2018 Mac Mini - Restarts at "Getting memory controller details..."

[Blast12345]

Hello! I have a 2018 Mac Mini with 32 gigs of Crucial memory running Big Sur.

I've been getting random crashes over the last month; it is entirely possible that this is related to Big Sur (as I started running the beta around that time) - but I'd like to at least rule out memory.

I downloaded Memtest86 8.4 and imaged it to two USB drives via terminal. Regardless of which I use, I the Mac restarts while initializing Memtest - specifically on "Getting memory controller details..."

Crash log on boot is:

panic(EFI): Exception 14 occurred: Page Fault Addr 0x78ABB000 0x03 write protect

(Full log)

I have to problems at this point:
1. The random crashes
2. Inability to run Memtest

The point of this post is to focus on #2 as that will allow me to (hopefully) determine #1 on my own.

I put the stock memory modules back into the Mini in hopes that I could load Memtest, but to no avail. It still crashes at the memory controller step.
The next thing I tried was removing all unnecessary peripherals; I only had my keyboard, monitor, and USB drive connected. Same issue.
This is quite peculiar because I can find posts of other people running Memtest on 2018 Minis. It seems likely that my memory is causing the the Memtest issue and that worries me. My mini's board died in early October (no idea why), so I've barely been running the new board for two months. It would really suck to have yet another mainboard problem.

I attempted to initialize Memtest on my 2017 Macbook Pro (also running Big Sur) as a sanity check and it passed the memory controller step. I didn't actually run the tests because simply being able to load into Memtest was enough to verify that my USB drive is properly formatted.

I poked around the forums a little bit, but certainly good do some more digging of my own. This thread seems to suggest that there may be a screen resolution issue; I run a 32:9 monitor - so I could totally see that being a problem. I'll try using my old 1080P monitor.

TLDR:
Memtest86 crashes on an initialization step regarding the memory controller.
I suspect it may do to my screen resolution; I will test that this evening.
In the meantime - I'd love to get some input!

[David (PassMark)]

Can you post the debug log from MemTest86
https://www.memtest86.com/tech_debug-logs.html

[Blast12345]

Certainly! A log is attached to this post.
I had actually tried to edit my original post and add it, but I couldn't since it was still pending moderator approval.

What I find striking is that it appears to be failing while printing out the processor cores. Here are some snippets from some tests I ran yesterday:

Mac Mini (failing)

2020-11-18 20:13:26 - Getting memory controller info
2020-11-18 20:13:26 - find_mem_controller - found Intel Coffee Lake (8086:3EC4) at 0-0-0
2020-11-18 20:13:26 - Intel Coffee Lake chipset init
2020-11-18 20:13:26 - MCHBAR_LO=FED10001
2020-11-18 20:13:26 - MCHBAR_HI=00000000
2020-11-18 20:13:26 - MCHBAR=FED10000
2020-11-18 20:13:26 - CAPID0_A=02012071 (ECC capable: no)
2020-11-18 20:13:26 - find_mem_controller - Intel Coffee Lake (8086:3EC4) at 0-0-0
2020-11-18 20:13:26 - find_mem_controller - Intel Coffee Lake ECC mode: detect: no, correct: no, scrub: no, chipkill: no
2020-11-18 20:13:26 - ECC polling disabled
2020-11-18 20:13:26 - Successfully located the PI MpService protocol.
2020-11-18 20:13:26 - BSP is Proc 0
2020-11-18 20:13:26 - This platform has 12 logical processors of which 12 are enabled.
2020-11-18 20:13:26 - ProcID Enabled Type Status Pkg Core Thread Hyperthread?
2020-11-18 20:13:26 - -------- -------- ---- -------- ----- ---- ------ ------------
2020-11-18 20:13:26 - 00000000 Y BSP 00000007 0 0 0 N
2020-11-18 20:13:27 - 00000001 Y AP 00000006 0 0 1 Y

Macbook Pro (working)

