05-10-2012 04:01 PM
937 ... 2122232425 ...
tools
  1. LeslieAnn's Avatar
    Are there any confirmed flashes of just the radio on a custom recovery?

    I'm pretty sure that it requires the hboot|radio option to be set, and it gets written on reboot, not during the flashing process, which requires AFAIK the stock firmware.

    (check your email )
    I've been watching, and again, I'm not entirely convinced one way or another.
    Using QPST is not a way to tell.
    04-09-2012 09:26 PM
  2. JerryScript's Avatar
    If you check the recovery code, you'll see that the radio is installed outside of recovery, most likely by the lower level firmware. See /bootable/recovery/firmware.c and /bootable/recovery/install.c

    Doesn't look like the radio can be flashed solo.
    04-09-2012 09:37 PM
  3. brotherswing's Avatar
    @Jerry and/or Leslie - I know CMWT bricks with the vm security update, what about xionia? I'm asking so I can clarify my guide. Thanks!

    Sent from the BACKside of Tapatalk
    04-09-2012 10:26 PM
  4. cole2kb's Avatar
    Dannemand, or anyone else who has done this: Has running the full stock update fixed the reboot / texting lag? (Meaning not flashing a custom ROM, just running the new stock software.) It was bad today for me, no reboots, but I basically have to send a text, then send another one in the same thread to the same person to get the first one to go through. Really frustrating.
    Dannemand likes this.
    04-09-2012 11:57 PM
  5. JerryScript's Avatar
    @Jerry and/or Leslie - I know CMWT bricks with the vm security update, what about xionia? I'm asking so I can clarify my guide. Thanks!

    Sent from the BACKside of Tapatalk
    AFAIK Xionia is safe. The only ones I know are dangerous are the touch recovery and multi rom recovery.
    04-10-2012 12:02 AM
  6. cole2kb's Avatar
    I have an important day tomorrow and need my phone, my best friend / brotha from anotha motha is having a baby (well, he isn't, his lady is, you know what I mean, lol.) I'm gonna be an uncle. Going to go back to full VM9 update stock tonight, I think, and see how it acts tomorrow.
    04-10-2012 12:25 AM
  7. LeslieAnn's Avatar
    If you check the recovery code, you'll see that the radio is installed outside of recovery, most likely by the lower level firmware. See /bootable/recovery/firmware.c and /bootable/recovery/install.c

    Doesn't look like the radio can be flashed solo.
    I would imagine that would take a much different line of code for the install, and yet the installer from Virgin doesn't make use of anything special.
    04-10-2012 02:56 AM
  8. economichitman's Avatar
    Ive started to experience random reboots since flashing the security update to get the radio to stick. I did a clean install of the 4/3 build of mirage iho. Experienced 3 reboots today, while using the phone or sending a text. I'm running what was a stock 2.2.1 ZV4 phone with the 61677 PRL update.
    Dannemand likes this.
    04-10-2012 03:08 AM
  9. JerryScript's Avatar
    I would imagine that would take a much different line of code for the install, and yet the installer from Virgin doesn't make use of anything special.
    I've been doing a lot of research on this the past few days, both in the code of AOSP and CM's recoveries, and using my google-fu. Everything seems to point to the radio being flashed outside of recovery.

    Examples:
    CDMA Hero: Fresh Rom – CDMA Android Development How To: Flash a Radio Update
    (behaves as I've described, except it reboots back into recovery after the radio flash)
    Thunderbolt: http://forums.androidcentral.com/bio...ml#post1320731
    (flashed in recovery like our update zip, but must use hboot to finish flashing the radio)
    Android-Dev-Phone: Android Dev Phone 1: Full Update Guide - CyanogenMod Wiki
    (flashing of radio occurs outside recovery during reboot process)

    I've also read documentation indicating the recovery does not have access to the radio, only the bootloader does, which of course runs before recovery is started.

    Together with the code examples I've previously cited, I'm pretty much convinced. Not to mention the donor phone and my sister-in-law's phone both exhibited the same reboot then message displayed about flashing radio when I ran the full update. On my phone and my sister-in-law's, just flashing your radio update zip did not work (based on the value of the soc0/build_id).

