Hardware/Driver/Firmware not compatible - journey to self enlightenment

[deadalus]

New rigs @ 30 Aug

AMD Ryzen
ASUS GT440 GPU
GS Skill 8GB Ram

History
1. Old AM3 system mobo failure
2. Acquire new AMD Ryzen system
3. Fresh install Linux Mint 18.1 64 using Live USB
4. Update graphic driver from Open GL to a recommended Proprietary driver from NVDia (Good for 2 weeks)
5. Update Firmware version (not sure is 4.10 or 4.11. but after the firmware update, the system were good for 1 week or so)
6. System getting unstable. On driver manager there is an unknown device, AMD microcode @#$%^^ (forgot the exact term)
7. At the later stage, totally not able to boot into Linux Mint. (OR the HDD is booting but no display on the monitor. Gods know...)
8. After 10++ attempt, manage to start Linux on Live USB and fresh installation.

Symptom:

Initial stage:
1. The system automatically reboot itself after few hours
2. System auto reboot itself after 20 minutes, and the interval is getting shorter. To a point it is unable to boot up even after a hard reset, power off and power on again.

Advance stage:
3. The system unable to load Linux mint.
4. It will reboot even inside the BIOS screen after less than a minute.

Terminal stage:
5. Unable to load the BIOS. OR it has not display anything on monitor.
6. Sometime CPU fan is spinning but nothing on the screen.
7. Sometime CPU fan is not spinning, and nothing on the screen.

I remember reading on Linux forum somewhere, that don't fix it if it ain't broken.

Will not do any more update after the root cause of previous crash has been identify...

[richyrich]

Have you checked at your motherboard/laptop manufacturer's website for any BIOS update ?

[ClixTrix]

If you can get into Bios, check CPU temperature and voltages. If it's too hot, check if your heatsink is mounted correctly. If voltages are wrong check to see if the PSU ATX and ATX-12V lines are fully seated in the motherboard. In fact, I'd check that last item first to make sure you have them fully seated in the board. If in doubt remove and reseat them.

My board suggests minimum 500 Watt PSU for Ryzen. Which Ryzen do you have as mine is 65W TDP, but most of the upper Ryzen's are 95W? Did you update/replace your PSU that you used on the old AM3?

[deadalus]

AMD Ryzen
ASUS GT440 GPU
GS Skill DDR4 8GB Ram 2400mhz
ASUS PRIME 350-A Motherboard
BIOS 0805 July 2017.

In case anyone is having the same symptoms and wondering how this end.

Eventually last year incident were resolved with changing to a working GPU driver. (if i remember correctly)

Merely after 10 months. The same shit happen again (i think)

July 2018 symptoms:

1. The computer will crash occasionally, power off on its own, at the frequent of twice a month, eventually weekly and more recently daily.
2. 5 days ago the system having problem to POST, unable to load BIOS. Out of 10 times power on/off restart, 1-2 occasion the system will load Mint.
3. 1 day ago. The system totally not POST, no BIOS. CPU will spin 2 secs and stop.
4. Changing RAM slot. Not working.
5. Change to another GPU card. Not working.
6. Unplug the HDD Sata2 cable from Motherboard. No BIOS either.
7. Plugging unplugging HDD GPU RAM 10+ times, on one occasion the system is loading Mint.
8. Restart and boot from Mint 19 Live USB. Erase old data and start fresh.
9. First restart went smoothly. Will monitor.

[sgtor]

Change the power cord, meaning the cable that plugs into the wall mains.

Try that for a while and if the problems persist try a new power supply.

Faulty power supplies are one of the most common causes of the crazy style of errors that you're describing.

You say it's a new rig though, can't you just return it and get a new one if it's under warranty?

Edit:
This is a long shot since it's a newer computer but maybe the CMOS battery is dead?
I've seen these problems in old computers with dead CMOS batteries too.

[deadalus]

thanks for the inputs. i do experience these unexplained crash with my previous 2 old system of >5years, due to dying CMOS battery, dying motherboard (capacitor bloat), fried GPU.

