Hyperthreading bug found in Intel Skylake and Kaby Lake chips could cause system crashes

A flaw has been found in Intel Skylake and Kaby Lake chips that could cause unpredictable system behavior under certain conditions when hyperthreading is enabled.

All Skylake and Kaby Lake processors, including embedded, mobile, HEDT, and related server chips, seem to be affected, with the exception of the new Kaby Lake-X chips (Skylake-X CPUs have the flaw) which have been listed by Intel as being fixed.  

The issue was singled out recently by Henrique de Moraes Holschuh, a Debian Linux developer, who wrote the following advisory:

“This advisory is about a processor/microcode defect recently identified on Intel Skylake and Intel Kaby Lake processors with hyper-threading enabled. This defect can, when triggered, cause unpredictable system behavior: it could cause spurious errors, such as application and system misbehavior, data corruption, and data loss.”

Intel is aware of the problem, and its errata list for its Skylake-X processors describes the problem as below:

"Under complex micro-architectural conditions, short loops of less than 64 instructions that use AH, BH, CH or DH registers as well as their corresponding wider register (eg RAX, EAX or AX for AH) may cause unpredictable system behaviour. This can only happen when both logical processors on the same physical processor are active."

While Holschuh is a Debian Linux developer, the problem afflicts all operating systems, including Windows.

Updated microcode drivers are currently available for Linux to fix the issue, but Windows users will have to rely on microcode updates through the motherboard manufacturer’s BIOS.

It’s not yet clear how widespread this use is. Certain applications may have specific code patterns that trigger the bug, but there isn’t yet a list of the software to avoid.

Still, it’s probably not that common, and Intel’s description of what triggers the bug is awfully specific.

Furthermore, the sixth-generation Skylake processors have been available for nearly two years without anyone complaining. Mark Shinwell, an OCaml toolchain developer, first discovered the flaw earlier this year, and Intel did issue microcode updates in response.

But short of a firmware update on the part of board makers, disabling hyperthreading is believed to be the best solution, especially if you’re dealing with sensitive data that cannot afford to be corrupted or lost. That said, most folks probably shouldn't worry too much and can carry on as usual.

Source: Ars Technica

Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.