Guest desalesouche Posted February 18, 2012 Report Share Posted February 18, 2012 :) About apps to sd: that's android native mechanism - not all apps support this. Thats normal behaviour. Link2sd for example should work, but untested. Dont know about dt a2sd. data2ext is expected to cause problems because parts of /system partition were linked to data in this rom - these shouldn't be moved to sd I believe. With Link2sd you can move individual apps (and their davlik-cache) to sd. This should not conflict with the partition setup of that rom. try it if you need a 2 sd solution for that rom quickly. Could you give some more detail about "after trying to install player" - which and how? i have mobo player apk and his codec for armv6 in stock on my computer....so i install it from sd card... but there is a problem with mobo player installation ...so i could run only few seconds on video with mobo player....then i try to see movie with normal player but i notice he couldn²t run mp4 with 25 frame/s he could only read mp4 with 24 frame/s (very good reading) and mp4 with 23frame/s ( but it jerks/not fluide).... that is true that i saw on post that link2sd is supposed to work fine... Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 18, 2012 Report Share Posted February 18, 2012 try just replace /system/wifi from um840. It will contain both driver and its firmware in that folder. Might hep if you didn't do it yet... Implemented that - replaced /system/wifi with one from um840. With your libhardware_legacy.so and respective symlink in /data, driver loads, no EEPROM exceptions, but the following is in logcat: E/WifiHW ( 1272): Unable to open connection to supplicant on "/data/misc/wpa_supplicant/sta": No such file or directory I'd guess ar6000 might need that aext version of supplicant to communicate with it. Try one from Zte Blade. The issue of not able to turn on wifi for the second time is solved by manually unload stuck wifi kernel module: rmmod ar6000 Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 18, 2012 Report Share Posted February 18, 2012 (edited) it's needed to rename init.huawei.rc ro init.qcom.rc in ramdisk to load that piece. (Or alternatively change qcom->huawei in mkbootimg's cmdline parameter when assembling boot image). There are many wifi related definitions there. This should also help bluetooth, but some diag utility fails on BT start. With that done it started to complain on unsupported ioctl . As I remember people added some placeholders somewhere to satisfy. dmesg: <4>[2016, wpa_supplicant] [ 322.920468] iwioctl: cmd=0x8b0c not allowed in this mode <4>[2016, wpa_supplicant] [ 322.925554] iwioctl: cmd=0x8b0c not allowed in this mode <4>[2016, wpa_supplicant] [ 322.945511] iwioctl: cmd=0x8b0c not allowed in this mode <4>[2016, wpa_supplicant] [ 322.946353] iwioctl: cmd=0x8b0c not allowed in this mode <4>[2016, wpa_supplicant] [ 322.956323] iwioctl: cmd=0x8b0c not allowed in this mode <4>[2016, wpa_supplicant] [ 322.957901] iwioctl: cmd=0x8b0c not allowed in this mode <4>[2016, wpa_supplicant] [ 322.960559] iwioctl: cmd=0x8b0c not allowed in this mode <4>[2016, wpa_supplicant] [ 322.981224] iwioctl: cmd=0x8b0c not allowed in this mode logcat: I/wpa_supplicant( 2016): CTRL-EVENT-STATE-CHANGE id=-1 state=2 BSSID=00:00:00:00:00:00 E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): MACADDR E/wpa_supplicant( 2016): ioctl[sIOCSIWPRIV] (cscan): -1 W/wpa_supplicant( 2016): Failed to initiate AP scan. E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): RSSI E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): LINKSPEED E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): RXFILTER-ADD 0 E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): BTCOEXSCAN-STOP I/wpa_supplicant( 2016): CTRL-EVENT-DRIVER-STATE HANGED E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): COUNTRY US E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): SCAN-PASSIVE E/WifiStateTracker( 1288): Wifi Driver reports HUNG - reloading. I/wpa_supplicant( 2016): CTRL-EVENT-TERMINATING - signal 15 received I/wpa_supplicant( 2016): CTRL-EVENT-TERMINATING - signal 0 received I/wpa_supplicant( 2016): CTRL-EVENT-STATE-CHANGE id=-1 state=0 BSSID=00:00:00:00:00:00 E/QC-DSS-LIB( 1207): unrecognized ifindex 9 I/dun_service( 1215): process rmnet event I/pcm_service( 1216): process rmnet event I/dun_service( 1215): rstate == DUN_RMNETSTATE_ERROR in dun_monitor_kevents I/pcm_service( 1216): rstate == PCM_RMNETSTATE_ERROR in pcm_monitor_kevents I/WifiStateTracker( 1288): mDhcpTarget has been reseted to null! E/WifiService( 1288): Failed to unload Wi-Fi driver. E/WifiHW ( 1288): wifi use default mac address E/WifiService( 1288): Failed to load Wi-Fi driver. Edited February 18, 2012 by pier11 Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 18, 2012 Report Share Posted February 18, 2012 (edited) Implemented that - replaced /system/wifi with one from um840. With your libhardware_legacy.so and respective symlink in /data, driver loads, no EEPROM exceptions, but the following is in logcat: E/WifiHW ( 1272): Unable to open connection to supplicant on "/data/misc/wpa_supplicant/sta": No such file or directory Thank you for that! It means I didn't integrate ar6000 files well