Mi4 Lte Ramdom Reboots [with Logs]

[buzdo]

@jc rey I did. I flashed 7.1.2.0 global stable, then xiaomu.eu 7.1.5.0...
For 7.1.5.0 (unless you manually strip it down further) you need 7.1.2.0 or higher for merged system partition, because it's too big to fit on ~600mb.
This is why I keep wondering if there could possibly be a relation with reboots with this particular KK and the rest of MM's. But it seems such a trivial thing.

I think noone with 3/64 version is getting reboots right? If this is the case then it's definitely hardware related.

@Gingernut I will root it and install that log catcher and hopefully get some info.

 

[jc rey]

yes my unaffected is 3G only 3/64
but I have an affected one 3/16 3G only.
difference is the battery brand, memory, this is why I suspect out of spec components.

 

[jc rey]

@jc rey I did. I flashed 7.1.2.0 global stable, then xiaomu.eu 7.1.5.0...
For 7.1.5.0 (unless you manually strip it down further) you need 7.1.2.0 or higher for merged system partition, because it's too big to fit on ~600mb.
This is why I keep wondering if there could possibly be a relation with reboots with this particular KK and the rest of MM's. But it seems such a trivial thing.

I think noone with 3/64 version is getting reboots right? If this is the case then it's definitely hardware related.

@Gingernut I will root it and install that log catcher and hopefully get some info.

the KK test I did is just with global 7.1.2, running fine more than 500h.

 

[Gingernut]

yes my unaffected is 3G only 3/64
but I have an affected one 3/16 3G only.
difference is the battery brand, memory, this is why I suspect out of spec components.

Most incompatibility problems can be patched in kernel code, driver updates etc but by the looks of it they don't have a clue as why this is happening.

 

[buzdo]

@Gingernut Any special settings I need to set in logcat extreme for this?

 

[Gingernut]

@Gingernut Any special settings I need to set in logcat extreme for this?

I don't think so, just press record and wait for a reboot. That's what I'm doing but since I installed it I've had a totally stable phone.

 

[buzdo]

Alright I'm recording, let's hope a reboot happens soon. From tomorrow on the phone will be roaming on a different network in another country so I guess it's possible that changes things perhaps.

Maybe mi4 doesn't like to be idle and likes processes running in the background hence the new stability.

 

[giobooo]

I had just One reboot after flashing with miflash and updating to xiaomi.eu but I have a strange problem: from time to time I have to reboot my phone because accelerometer stops working. Sensor just shows no data and I can't go in landscape unless I reboot. I've find out that this is a common problem on redmi phones. This is also a random behaviour...

 

[ko_x1]

I have same problem with Xiaomi Mi4, Miui 7.5.2.0, android: 6.01. Random reboot when it idle but not often( one / 1-2 days).
So can it solve? It about hardware or software?

 

[NOECUBI]

I have same problem with Xiaomi Mi4, Miui 7.5.2.0, android: 6.01. Random reboot when it idle but not often( one / 1-2 days).
So can it solve? It about hardware or software?

I think it's a software problem caused by the hardware. In kk I never had a reboot.
The solution may be a software patch in the kernel....
I hope developers solve the issue

Sent from my MI 4LTE using Tapatalk

 

[buzdo]

Well it looks like I've captured a reboot, at least I hope so. Set it went to sleep and it recorded a few hours. I saw the phone was rebooted in the morning, it's almost on a 24 hour cycle... Tbh I don't know if this log will help at all.

Oh and keep in mind this is on custom xiaomi.eu 7.1.5.0. KK so the devs at official might not consider it.

 

[Gingernut]

Well it looks like I've captured a reboot, at least I hope so. Set it went to sleep and it recorded a few hours. I saw the phone was rebooted in the morning, it's almost on a 24 hour cycle... Tbh I don't know if this log will help at all.

Oh and keep in mind this is on custom xiaomi.eu 7.1.5.0. KK so the devs at official might not consider it.

Thank you for the log, been glancing over it but nothing jumps out at me. I've change, in the log app setting, the priority level from Info to Debug so more information is logged.

I'm sure the devs will say your phone is on an obsolete version.

But why not update to Marshmallow if you have reboots anyway?

Edit: Do you have Maps installed?

 

[Gingernut]

Rebooted and caught a log.

In the process of sending it to Miui.

 

[buzdo]

What exacly is qdhwcomposer unfortunately this is way out of my league. Looks to me like both reboots happen with this.

No I don't have Maps installed. Tbh I went with KK cause it doesn't 'kill' 1 core and it's stabler and faster.

 

[Gingernut]

What exacly is qdhwcomposer unfortunately this is way out of my league. Looks to me like both reboots happen with this.

No I don't have Maps installed. Tbh I went with KK cause it doesn't 'kill' 1 core and it's stabler and faster.

One core gets killed because Miui have configured it in the thermal-engine-8974.conf, but you can change it easily with root.

qdhwcomposer looks to be part of the graphics chip driver.

