52
Lets Confirm Who Has the Ghost Touch Issue

  1. BrutalisCZ
    Cupcake Jun 4, 2019

    Last edited: Jun 4, 2019

  2. OPOBEAST
    Honeycomb Jun 4, 2019

    OPOBEAST , Jun 4, 2019 :
    I have to say, I didn't experience ANY sort of ghost touches on my 7 Pro yet. Not even in the mentioned App CPU-Z.
     

  3. jdpeters79
    Gingerbread Jun 4, 2019

    jdpeters79 , Jun 4, 2019 :
    Yeah, it's been communicated that turning off NFC definitely eliminates the problem as a temporary solution. There's a lot of good info in this whole thread about different tests people have run. The next update from OnePlus will address the problem.
     

  4. BrutalisCZ
    Cupcake Jun 4, 2019

    BrutalisCZ , Jun 4, 2019 :
    try reading card with NFC on - it works on some credit cards, should make a sound .
     

  5. deuse
    Jelly Bean Jun 5, 2019


  6. xenom0rph
    Gingerbread Jun 5, 2019

    xenom0rph , Jun 5, 2019 :
    they said upcoming weeks... plural weeks from this week.
     

    jlopez03 likes this.
  7. Jeresone
    Cupcake Jun 5, 2019

    Jeresone , Jun 5, 2019 :
    [​IMG]
    Here is my new phone broblem. I could only see the homescreen without the top status bar. I couldn't record the video but here is the screenshot (I took multiple). It shows totally something else what is in my screen. I only could see my home screen without the status bar. Nothing else. And I did even disabled the NFC. This thing is weird.

    https://ibb.co/9rN8D9Y
    [​IMG]
     

  8. patts75
    Donut Jun 5, 2019

    patts75 , Jun 5, 2019 :
    I belive there's a bug in the launcher? And updated launcher was released yesterday with a statusbar fix.. Try it out.
     

  9. S1558431183463
    Cupcake Jun 5, 2019


  10. Reworker71
    Gingerbread Jun 5, 2019


  11. MrBiron69
    Cupcake Jun 5, 2019

    MrBiron69 , Jun 5, 2019 :
    I literally only get ghost touches when using CPU-Z with NFC on. As I only installed the app to test it, it isn't an issue for me.
     

  12. S1558431183463
    Cupcake Jun 5, 2019


  13. edrose
    Froyo Jun 5, 2019

    edrose , Jun 5, 2019 :
    Thank you so much for going into more detail! Whilst I've done some quite low level work on embedded stuff before, proper kernels and low level OS stuff is not my area. I was trying to learn about kernel memory allocation since the memory leak revolved around the use of kzalloc, but I simply don't have time right now to dive too deeply into it. It does seem though that the memory leak is not related to the ghost touches however, since custom kernels without the memory leak still experience ghost touches.

    Do you have any idea about the below log messages? The first vmscan one came out of the kernel log at almost the exact same time as ghost touches were occuring, which I'm sure isn't a coincidence. Could be an effect of the touches however rather than a cause. The audit one also appeared with the touches, although less frequently.

    Code:
    05-28 11:36:34.503     0     0 E [20190528_11:36:34.498334]@3 vmscan: putback:0
    
    Code:
    05-28 11:36:09.937     0     0 W [20190528_11:36:09.936333]@6 audit: audit_lost=447 audit_rate_limit=5 audit_backlog_limit=64
    05-28 11:36:09.937     0     0 E [20190528_11:36:09.936338]@6 audit: rate limit exceeded
    
    They don't seem to be standard Linux messages which makes finding out exactly what they are quite tricky. Vmscan is something to do with swap memory if I'm not mistaken, and would putback:0 be page zero? So undefined or kernel space memory? I was linking the memory leak to these messages, and these messages to the touches but I may have been barking up the wrong tree completely. Having someome with more expertise looking at these logs would be interesting though!
     

    pgrey and xythiii like this.
  14. Mike_Routso
    Donut Jun 5, 2019

    Mike_Routso , Jun 5, 2019 :
    ghost touch on famous messengers:
    -facebook messenger (a lot of them)
    -whatsapp messenger (a few)
    -viber (none at the moment)

    *facebook messenger is make sometimes video call with ghost touch [e]1f62a[/e]

    why this big problem is not fixed yet??
     

  15. BrutalisCZ
    Cupcake Jun 5, 2019

    BrutalisCZ , Jun 5, 2019 :
    Ppl still think this is a software problem... IMO I don't sadly think it is...the NFC on the phone is weird, they have experience, it never happened before - obv they used some kind of different NFC on 7 pro and it's causing problems...like I said before - no phone works for opening door at my company, no iPhone, no OP6/T no OP5T, no Samsung note 9 - but OP7 PRO does work.. so there has to be something different.
    They also did not glue the screen to the glass as 99% of manufactures do..
    I hope it is a software issue,it's just seem like it's not...
     

  16. insomnia77
    Froyo Jun 5, 2019

    insomnia77 , Jun 5, 2019 :
    I cant understand why they are still selling this phone? At the very least they should halt sales until the ghost touch problem is fixed or put a disclaimer letting people know about it before they purchase it.
     
    Last edited: Jun 5, 2019

    everkurt likes this.
  17. everkurt
    Gingerbread Jun 5, 2019

    everkurt , Jun 5, 2019 :
    That's right.
    I told it to the staff. Also, I made a thread.
    But opinions are ignored.
     

  18. Oxie
    Cupcake Jun 5, 2019

    Oxie , Jun 5, 2019 :
    Because it's making them money, of course. Not everyone seems to notice the issue, so at least some people will keep the phone (and so will let OnePlus keep the money). Also probably why they haven't made any official announcement: So that people don't know about it before purchase, unless they go through the forums and happen upon this thread.
     

  19. seajamiet
    Donut Jun 5, 2019

    seajamiet , Jun 5, 2019 :
    Bloody hell, excellent summary and not based on pure speculation and opinions like my post! Let's hope for this to be the case!!
     

  20. Derple
    Froyo Jun 5, 2019

    Derple , Jun 5, 2019 :
    I discovered after this post that the memory leak in synaptics_get_touch_points() is not the cause of the ghost touches. This was tested with the ElementalX custom kernel, which merged the memory leak fix found on the OP7Pro kernel github page. This does not mean it's not software related, it just that the theory about it being a memory leak in that function causing it is basically debunked.
     

    Dthaki likes this.