enough because in my experiment, logcat did not show anything. Now the above error reminds me of where we were with um840 once. I will go ahead and play with wpa_supplicant and its config and with wext / awext... Should yield some results... edit: ... ok... waiting for your edits above :) Edited February 18, 2012 by dr.flo Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 18, 2012 Report Share Posted February 18, 2012 it's needed to rename init.huawei.rc ro init.qcom.rc in ramdisk to load that piece. (Or alternatively change qcom->huawei in mkbootimg's cmdline parameter when assembling boot image). There are many wifi related definitions there. This should also help bluetooth, but some diag utility fauls on BT start. what wpa_supplicant did you use? stock rom one? I expect this should be replaced by cm7 one. The errors somewhat remind me of TomGs problems before he noticed that he needs to switch from WEXT to AWEXT. Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 18, 2012 Report Share Posted February 18, 2012 (edited) what wpa_supplicant did you use? stock rom one? I expect this should be replaced by cm7 one. The errors somewhat remind me of TomGs problems before he noticed that he needs to switch from WEXT to AWEXT. yes, I used stock supplicant. not sure that stock libhardware_legacy will happily talk to awext supplicant, but who knows. Also cm7 supplicant might work with sockets on different paths in filesystem - something to keep in mind and adjust in ramdisk and wpa_supplicant.conf Edited February 18, 2012 by pier11 Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 18, 2012 Report Share Posted February 18, 2012 (edited) yes, I used stock supplicant. not sure that stock libhardware_legacy will happily talk to awext supplicant, but who knows. Also cm7 supplicant might work with sockets on different paths in filesystem - something to keep in mind and adjust in ramdisk and wpa_supplicant.conf CM7 sources indicate that libhardware_legacy part of wifi is not influenced by supplicant version.... hm just thinking that we're currently using cm7 wpa_supplicant on froyo (cm6) which means wpa_supplicant can be exchanged between gingerbread and froyo. Which version of wpa_supplicant exactly are you currently using on um840? origin cm6 TomG/TXG or origin cm6-clean (then it is cm7!) Are paths of um840 the same as in ideos x3? I guess this is more or less standard and might just be different in CMx for some reason. Edited February 18, 2012 by dr.flo Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 18, 2012 Report Share Posted February 18, 2012 Which version of wpa_supplicant exactly are you currently using on um840? origin cm6 TomG/TXG or origin cm6-clean (then it is cm7!) origin cm6 TomG/TXG Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 18, 2012 Report Share Posted February 18, 2012 Are paths of um840 the same as in ideos x3? I guess this is more or less standard and might just be different in CMx for some reason. original paths are likely the same, but when I added init.qcom.rc in um840 ramdisk from our roms they stopped to be such. wpa_supplicant are interchangeable b/w cm6 and cm7 only I think, as both are of awext type. And libhardware_legacy is adopted to such in both roms. Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 18, 2012 Report Share Posted February 18, 2012 did - use wifi folder of um840 - use wpa_supplicant of cm6/cm7 - enable init.huawei.rc in logcat wpa_supplicant complained about unsupportet driver "wext" - as expected - then changed -dwext to -dawext in init* Now I have E/wpa_supplicant( 1909): Failed to initialize control interface 'DIR=/data/misc/wpa_supplicant GROUP=system'. E/wpa_supplicant( 1909): You may have another wpa_supplicant process already running or the file was E/wpa_supplicant( 1909): left by an unclean termination of wpa_supplicant in which case you will need E/wpa_supplicant( 1909): to manually remove this file before starting wpa_supplicant again. looks like basically exchanging awext / wext versions is ok. Some confusing about paths still left over. I remember I had this same error once when experimenting with wifi on u8150-froyo Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 19, 2012 Report Share Posted February 19, 2012 E/wpa_supplicant( 1909): Failed to initialize control interface 'DIR=/data/misc/wpa_supplicant GROUP=system'. that means parties now don't even see each other. Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 19, 2012 Report Share Posted February 19, 2012 (edited) hm.... got some more wifi (but not yet working).. - hacked libhwardware_legacy - um840 wifi folder - dhd.ko symlink - umodified ramdisk, but init.huawei.rc active aka the same setup that yields E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): MACADDR E/wpa_supplicant( 2016): ioctl[sIOCSIWPRIV] (cscan): -1 W/wpa_supplicant( 2016): Failed to initiate AP scan. E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): RSSI E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): LINKSPEED E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): RXFILTER-ADD 0 E/wpa_supplicant( 2016): wpa_driver_priv_driver_cmd failed (-1): BTCOEXSCAN-STOP then replace wpa_supplicant with pulse eclair version. it is "wext"! I'm