2020-11-18 22:25:32 - Getting memory controller info
2020-11-18 22:25:32 - find_mem_controller - found Intel Kaby Lake (8086:5910) at 0-0-0
2020-11-18 22:25:32 - Intel Kaby Lake chipset init
2020-11-18 22:25:32 - MCHBAR_LO=FED10001
2020-11-18 22:25:32 - MCHBAR_HI=00000000
2020-11-18 22:25:32 - MCHBAR=FED10000
2020-11-18 22:25:32 - CAPID0_A=02012059 (ECC capable: no)
2020-11-18 22:25:32 - find_mem_controller - Intel Kaby Lake (8086:5910) at 0-0-0
2020-11-18 22:25:32 - find_mem_controller - Intel Kaby Lake ECC mode: detect: no, correct: no, scrub: no, chipkill: no
2020-11-18 22:25:32 - ECC polling disabled
2020-11-18 22:25:32 - Successfully located the PI MpService protocol.
2020-11-18 22:25:32 - BSP is Proc 0
2020-11-18 22:25:32 - This platform has 8 logical processors of which 8 are enabled.
2020-11-18 22:25:32 - ProcID Enabled Type Status Pkg Core Thread Hyperthread?
2020-11-18 22:25:32 - -------- -------- ---- -------- ----- ---- ------ ------------
2020-11-18 22:25:32 - 00000000 Y BSP 00000007 0 0 0 N
2020-11-18 22:25:32 - 00000001 Y AP 00000006 0 0 1 Y
2020-11-18 22:25:32 - 00000002 Y AP 00000006 0 1 0 N
2020-11-18 22:25:32 - 00000003 Y AP 00000006 0 1 1 Y
2020-11-18 22:25:32 - 00000004 Y AP 00000006 0 2 0 N
2020-11-18 22:25:32 - 00000005 Y AP 00000006 0 2 1 Y
2020-11-18 22:25:32 - 00000006 Y AP 00000006 0 3 0 N
2020-11-18 22:25:32 - 00000007 Y AP 00000006 0 3 1 Y
2020-11-18 22:25:32 - This platform has 8 logical processors of which 8 are enabled.
2020-11-18 22:25:32 - Number of hyperthreads detected: 4
2020-11-18 22:25:32 - Disabling all hyperthreads
2020-11-18 22:25:32 - Disabling hyperthread processor 1
2020-11-18 22:25:32 - Disabling hyperthread processor 3
2020-11-18 22:25:33 - Disabling hyperthread processor 5
2020-11-18 22:25:33 - Disabling hyperthread processor 7
2020-11-18 22:25:33 - Testing MP support
2020-11-18 22:25:33 - MPSupportTestMPServices - AP dispatch test
2020-11-18 22:25:33 - MPSupportTestMPServices - Starting AP#2
2020-11-18 22:25:33 - MPSupportTestMPServices - Starting AP#4
2020-11-18 22:25:33 - MPSupportTestMPServices - Starting AP#6
2020-11-18 22:25:34 - MPSupportTestMPServices - Waiting for AP's to complete execution
2020-11-18 22:25:34 - MPSupportTestMPServices - CPU #2 finished execution
2020-11-18 22:25:34 - MPSupportTestMPServices - CPU #4 finished execution
2020-11-18 22:25:35 - MPSupportTestMPServices - CPU #6 finished execution
2020-11-18 22:25:35 - MPSupportTestMPServices - AP dispatch test complete
2020-11-18 22:25:35 - MP test passed. Setting default CPU mode to PARALLEL
2020-11-18 22:25:35 - Applying configurations
2020-11-18 22:25:35 - [CONFIG] This platform has 8 logical processors of which 4 are enabled.
2020-11-18 22:25:35 - Applying configurations complete
2020-11-18 22:25:35 - Console size = 80 x 25
2020-11-18 22:25:35 - Checking for graphics mode support
2020-11-18 22:25:35 - Graphics mode available
2020-11-18 22:25:35 - Enabling graphics mode
2020-11-18 22:25:35 - Get screen size
2020-11-18 22:25:35 - Current screen size: 3360 x 2100
2020-11-18 22:25:36 - Screen resolution is too high (3360 x 2100). Attempting to set new screen size.
2020-11-18 22:25:36 - Screen size = 3360 x 2100
2020-11-18 22:25:36 - Char width=8 height=19
2020-11-18 22:25:36 - Loading images

Update:
I tried to initialize Memtest with only a 1080P monitor, Mac keyboard, and USB drive connected. Fails in the exact same spot.

[David (PassMark)]

That is a strange (random looking) place to crash.

"Exception 14 occurred", means there was a paging fault. i.e a memory address was accesses by the CPU that was not available and could not be made available (e.g. by swapping it from disk).

