"Hard" reboot... nonsense, right?

[Lex]

Deleting settings happens in nanoseconds.

I sure hope they are doing a secure wipe of files / partition(s) with personal info, not just file deletes. That will take a lot longer.

Love the thread though, and am amazed that some cars / configs can no longer reboot at all

 

[CK_Stuggi]

I sure hope they are doing a secure wipe of files / partition(s) with personal info, not just file deletes. That will take a lot longer.

Love the thread though, and am amazed that some cars / configs can no longer reboot at all

Considering all those settings probably fit in files of very few kilobytes (web addresses, WiFi credentials, location coordinates, Homelink numerical codes and seat position settings are essentially text files, maybe hashed, but nonetheless), even a ten times secure random overwrite should not take long.

 

[bcsteeve]

Only if "long" is measured in fractions of a second. I'm sure you're right that it would be a tiny amount of information. It isn't stored on a magnetic drive anyway, so I don't think privacy concerns are an issue in the slightest.

Then again... if my seat position data falls into the wrong hands.... fuk, can you imagine? Someone might be able to engineer a pair of jeans using my measurements! I shutter to think.

 

[CK_Stuggi]

So... Just for the fun of it, this morning I did press the break for a reboot of the MCU.

The reason being the argument that a) Tesla Assistance does recommend the practice and b) the pedal press can be used to flag sensors used for driving to be reset as well.

Considering the reset of a sensor to be barely more than also cutting its power, this would not affect the duration of the reboot.

Now as to my result; the moment I rebooted the MCU, a "touchscreen unresponsive, hold both..." warning popped up, and sure enough, the touchscreen was unresponsive and didn't recover (I waited past requisition of the LTE signal). It never did that on a non-brake reset before for me.
So, brake-reset again, everything back to normal.
This might be anecdotal. But it certainly seems like a strange coincidence.

 

[bcsteeve]

I can agree that the pedal press *might* be used for a flag of some sort... but for the reasons I gave before, I really doubt it is. You're possibly right that time-to-reboot may not be definitive, but I still think my points stand: 1) if you are going to differentiate and so something different with the sensors, then it would still take longer because surely if you're bothering to do that, then the processor is going to do some calibration checks of the sensors. 2) (more importantly) as a designer, you're not going to choose the brake as a trigger for something "different" if "different" actually matters.

To clarify on 2... if "different" is really different - and that difference matters - then you aren't going to choose a signal that is going to be thoughtlessly triggered or not triggered so easily. You'll pick something that's not likely to be casually pressed. More likely, you'd use an alternative sequence. For example, continuing to hold the buttons through the reset so that they are pressed upon boot.

If it really doesn't matter, on the other hand... ie. it makes no difference in boot time (it doesn't) and there's no harm in doing a "hard" vs "soft"... then they would just make every reset a hard reset.

From all evidence, I'm standing by my assertion that there's no difference between the supposed "soft" and "hard" resets, in particular I think the brake element is complete nonsense. The only thing I'll add to my findings is that the time to leave the car in the powered-off state may make a difference in some problems. I've confirmed with Tesla that the power-off through the screen is just cutting off the high voltage systems and not the 12v that runs the computers, etc. Possibly that could be so the computer can specifically reset some sensors after the HV has been off for X cycles (for example to allow capacitors to discharge). There have been consistent reports that Tesla has instructed the driver to use that power off option (and to not touch ANYTHING) for 3+ minutes, and that has fixed problems relating to AP sensors.

 

[WannabeOwner]

I've read all this, a year after the last post - because I followed a a link from another thread. I'm a Software Guy and fully agree .. except for one point:

If it really doesn't matter, on the other hand... ie. it makes no difference in boot time (it doesn't) and there's no harm in doing a "hard" vs "soft"... then they would just make every reset a hard reset

Except: if a Reset does something different when you are stationary, which must never be done when moving. It might not take any (measurably) longer time, but it would be "more". That's the only justification for "press the brake when you initiate the reset" that I can think of, and possibly also "keep pressing the brake" so that after the reboot the "startup" does some extra "We are definitely stationary" conditional code. Might only take a few milliseconds ...

I don't build software like that, I would put an option up for the user 1. "Big reset", 2. "Small reset" ... or just have the car figure out if we are in PARK or not ... I mean ... the car DOES know if we are in PARK, right?

 

[csimgo]

No matter if I do the pedal+button reset or just button reset, and power off the car, still not showing streetlines or recognizing other cars on the road.

 

[jimbo69ny]

Just came across this thread. I have a pesky error message and my car will not drive. Closest SC is 5 hours each way and I really prefer not to deal with them. I have tried the "soft" resets above but not the factory (hard) reset yet.
Instead I disconnected my 12v battery about 2 hours ago and I am hoping that clears it. I havent reconnected it yet.

Some people say doing this will clear a code and some people say it wont. I'm crossing my fingers.

 

[dhrivnak]

A few times I've come across posters mentioning a "hard" reboot vs a soft or normal one. Some go to painstaking detail in outlining the crazy voodoo steps required to engage this magical hard reboot. Some even purport to say there are times when such a reboot is "necessary".

