Tag: Troubleshooting

MSI B650 Tomahawk EZ Debug LED Memory and CPU Issues – Early Adopter Woes One Year Later

Robert Lohr1 Comment

I have recently treated myself to an AM5 system on an MSI B650 Tomahawk WIFI. The R5 7600 has access to 32 GB of G.Skill 6000 MT/s CL32 memory, which supports AMD’s EXPO technology. Enabling EXPO worked fine for the first hour before the trouble began.

At first, the SSD wasn’t detected anymore. I wanted to reinstall Windows because of licensing issues since I chose the wrong edition at first. Then, changing settings in BIOS did not restart the computer. The BIOS just froze up. Finally, the CPU and memory EZ Debug LEDs were both active, indicating a fried CPU and RAM.

That was a fun experience.

The Debugging

For reference, here is the layout of the debugging LEDs.

I am showing you this because the image of the error was under bad lighting conditions.

Whatever I tried, the system appeared to be dead. No more BIOS, no CMOS reset, nada. A CMOS reset had no effect even when the system would still boot into BIOS. Cutting the power… you guessed it: did nothing.

In a last hurrah, I attempted to boot with only one memory stick installed. Maybe it was just a single DIMM that unleashed hell. Lo and behold, the system booted. Testing the other stick also booted. Trying both RAM sticks again also booted. What the devil?

Changing the RAM configuration from two sticks down to one triggered something in the BIOS to reconfigure itself. A BIOS often shows a message when changes to the previously installed hardware are detected. And suddenly, a BIOS reset also worked.

The Fix

This time, I ran the RAM at 4800 MT/s while I installed Windows and tested a game for good measure. After ensuring the system was stable, I dared to perform a BIOS update.

And this was the solution to the problem. My MSI B650 Tomahawk came with BIOS version E7D75AMS.160; the latest version is E7D75AMS.170. The changelog mentions several improvements regarding RAM support.

An unreleased beta version also mentions stability improvements with EXPO enabled. I assume this also made its way into the latest release.

Since then, the system has been stable. I have run several benchmarks and played Starfield for about four hours straight without any stability problems.

The Disappointment

I am shocked that users can still run into compatibility issues like this about a year after the release of the Ryzen 7000 processors. The original installed BIOS was published in May 2023. That was already about eight months after the release of the AM5 platform, and it still had issues.

I hope this was helpful and saved you some time debugging a similar issue on the same motherboard or even a different one that behaves the same.

Thank you for reading.

Windows 11 Host VirtualBox Install Error “This app can’t run because it causes security or performance issues”

Robert LohrLeave a comment

I have recently upgraded to Windows 11 out of curiosity. Despite the more or less negative first impression, I decided to continue to use it. One of the first applications I install is VirtualBox to try out different Linux flavors and stay current in that world. To my surprise, the VirtualBox installer (version 6.1.26) would not start. Windows was complaining about issues with this software.

> This app can’t run because it causes security or performance issues > on Windows. A new version may be available. Check with your software > provider for an updated version that runs on this version of > Windows.

Well, I checked because it was the latest version of VirtualBox. I found hints on the Internet that VirtualBox does run on Windows 11, albeit without indicating what these persons had done.

A little bit discouraged, I clicked the "Learn more" button. You never know; it might actually be helpful – or a complete waste of time. In this instance, it was of great help. It redirected me to the following Microsoft page discussing the "A driver can’t load on this device issue". It also contains a very convenient link to the corresponding location in the Windows Defender application. Somehow I cannot reproduce that link for your convenience so you must visit Microsoft’s site yourself.

Be aware. There may be a security risk associated with disabling this setting. I have not yet dug deeper to ascertain the whole picture. I figured it must have been disabled or not existed on Windows 10 at all, and I was fine there. Windows will ask you several times to grant administrative rights to perform the operation and require a reboot.

After that, VirtualBox was installed and ran just fine.

Curious, I wondered if I could disable the setting once VirtualBox was installed.

Well, I could not. Windows will try and fail. If you click "Review incompatible drivers", it will show you which component prevents the change. And sure enough, it is VirtualBox.

We will see if Oracle’s VirtualBox team can figure this out, but I would assume so. For now, this works for me.

I hope this has helped you. Thank you for reading.

Debian Testing “Bullseye”: The Repository Does Not Have a Release File

Robert Lohr8 Comments

After installing Debian Testing "Bullseye" mid March 2020 I got an error trying to run apt update.

E: The repository 'http://security.debian.org./debian-security 
bullseye/updates Release' does not have a Release file.