According to Google there are exactly zero instances of this error being reported before from UEFI, for any software. So it is super rare. If this was on a brand new Mac, then it would be thinking it was a software fault. But for a fairly common 2018 machine model to report this error for the first time, likely means that it is something special about your machine. i.e. a hardware fault.

[Blast12345]

Huh. How strange. I would have expected more random behavior if it were a hardware fault. I've attempted to run Memtest 10+ times and it always fails in the exact same spot.
Perhaps something is wonky on my 3rd thread (the 2nd physical core) since that is where it is always failing? I'll keep my eye on future system crashes to see if there is a pattern as to the core.

Even so - I don't know if the issue preventing me from running Memtest is the same thing causing my system crashes. Perhaps I have two problems at play; it's hard to tell, so I will just need to continue being observant and gathering evidence.

I've reseated my memory many times over the course of this testing process, so maybe I'll be lucky and a loose module was my issue. Probably not, but I can dream

In the meantime - it's less than ideal, but I can boot into Safe Mode and run Rember.

[David (PassMark)]

Can you test with half the RAM (16GB) then test the other 16GB.

Update: A similar sounding issue was just posted in another thread. Not sure if it is the same issue or not at this point.

[keith]

Can you open 'blacklist.cfg' with a text editor under EFI/BOOT/ of the USB flash drive and replace with the following:

#
# MemTest86 baseboard blacklist file
#
# This file contains a list of baseboards and corresponding UEFI BIOS firmware that are known to have issues that prevent MemTest86 from booting properly.
# Adding a baseboard to the list shall restrict some functionality, but doing so may allow MemTest86 to boot properly.
#
# Blacklist format:
#
# Each blacklisted baseboard is stored on a separate line with the following format:
#
# <baseboard>,<BIOS version>,<EXACT|PARTIAL>,<restriction flag>
#
# <baseboard> is the case-sensitive baseboard string in double quotes (eg. "X99-DELUXE")
#
# <BIOS version> is the first BIOS version (string in double quotes) that no longer exhibits the issue. If no fix is available, specify ALL.
#
# <EXACT|PARTIAL> determines whether exact or partial matching is used on <baseboard string>.
#
# <restriction flag> determines the restriction policy to impose if there is a match. This can be one of the following values:
#
# RESTRICT_STARTUP : Display a warning message before MemTest86 boots
# RESTRICT_MP : Do not perform the multiprocessor test during startup, and set the default CPU mode to SINGLE
# DISABLE_MP : Completely disable multiprocessor support, restricting the CPU mode to SINGLE only
# DISABLE_CONCTRL : Disable console control protocol, which may be needed for some older systems (eg. iMac 7.1)
# FIXED_SCREENRES : Do not modify the default screen resolution, which is known to cause issues on some Lenovo systems
# RESTRICT_ADDR : Set the lower address limit to 0x100000 by default, which may be needed for some Mac systems (eg. Late 2013 27" iMac)
# TEST12_SINGLECPU : Run Test 12 in SINGLE CPU Mode. Workaround for CPU threads hanging in PARALLEL mode.
# DISABLE_LANG : Disable language support and font installation, which is known to cause issues on some Dell systems