I say it is all nonsense. [skip to end for a short version]

Here's a particular and recent thread on the subject (different forum, sorry) where you'll see several people explaining that there's a difference. One even links to a different thread, presumably as evidence that there is in fact a difference... but that thread contains more of the same: pure conjecture and a bunch of - I believe - nonsense. There are many MANY other threads, blogs, etc all spreading the same [mis]information.

The general "consensus" seems to settle on two different ways to achieve the same magical state: 1) Press the brake... no wait, press it "hard", while doing the common reset and you'll do a "hard" reset; 2) hold the reset switch down for 20, no 30, no 45 seconds to get to the "hard" reset.


Neither of these made much sense to me, but both were plausible. First, the time component: We've all had experiences when some low-tier tech support rep at our ISP says to unplug the router and wait 30 seconds, so that must be how to do it right? The time element is ONLY to clear a capacitor, ensuring power has fully been turned off. Once a reset is triggered, it is triggered. You can't trigger it "harder" by holding it down longer. Now... it is possible to trigger a true hard reset by having the button pressed at the beginning of a power on vs. while powered on, so holding it down through the complete cycle *can* make a difference. But does it on the Tesla? We'll see. As far as the brake goes: pressing it hard vs. not had to be complete nonsense, but it could be possible that the brake is a secondary hardware input signaling the computer to do a "hard" reset vs. a quickie... to me that made little sense because if there truly is a difference between "hard" and normal resets, surely they'd want a less conspicuous trigger than something you probably press out of muscle memory whenever you're sitting in the driver seat and not driving. I mean, if there REALLY is a difference, they don't want you doing one vs the other all willy-nilly, right?

But... I'm not one to put my thoughts as correct and others as nonsense without at least SOME experimentation, so that's what I did.

In all my years working with electronics, I have never ever EVER experienced a system where one type of startup vs. another didn't affect the time to some degree. For example, you do a soft restart of your Windows system vs. a full power down, power on... one takes longer than the other, right? If you unplug your router and plug it back in, that takes a different amount of time than if you do so while holding the reset, right? A "hard" reset is simply going to take longer than a soft one... right? If we can't agree on that, then you're dismissing my entire premise so no need to read on.

I went out to my car with stopwatch in hand. I engaged the reset of the instrument cluster (holding down both buttons above the thumbwheels) without the brake pressed and the display went down at 14 seconds. I immediately let go. The display started back up 11 seconds later and was fully on 4 seconds after that. I repeated this TEN TIMES, because if this wasn't consistent than there was no way I could convince myself that witnessing a "hard" reset was or wasn't real. It was precisely consistent, with each and every one taking the same timing plus or minus half a second. Then I repeated it with the brake pressed with normal pressure... same result. Then I pressed the brake hard... same result. Then I pressed the brake SUPER HARD - just in case that really was the trick... same result. No variance at all. There's no way a magical "hard" reset was being performed. Then I tried again but this time I never let go of the buttons. Again, the display went out after 14, on 11 later and fully on 4 after that... I continued to hold... for 5 boring minutes. Nothing. That was simply 5 minutes I'll never get back. Myth busted. I then repeated the same nonsense with the 17" display (holding down both thumbwheels). This one resets after only 4 seconds, takes 21 to come back on and another 12 to be ready to rock. Brake, no brake, hard brake, keep holding... no variance whatsoever.

So then out of curiosity I decide to do a real factory reset. To me, that's a "hard" reset. Because if that too took more or less the same amount of time then maybe I'm wrong and maybe hard vs soft just doesn't make enough time difference to notice. The funny part is that I didn't want to inconvenience my wife to reset her profile, so i put it in her profile before doing the reset. Now let me set the stage for you: I'm 6'4 and closer to 300lbs than I'd like to be. My wife is 5'1 and a tiny little thing. So here I am contorting myself as best I can with her profile squeezing me like a grape while I do this reset. I'm thinking instead of the 30 seconds or so that it might be a minute. No such luck... I was smashed up for over 6 minutes while it did its thing. Now THAT is a hard reset! Lest you think that there's actual demons in the box that have to paint over the memory sections one stroke at a time... that isn't explained simply by settings getting deleted. Deleting settings happens in nanoseconds. That time difference has to be due to a complete kernel reload.

Conclusion? Yes, there is a "hard" reset. No, it isn't dependent on how hard (or even if) you press the brake or whatever other nonsense voodoo. Its right in the settings screen.



Am I wrong? By all means, prove it.


[TL/DR Version]
- Instrument cluster takes consistent 14 seconds to engage reboot, 11 seconds to power back on and 4 seconds to normal state. This is regardless of whether or not brake is pressed or buttons held extended time.
- Similarly, 17" display is 4 + 21 + 12. Again, no difference with brake, etc.
- Factory reset takes several minutes.
- Conclude that no "hard" reset exists, other than the factory reset.

No I think you are spot on. I have tried both "soft" and "hard" resets and I am convinced they are the same and take the same time and have yielded the same results. So for now I just press both thumb wheels and let'r reboot.

