Fix KERNEL_DATA_INPAGE_ERROR on Windows 11

Back up your files before you do anything else. This blue screen is different from most — KERNEL_DATA_INPAGE_ERROR means Windows tried to read a page of data from your storage drive and the read failed. Very often that’s a drive physically dying, and every fix in this guide involves reading and writing to that potentially failing drive, putting more stress on hardware that might be marginal. A failing drive can survive long enough for a calm backup, then die completely during a CHKDSK scan because CHKDSK hammers every sector.

If Windows boots long enough to reach the desktop, plug in a USB drive and start dragging your important folders over — Documents, Photos, Desktop, whatever you can’t replace. If the machine crashes mid-copy, it restarts and you continue where you left off. A partial backup is infinitely better than nothing when the drive finally gives up. Our data recovery guide covers more options if you can’t access files normally.

Check the Drive First

Download CrystalDiskInfo and read the SMART data your drive monitors about its own health. Three numbers matter: Reallocated Sectors Count should be zero — anything above means the drive has started failing. Current Pending Sector Count shows sectors that threw read errors but haven’t been confirmed bad. Reported Uncorrectable Errors is sectors the drive couldn’t read at all and couldn’t remap.

Yellow “Caution” or red “Bad” status? Stop running repair commands. The drive is dying and CHKDSK will make it worse. Focus on getting your data off, then replace the drive. A 1TB NVMe runs $60-80 in 2026. Our hard drive guide covers the full diagnostic.

If CrystalDiskInfo shows blue “Good” with all counts at zero, this probably isn’t hardware — keep reading.

A common case: an HP Pavilion blue-screens every time Lightroom opens. CrystalDiskInfo shows hundreds of reallocated sectors on a two-year-old Seagate 2TB. The drive is dying one sector at a time, and whenever Windows hits the chunk of paging file that lives on a dead sector, the kernel panics. The right move is to get the files off before the drive fails completely and swap in a fresh NVMe. Repeatedly rebooting and running CHKDSK, as many guides suggest, can push a failing drive over the edge and cost you the data.

The 24H2 and Western Digital Bug

If you updated to Windows 11 24H2 and have a WD_BLACK SN770, WD Blue SN580, WD Blue SN5000, or SanDisk Extreme M.2 in the 2TB size — there’s a known bug where Windows requests a larger Host Memory Buffer allocation than these drives can handle. Every diagnostic comes back clean, CrystalDiskInfo shows Good, CHKDSK finds nothing, MemTest passes. The drive is healthy, it’s a protocol-level incompatibility.

Fix: open WD Dashboard from westerndigital.com, go to Firmware tab, install the update. Three minutes and a reboot. Or the registry workaround: open regedit, navigate to HKLM\SYSTEM\CurrentControlSet\Control\StorPort, create a DWORD value called HMBAllocationPolicy, set it to 0, reboot.

Software Fixes

Drive is healthy and you’ve ruled out the WD bug. Run DISM first, then SFC in admin Command Prompt:

DISM /Online /Cleanup-Image /RestoreHealth

Wait for that to finish, then:

sfc /scannow

SFC and DISM fix corrupt system files in some kernel data inpage cases where the drive is healthy — less often than with most blue screen errors because this one is more often hardware-driven, but it’s fast and free.

Then chkdsk C: /f /r — it needs to run at next restart. The /r flag scans for bad sectors and attempts recovery. On a 1TB HDD expect two to three hours, SSDs usually under 30 minutes. If CHKDSK finds bad sectors, check CrystalDiskInfo again afterward — if Reallocated Sectors Count went up, the drive is degrading and you should plan a replacement even if the crashes stop temporarily.

Check RAM too. MemTest86 from USB, at least four passes overnight. Windows Memory Diagnostic is basically useless — it runs two quick passes and misses intermittent faults. For example, a system with an ASUS TUF board and two sticks of Corsair DDR4-3600 might throw this error only during video rendering. Memory Diagnostic says fine; MemTest86 finds errors in pass 4, all in the same address range on one stick. Pull that stick and the crashes stop — and RAM like this is usually covered by the manufacturer’s warranty.

Also check your paging file: right-click Start, System, About, Advanced system settings, Performance Settings, Advanced tab, Change under Virtual Memory. Make sure “Automatically manage paging file size” is checked. Some older optimization guides recommend shrinking it to 1GB, and an undersized paging file alone can cause this crash.

Reading the Error Code

The blue screen itself tells you more than just the stop code name. Disable automatic restart first — right-click Start, System, About, Advanced system settings, Startup and Recovery Settings, uncheck “Automatically restart.” Next crash, the screen stays up long enough to read.

0xC000009C = bad sectors, drive returned a CRC error, hardware replacement needed. 0xC000009D = general I/O device error, could be dying drive or loose SATA cable. 0xC0000185 = can’t establish communication with the storage device at all, check cables on desktop or reseat the NVMe on a laptop. 0xC000009A = insufficient system resources, usually a paging file that’s too small or a drive completely out of space — easiest to fix. 0xC000016A = corrupted filesystem structure, CHKDSK has a good chance here.

If you didn’t catch the code, Event Viewer after reboot — Windows Logs, System, look for Source “Disk” with Event IDs 7, 11, 51, or 153. Event ID 11 specifically means your boot drive is having read failures.

If none of this worked — drive healthy, RAM passes, SFC clean, CHKDSK clean, paging file automatic — the problem is deeper. Sometimes it’s a storage controller driver conflict, sometimes a SATA cable that tests fine at rest but loses contact under vibration, sometimes an NVMe thermally throttling because it’s sandwiched against a GPU backplate with no heatsink. Our BSOD Fix pulls your crash dump files from C:\Windows\Minidump and analyzes them with WinDbg to identify the exact faulting module — average session runs about 30 minutes because the dump usually points straight at the answer. If the drive is failing and you need files recovered, we can image it sector by sector before it degrades further.