Variations in CPU stepping leading to crashes on the same device.
Variations in CPU stepping leading to crashes on the same device.
I'm setting up six identical machines for industrial control. They use older 9th generation i5-9400 processors on the same industrial boards from As Rock. I'm running Windows 10 32-bit, which is the latest version available to support the control hardware. After the initial build and testing, I cloned and verified each drive so I didn't need to install Windows for every unit (since the hardware is uniform). The first build features a CPU with stepping number 13. Machines with stepping 10 boot normally, but some with stepping 13 fail to start due to a kernel-mode trap BSOD. Examining the dump file, the issue seems linked to a driver loaded for a hardware license dongle—disabling it allowed booting, but without the driver, the control software can't run. The driver is quite outdated, so upgrading isn't an option. Swapping processors works fine; Windows boots normally, which suggests the problem is tied to this specific driver on stepping 13. Have anyone encountered similar issues when swapping CPUs of the same model but different stepping numbers? I'm wondering if Windows doesn't automatically detect the processor during startup, despite being designed for flexibility across machines. I'm trying not to reinstall Windows for these processors since it was difficult initially and goes against the idea of using any unit for backup purposes. Below are some details from proc/cpuinfo on a live Linux USB stick if that helps. Cheers! Working CPU processor: 0 vendor_id : GenuineIntel cpu family : 6 model : 158 model name : Intel® Core™ i5-9400 CPU @ 2.90GHz stepping : 13 microcode : 0xb0 cpu MHz : 799.978 cache size : 9216 KB physical id : 0 siblings : 6 core id : 0 cpu cores : 6 apicid : 0 initial apicid : 0 fpu : yes fpu_exception : yes cpuid level : 22 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp md_clear flush_l1d bugs : spectre_v1 spectre_v2 spec_store_bypass swapgs itlb_multihit srbds mmio_stale_data retbleed eibrs_pbrsb bogomips : 5799.77 clflush size : 64 cache_alignment : 64 address sizes : 39 bits physical, 48 bits virtual power management: Non booting CPU processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 158 model name : Intel® Core™ i5-9400 CPU @ 2.90GHz stepping : 10 microcode : 0xb4 cpu MHz : 800.415 cache size : 9216 KB physical id : 0 siblings : 6 core id : 0 cpu cores : 6 apicid : 0 initial apicid : 0 fpu : yes fpu_exception : yes cpuid level : 22 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb invpcid_single ppb sbbs
I'm checking this on a mobile device, so the display looks different than usual. Your recent bug fixes for non-booting CPUs include the meltdown fix, but the ones that start up have it missing. That’s the main difference I notice, and it might be related to your very old driver.
Thanks @ApolloX75, your insight is helpful. It seems the older driver struggles with the newer CPU because the Microsoft fix for meltdown doesn’t apply to it. The updated CPU works without that requirement. I spent some time researching and found a Microsoft article about this for Windows 7 32-bit when the initial fix was released. Now it appears this is part of Windows 10, so I’m left with using a specific CPU step but hoping a workaround exists first. It’s frustrating that the maker didn’t release an update to address this since they considered the devices legacy. One organization even rewrote their own driver for the 64-bit version to bypass the problem.
A simple fix was found by turning off the meltdown protection in the registry (or using inspectre). This resolved the issue, though it might still be vulnerable to Meltdown attacks. The setup is a control PC located inside a CNC machine with no network access—no AWS or Azure involved. Thanks for your assistance, ApolloX75! This was a key step toward a solution.