Hello,

I'm running into a problem with my DLL procedure where it crashes the program on the wrmsr instruction. I'm trying to follow the Intel procedure to read the CPU's microcode revision.

The procedure is as follows:

[code]
proc ReturnCPURevision uses rcx ;from Intel's IA-32 Volume 3A Ch9 9.11.7.1
push rcx
;fasm manual: wrmsr writes the contents of registers EDX and EAX into the 64-bit MSR of the address specified in the ECX register
;EAX+EDX -> MSR; MSR[ECX];
xor eax,eax ;clear EAX
xor edx,edx ;clear EDX
mov ecx,08BH ;IA32_BIOS_SIGN_ID
wrmsr ;load 0 to MSR at 8BH
mov eax,01H
cpuid
mov ecx,08BH ;IA32_BIOS_SIGN_ID
rdmsr ;Read Model Specific Register
;"If there is an update active in the processor, its revision is returned in the EDX register after the RDMSR instruction executes"
pop rcx
;mov [rcx],edx
ret
endp
[/code]

Any help would be appreciated as I can't seem to find why this doesn't work.

jmurray: You will need to make a driver for Windows and execute the code within the driver context.

A normal user mode program has no chance to do this.

Writing and installing a driver can range from easy to hard depending upon your version of Windows. XP and below you have a much easier time. Vista and above you will find many barriers.

Thank you both for replying.

I found it a bit odd that the "Intel Processor Identification Utility" program seemed to be able to view the CPU's revision. And it didn't mention requiring a driver. I'm now presuming that this is read from the registry instead

Thank you for your help. I had hoped it was possible to accurately read it from software somehow.

Agner Fog's driver might jump-start your effort if you want to pursue this ...
https://www.agner.org/optimize/#testp

In my own experience, without driver signing on Windows 10 it requires turning off the check at boot every time the driver is needed. Driver signing is expensive. For this type of work, Windows is becoming more effort than it's worth. Yet, your use case might differ. Intel can afford driver signing.