proc/cmdline Dec 31 12:41:23 init=/sbin/preinit ubi.mtd=rootfs root=ubi0:rootfs rootfstype=ubifs rootflags=bulk_read,no_chk_data_crc rw mtdoops.mtddev=log console=tty0 console=ttyO2 omapfb_vram=7M omapfb.mode=lcd:848x480-16 nokia-modem.pm=0 Dec 31 12:41:47 $ dpkg -S /sbin/modprobe Dec 31 12:41:47 module-init-tools: /sbin/modprobe Dec 31 12:41:53 going to look into source Dec 31 12:44:14 freemangordon: we have old version of module-init-tools without support for /proc/cmdline :-( Dec 31 12:44:18 this should be fixed Dec 31 12:45:47 freemangordon: so create /etc/modprobe.d/nokia-modem.conf with options nokia-modem pm=0 Dec 31 12:46:07 I will fix default value in linux-n900 git tree Dec 31 12:46:22 no, don't Dec 31 12:46:34 we'd better have .conf file Dec 31 12:47:10 ok, conf file Dec 31 12:48:56 soon or later we should make a support package, will put it there Dec 31 12:57:11 Dec 31 14:56:09 Nokia-N900 cellular: ssc[3314]: CMT self tests passed. Dec 31 12:57:22 Pali: ^^^ Dec 31 12:57:37 great :-) Dec 31 12:57:47 can you test SIM card? Dec 31 12:57:59 yup. but conf file is not used in preinit :( Dec 31 12:58:50 or I did not put the correct contents in it Dec 31 13:01:49 Pali: also, emmc is not mounted Dec 31 13:02:07 /opt? Dec 31 13:02:17 MyDocs Dec 31 13:02:27 maybe fsck is running? Dec 31 13:03:29 dunno Dec 31 13:04:42 ps | grep fsck Dec 31 13:04:47 looks correc, ain't? "options nokia-modem pm=0" Dec 31 13:04:59 yes Dec 31 13:05:20 preinit modprobe does not use it Dec 31 13:06:09 anyway, will stop for a while, until the battery is charged a bit Dec 31 13:06:25 Pali: any idea how to test isp? Dec 31 13:06:34 isp = camera? Dec 31 13:06:40 no, charger detection Dec 31 13:07:05 cat /sys/class/power_supply/... Dec 31 13:07:23 err, cat --> ls Dec 31 13:07:28 it should be there Dec 31 13:09:54 Pali: wanna remote access to the device? Dec 31 13:10:06 no need... Dec 31 13:10:26 ~ # ls /sys/class/power_supply/ Dec 31 13:10:27 bq24150a-0 bq27200-0 rx51-battery Dec 31 13:10:34 so isp probe fails Dec 31 13:10:44 that error message in dmesg Dec 31 13:11:30 hmm, modprobe succeed Dec 31 13:12:13 but charging still does not work Dec 31 13:13:50 and it is still not liset in power_supply Dec 31 13:15:47 isp1704_charger isp1704: init gpio 67 Dec 31 13:16:02 modprobe is just loading module Dec 31 13:16:08 rmmod and modprobe again Dec 31 13:16:09 yeah Dec 31 13:16:16 this is the result Dec 31 13:16:29 $ ls /sys/class/power_supply/ Dec 31 13:16:41 [ 514.637176] isp1704_charger isp1704: init gpio 67 Dec 31 13:16:42 [ 514.647979] isp1704_charger isp1704: failed to register isp1704 with error -19 Dec 31 13:16:47 no isp there Dec 31 13:16:57 ah :-( Dec 31 13:17:44 need to debug init routine Dec 31 13:20:25 Pali: will put some printk to see where it fails Dec 31 13:21:12 where does the output of printk go? dmesg? Dec 31 13:21:17 freemangordon: use this http://pastebin.com/GCViTeVF Dec 31 13:21:25 kerio: yes dmesg Dec 31 13:21:48 Pali: ok Dec 31 13:23:05 Pali: not all error cases are covered Dec 31 13:23:17 thats enough Dec 31 13:23:38 like of_get_named_gpio and devm_gpio_request_one Dec 31 13:23:39 ok Dec 31 13:24:36 hm.. there can be problem too? Dec 31 13:25:26 why not? Dec 31 13:25:34 if you check for error... Dec 31 13:27:03 Pali: [ 1130.078521] isp1704_charger isp1704: isp1704_test_ulpi failed Dec 31 13:30:57 hmm, "omap_i2c 48072000.i2c: controller timed out" Dec 31 13:35:08 -19 is -ENODEV Dec 31 13:35:40 freemangordon: so isp1704_test_ulpi fails at some -ENODEV Dec 31 13:40:54 will put pr_err in that function Dec 31 13:48:59 Pali: isp1704_read(isp, ULPI_SCRATCH); does not return 0xAA Dec 31 13:49:12 and what? Dec 31 13:50:08 freemangordon: can you compare result from that isp1704_read with some 3.x kernel? Dec 31 13:50:12 just a second, to put some more traces Dec 31 13:56:57 Pali: musb-hdrc musb-hdrc.0.auto: musb_init_controller failed with status -517 Dec 31 13:57:10 and "HS USB OTG: no transceiver configured" before that Dec 31 13:57:29 right, isp will not work without musb Dec 31 13:58:35 something in the .config? Dec 31 13:59:28 no idea... Dec 31 13:59:44 do you have musb in /sys/ ? Dec 31 14:00:06 # CONFIG_USB_ULPI_BUS is not set Dec 31 14:00:08 do you have g_nokia gadge active? Dec 31 14:00:27 I think CONFIG_USB_ULPI_BUS is not relevant Dec 31 14:00:46 this is from 4.2 Dec 31 14:01:09 was it enabled in 4.2? Dec 31 14:01:25 no idea Dec 31 14:01:29 but see http://cateee.net/lkddb/web-lkddb/USB_ULPI_BUS.html Dec 31 14:01:30 look into git Dec 31 14:02:14 hm... looks like ulpi code is now in kernel Dec 31 14:02:37 it was not enabled Dec 31 14:02:48 I.ll enable that Dec 31 14:02:50 this probably duplicates isp charger code... Dec 31 14:02:56 ah Dec 31 14:04:25 hmm, CONFIG_USB_ULPI=y Dec 31 14:23:28 Pali: could it be that dtb is broken? Dec 31 14:23:36 maybe Dec 31 14:23:52 how to check? Dec 31 14:24:07 as I see usb-phy configured in the dts file Dec 31 14:24:54 oh, wait Dec 31 14:25:01 it should be in /sys? Dec 31 14:26:57 hmm, we have usb_otg_hs Dec 31 15:00:23 Pali: shouldn't isp1704 return -EDEFER in case it cannot init ulpi? Dec 31 15:01:35 there is no explicit defer Dec 31 15:02:04 Pali: as -517 is -EPROBE_DEFER Dec 31 15:02:15 so actually musb is ok Dec 31 15:02:21 ok Dec 31 15:02:29 them problem is in isp1704 Dec 31 15:02:39 yes Dec 31 15:02:42 try 3.x kernel what that method returns and compare Dec 31 15:02:51 I don;t have 3.x kernel Dec 31 15:03:33 but what I see is that isp1704_read(isp, ULPI_SCRATCH); returns 0x00 instead of 0xAA Dec 31 15:06:08 Pali: maybe isp should power up the usb? Dec 31 15:06:34 no idea whats with that Dec 31 15:06:42 you can try to compile g_nokia as external module Dec 31 15:06:47 now it is compiled into zImage Dec 31 17:06:57 Pali: if I remove the check for 0xAA return value, the driver registers ok and cherging works Dec 31 17:07:03