Quick answer

IRQL_NOT_LESS_OR_EQUAL (0x0000000A) and DRIVER_IRQL_NOT_LESS_OR_EQUAL (0x000000D1) occur when kernel code or a third-party driver touches paged or invalid memory while running at DISPATCH_LEVEL (IRQL 2) or higher. Do not blame ntoskrnl.exe—it is the Windows kernel catching the fault. Undo your most recent driver, anti-cheat or antivirus change, test with XMP/EXPO disabled for 48 hours if no driver is named, and inspect Parameter 3 and Parameter 4 in Event Viewer or WinDbg.

Fix steps

4 steps

Diagnostic workbench
Contents

At a glance

Applies to
Windows 11 and Windows 10, stop codes 0x0A and 0xD1
ntoskrnl.exe?
Victim frame in the Windows kernel—look for a third-party .sys or RAM instability
First check
Any .sys file named on the blue screen + last driver/anti-cheat update
Hardware check
Turn XMP/EXPO off for 48 hours before running MemTest86 or buying RAM

Provenance

  • Nov 21, 2026Published

IRQL_NOT_LESS_OR_EQUAL (0x0000000A) and its sibling DRIVER_IRQL_NOT_LESS_OR_EQUAL (0x000000D1) share the same rule in Windows internals: once a CPU core raises its Interrupt Request Level (IRQL) to DISPATCH_LEVEL (2) or higher, the Windows memory manager is forbidden from pausing that thread to fetch paged-out data from your SSD. Touching a paged-out address—or dereferencing a NULL, freed, or bit-flipped pointer—at raised IRQL halts Windows immediately.

The two codes give you an instant clue before you touch a single setting:

  • DRIVER_IRQL_NOT_LESS_OR_EQUAL (0xD1, decimal 209): The faulting instruction sat directly inside a device driver (.sys). Check what failed on the blue screen (What failed: xxx.sys), update or roll back that driver, and you are usually done.
  • IRQL_NOT_LESS_OR_EQUAL (0xA, decimal 10): The trap fired inside core Windows (ntoskrnl.exe), either because a driver handed the kernel a bad pointer or because unstable RAM / XMP / EXPO / CPU undervolting flipped a bit in a kernel pointer.

Step 1. Check the .sys name and read Parameters 1–4

Open Event Viewer > Windows Logs > System, locate the Kernel-Power (Event ID 41) or BugCheck (Event ID 1001) entry from the crash time, and inspect the four parameters documented on Microsoft’s 0xA and 0xD1 reference pages:

ParameterWhat Windows recordedHow to read it
Parameter 1 (Arg1)Memory address that couldn’t be accessedUnder 0x1000 (for example 0x0 or 0x28): Microsoft says this is almost always a NULL pointer dereference (a software/driver bug, rarely bad RAM). Wild high address with a single odd bit: points toward use-after-free or hardware bit-flipping
Parameter 2 (Arg2)IRQL at the moment of the faultCommonly 0x2 (DISPATCH_LEVEL)
Parameter 3 (Arg3)Operation bitfield0x0 = Read fault · 0x1 = Write fault · 0x2 or 0x8 = Execute fault (tried to execute paged or non-executable memory)
Parameter 4 (Arg4)Instruction address that made the bad accessIn WinDbg (!analyze -v), maps to the exact driver function if 0xD1

Why ntoskrnl.exe is the messenger, not the cause

Tools that read only the top stack frame routinely highlight ntoskrnl.exe. Do not chase “corrupt ntoskrnl.exe” guides first. ntoskrnl.exe is the Windows kernel; when a network filter, anti-cheat module, audio driver, or unstable memory stick corrupts a structure, ntoskrnl.exe is simply the first code to trip over it.

Step 2. If a .sys driver is named (or crashes started after an update), fix that software stack

When 0xD1 prints a .sys filename—or 0xA started right after installing a game, VPN, peripheral app, or driver update—start here:

Faulting .sys familyComponentFix
Netwtw*.sys, rtwlane*.sys, e2fnexpress.sys, ndis.sys, tcpip.sysIntel / Realtek / Killer Wi-Fi or Ethernet, or a third-party VPN / firewall packet filterUninstall third-party VPN/network-booster utilities first (cfosspeed, Killer Intelligence Center), then install the board maker’s current LAN/Wi-Fi driver
nvlddmkm.sys, amdkmdag.sys, igdkmdn64.sys, dxgmms2.sysNVIDIA, AMD, or Intel graphics driverClean-reinstall a WHQL driver following our GPU driver crash & DDU guide
vgk.sys, EasyAntiCheat_EOS.sys, BEDaisy.sysKernel anti-cheat (Riot Vanguard, Easy Anti-Cheat, BattlEye)Repair or update the anti-cheat (Vanguard error guide, EAC repair guide) and check whether Windows Memory Integrity (HVCI) conflicts with an older hardware driver
RTKVHD64.sys, Nahimic*.sys, rzendpt.sys, lghub*.sysAudio or peripheral / RGB kernel driversUpdate motherboard Realtek audio or uninstall peripheral macro/RGB suites to test

Stuck in a boot loop? Interrupt Windows boot three times to reach Windows Recovery > Troubleshoot > Advanced options > Startup Settings > Restart, press 4 or F4 for Safe Mode (keep your BitLocker recovery key handy if disk encryption is active), and roll back or uninstall the driver from Device Manager or Installed apps.

