Root [ROM] Quattrimus JB (CM10, AOKP41)

[tdm]

Kids reported another spurious reboot. It's not the same as the last one though. Here's the relevant lines in /proc/last_kmsg:

[ 7674.264465] saw_set_voltage: 8901_s1: msm_spm_set_vdd failed -5
[ 7674.264892] decrease_vdd: vdd_sc (cpu1) decrease failed (-5)
[ 7674.265106] cpufreq: cpu1 init at 1674000 switching to 1512000
[ 7692.417480] SysRq : Emergency Remount R/O
... followed by about two dozen lines of shutdown stuff.

EIO is 5 (so -EIO is -5). There are two ways for msm_spm_set_vdd() to return -EIO: (1) the if() in that function fails, or (2) msm_spm_drv_set_vdd() returns -EIO. We can rule out (2) since the error path in msm_spm_drv_set_vdd() logs an error and we don't see it. So it must be (1). Here's the if():

if (!atomic_read(&msm_spm_set_vdd_x_cpu_allowed) &&
unlikely(smp_processor_id() != cpu)) {
goto set_vdd_x_cpu_bail;
}

Now what the heck is that code doing and why did it fail...?

 

[reeseboy]

Now what the heck is that code doing and why did it fail...?

Started the new years party too early and drank itself into a coma is my guess

 

[tdm]

And my nitro just did the same exact thing with the same exact error messages. Hmmm .. I was thinking it was overclock related but now I'm thinking it's not. At least it's reproducible.. Just let the kids play games for an hour or two... Lol

 

[IMUcarmen]

And my nitro just did the same exact thing with the same exact error messages. Hmmm .. I was thinking it was overclock related but now I'm thinking it's not. At least it's reproducible.. Just let the kids play games for an hour or two... Lol

for what it's worth, I've been playing bridge construction kit lately and i get reboots while playing it sometimes. I'm not a huge gamer so I can't report on gaiming in general.

 

[ortrigger]

Now what the heck is that code doing and why did it fail...?

What is it remounting? It almost looks to me that the remount is causing the cpu to lock up.

 

[littlenanner85]

I have been getting failed to connect wifi pop up multiple times. It does it on different wifi sources too. A Reboot fixes it.

 

[MrB206]

I have been getting failed to connect wifi pop up multiple times. It does it on different wifi sources too. A Reboot fixes it.

Do you have the auto connection and network notification turned on? I recall seeing that error when I had network notification turned on.

 

[tdm]

What is it remounting? It almost looks to me that the remount is causing the cpu to lock up.

It's not entirely clear exactly what's going on. But it appears that failure to set a hardware voltage regulator causes the kernel to do an orderly shutdown and reboot.

The remount message is part of that orderly shutdown. Actually the function that produces this message is intended to be invoked from the so-called "magic sysrq key", but since we have no physical console, something is obviously invoking the sysrq functions via other means.

I'll be instrumenting the kernel to try to find out more. The kids are well on their way to having careers as professional software testers... it's actually my 7 year old's dream job, to be a video game tester, since he read about it in a kids magazine.

 

[phillip_viado]

I'm just wondering, will we ever see cm10.1 (4.2) on the spectrum?

 

[death2all110]

I'm just wondering, will we ever see cm10.1 (4.2) on the spectrum?

No.

 

[ortrigger]

It's not entirely clear exactly what's going on. But it appears that failure to set a hardware voltage regulator causes the kernel to do an orderly shutdown and reboot.

The remount message is part of that orderly shutdown. Actually the function that produces this message is intended to be invoked from the so-called "magic sysrq key", but since we have no physical console, something is obviously invoking the sysrq functions via other means.

I'll be instrumenting the kernel to try to find out more. The kids are well on their way to having careers as professional software testers... it's actually my 7 year old's dream job, to be a video game tester, since he read about it in a kids magazine.

Did you enable adaptive voltage scaling when you compiled this kernel? I could be completely off track but could it be a bug in the code for that?

 

[tdm]

Did you enable adaptive voltage scaling when you compiled this kernel? I could be completely off track but could it be a bug in the code for that?

I'm not aware of a separate option to enable or disable voltage scaling.

 

[ortrigger]

I'm not aware of a separate option to enable or disable voltage scaling.

Okay. I found it while exploring the menuconfig in the kernel I'm working on. I guess that rules out my idea. Are there actual voltage tables in this kernel? I haven't been able to find them myself. Maybe it's something relating to them?

 

[tdm]

Okay. I found it while exploring the menuconfig in the kernel I'm working on. I guess that rules out my idea. Are there actual voltage tables in this kernel? I haven't been able to find them myself. Maybe it's something relating to them?

Yes acpuclock-8x60.c has voltage tables. It apparently works most of the time but there seems to be a bug in a corner case or race condition or something.

 

[ortrigger]

Yes acpuclock-8x60.c has voltage tables. It apparently works most of the time but there seems to be a bug in a corner case or race condition or something.

I'll look into that and see if I can see anything.

 

[MrB206]

Just came across this after I wiped data and flashed beta 4. Anyone else see this error?

 

[ortrigger]

Just came across this after I wiped data and flashed beta 4. Anyone else see this error?

View attachment 46562

I haven't seen that one. Did you try wiping data and cache for play store?

 

[legmando33]

Just came across this after I wiped data and flashed beta 4. Anyone else see this error?

View attachment 46562

I've had that happen to my on my s3. Did you change DPI?

 

[MrB206]

I've had that happen to my on my s3. Did you change DPI?

Yea, but I changed the dpi before and didn't see that error. Lemme set it back to 320 and see what happens.
*edit* no dice to both.

 

[legmando33]

The only way I've ever fixed it for myself was to dirty flash the ROM.

 

[reeseboy]

Just came across this after I wiped data and flashed beta 4. Anyone else see this error?

View attachment 46562

I've gotten that a few times when I started plying around with dpi. The way I fixed it was resetting the dpi to 320 and then wiping the play store data and rebooting. It took a few times but it worked finally.

 

[IMUcarmen]

Just came across this after I wiped data and flashed beta 4. Anyone else see this error?

View attachment 46562

yea I've been getting that too. With google search, and a few other apps as well.

 

[IMUcarmen]

I'm just wondering, will we ever see cm10.1 (4.2) on the spectrum?

tdm said a month or so ago that it may happen. I could be off base here, but I think the focus is to get what we got now good and stable and then try to get 4.2 running.

 

[littlenanner85]

I did have notification turned on but auto connection turned off. I will try it with it off thanks.

Edit: still after awhile it fails to connect. Recognizes it but will not connect.

 

[jfatica]

On the topic of reboots..

I've had 5 spectrums since April, and they've all had random reboots.. some more frequently than others. My latest one has been stable for about 3 weeks. I was only able to pinpoint and repeat one of the issues I had with a past phone, it was the car sensor randomly triggering and loading the car dock app...then causing a reboot.

This has happened on stock GB, Black Plague, Broken Out, and stock ICS. So far no problems on CM10. Although now that I'm hearing reports of reboots I'm a little nervous.

I think this is just a cheaply made phone.