    Have I missed something that shows it being flashed successfully within recovery? You know I like hot-sauce if you make me eat my words!
    Dave Begue and Dannemand like this.
    04-10-2012 05:01 AM
  10. Dannemand's Avatar
    I did a bunch of testing yesterday, and I'm not finished yet. (Getting ADB again after messing around with drivers to get QPST working is what took the longest )

    On the subject of radio flashing: I personally don't think there is any doubt that the radio only flashes on the first reboot after flashing the stock recovery in update.zip. Wraith's original mention of that "flashing radio" message was the first clue. And Jerry's studies seem to bear that out. Thanks again, Jerry!

    About reproducing the reboots in IHO after updating: I found that the best way was to cycle airplane mode, then send an SMS to myself, then keep cycling these two actions. After one or two cycles, SMS messages will take longer and longer to send and receive -- eventually messages never arrive until the phone is rebooted. Sometimes it'll reboot after a few cycles. In my old days of assembler and C programming, I would instantly look for a memory leak, particular one with a mismatched number of arguments in a library call. (Don't know if that's even possible in modern languages). Jerry mentioned the ril interface. That kind of thing would be my first suspect as it's likely to have been updated along with ZV9. I haven't been coding in many, MANY years (I'm an old dude) so I'm afraid I cannot provide more specifics beyond this kind of speculation.

    About my stability tests: Nothing I've tried so far has brought stability in IHO. That includes fresh IHO installs (MiRaGe, I assume it's the same on all) and double-flashing of the VMU update.zip (factory restores in between). However, having run the new stock since yesterday, I have NOT been able to reproduce any instability, reboots or lag with airplane mode and/or SMS -- or anything else for that matter.

    Other than that, stock isn't a whole lot of fun -- particularly since I'm unrooted Now that I finally have ADB again, I'll be rooting this afternoon and try again with a fresh IHO install.

    I don't know if this is any help in steering you devs in the right direction. I'll report more when I am done testing, hopefully tonight.

    Thanks in advance!
    tdm likes this.
    04-10-2012 11:09 AM
  11. gunson's Avatar
    It looks like we are on the same testing schedule. I too used the stock VM9 rom all day yesterday and had no reboots. Im going to root and stay on this rom to see if I get any reboots after rooting. Good luck!
    Dannemand likes this.
    04-10-2012 11:15 AM
  12. Dave Begue's Avatar
    (Getting ADB again after messing around with drivers to get QPST working is what took the longest )

    Now that I finally have ADB again, I'll be rooting this afternoon and try again with a fresh IHO install.



    Thanks in advance!
    uhmm - dannemand - I am for sure scared of adb now !!
    04-10-2012 11:19 AM
  13. Dannemand's Avatar
    uhmm - dannemand - I am for sure scared of adb now !!
    LOL Maybe I shouldn't have posted that...

    The only thing bad about ADB is to get the drivers working in Windows. It can be a breeze or it can be a hassle. XP is generally a lot easier -- Win7 32-bit maybe too. Win7/64 can be tricky. Once the drivers are running, ABD isn't bad at all.

    I found that in the newer version of the LG Unified Drivers (3.2) the ADB Sooner driver is not properly signed, which is what causes problems in Win7/64. And once 3.2 has been installed, no amount of uninstalling will get it working again, even with the older drivers. Of course you can dig deep and find the right registry entries to remove. In my case I just restored an earlier image and started over with 2.1 drivers -- no problem with those.
    Dave Begue likes this.
    04-10-2012 11:36 AM
  14. Dannemand's Avatar
    It looks like we are on the same testing schedule. I too used the stock VM9 rom all day yesterday and had no reboots. Im going to root and stay on this rom to see if I get any reboots after rooting. Good luck!
    Thank you! Yes, I did see your earlier post just now when I caught up with this thread. Sorry I was out of the loop for a while.

    @cole, same there: Did see your post from last night now. Thank you!