The current system is rather new, only HDD/GPU were from the old rig. PSU, CPU, RAM, motherboard and etc are all new.

For now i will do more start up with existing hardware set up (fresh mint 19 install) and see if it will crash again.

Since the warranty is expired by 30 August, the need to identify the root cause is there. Sending it to pc vendor will incur certain processing fee (i.e motherboard back to ASUS normally there is postage charge 20-30 bucks) and only as last resort.

The old dual core running very well with Mint 16-17.3 until motherboard spoiled. With the hassle now with the new quad core, i do wonder it might be a better choice if i were to spend less on old dual core CPU build up last year.

[deadalus]

It crash again on 2nd start up after good for 20 minutes It sends the GPU fan running wild with intermittent blinking screen before the system force shut down. (CPU fan still spin though)

Plugging in the spare GPU which is an Ati Radeon card fail to boot the system up either, CPU fan spin for 2secs and stop.

Should have checked the GPU driver version when the system is still up. (i remember there is Nvdia & OpenGL driver available for selection)

What is the odd that 2 previously known working GPU were damage at the same time.

Unless i can miraculously power up the system again, getting a brand new GPU seems to be the option left.

[deadalus]

Did send the computer to vendor for checking. They manage to power up the system after initial fail start (CPU fan spin & stop), by unplugging and plugging the RAM in different slot of the dual module. The system stay on for 10 minutes and were later declare as no problem. Knowing that the crash normally kicks in much later, bought a new GPU as plan B.

After returning home, update the the GPU driver from open source nouveau driver to recommended Nvdia 390.48 driver. Run a movie and monitor the temperature, CPU, GPU, HDD were all at the 40C range which looks perfectly normal. The screen were freezed at 1hr05 mark.

Forced restart the system and boot into the OS again, within 10 minutes the system crash and restart itself, the last program in use is Firefox.

From past experience, this looks like a driver issue. If it is hardware failure, either it will not start at all, or it will keep restarting itself due to over heating. Neither the case here.

Compare to the previous release, the Mint 19 driver manager provide much lesser selection on driver, 3 to be specific: Nvdia driver 390, 340, nouveau open source driver. Other than that user need to change the driver via the terminal..

Few more option on the table:

1. Install Mint 15-17 and see which version works. (but i expect problem with the internet browser eventually if it is not keep up to date)

2. Put on a new GPU and finger cross it will work.

3. Move on to other distro i.e Fedora

[deadalus]