 

[buzdo]

I see. Has anyone tried turning proximity sensor off yet?

 

[Gingernut]

Did a TWRP backup of Global Beta 6.7.29 and downgraded to Xiaomi.eu KitKat 7.1.5 to test, wish it was a least Lollipop though.

First off the thermal-engine-8974.conf has some extra battery tempreature code at the end to scale CPU speed vs battery temp, basically it down clocks or shuts down CPU cores based on battery termpreature.

Also there is another config file, thermald-8974.conf, with more sensor code to scale CPU/GPU speed vs temp, similar to thermal-engine-8974.conf, which I believe is not present in the Android 6 based builds.

 

[jc rey]

6.7.29 global, I ve got my first reboot on my unaffected device ...
pluged to charger all night, rebooted when I unplugged, any clue ?

 

[Gingernut]

6.7.29 global, I ve got my first reboot on my unaffected device ...
pluged to charger all night, rebooted when I unplugged, any clue ?

Sorry but no idea here, I've had reboots on every rom so far based on Android 6 under different scenarios.

Yesterday I had two after uninstalling/installing apps.

I've contacted Patrick at Miui, seems he's a forum admin, and asked him for any news on the reboot issue.

 

[buzdo]

Is xiaomi.eu 7.1.5.0 stable based on china KK rom from that period right? I'm wondering what's so different in that rom compared to last KK 7.1.2.0 global stable rom where reboots supposedly don't happen?

 

[Gingernut]

Is xiaomi.eu 7.1.5.0 stable based on china KK rom from that period right? I'm wondering what's so different in that rom compared to last KK 7.1.2.0 global stable rom where reboots supposedly don't happen?

I think so but not sure what's changed

 

[jesuslg123]

After few weeks trying differents MM roms, I'm very dissapointend, here my symmary:

• CyanogenMod 13: Sensors does not work by default, need to flash it. IR Blaster does not work with out change SE state to permisive, also DOS (Dead of Sleep) bug is present and very annoying.
• Mokee: It's based cn CM13, so Sensors does not work by default, need to flash it. IR Blaster does not work with out change SE state. DOS (Dead of Sleep) bug is present.
• Ivan AOSP (Actual rom): It has random reboots. (I have not try IR Blaster but I think is the same as other roms)

I'm starting to think that MIUI is the best solution, even with random reboots and performance decrease that it has...

If anyone knows more about them or can correct something... let me know!

Edit: "and also I have some problems with power button, it does not respond always, but it is not dead, I can unlock it with proximity sensor and AC Display"

I have discovered the issue is caused by AC Display app, removing it, it works fine.

Well, I have update the summary, the power button not responding issue It was caused by an app, not the OS itself.

Anyway, I have re-flashed the IVAN rom following his tutorial to create the system partition and set the recovery image he created and for first time seems really stable. I did this 2 days ago, short time by now, but no ramdon reboots on the last 2 days, before I had 1 per day.

So system partition issue sound to me factible... I will keep testing a few days before post more about it.

EDIT: Also IR is working nicely. I think this is the best rom...

Also, I'm testing loock screen off/on to see if makes different.... sound completely strange but I have seen this a possible issue...
Somebody with lockscreen not secured by pin or other pattner had reboots?

 

[buzdo]

I guess there could be extra band frequencies enabled for china's version among other things as it's also meant for LTE-CT version too, while global supposedly isn't. Also these versions are the very last before MM came along. I do know they were preparing for MM by merging partitions with those versions.

Where do you check kernel version in the rom package? I would like to check if xiaomi.eu 7.1.3.0 is based on the same china rom as 7.1.5.0.

 

[jc rey]

Is xiaomi.eu 7.1.5.0 stable based on china KK rom from that period right? I'm wondering what's so different in that rom compared to last KK 7.1.2.0 global stable rom where reboots supposedly don't happen?

another good confirmation for 7.1.2. I discussed today to a work mate with affected device (Mi4 3/16 overseas) under all marshmallow. I installed him 7.1.2 global (full miflash, all stock) before going on vacation beginning of july, almost 4 weeks ago. He confirmed that he did not have a single reboot while he was close to throw his device before.

 

[jc rey]

Well, I have update the summary, the power button not responding issue It was caused by an app, not the OS itself.

Anyway, I have re-flashed the IVAN rom following his tutorial to create the system partition and set the recovery image he created and for first time seems really stable. I did this 2 days ago, short time by now, but no ramdon reboots on the last 2 days, before I had 1 per day.

So system partition issue sound to me factible... I will keep testing a few days before post more about it.

Also, I'm testing loock screen off/on to see if makes different.... sound completely strange but I have seen this a possible issue...
Somebody with lockscreen not secured by pin or other pattner had reboots?

yes my unaffected that just rebooted (!!!!!) is not pin or password protected, just swipe.