    I am still puzzled as to why some users are not seeing these issues at all, if indeed there is an incompability betwen IHO and the new radio. Maybe it is somewhere else entirely that we haven't thought of...
    04-10-2012 11:40 AM
  15. drezliok's Avatar
    I did a bunch of testing yesterday, and I'm not finished yet. (Getting ADB again after messing around with drivers to get QPST working is what took the longest )

    On the subject of radio flashing: I personally don't think there is any doubt that the radio only flashes on the first reboot after flashing the stock recovery in update.zip. Wraith's original mention of that "flashing radio" message was the first clue. And Jerry's studies seem to bear that out. Thanks again, Jerry!

    About reproducing the reboots in IHO after updating: I found that the best way was to cycle airplane mode, then send an SMS to myself, then keep cycling these two actions. After one or two cycles, SMS messages will take longer and longer to send and receive -- eventually messages never arrive until the phone is rebooted. Sometimes it'll reboot after a few cycles. In my old days of assembler and C programming, I would instantly look for a memory leak, particular one with a mismatched number of arguments in a library call. (Don't know if that's even possible in modern languages). Jerry mentioned the ril interface. That kind of thing would be my first suspect as it's likely to have been updated along with ZV9. I haven't been coding in many, MANY years (I'm an old dude) so I'm afraid I cannot provide more specifics beyond this kind of speculation.

    About my stability tests: Nothing I've tried so far has brought stability in IHO. That includes fresh IHO installs (MiRaGe, I assume it's the same on all) and double-flashing of the VMU update.zip (factory restores in between). However, having run the new stock since yesterday, I have NOT been able to reproduce any instability, reboots or lag with airplane mode and/or SMS -- or anything else for that matter.

    Other than that, stock isn't a whole lot of fun -- particularly since I'm unrooted Now that I finally have ADB again, I'll be rooting this afternoon and try again with a fresh IHO install.

    I don't know if this is any help in steering you devs in the right direction. I'll report more when I am done testing, hopefully tonight.

    Thanks in advance!
    i tried to recreate your reboots on my phone. it didn't reboot, but I had to yank the battery.
    Dannemand likes this.
    04-10-2012 11:46 AM
  16. Eollie's Avatar
    I am still puzzled as to why some users are not seeing these issues at all, if indeed there is an incompability betwen IHO and the new radio. Maybe it is somewhere else entirely that we haven't thought of...
    Maybe they just arent seeing the reboots. As in not noticing them. I suggest putting in a pleasant bootsound for this purpose. Without any sound from the phone if it is in your pocket or the screen locked on your desk you might not notice it. When I applied the full update I seen the applying radio update message too.

    --------------------edited to add-----------------------

    Another way to check and see if you are getting silent reboots is hit settings>about phone>status then scroll to up time. If you know you havent rebooted in a while and the times dont match....ding ding ding you have gotten a reboot.
    Dannemand likes this.
    04-10-2012 11:50 AM
  17. anthonycr's Avatar
    Just thought i'd post this here too: I'm trying to rebuild bumblebee, but there's something in the update that breaks screen rotation with non-stock kernels...

    Sent from my VM670 using Tapatalk 2
    04-10-2012 11:55 AM
  18. Dannemand's Avatar
    i tried to recreate your reboots on my phone. it didn't reboot, but I had to yank the battery.
    Thank you. Yeah, that's close enough. Pretty much anything can happen if it's a memory leak or stack overflow.
    04-10-2012 11:55 AM
  19. dobbs69's Avatar
    The way I can tell if I get reboots is an unclean shut down notice from NO FRILLS CPU

    Sent from my LG-VM670 using Tapatalk
    04-10-2012 12:34 PM
  20. pbailey212's Avatar
    Has anyone tried to flash the radio by its self using fastboot?

    Sent from my Kindle Fire using Tapatalk 2 Beta-5
    04-10-2012 12:57 PM
  21. tdm's Avatar
    From the comments posted here, especially by Dannemand, it is starting to look like the radio update may have changed the baseband RIL API. Not enough to stop it from working, just enough to make it unstable.

    Unfortunately, getting to the bottom of this is going to be a pretty big technical challenge. One thing I would suggest doing is to copy rild and all the ril libs from the VM updated system into your IHO system and test to see if it stabilizes. Of course, you need to be pretty technical to do this. I'm hoping someone on this thread is up to the task.