 

[drtimhill]

..... Conclusion? Yes, there is a "hard" reset. No, it isn't dependent on how hard (or even if) you press the brake or whatever other nonsense voodoo. Its right in the settings screen. .....

But wait .. was the car facing west when you did these tests????

 

[Lex]

Interesting thread resuscitation here.. I've got some more of my own personal experience to report:

1) this year I did perform a factory reset on my Model S, and not only do I suspect it doesn't do any sort of secure wipe of the drive, but it actually leaves some personal data on the system after the reset. Yikes. From the surface it seemed to only be media player related if that makes you feel better, but even these items can reveal sensitive information about a person.

2) There seems to certainly be a difference between the "soft" reset of just pressing down the two scroll wheels for a couple of seconds, versus the "hard" reset of holding them down until the Tesla logo appears. I had not been bothering to hold them down but my MCU didn't come back once after a "soft" reset so I had to hold them down to get MCU back. Now out of paranoia I always hold them down until logo.

I don't think the brake has anything to do with it. They probably just don't want customers to drive around as they reset their cars (which for me now is all day long to try and get MCU1 browser to work again, and now I have to hold down the scroll wheels for a while too. sigh).

 

[kavyboy]

Remembering this thread, I asked the mobile tech about this a couple of months ago after he disconnected the 12V battery and we were waiting around. According to him, there's the process of disconnecting the 12V battery and letting the system shut down. There's also the steering wheel reboot(s). That's it. There's no special procedure and no other kind of reboot.
The guy had been doing this for years and seemed to know his stuff.

 

[scottf200]

... I've confirmed with Tesla that the power-off through the screen is just cutting off the high voltage systems and not the 12v that runs the computers, etc. ...

'17 X 100D --
I was surprised recently doing some test with the 'CANbus' tool ScanMyTesla and saw 57 stats still being collected as I was sitting in the drivers seat after doing the POWER OFF menu option.
Include data for both the HV and 12v batteries.

 

[bcsteeve]

Interesting how this thread got revived!

@scottf200 I'm not really surprised to see that, simply because of how quickly the car sparks to life after that supposed "power off". I long suspected that it did something very specific relating to the high voltage lines, but was in no way a powering down of the car's systems.

 

[ThisIsTrue]

But wait .. was the car facing west when you did these tests????

Quit trying to deceive people. As we all know, the car must be pointed directly at Fremont!

 

[kavyboy]

Please develop this, but for Fremont instead. Perhaps we'll get a new Easter Egg.
https://play.google.com/store/apps/details?id=com.apptome.navigation.meckapil&hl=en_US

 

[drtimhill]

Please develop this, but for Fremont instead. Perhaps we'll get a new Easter Egg.
https://play.google.com/store/apps/details?id=com.apptome.navigation.meckapil&hl=en_US

New voice command "navigate to origin" immediately adds navigation route back to Fremont

 

[Chaserr]

From the surface it seemed to only be media player related if that makes you feel better, but even these items can reveal sensitive information about a person.

Media player info is actually stored outside of the car. Slacker / Spotify accounts are created for your car specifically, and the account is kept so when you reset it logs back in and pulls back down your favorites. We used to be able to call tesla and ask for our Slacker login so we could make changes on a computer. I don't know if they still do it but that's why I know the settings are stored in the cloud somewhere.

 

[CaliMS]

Not claiming any special knowledge on voodoo reboot methods, but this is how Tesla sells it:

I had an issue with my media player yesterday and called Tesla. I told them I had already done the "dual wheel button reboot". They told me to do a full reboot and emailed me these instructions:

How to Power Cycle a car:

• Navigate to Controls (Car icon in bottom left corner of touchscreen) > Safety & Security > Power Off
• Once the car has been turned off for around 2 minutes, either open the door to the car or put your foot on the brake to turn the car back on

When should a Power Cycle be used?
If the car is displaying alerts on their touchscreen that don't seem to have a clear cause, a Power Cycle is recommended. However, alerts that directly point to an issue, such as a 12v Battery Replacement, should be diagnosed and addressed by Tesla.

 

[Duke-U]

Interesting discussion. May I add another consideration to it.
Pressing the scroll wheel creates a data message on the CAN-bus. This has been used e.g. for switching a HDMI switch from rear camera to my media box. It also differentiates between a short press and a long press (I need to long press the menu button to switch it).

Another thing is that there are many CAN-bus connected subsystems, little computers on their own, like the window control box or the HVAC. So keeping these steering buttons/wheels pressed together with a brake pedal press could well cause some resets of subsystems on the drive canbus that we have no human interaction with. Would be interesting to see what CANbus message is send around when you initiate a reboot.

Also, with replacing my eMMC recently, I got to see the logs and config files in there. I was really amazed what is sent to the mothership. One thing I saw is that the limiting of supercharging speed on my 90 pack was initiated by a person logging in on the car.

All this is pure speculation and in line with Tesla’s very poor customer communications. Maybe @verygreen can shine a light on this?