From d0c85e36e4de67af628d54e9ab577cc3fad7796a Mon Sep 17 00:00:00 2001 From: Christian Krinitsin Date: Thu, 3 Jul 2025 07:27:52 +0000 Subject: add deepseek and gemma results --- results/classifier/deepseek-2/reasoning/hypervisor/1844946 | 11 +++++++++++ 1 file changed, 11 insertions(+) create mode 100644 results/classifier/deepseek-2/reasoning/hypervisor/1844946 (limited to 'results/classifier/deepseek-2/reasoning/hypervisor/1844946') diff --git a/results/classifier/deepseek-2/reasoning/hypervisor/1844946 b/results/classifier/deepseek-2/reasoning/hypervisor/1844946 new file mode 100644 index 00000000..ae3b9221 --- /dev/null +++ b/results/classifier/deepseek-2/reasoning/hypervisor/1844946 @@ -0,0 +1,11 @@ + +Alright, let me try to figure out how to classify this bug report. The user mentioned they're using macOS with QEMU and the HVF accelerator. They set up a simple configuration but when they enable HVF, their WinXP VM hangs during boot. Without HVF (using TCG), it works fine. + +Looking at their analysis, they pointed out two commits: b7394c8394 which replaced explicit VMCS checks with hflags/hflags2, and 92d5f1a414 which changed the behavior of these flags. They believe that this second commit broke HVF's interrupt handling, causing WinXP to fail. + +They also noted that KVM and HAXM weren't affected because they still do explicit checks, but HVF relies on the new approach which is now flawed. + +So, the issue seems related to how QEMU handles virtualization flags when using HVF. Since it's about interrupt handling and hflags, which are part of the hypervisor's management, I think this falls under the 'hypervisor' category. + + +hypervisor \ No newline at end of file -- cgit v1.2.3