summary refs log tree commit diff stats
path: root/results/classifier/gemma3:12b/hypervisor/1855617
diff options
context:
space:
mode:
authorChristian Krinitsin <mail@krinitsin.com>2025-07-03 07:27:52 +0000
committerChristian Krinitsin <mail@krinitsin.com>2025-07-03 07:27:52 +0000
commitd0c85e36e4de67af628d54e9ab577cc3fad7796a (patch)
treef8f784b0f04343b90516a338d6df81df3a85dfa2 /results/classifier/gemma3:12b/hypervisor/1855617
parent7f4364274750eb8cb39a3e7493132fca1c01232e (diff)
downloademulator-bug-study-d0c85e36e4de67af628d54e9ab577cc3fad7796a.tar.gz
emulator-bug-study-d0c85e36e4de67af628d54e9ab577cc3fad7796a.zip
add deepseek and gemma results
Diffstat (limited to 'results/classifier/gemma3:12b/hypervisor/1855617')
-rw-r--r--results/classifier/gemma3:12b/hypervisor/18556175
1 files changed, 5 insertions, 0 deletions
diff --git a/results/classifier/gemma3:12b/hypervisor/1855617 b/results/classifier/gemma3:12b/hypervisor/1855617
new file mode 100644
index 00000000..97f91b8f
--- /dev/null
+++ b/results/classifier/gemma3:12b/hypervisor/1855617
@@ -0,0 +1,5 @@
+
+savevm with hax saves wrong register state
+
+I use qemu-i386 with IntelHaxm on Windows 10 x64 host with Windows 7 x86 guest. I run the guest till OS loads and create a snapshot with savevm, then close qemu, run it again and try to load the snapshot with loadvm. The guest crashes or freezes. I dumped registers on snapshot creation and loading (in Haxm) and found that they are different.
+When returning from Haxm in hax_vcpu_hax_exec, there is no regular register read. I found hax_arch_get_registers function which reads registers from Haxm and is called from a synchronization procedure. I placed a breakpoint on it, ran qemu and found that it is hit one time during guest OS boot. Exactly these registers where saved in the snapshot.
\ No newline at end of file