able to scan for networks now. They get listed correctly. I just cannot connect to my wpa2 secured wifi. Can connect to unsecured or 128 bit WEP though... Minor other issue: cannot turn wifi off - it tries to unload "dhd" which it doesn't find. I can't hexedit dhd into ar6000 of course because the latter is longer. Could the name be changed inside ar6000.ko (is it coded there, or is it simply the filename without .ko)? Edited February 19, 2012 by dr.flo Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 19, 2012 Report Share Posted February 19, 2012 Minor other issue: cannot turn wifi off - it tries to unload "dhd" which it doesn't find. I can't hexedit dhd into ar6000 of course because the latter is longer. Could the name be changed inside ar6000.ko (is it coded there, or is it simply the filename without .ko)? interesting progress if it tries to offload "dhd.ko", it can probably be replaced to just "ar6000" (without ".ko"), as rmmod works both ways - with .ko and without... Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 19, 2012 Report Share Posted February 19, 2012 interesting progress if it tries to offload "dhd.ko", it can probably be replaced to just "ar6000" (without ".ko"), as rmmod works both ways - with .ko and without... It seems to work now. I could not change the string inside libhardware, it was just three chars long. But I found the corresponding string "ar6000" inside ar6000.ko and changed that to "dhd" - now lsmod lists "dhd" and rmmod dhd works as it should ;) I don't know why, but it seems this fixed wpa issue! I could connect to my network. Also mac address is now the usual default address 00.18.82 .... - before it was all zeros. Maybe all zeros are not valid with wpa. Will update the rom download once experiements finished... Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 19, 2012 Report Share Posted February 19, 2012 (edited) currently uploading: pulse-x3-20120219.zip, download from first post. - added wifi - added gps wifi works just fine with eclair's original wpa_supplicant, patched libhardware_legacy.so and patched ar6000.ko. I renamed patched ar6000.ko to ar6000_dhd.ko to remind us that it's a little different. Also, dhd.ko symlink is now back to /wifi not in /data any longer (added it in ramdisk). also added gps driver from pulse cm7. note: wifi has resume issue as currently in every ported rom that needs kernel .32 to run Edited February 19, 2012 by dr.flo Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 19, 2012 Report Share Posted February 19, 2012 thanks to wifi I just noticed that this rom will play fluid HQ video in youtube app. No normal video. Link to comment Share on other sites More sharing options...
Guest desalesouche Posted February 19, 2012 Report Share Posted February 19, 2012 (edited) hi and first a big thx...:) 1/When scanning avalaible network.....error while searching.....but can select automatically 2/when i dl on market cliking on square (allow automatic update) not working.. 3/usb tethering is not working... this trick is not working about helping http://forum.xda-dev...071&postcount=4 4/1/tried to play a borgia mp4video.....very very slow playing and no audio 2/playing apollo13 mp4video.....just fine... audio and video run excellent ... 3/playing avi video with mobo player run slow...no sound.....then video stop... 5/ playing 8 sec of mp3 music then stop 6/news and weather app can²t give weather... edit...i tried to replace all libOmx* from rom um840-20120218 not working Edited February 19, 2012 by desalesouche Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 20, 2012 Report Share Posted February 20, 2012 (edited) Yes this rom still has many issues. It seems msm7227 is not truly binary compatible with ours. When watching logcat, there are often crashes caused by "illegal opcode". Is there any official gingerbread rom known for a msm7225 device (like u8150?) - I didn't find anything yet. My impression also is that after adding the huawei init in ramdisk, there are even more errors in logcat. Init scripts and possibly build.prop would require some major cleanup. Maybe we should first concentrate on the um840 froyo port and solve remaining issues there. I would really appreciate having a fully functional ported stock rom. The chance of reaching that goal is highest for the um840 port I believe. Once that one runs without issues, we could have a look at gingerbread again... Edited February 20, 2012 by dr.flo Link to comment Share on other sites More sharing options...
Guest desalesouche Posted February 20, 2012 Report Share Posted February 20, 2012 youtube video is smooth in gingerbread.....possible to port that smooth to um840 rom?.... Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 20, 2012 Report Share Posted February 20, 2012 (edited) youtube video is smooth in gingerbread.....possible to port that smooth to um840 rom?.... It is also smooth on froyo (cm6-clean, when adding libOmxH264Dec.so from um840 rom). Issue might be in some other *omx*, *Omx* or *stagefright* file. Or even in build.prop as it sets some stagefright related options I believe... stagefright is part of multimedia framwork. Edited February 20, 2012 by dr.flo Link to comment Share on other sites More sharing options...
