-
Posts
212 -
Joined
-
Last visited
Everything posted by fabiobassa
-
@Luis have you tried the command startx is unusual that the x11 doesn't start automatically, maybe a fesh installation or jock's suggestions
-
@nokirunner you should put in the env also an olive, some martini and mix, not shake ok other than this , yes would be anice idea have both , compliments for the idea
-
@Reddwarf @jock will answer more completely to this question, but I guess the short answer is : NO
-
Commercial names of those board = giving number on the lottery !! A little , but little better approach is name ON THE PHYSICAL board or (the best approach indeed ) is trying to get those basic infos: 1) nand or emmc ( multitool will reveal it) 2) 3228 ( a/b doesn't really matter at this point since the difference is a bunch of MHZ speed but less than 10% ) or 3229 ( efuse command will reveal it ) 3) ddr2 or ddr3 ( at the writing moment only uart will reveal )
-
Thank you so much @Alex83 yes I had even a vpn on a broken WiFi chip ubiquity nano loco CPE with openwrt and somewhere I read that wireguard could perform better. Anyway the performance was ideal to do in real environment ( two different remote ips and not in perfect environment such as intranet) but anyway your results are awesome. Ty
-
@Alex83 ok it is a real 3229 thanks for efuse 00000000 52 4b 23 92 >>>> 2392 >>> 3329 This iperf is done trhought wireguard ?
-
@Alex83 thank you so much for reporting your test on wifi speed Please would also report you experience with wireguard and/or any other vpn method since my expereinces were NOT so happy , neither with openvpn neither with wireguard. I must investigate if depends by my isp carrier or/and wrong configurations. thanks in advance EDIT: share also the efuse infos and if you have some board photos
-
@Alex83 the only way to declare the processor inside is this command, as we said many times hexdump -C /sys/bus/nvmem/devices/rockchip-efuse0/nvmem And even with this command we still unsure if a 3228a or a 3228b; so if you say that is a "3229 for sure" ( and not because you read by the stamp on processor.......) I believe you, even if you have the same boards of the photos i doubt it is a 3229 but mostly a 3228. this could be also proved by the fact that rk322x-box (Generic RK322x TV Box board) is working With this chinese people never say never thought and I surrender to all their fake statements ( just an example some producer say that an all winner has 4 giga ram even if is well know that h6 has PHYSICAL limitation to 3 giga....) EDIT: I have a board called sardine... wll advertized as a 3229 chip on the processor also 3229 this is my efuse: hexdump -C /sys/bus/nvmem/devices/rockchip-efuse0/nvmem 00000000 52 4b 23 82 >>>>>> 3228 and since some erratic problems quite sure a 3228 A
-
@Alex83 if NONE of profiles with name " rk3229" brings life to your board is simply because your boad probabily isn't a 3229 but a 3228a or 3228b The rk322x-box (Generic RK322x TV Box board) because it is just generic The mxq you have is a 3228 with ssv6051 so try any of the other combinantions but with 3228 in name and yes nands don't like to many rewrites but you cannot imagine ANDROID how much stress them p.s did you try the moltipurpose tool and steP-nand tool to help you identify the board ?
-
@paulml the problem of erratic mac was solved in mainline uboot since on those "lovely" nands we use the legacy uboot we have the problem since the mac address usually is set in vendor partition or/and folder of android Is planned to patch again the uboot ( yes, patch over patch) but try consider that have nand working was a big effort and ARTIFACT so may be following this other strategy can help : 1) log in via ssh into board at his actual ip ( it should have one via dhcp , whatever it is) and take note of actual mac (ifconfig will tell you ) 2) on the cli digit nmtui ( this is the cli interface to network-manager) 3) go to edit a connection , you must MOVE with TAB KEY ON KEYBOARD across the fields 4) you probabily will have more than one wired connection ( 1...2 and so on) 4) delete the interfaces that DOESN'T CORRESPOND AT actual mac 6) edit the one that has a correspondent mac to ifconfig, actual mac 7) there is a fiels named ETHERNET and a " show" by clicking on that tab you will have a new field: cloned mac address well just digit in that field the same mac address you have in the upper field eth0 ( xx:xx:xx:xx:xx:xx) then go to the end and click ok, then back then quit this whole procedure can be also made in graphical environment if you have desktop by calling the edit connection on network-manager. in this way you will have a stable mac address over reboot remember to DELETE eventual interface that doesn't correspond to ACTUAL mac With nmtui you can also join a wifi with appropriate interface ( wifi and NOT p2p ) @Mark Khevin Rogacion yes you right , erasing with ef was an un-needed step but now no more mistakes: people with original android on nand must ONLY boot and call steP-nand tool and everything will be ok
-
@Ztrawberry Ty for joining us in this crazy crazy adventure in these more crazy crazy days ! Unless there is some physical damage those board are quite unbricable thanks to bootroom ( maskrom) mode . Ty for your offer but we already have those mx4vr both 01 and 02 What we REALLY would appreciate now is someone willing to rewrite the driver for ssv6x5x WiFi ( we have the one for ssv6051, we need the dual band ) !!!