Unfortunately, the Debian maintainers managed to let a bug creep into the /etc/apt/sources.list. It’s called "Testing" for a reason, I guess.

The offending lines are this.

deb http://security.debian.org/debian-security/ bullseye/updates main
deb-src http://security.debian.org/debian-security/ bullseye/updates main

Note "bullseye/updates", which is where the error is. Change those two lines to look like this.

deb http://security.debian.org/debian-security/ bullseye-security main
deb-src http://security.debian.org/debian-security/ bullseye-security main

After that, the update will work. I have noticed that in a later version of the installer this bug has been fixed.

Ryzen Master not Resetting to “Auto” Control Mode

Robert LohrLeave a comment

Recently I set out to figure out how much clock speed I can squeeze out of my Zen+ based Ryzen 5 2600. To make life easier I figured I use Ryzen Master so I can change the settings while I’m in Windows so I don’t have to reboot every time I increase the clock speed. This has worked nicely until the point where I figured the viable maximum was. The next step was to dial those numbers "into hardware", meaning setting the options in the BIOS so that Ryzen Master is not required any more. And this is where my issues started to appear.

The Error

First, here’s a screenshot of the message Ryzen Master was giving me. After that I’ll explain what had happened.

In order to set the CPU multiplier you have to change from automatic to manual mode in Ryzen Master. I wanted to reset all options to their defaults after setting the overclock in the BIOS, but I always kept getting the message that Ryzen Master wants to restart Windows because the setting was changed to "Manual" – which it wasn’t, but more on that later. So I did as it asked multiple times with the same outcome every time. Effectively, I was doing a boot loop manually.

So, how did I get there?

The Journey

In brief:

  • Find a stable overclock in Windows using Ryzen Master.
  • Reboot to BIOS and set the overclock closer to the hardware.
  • Reboot to Windows and reset everything in Ryzen Master.
  • Manual "Boot Loop" a few times.
  • Notice CPU always at 4GHz, no more Cool’n’Quiet operation mode.
  • Undo overclock in BIOS.
  • Still see overclock in Windows.
  • Uninstalling Ryzen Master.
  • Still see overclock in Windows.
  • Ryzen Master still not resetting.
  • Manual "Boot Loop" a few more times.
  • Getting pissed and searching the Internet – apparantly I was not alone.
  • More reboots and tests with BIOS settings.

The Fix

It was the frickin’ BIOS! Ryzen Master was not to blame.

I have an ASRock B450 Gaming mITX mainboard with the latest non-Matisse (Ryzen 3000) BIOS. It is not recommended to upgrade unless a Ryzen 3000 is installed. There’s a weird bug in the BIOS that still applies the overclock even if the setting is set to "Auto by AMD CBS" (or something like that). There were two things that helped:

  1. Load BIOS defaults.
  2. Enable manual control and set the correct CPU base frequency at 3400MHz.

When applying the overclock with 4000MHz it effectly ran at 4GHz every time, even in idle. When setting 3400MHz it properly clocked down and also boosted as a R5 2600 should. The same setting only with a different clock value produced a different behavior. And unless the BIOS defaults are loaded the "Auto" mode doesn’t do what you expect – if you’ve set an overclock previously.

Curiously enough, booting Fedora Linux from an USB stick did properly scale the CPU frequency based on the load, even with the overclock applied. Apparently only Windows or AMD’s drivers didn’t manage to do that. Booting a Linux helped me to rule out Ryzen Master as the root of the always applied overclock although the BIOS setting was set to the default Auto mode.

The takeaways:

  • Don’t overclock on this mainboard.
    • The OC options for the CPU are laughable at best. No way to set the multiplier per core.
  • Next time buy a higher-end mainboard for overclocking (ITX is expensive though…).

The Pitfall of Trying to Be Too Smart When Using dllexport/dllimport

Robert LohrLeave a comment

As a seasoned C++ developer I should’ve been aware of this which makes it a little bit embarrassing. But, since this issue has cost me several hours of searching through the Internet over the course of two or three days, I thought it might be worth sharing. Maybe somebody else is trying to be too smart or just doesn’t know better.

The problem? It is summarized in short in this StackOverflow question that I posted. With this blog post I’ll be a bit more elaborate and show some details.
Read More »

Grails Startup Error: ASSERTION FAILED at JPLISAgent.c

Robert LohrLeave a comment

Grails always has something new to offer, not only features but also random bugs (unfortunately). Out of the blue I couldn’t even let Grails show its own version (grails –version) let alone perform any other action. This is all I got for each and every command I tried to execute.