Guest desalesouche Posted February 20, 2012 Report Share Posted February 20, 2012 (edited) It is also smooth on froyo (cm6-clean, when adding libOmxH264Dec.so from um840 rom). Issue might be in some other *omx*, *Omx* or *stagefright* file. Or even in build.prop as it sets some stagefright related options I believe... stagefright is part of multimedia framwork. hi.i did already try to change *omx file* and *Omx* files...no positive result but i might have done mistake(s)....will try with *stagefright*files.... i have read about "smooth" that libopencore could be in relation with "smooth"...(changing or rewriting libopencore_common.so) what do you think? Edited February 20, 2012 by desalesouche Link to comment Share on other sites More sharing options...
Guest lloco73 Posted February 20, 2012 Report Share Posted February 20, 2012 Maybe we should first concentrate on the um840 froyo port and solve remaining issues there. I would really appreciate having a fully functional ported stock rom. The chance of reaching that goal is highest for the um840 port I believe. Once that one runs without issues, we could have a look at gingerbread again... I totally agree with you. UM840 froyo rom development is very promising and the focus should be on it. Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 22, 2012 Report Share Posted February 22, 2012 With re to graphics: Thinking if replacing Gallery3D.apk from cm7 would help? As this thing kinda "opens" hardware 3d capabilities when run. Can be seing in dmesg. Maybe X3 version of the app was compiled requiring openGL v2 ?.. Link to comment Share on other sites More sharing options...
Guest dr.flo Posted February 22, 2012 Report Share Posted February 22, 2012 (edited) Gallery3D.apk from cm7 crashes in exactly the same fashion (as far as I can judge; looks the same in logcat at least) as stock gallery. looks like this: I/DEBUG ( 1193): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** I/DEBUG ( 1193): Build fingerprint: 'Huawei/U8510/hwu8510:2.3.5/HuaweiU8510/C364B856:user/ota-rel-keys,release-keys' I/DEBUG ( 1193): pid: 3370, tid: 3370 >>> com.cooliris.media <<< I/DEBUG ( 1193): signal 4 (SIGILL), code 1 (ILL_ILLOPC), fault addr ad330eac I/DEBUG ( 1193): r0 405749d0 r1 40535468 r2 acaa3fbc r3 00000001 I/DEBUG ( 1193): r4 0000ac90 r5 405749c0 r6 405354a8 r7 405354a8 I/DEBUG ( 1193): r8 405354b4 r9 405354ac 10 405354b0 fp 40535468 I/DEBUG ( 1193): ip aca9e654 sp be92e3b0 lr aca4460b pc ad330eac cpsr 00000010 I/DEBUG ( 1193): cannot get registers: I/O error I/DEBUG ( 1193): #00 pc 00030eac /data/sys_part/lib/libandroid_runtime.so I/DEBUG ( 1193): #01 pc 00011df4 /data/sys_part/lib/libdvm.so I/DEBUG ( 1193): #02 pc 00030d70 /data/sys_part/lib/libandroid_runtime.so I/DEBUG ( 1193): I/DEBUG ( 1193): code around pc: I/DEBUG ( 1193): ad330e8c e0890107 e2869004 e08ba108 e1a0100a I/DEBUG ( 1193): ad330e9c e58da008 e286800c e286a008 e2817040 I/DEBUG ( 1193): ad330eac edd17a00 edd65a00 ed996a00 edda6a00 I/DEBUG ( 1193): ad330ebc ed987a00 ee675aa5 ee277a87 ee276a86 I/DEBUG ( 1193): ad330ecc ee676aa6 e1a03006 e3a02000 e081e002 I/DEBUG ( 1193): What I don't like about it that this is obviously caused by an illegal opcode -> binary incompatibility of msm7227 and msm7200a likely. libdvm.so is ours (cm7) already, original one wouldn't boot at all. libandroid_runtime cannot be replaced with the one from cm7 - will not boot. edit: maybe I should try libandroid_runtime of cm7 v 2.3.5 (dtox) instead of cm 2.3.7... I'm not good at android debugging - maybe there is a way to really find out which lib holds the illegal opcode inside.... Edited February 22, 2012 by dr.flo Link to comment Share on other sites More sharing options...
Guest pier11 Posted February 22, 2012 Report Share Posted February 22, 2012 (edited) libdvm.so - is it Dalvic virtual machine itself?.. Can it be turned off in build.prop, at least experimentally. I'm even worst at figuring libs dependencies... other experiment is to try to run X3's Gallery3D.apk on cm7 to prove it's complete innocence :) It is a good idea to use android versions as close as possible, to avoid API incompatibilities... Edited February 22, 2012 by pier11 Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now