"Mac-F42C88C8",ALL,EXACT,RESTRICT_STARTUP
"Mac-F4238CC8",ALL,EXACT,DISABLE_CONCTRL
"Mac-F4208DC8",ALL,EXACT,DISABLE_CONCTRL
"Mac-27ADBB7B4CEE8E61",ALL,EXACT,RESTRICT_ADDR
"Mac-7BA5B2DFE22DDD8C",ALL,EXACT,DISABLE_MP
"Mac-27AD2F918AE68F61",ALL,EXACT,TEST12_SINGLECPU
"80AF",ALL,EXACT,RESTRICT_MP
"Z97MX-Gaming 5",ALL,EXACT,RESTRICT_MP
"Z170MX-Gaming 5",ALL,EXACT,RESTRICT_MP
"Z170X-Gaming 3",ALL,EXACT,RESTRICT_MP
"Z170X-Gaming 7",ALL,EXACT,RESTRICT_MP
"Z170X-Gaming GT",ALL,EXACT,RESTRICT_MP
"Z170X-UD3-CF",ALL,EXACT,RESTRICT_MP
"Z170-HD3P",ALL,EXACT,RESTRICT_MP
"990FXA-UD3",ALL,EXACT,RESTRICT_MP
"970A-DS3P",ALL,EXACT,RESTRICT_MP
"X99-Ultra Gaming-CF",ALL,EXACT,RESTRICT_MP
"M5A97 R2.0",ALL,EXACT,RESTRICT_MP
"M5A97 EVO R2.0",ALL,EXACT,RESTRICT_MP
"M5A99FX PRO R2.0",ALL,EXACT,RESTRICT_MP
"M5A99X EVO R2.0",ALL,EXACT,RESTRICT_MP
"X99-A","3701",EXACT,RESTRICT_MP
"X99-A/USB 3.1","3701",EXACT,RESTRICT_MP
"X99-A II","1701",EXACT,RESTRICT_MP
"X99-A","1701",PARTIAL,RESTRICT_MP
"Sabertooth X99","3701",EXACT,RESTRICT_MP
"SABERTOOTH 990FX R2.0",ALL,EXACT,RESTRICT_MP
"SABERTOOTH 990FX R3.0",ALL,EXACT,RESTRICT_MP
"STRIX X99 GAMING","1701",EXACT,RESTRICT_MP
"X99-DELUXE","3701",EXACT,RESTRICT_MP
"X99-DELUXE II","1701",EXACT,RESTRICT_MP
"RAMPAGE V EXTREME","3701",EXACT,RESTRICT_MP
"MAXIMUS VIII RANGER",ALL,EXACT,RESTRICT_MP
"Z10PE-D8 WS",ALL,EXACT,RESTRICT_MP
"X9DRW",ALL,EXACT,RESTRICT_MP
"X9DRW-3LN4F+/X9DRW-3TF+",ALL,EXACT,RESTRICT_MP
"X9DRi-LN4+/X9DR3-LN4+",ALL,EXACT,TEST12_SINGLECPU
"X9DR3-F",ALL,EXACT,TEST12_SINGLECPU
"X9SRL-F",ALL,EXACT,RESTRICT_MP
"X9DRD-iF",ALL,EXACT,RESTRICT_MP
"151-BE-E097",ALL,EXACT,RESTRICT_MP
"131-HE-E095",ALL,EXACT,RESTRICT_MP
"131-HE-E095-KR",ALL,EXACT,RESTRICT_MP
"151-HE-E999-KR",ALL,EXACT,RESTRICT_MP
"PRIME B350-PLUS",ALL,EXACT,RESTRICT_MP
"PRIME X370-PRO",ALL,EXACT,DISABLE_MP
"CROSSHAIR VI HERO","1602",EXACT,DISABLE_MP
"CROSSHAIR V FORMULA-Z",ALL,EXACT,DISABLE_MP
"ROG STRIX X370-F GAMING","4012",EXACT,DISABLE_MP
"OEMSH Product",ALL,EXACT,DISABLE_CONCTRL
"LNVNB161216",ALL,EXACT,FIXED_SCREENRES
"GA-990FX-GAMING",ALL,EXACT,RESTRICT_MP
"X399 SLI PLUS (MS-7B09)",ALL,EXACT,RESTRICT_MP
"0F685W",ALL,EXACT,DISABLE_LANG
"ProLiant DL325 Gen10",ALL,EXACT,DISABLE_MP
"ProLiant DL385 Gen10",ALL,EXACT,DISABLE_MP
"ProLiant DL385 Gen10 Plus",ALL,EXACT,DISABLE_MP
"06CDVY",ALL,EXACT,FIXED_SCREENRES
"Surface Book 2",ALL,EXACT,FIXED_SCREENRES
"Surface Pro",ALL,EXACT,FIXED_SCREENRES
"Surface Pro 7",ALL,EXACT,FIXED_SCREENRES
"MACH-WX9-PCB",ALL,EXACT,FIXED_SCREENRES

Then run MemTest86 again.

[Blast12345]

Ah! Fantastic news - I was able to full initialize Memtest once I updated the blacklist with the provided text.

I've not run a test yet, so perhaps I'll stumble across some other issues, but this looks very promising. I'll follow up once I've given it a run.

[Blast12345]

Quick update - it ran without issue. So I think it is safe to say this issue is completely solved.

Thank you!