-
@nokirunner Believe me the name is reaaaallllyyyy appropriate, after days and days of trial and errors of course it works on emmc and ( if you have) on nand boards @Maker39 well, with all those different commercial name, brands marked on circuits, and producers I am afraid some days will come out someone saying " ON MY BOARD NOT WORKING" but I guess we can say that 85/90% of total cases are now covered. New hardware disposition/arrangments are to be studied ( seriously hope not, lol )
-
@paulml your board is a 3228b with 1 giga ram and nand 1) once more , and unfortunately at the moment , your board has a nand storage 2) as long as you run by sd card and SKIP the approach to flash internal storage ( since it is nand) everything will be ok 3) simply don't touch internal android installation and run from sd, for the moment tool for manage even nand is on the way :-)
-
@MMarcio yes you too have a nand, but I already said this to you: mostly r329q have nands... ( mostly ) But on the other hand.. they have wonderfull ddr3 at 600 mhz very powerfull, while some boards with emmc have " only " ddr2 at 330 mhz
-
hello @Mark Khevin Rogacion and ty for joining in this adventure Yes you right because your 329q has a nand and not an mmc But @jock is actively and strongly working on a simple tool that in future will avoid all this mis-confusing due to different commercial names . So kindly wait just some little time that the new tool will be perfectly functional
-
@MMarcio what it means no one of the board is recognized by pc ???? Do you plug into OTG port while pressing the little switch in the audio video hole? PRESS that little switch and AT same time plug male male usb Is not possible that 3 on 3 are not recognized by pc
-
@nokirunner you do a lot of things but you don't read CAREFULL what people suggest to you It is easy you started at 400 and even at 500.... your ddr is up to 767. might be 800 and it is a ddr3 so oveclocking will occur in the kernel itself with the right dmc actived in dtb. What mostly you say in idbloader will be OVERWRITTEN in kernel or/and dtb. It is just a bootstrap this is the reason better stay on low band of mhz. This occurs with legacy kernel. With main line it is different and @jock can explain better. Try to overclock this on a board with ddr2 and tell me the conseguences . It is EXACTLY what happened to you when you bricked the board, flashe a WRONG idbloader with wrong settings Again, and for last time: those board come with a lot of ddr some have ddr2 some ddr3 mostly very economic and second market ddrs and not realiable not even on marks printed on case for a stable functioning test it at various speed and stay under maximum speed or just use original idb .If dmc is in dtb , maximum speed will be AUTOMATIC reached any way is enough speaking or ddrs , important is the board is up and running
-
NO... they start to became erratic and malfunctioning. With DDR3 there is a command in kernel that can try WHEN KERNEL is up to modify on the fly the ddr frequency, a sort of DDR scaling. It si totally automatic and works ONLY with original loader-uboot-trust combination With boards with DDR2 the will just work at 330 ... something ridiculous So PLEASE.. don't overclock DDR
-
@nokirunner Is very ok that you try to learn and this is the right approach . So let me explai what is best approach with those boards Unfortunately those economic vendors install the first avaible CHEAPEST bank of ddr on the market at moment of production, ddr ofsecond hand, semi-broken. Is economia di larga scala .... every cent is gain MIGHT be even if DDR are TARGETED 400 mhz they simply REFUSE to work at 400 MHZ or even at 350 Many idbloader just start at lowest frequency possible, that is 300 MHZ To be sure to have success do this : go find this program on the net: RK3xxx_firmware_tools_5.23.1_by_SergioPoverony.zip or even this but is in dos imgRePackerRK_106.zip Those two tools can extract your androind rom, and go deep and extract even the early stage that means idbloader for your board the idbloader should be RK322XMiniLoaderAll_V2.47_spectek_en_ddr2_rd_odt_171209.bin and if you go deeper and deeper you will find it's blob: rk322x_ddr_330MHz_v as you can see the ddr is initialized at 330 MHZ when this very first step is done and the ddr is recognized, all further step can occour. My 2 cents opinion : use the images already done from jock without trying to mixt things or... try even to mixt things per seguir virtute e conoscenza but be ready to some frustrating unsuccess because of wrong early stage loader
-
@nokirunner could be a good starting point study this document, for your experiments if on internal storage you have a idbloader it will search then uboot on sd card. If you flash the wrong idbloader into internal storage , boot no more occurs and you have to short pins http://opensource.rock-chips.com/wiki_Boot_option
-
@MMarcio hello there , your question is a bit generic and cannot have a unique answer. what processor have you ? amount of ram? is your board equipped with nand or emmc? This particular topic of course is for rk 3228 3228b 3229 could be usefull the original name of rom and some board photos and anyway the process of installing on internal storage is last step and eventually totally automated @nokirunner well , give us the time of investigating on those drivers
-
@Maker39 I am very sorry maker39 I didn't figure libreelc and remotes on those boards. me and jock were focused on armbian and X11 port on 322x and indeed we have shared some patches and ideas and part of codes on the excelent work of the author of that post but never really used libreelec on my boards. Usually they are kind to answer so just wait maybe they are busy with real life. Personally cannot help
-
Ahahah allora te lo dico in italiano Sto gran figlio di driver è una 💩 Ma a qualche parte arriveremo....
-
Unfortunately things are not so easy as the can appear: the driver for ssv6051 is 5.037.020 byte long the driver for ssv6x5x is 6.791.040 byte long the vermagic of jock's armbian compilation is 4.4.194-rk322x SMP mod_unload ARMv7 p2v8 the vermagic of android's compilation is 4.4.126 SMP preempt mod_unload modversions ARMv7 p2v8 could also be that as you say is just one driver stripped down but the result is that is not working the ssv6051 on ssv6252 needings . And because of different vermagic even insmod -f (force) simpli refuses to load Could be a good apport finding the sources ( whole sdk ) of the android rom. Full sdk including the kernel source Any way @jock is the coder man, I am just the hardware supplier so wait till he enlight us Where are you located @nokirunner ?