the system log before the screen freeze while running a test movie


Aug 05 19:25:05 amar-System-Product-Name systemd[1]: Stopping Clean the /media/amar/E85A75555A75218C mount point...
Aug 05 19:25:05 amar-System-Product-Name systemd[1]: Stopped Clean the /media/amar/E85A75555A75218C mount point.
Aug 05 19:25:05 amar-System-Product-Name udisksd[754]: mountpoint /media/amar/E85A75555A75218C is invalid, cannot recover the canonical path
Aug 05 19:25:05 amar-System-Product-Name udisksd[754]: Cleaning up mount point /media/amar/E85A75555A75218C (device 8:17 is not mounted)
Aug 05 19:25:05 amar-System-Product-Name ntfs-3g[2841]: Unmounting /dev/sdb1 ()
Aug 05 19:25:05 amar-System-Product-Name udisksd[754]: Unmounted /dev/sdb1 on behalf of uid 1000
Aug 05 19:25:23 amar-System-Product-Name kernel: usb 1-8: USB disconnect, device number 2
Aug 05 19:25:23 amar-System-Product-Name upowerd[1251]: unhandled action 'unbind' on /sys/devices/pci0000:00/0000:00:01.3/0000:03:00.0/usb1/1-8/1-8:1.0
Aug 05 19:25:23 amar-System-Product-Name upowerd[1251]: unhandled action 'unbind' on /sys/devices/pci0000:00/0000:00:01.3/0000:03:00.0/usb1/1-8
Aug 05 19:26:01 amar-System-Product-Name systemd[1]: Starting Cleanup of Temporary Directories...
Aug 05 19:26:01 amar-System-Product-Name systemd[1]: Started Cleanup of Temporary Directories.
Aug 05 19:27:15 amar-System-Product-Name rtkit-daemon[1297]: Supervising 2 threads of 1 processes of 1 users.
Aug 05 19:27:15 amar-System-Product-Name rtkit-daemon[1297]: Successfully made thread 6083 of process 1296 (n/a) owned by '1000' RT at priority 5.
Aug 05 19:27:15 amar-System-Product-Name rtkit-daemon[1297]: Supervising 3 threads of 1 processes of 1 users.
Aug 05 19:27:15 amar-System-Product-Name rtkit-daemon[1297]: Supervising 2 threads of 1 processes of 1 users.
Aug 05 19:27:15 amar-System-Product-Name rtkit-daemon[1297]: Successfully made thread 6088 of process 1296 (n/a) owned by '1000' RT at priority 5.
Aug 05 19:27:15 amar-System-Product-Name rtkit-daemon[1297]: Supervising 3 threads of 1 processes of 1 users.
Aug 05 19:28:03 amar-System-Product-Name AptDaemon[1611]: INFO: Quitting due to inactivity
Aug 05 19:28:03 amar-System-Product-Name AptDaemon[1611]: INFO: Quitting was requested
Aug 05 19:28:03 amar-System-Product-Name org.debian.apt[766]: 19:28:03 AptDaemon [INFO]: Quitting due to inactivity
Aug 05 19:28:03 amar-System-Product-Name org.debian.apt[766]: 19:28:03 AptDaemon [INFO]: Quitting was requested
Aug 05 19:33:02 amar-System-Product-Name systemd[1]: Starting Message of the Day...
Aug 05 19:33:02 amar-System-Product-Name systemd[1]: Started Message of the Day.
Aug 05 20:00:01 amar-System-Product-Name CRON[9918]: pam_unix(cron:session): session opened for user root by (uid=0)
Aug 05 20:00:01 amar-System-Product-Name CRON[9919]: (root) CMD (timeshift --check --scripted)
Aug 05 20:00:02 amar-System-Product-Name crontab[9952]: (root) LIST (root)
Aug 05 20:00:02 amar-System-Product-Name crontab[9953]: (root) LIST (root)
Aug 05 20:00:02 amar-System-Product-Name CRON[9918]: pam_unix(cron:session): session closed for user root
Aug 05 20:02:02 amar-System-Product-Name systemd[1]: Started Run anacron jobs.
Aug 05 20:02:02 amar-System-Product-Name anacron[10187]: Anacron 2.3 started on 2018-08-05
Aug 05 20:02:02 amar-System-Product-Name anacron[10187]: Normal exit (0 jobs run)
Aug 05 20:17:01 amar-System-Product-Name CRON[11937]: pam_unix(cron:session): session opened for user root by (uid=0)
Aug 05 20:17:01 amar-System-Product-Name CRON[11938]: (root) CMD ( cd / && run-parts --report /etc/cron.hourly)
Aug 05 20:17:01 amar-System-Product-Name CRON[11937]: pam_unix(cron:session): session closed for user root

[deadalus]

Gut feeling > Driver is the culprit that causing all these instability.

Did a fresh Mint 19 installation, wipe out all previous system file.

Change to 390 Nvdia driver. Keep the update manager untouched and using the stock firmware that comes with the Mint19 ISO.

Let see.

[ClixTrix]

"BIOS 0805 July 2017"

I would highly recommend you update BIOS to current for your motherboard.

[sgtor]

First off I admire your persistence to stay with Linux and to keep trying to find the problem. Now you need to get into a mindset of process of elimination. Trying one thing at time with it's purpose being to prove it's not the problem. This way you narrow it down and will eventually find the real problem. I still think you should try my suggestion with the power cord and then the power supply. If you simply don't want to follow any kind of process like I'm suggesting then the one thing you can try is your theory of it being the driver.