    I hear lots of talk about kernel/baseband compatibility in the LS670 community. So it would be nice to compare the new and old kernel sources. But unfortunately, it seems that LG has not released the kernel sources for the update yet. I am hoping that they do -- they seem to be pretty good about releasing the sources for the LS670 updates, so perhaps they're backlogged, or perhaps they just haven't gotten a request yet.
    Dannemand likes this.
    04-10-2012 01:06 PM
  22. mrg666's Avatar
    About reproducing the reboots in IHO after updating: I found that the best way was to cycle airplane mode, then send an SMS to myself, then keep cycling these two actions. After one or two cycles, SMS messages will take longer and longer to send and receive -- eventually messages never arrive until the phone is rebooted.
    I have done six cycles exactly as you have described - toggle airplane mode, send an SMS, wait for it to be sent, toggle again, repeat. No crashes or freezes, all of the messages were sent and received. I would recommend you to do the update/radio flash procedure again. There might be something incorrectly or partially flashed. My uptime is more than 19 hours right now. I am not able to reproduce these reboots and there is no IHO/CM7 incompatibility with the ZV9 radio as far as I can see.
    JerryScript likes this.
    04-10-2012 01:26 PM
  23. mrg666's Avatar
    From the comments posted here, especially by Dannemand, it is starting to look like the radio update may have changed the baseband RIL API. Not enough to stop it from working, just enough to make it unstable.

    Unfortunately, getting to the bottom of this is going to be a pretty big technical challenge. One thing I would suggest doing is to copy rild and all the ril libs from the VM updated system into your IHO system and test to see if it stabilizes. Of course, you need to be pretty technical to do this. I'm hoping someone on this thread is up to the task.

    I hear lots of talk about kernel/baseband compatibility in the LS670 community. So it would be nice to compare the new and old kernel sources. But unfortunately, it seems that LG has not released the kernel sources for the update yet. I am hoping that they do -- they seem to be pretty good about releasing the sources for the LS670 updates, so perhaps they're backlogged, or perhaps they just haven't gotten a request yet.
    I have compared the RIL libraries in the update and in our proprietary files in the IHO repo. They are the same.

    Edit: these are the files I checked. They have the same MD5sum.
    /system/lib/libril-qc-1.so
    /system/lib/libril-qcril-hook-oem.so
    /system/lib/libril.so
    /system/bin/rild
    04-10-2012 01:33 PM
  24. Dannemand's Avatar
    I have compared the RIL libraries in the update and in our proprietary files in the IHO repo. They are the same.

    Edit: these are the files I checked. They have the same MD5sum.
    /system/lib/libril-qc-1.so
    /system/lib/libril-qcril-hook-oem.so
    /system/lib/libril.so
    /system/bin/rild
    Thanks a lot, mrg666, that takes that one off the table. So much for my hunches

    I absolutely believe you that some (hopefully most) users are not having any issues, which would point to flash errors -- as always. The only reason I've raised this is because those who ARE having issues all seem to hit them at the exact same tasks involving the radio. As I mentioned before, it's not the same old issues (including random ones) that we're used to seeing in the IHO thread.

    For my part, I am one of your greatest fans, and I am not in the habit of whining about stability or bugs -- as I hope you'll agree. I approach my updating conservatively, and I know where to look when things get wonky. This one just has a different and more consistent smell to it.

    All that said, my latest fresh MiRaGe install today actually DOES seem to be stable -- after doing my third full VMU update (twice, of course, with factory wipe before both, as I did the other times). At least I've not been able to reproduce reboots or extreme SMS delays yet.

    Of course my focus now is to try and see if I can reproduce the issues again. I still believe there is a reason why a larger percentage than usual experience the exact same radio related issues after making this radio update.

    Thanks again!
    mrg666 likes this.
    04-10-2012 02:21 PM
  25. economichitman's Avatar
    Those with reboot issues, what radio version did you have before flashing? I have a launch day v, zv4 radio 2.2.1 os
    04-10-2012 02:26 PM
937 ... 2122232425 ...
LINK TO POST COPIED TO CLIPBOARD