Step 3. No single driver named? Run the 48-hour JEDEC memory & stock-CPU test

Microsoft’s 0xA guide specifically calls out bit-flipping as a source of bad pointers at DISPATCH_LEVEL. Whenever 0xA crashes under gaming or compile load with ntoskrnl.exe at the top of the dump—or alternates with PAGE_FAULT_IN_NONPAGED_AREA (0x50), MEMORY_MANAGEMENT (0x1A), or KERNEL_SECURITY_CHECK_FAILURE (0x139)—suspect memory or CPU voltage margins before Windows itself:

  1. Enter BIOS and turn XMP or EXPO off (Auto / JEDEC defaults—see how XMP and EXPO work). If you also run an aggressive CPU undervolt (AMD Curve Optimizer negative offset or Intel voltage offset), temporarily return CPU voltage to stock at the same time.
  2. Use the PC normally for 48 hours across the exact games or workloads that triggered 0xA.
  3. Crashes vanish at JEDEC? Your RAM sticks aren’t necessarily dead—your CPU’s memory controller or motherboard BIOS couldn’t hold that overclock profile (especially common when running four DDR5 sticks instead of two). Update your motherboard BIOS (AGESA / Intel microcode) before re-testing a stepped-down memory speed.
  4. Crashes continue even at stock JEDEC? Boot a USB stick with PassMark MemTest86 and run the single-stick A2 vs B2 slot isolation matrix in our MEMORY_MANAGEMENT (0x1A) guide before spending money on a new DDR5 kit.

Step 4. Read the full call stack in WinDbg and repair corrupted OS files

If JEDEC defaults didn’t change anything and Event Viewer didn’t name a .sys module:

  1. Install WinDbg from the Microsoft Store, open the newest .dmp copied from C:\Windows\Minidump, and run:
    !analyze -v
    k
    
  2. Scan down the k stack trace below nt!KiBugCheckDispatch and nt!KiPageFault. Any third-party .sys module sitting mid-stack right before the page fault is your prime suspect—even when SYMBOL_NAME at the top still prints nt!Mi... or nt!Ex....
  3. After updating or removing the faulting driver (or stabilizing memory), repair any OS file corruption left behind by hard crashes from an elevated Terminal (Admin):
    DISM /Online /Cleanup-Image /RestoreHealth
    sfc /scannow
    

A caution on Driver Verifier (verifier.exe)

Microsoft documents Driver Verifier on its 0xA page because it catches drivers touching paged pool at DISPATCH_LEVEL red-handed. Leave it off until Steps 1–4 fail, never select “Automatically select all drivers installed on this computer,” and create a System Restore point first: verifying only non-Microsoft third-party .sys modules—and running verifier /reset from Safe Mode or Command Prompt as soon as a minidump is written—prevents a permanent verification boot loop.

Frequently asked questions

What is the difference between IRQL_NOT_LESS_OR_EQUAL (0xA) and DRIVER_IRQL_NOT_LESS_OR_EQUAL (0xD1)?
Both mean code running at an elevated Interrupt Request Level (typically DISPATCH_LEVEL, IRQL 2, or higher) tried to touch pageable or invalid memory, where page faults are forbidden. Windows logs 0xD1 when the faulting instruction address sits directly inside a loaded device driver (.sys) module, and 0xA when the fault surfaced inside core kernel code—often after a driver passed a bad pointer or unstable RAM flipped a bit.
Why do BlueScreenView and WinDbg blame ntoskrnl.exe for 0xA?
ntoskrnl.exe is the Windows NT OS kernel itself. When a third-party driver passes a NULL or freed pointer to a kernel routine—or an unstable DDR5 XMP/EXPO profile corrupts a pointer in flight—ntoskrnl.exe is the code executing when the CPU traps the invalid address. Replacing or ‘fixing’ ntoskrnl.exe is almost never the answer; finding what fed it the bad address is.
Can XMP, EXPO, or bad RAM cause IRQL_NOT_LESS_OR_EQUAL?
Yes. Microsoft’s 0xA documentation explicitly notes that the invalid address in Parameter 1 can come from a bad pointer caused by bit-flipping. If no single third-party .sys driver repeats across your crashes—or the crashes started after enabling XMP/EXPO or populating four DDR5 slots—test at default JEDEC memory speeds for 48 hours first.
Should I run Driver Verifier to find the bad driver?
Keep Driver Verifier off a daily PC unless normal minidump inspection fails and you create a System Restore point and know how to enter Safe Mode first. Microsoft notes that Driver Verifier adds live overhead and intentionally crashes offending drivers immediately—which can put a machine into a boot loop if a boot-start driver faults.
How do I get out of an IRQL_NOT_LESS_OR_EQUAL boot loop?
Let Windows fail to boot three times (or interrupt startup with the power button) to enter Windows Recovery Environment, then select Troubleshoot > Advanced options > Startup Settings > Restart and press 4 or F4 for Safe Mode. From Safe Mode, uninstall or roll back the driver, VPN client, RGB tool or security software installed right before the loop started.

Sources and corrections

Last updated
Methodology
See our methodology for research and review standards. It draws on 5 cited sources, listed below, each checked against the original page on the date above. Parameter definitions (Parameters 1 through 4, IRQL levels, read/write/execute bitfields, NULL-pointer vs bit-flipping guidance, and KiBugCheckDriver) come directly from Microsoft Learn’s Bug Check 0xA and Bug Check 0xD1 pages opened November 21, 2026 (UTC), paired with Microsoft Support’s Safe Mode, System File Checker and stop-code guides.
Corrections
Spotted an error? Email a correction with the affected line and what you found. The draft includes this page’s address. Verified changes are recorded here. Contact details.

About the author

I am Lowell K. Wood IV. I research, write and maintain everything on TechFuelHQ from St. Louis County, Missouri.