Burn yourself a Linux Mint 18.3 CD or USB and boot up with it. Watch a movie and see what happens.
You'll then know for sure if it's a driver issue.

Your vendor techs were correct to check and reseat the memory. I had a problem once after I moved across the country and tried to use my computer in my new place. I actually forget what the problem was but I started running diagnostics and it showed memory errors. So I opened up the case and one of the memory sticks was slightly loose on one end. I only discovered this when I pushed the sticks in to make sure they were tight. Visually they looked fine and seated properly. This was years ago when the memory slots didn't have that latch to hold them in place.

My point is this, looking at that memory, it seemed fine yet it was still able to cause such problems.

If you were using windows then yes I can understand a driver causing corrupt memory and BSOD and the need to reboot, etc etc.
You're using Linux though. I find it hard to believe that a video driver in Linux can cause such problems as random reboots.

One of the ways the video driver could cause a reboot would be for it to overwrite the memory where the OS is sitting. Another way would be to cause the video card to pull too much power from the system for a few seconds thus causing a reboot. I'm sure there are other ways that I can't think of right now.

Do you see this being likely with the Linux OS? It's not impossible sure, but IMO it's unlikely.
I think you need to start testing your hardware and eliminating possible problems until you find the culprit.
First though try the Mint 18.3 USB like I suggested and see if that is the problem.

Edit:
I just realized the error in my post. If you try USB first and it is a hardware problem it will still reboot on you and you still won't know for certain what caused the error. On the other hand if it doesn't reboot then you will have proved that it may be the driver. I said try that that first because you are so intent on it being the driver.

[sgtor]

Advance stage:
{snip}
4. It will reboot even inside the BIOS screen after less than a minute.

I just went back and reread your first post. This is proof enough that it is not the video driver. I'm sorry to say it is hardware related. I know you may not want to hear that but it is.

You've been back to the vendor already while it was still under warranty and they failed to fix it so I would go back and demand they do fix it even if the warranty is expired.
There's a few things that could be wrong with it.
1. power cord or power supply
2. faulty memory
3. faulty motherboard

Whatever is new is the likely culprit so it's probably the motherboard. If you have the same memory and power supply, cord, etc then it's likely not any of them. However a good question to ask is what caused your old MB to fry?

[deadalus]

First off I admire your persistence to stay with Linux and to keep trying to find the problem. Now you need to get into a mindset of process of elimination.

thanks for the inputs.

as of now i have 3 GPU with me, and will have no problem if 1 or 2 of the older one went faulty. Just bought a new GPU card 2 weeks ago.

my theory is simple. if the GPU chips is fried up, no screen and no boot. if the GPU fan is faulty, the system will reboot when the card is over heating and this is as good as gone. the way that the computer is behaving now is beyond my understanding.

for some reason, most the the crash is not capture in the system log of Mint 19. on 1 of the crash that being recorded in system log, it points to the GPU driver.

the 600W PSU is 11 months bought together with the Ryzen CPU and is fairly new. power cord is supplied together with the PSU. if needed, can borrow some power cord for testing.

[deadalus]

Advance stage:
{snip}
4. It will reboot even inside the BIOS screen after less than a minute.

There's a few things that could be wrong with it.
1. power cord or power supply
2. faulty memory
3. faulty motherboard

Whatever is new is the likely culprit so it's probably the motherboard. If you have the same memory and power supply, cord, etc then it's likely not any of them. However a good question to ask is what caused your old MB to fry?

Memory & motherboard faulty is difficult to validate when the system can be power-on randomly. Since the crash can take place randomly 10 minutes after start up and may also last 4-5 hours before it crash. The PC shops does not seems willing to spend that much of time. I could left my computer at the shop, but unless i stay there to withness the entire trouble-shooting session, there is no telling whether the vendor only spend 20 minutes or actually let the system run for 10 hours.