*** java.lang.instrument ASSERTION FAILED ***: "!errorOutstanding" with 
message transform method call failed at 
../../../src/share/instrument/JPLISAgent.c line: 844 
Exception: java.lang.StackOverflowError thrown from the 
UncaughtExceptionHandler in thread "main"

The following runtimes were used:

  • Windows 7 64bit
  • Java 7 Update 71 64bit
  • Grails 2.4.4

Of course I removed all temporary files and folders, the “target” folder of the project and the “.grails”, “.groovy”, “.m2” and “.ivy” folders in the user directory. Nothing helped. Some say it has to do with forking processes, but playing with those settings didn’t change a thing. After all, the error happens way earlier.

Then I came across a post that mentioned to create the “.inputrc” file (on a Linux system) because through debugging it was found that Grails tries to access this files. Well, I’m not using Linux, but since I was already in the “helplessly desperate” phase I wanted to give it a shot. Surprisingly, this file already existed.

Solution: I deleted “.inputrc” from the user folder et voilà, Grails worked again.

ASRock P67 Pro3 BIOS Error Code 97

Robert Lohr1 Comment

Recently I wanted to install a Gigabyte Radeon HD7870 in an older PC with an ASRock P67 Pro3 Mainboard. The surprise was big when the monitor didn’t show an image and the computer didn’t boot. Instead, the mainboard’s debug panel showed the error code 97. According to the manual this means “Console Output devices connect”. Not connectED but more likely in the process of initializing the graphics card and failing while doing that.

There’s an easy fix for that. Installing the latest BIOS version (3.30, installed at the time was 2.02) resolved the issue and the computer booted without problems.

Grails Upgrade 2.3 to 2.4: Validation Errors

Robert LohrLeave a comment

Recently I have updated our Grails 2.3 based web application to Grails 2.4. Although the 2.3 release was working fine, one doesn’t want to fall too far behind. I know out of experience that this can happen very fast. If you wait too long, then at some point the migration to a newer version is almost like starting from scratch, instead of just updating a few lines of code to accommodate for deprecated APIs. The biggest problem I encountered going to version 2.4 was a behavioral change regarding the validation.
Read More »

The Bumpy Road to PC 5.1 Surround Sound

Robert LohrLeave a comment
When building the PC for gaming on the TV one thing I had in mind was leveraging the already existing 5.1 sound system. After the move from the TV screen back to a desktop monitor I thought my headphones would suffice for the time spent playing games. At first that assumption turned out to be true, however, not only did I use the headphones for gaming but also when watching TV shows. In the evening, after work, I wanted to enjoy the audio but had no interest in disturbing my neighbors. After a while this led to the headphones becoming quite uncomfortable for all those hours wearing them, especially during the weekend gaming session when having them on the head for several hours.

So, what does a tech-nerd do about that? Buy himself a dedicated sound system for the PC, he does!

Read More »

Mysterious Case of iPod Shuffle Always Shuffling

Robert LohrLeave a comment
Just recently I experienced something odd. My good old iPod Shuffle seemed to ignore the “play in order” setting on the device and instead elected to do as the name implies and shuffle the songs.
What preceded that sudden change of behavior was a failed attempt to sync with iTunes that actually turned into restoring the firmware of the device. For whatever reason, iTunes detected a problem and its only solution was to wipe the iPod. So I did that and afterwards I had that shuffling play-in-order mode.
What actually happened is that with the restore of the firmware the default sorting of tracks on the device changed to be by their track-id. That means it sounded random to me but was in order for the Pod-thing. To solve my problem I opened the “What’s on my iPod” screen and clicked on the “Name” column to sort by artist name.

I “saved” this setting by syncing and then I was done.

(Mountain) Lion automatically wakes from sleep

Robert LohrLeave a comment

Since I first got my iMac I always wondered why the machine would suddenly just wake up from Sleep in the middle of the night. I didn’t touch it since at night I prefer to sleep (or else I must be sleepwalking). I initially found a way to fix this but it came up again with the update to Mountain Lion (which just reset the preference).
Go the System Preferences and hit Energy Saver. There you’ll uncheck Wake for Wi-Fi network access (in German: Ruhezustand bei Netzwerkzugriff beenden). I figured that must be the reason because at night my router automatically turns off wireless and just before I get home from work it turns it back on. For me it did the trick.
Here are two images, one from my machine (in German) and one found on Google for an english machine.