NV4x 12th Gen v1.8.0: Random silent crashes with no kernel logs or crash dump

[scottwallacesh]

Component

EC firmware

Device

NovaCustom NV4x 12th Gen

Dasharo version

v1.8.0 (2026-03-30)

Dasharo Tools Suite version

No response

Test case ID

No response

Brief summary

The laptop crashes randomly and silently multiple times per day with no warning. No kernel logs are written, no crash dump is produced, and the system reboots as if a hard reset occurr

How reproducible

Constantly - multiple times per day, completely at random with no apparent pattern (no specific workload, application, or activity triggers it).

How to reproduce

Boot laptop and use "as normal".

Expected behavior

Stable operation without random crashes.

Actual behavior

System "randomly" reboots with notice or indication of a problem.

Screenshots

No response

Additional context

Device: NovaCustom NV4x 12th Gen (NV41)
Dasharo version: v1.8.0 (2026-03-30)
EC firmware: 2025-10-31_af6bd04
OS: NixOS 25.11
Kernel: 6.12.83
Graphics: Intel-only (i915)

Kernel log observations:
intel-hid INT33D5:00: failed to enable HID power button
This appears on every boot and may be relevant.

Current kernel parameters:
i915.enable_psr=0 loglevel=4 lsm=landlock,yama,bpf crashkernel=128M nmi_watchdog=panic softlockup_panic=1

Additional notes:
The fact that kdump never fires and no logs are written strongly suggests the crash is occurring at the hardware or EC level rather than within the kernel. The system reboots rather than powering off. This issue was present before the v1.8.0 firmware update (on the previous Dasharo version) and has continued after it.

Solutions you've tried

• Upgraded from kernel 6.12.63 to 6.12.83 - no improvement.
• Added kernel parameter i915.enable_psr=0 - no improvement.
• Added crashkernel=128M and boot.crashDump.enable = true - no crash dump is ever produced in /var/crash, suggesting the crash happens below the kernel level.
• All relevant panic sysctls enabled (kernel.hardlockup_panic=1, kernel.softlockup_panic=1, kernel.panic_on_oops=1 etc.) - none trigger.
• Persistent systemd journal enabled - journal from previous boot contains nothing useful around the time of crash.

[scottwallacesh]

It looks like clamping the CPU C-state to 1 has resolved the crashing.

i.e. setting the kernel parameter intel_idle.max_cstate=1

This indicates a potential problem with the EC firmware. Any comments/recommendations?

[scottwallacesh]

It looks like fwts found some issues:

FAILED [HIGH] AMLAsmASL_MSG_SYNTAX: Test 1, Assembler error in line 5278
Line | AML source
--------------------------------------------------------------------------------
05275|                                 Store (0x0C, Local2)
05276|                                 While (LGreater (Local2, Zero))
05277|                                 {
05278|                                     If (LEqual (Zero, Break)){}
     |                                                          ^
     | Error 6126: syntax error  
05279|                                     Sleep (0x19)
05280|                                     Decrement (Local2)
05281|                                 }
================================================================================

ADVICE: (for Error #6126, ASL_MSG_SYNTAX): The disassembled code cannot be
reassembled using the strict IASL compiler as it contains syntax errors.

FAILED [HIGH] AMLAsmASL_MSG_SYNTAX: Test 1, Assembler error in line 5423
Line | AML source
--------------------------------------------------------------------------------
05420| 
05421|                             Increment (\_SB.PCI0.TDM0.HR.DFP0.PWR)
05422|                         }
05423|                         Else
     |                           ^
     | Error 6126: syntax error  
05424|                         {
05425|                             Return (Zero)
05426|                         }
================================================================================

ADVICE: (for Error #6126, ASL_MSG_SYNTAX): The disassembled code cannot be
reassembled using the strict IASL compiler as it contains syntax errors.

FAILED [HIGH] AMLAsmASL_MSG_SYNTAX: Test 1, Assembler error in line 5444
Line | AML source
--------------------------------------------------------------------------------
05441|                     })
05442|                 }
05443|             }
05444|         }
     | ^                                           
     | Error 6126: syntax error and premature End-Of-File
05445|     }
05446| 
05447|     Scope (\_SB.PCI0.I2C0)
================================================================================

ADVICE: (for Error #6126, ASL_MSG_SYNTAX): The disassembled code cannot be
reassembled using the strict IASL compiler as it contains syntax errors.

FAILED [HIGH] SyntaxCheckIASLCompilerAborted: Test 1, Compilation aborted early
due to a parser detected syntax error.

ADVICE: Some subsequent errors may not be detected because the compiler had to
terminate prematurely. If the compiler did not abort early then potentially
correct code may parse incorrectly producing some or many false positive errors.

FAILED [HIGH] SyntaxCheckIASLCompilerAborted: Test 1, Compilation aborted early
due to a parser detected syntax error.

ADVICE: Some subsequent errors may not be detected because the compiler had to
terminate prematurely. If the compiler did not abort early then potentially
correct code may parse incorrectly producing some or many false positive errors.

Table SSDT (1) reassembly: Found 26 errors, 0 warnings, 0 remarks.

Test 18 of 209: Test _CRS (Current Resource Settings).
FAILED [MEDIUM] Method_CRS64BitQWordAddrSpaceMaxLessThanMin: Test 18,
\_SB_.PERC._CRS QWORD Address Space Descriptor minimum address range 0xc0000000
is greater than the maximum address range 0xfffffff.

ADVICE: See section '6.4.3.5 Address Space Resource Descriptors' table 6-179 of
the ACPI specification for more details about how the _MIF, _MAF and memory
range and granularity rules apply. Typically the kernel does not care about
these being correct, so this is a minor issue.

FAILED [MEDIUM] Method_CRS64BitQWordAddrSpaceLengthInvalid: Test 18,
\_SB_.PERC._CRS QWORD Address Space Descriptor length 0x10000000 does not match
the difference between the minimum and maximum address ranges
0xc0000000-0xfffffff.

ADVICE: See section '6.4.3.5 Address Space Resource Descriptors' table 6-179 of
the ACPI specification for more details about how the _MIF, _MAF and memory
range and granularity rules apply. Typically the kernel does not care about
these being correct, so this is a minor issue.

Test 196 of 209: Test _BQC (Brightness Query Current Level).
ACPICA Exception AE_AML_DIVIDE_BY_ZERO during execution of method DRCL
FAILED [CRITICAL] AEAMLDivideByZero: Test 196, Detected error 'Division by zero'
when evaluating '\_SB_.PCI0.GFX0.LCD0._BQC'.

ADVICE: A division by zero error occurred while executing the AML Divide op-code

[scottwallacesh]

Full report:
fwts_results.log

[scottwallacesh]

I'm currently clamped with the cstate at 4 (intel_idle.max_cstate=4) and this also seems stable.

[tlaurion]

I'm currently clamped with the cstate at 4 (intel_idle.max_cstate=4) and this also seems stable.

Random unrecoverable issues without kernel crashing are the worst to troubleshoot, but point to possible ram stick issues.

memtest+ over night and see if it's ram related?

[Sh4dowzinha]

I have the exact same model and firmware, and no issue on my part. Maybe it could really be a ram issue

[PeterLacknase]

Dito. Can't replicate the problem on NV4xPZ 1240p with 1.8.0.