Apart from the GPU Asus 1GB & the HDD 1TB, the rest are "newly" bought in 2017 Aug.

The previous old system of 8 years have 1 or 2 bloated capacitor on the motherboard. Totally will not power on.

[deadalus]

Some update:

--Same hardware as Aug 2017--
AMD Ryzen
ASUS GT440 GPU
GS Skill DDR4 8GB Ram 2400mhz
ASUS PRIME 350-A Motherboard
600W PSU
BIOS 0805 July 2017.

- 9 August. Fresh mint 19 installation from Live USB.
- Update the Display driver to Nvdia 390 proprietary driver. All update in the Update manager is untouched.
- After install Nvdia 390, restart the system.
- System crash after 20 minutes or so. The crash is not shown in the system log.

After above mentioned events, 9 August until 16 August, under a daily usage of 3-4 hours, there is no incidence of any crash take place.

System: Host: neo-System-Product-Name Kernel: 4.15.0-20-generic x86_64
bits: 64 gcc: 7.3.0
Desktop: Cinnamon 3.8.8 (Gtk 3.22.30-1ubuntu1)
Distro: Linux Mint 19 Tara
Info: Processes: 202 Uptime: 16:02 Memory: 3844.0/7967.7MB
Init: systemd runlevel: 5 Gcc sys: 7.3.0
Client: Shell (bash 4.4.191) inxi: 2.3.56

[richard-g8jvm]

Hiya
You didn't say which Ryzen CPU you are using ,
You will definitely need to update the bios on that mobo, as there were a few problems that even new mobos are being checked for the latest bios.
If you are using an additional graphics board, and not the integrated graphics on the motherboard, once the bios is updated, you
should be OK
The first generation Ryzen CPUs did have a bug, that was cleared with the current generation , AMD5 ryzen 2xxx series,
You mentioned Nvidia so I'm guessing an additional graphics card, the 2xxxG CPUs have vesa 11 additional cores.
AMDGPU is not fully supported in the current kernel with LM19 , it IS in Kernel -4.17 which is available and I use, as I use a Ryzen2400G.
Also can you clarify which Asus motherboard you are using, I checked the asus site for the bios versions for that board you quoted and its not listed
All the prime mobos listed start with a letter. and the are no AMD boards listed that dont start with a letter.
HTH

[deadalus]

model: PRIME B350M-A

Ryzen 1200, the 1st gen

a bit hesitant on bios update, the older mint version 15-18 were known to work with old computer of 10 years+ (it used to work on my old desktop until the hardware broken down) without having to equip with the latest BIOS.

[richard-g8jvm]

Hi
the Bios version thats validated against your CPU is version 0405 Feb 17 2017
the subsequent updates have been for performance update and later processors
https://icr-am.asus.com/webchat/icr.htm ... 1534173337
Dont know why it sends you to the web chat, but you dont need to talk to anyone.
From some of the other lists there is a bug on some of the asus boards with the B350 chip set.
That has been fix in later bios updates.
Asus are pretty good at keeping backward compatibility, un like MSI, I had to return a new mobo and swap it for an Asus one.
I would go ahead and up date to the latest version, which would eliminate those known problems , and speed setting on the early Ryzen processors.
HTH

[sgtor]

Every suggestion in this thread has been good and would lead to solving the problem.
The real problem though is you always have an answer for why none of it will work or why you don't want to try it.

So it seems you already know how to fix it. Why isn't it fixed yet?

After reading Richard's post I agree with him and that should be your first move.
Some of first posts also said update your BIOS. It turns out there are two problems and maybe that was confusing you.
1. hardware problem with the BIOS(Technically a firmware problem but you get my point)
2. Software problem with the LM19 kernel

You say you're reluctant to do the BIOS update and I can understand that, so take it back to the vendor and have them do it. Your window of time on the warranty is running out so do it soon.

That's pretty much all I have to say for this thread. The answers are there for you to solve the problem, it's up to you to take action on it now.