1 files changed, 43 insertions, 0 deletions
diff --git a/results/classifier/118/peripherals/752 b/results/classifier/118/peripherals/752
new file mode 100644
index 000000000..be421d4e8
--- /dev/null
+++ b/results/classifier/118/peripherals/752
@@ -0,0 +1,43 @@
+peripherals: 0.931
+device: 0.930
+graphic: 0.821
+VMM: 0.717
+debug: 0.664
+performance: 0.643
+semantic: 0.638
+arm: 0.538
+architecture: 0.522
+vnc: 0.503
+register: 0.493
+PID: 0.465
+virtual: 0.412
+risc-v: 0.378
+permissions: 0.370
+socket: 0.368
+boot: 0.346
+kernel: 0.263
+mistranslation: 0.261
+network: 0.248
+ppc: 0.229
+hypervisor: 0.228
+TCG: 0.183
+user-level: 0.156
+x86: 0.136
+assembly: 0.113
+i386: 0.108
+files: 0.053
+KVM: 0.032
+
+vmmouse device gets attached twice, one without i8042 associated
+Description of problem:
+I'm developing [a driver for the VMware mouse device](https://github.com/NattyNarwhal/vmwmouse). I know this works properly on VMware, but I'm trying it in QEMU too.
+
+[My full notes](https://github.com/NattyNarwhal/vmwmouse/issues/1), but most relevant is:
+
+* a vmmouse instance gets initialized twice (confirmed in qtree), one with i8042 the first time, one without the second time
+* the second vmmouse instance is the one receiving the events, passing them to the i8042 device's fake event handler
+* obviously, a crash because ISAKBDDevice should never be null
+Steps to reproduce:
+1. Load VMware mouse driver
+2. Move cursor (I recommend waiting until Windows loads before doing so, it is very easy to corrupt the guest filesystem if you do it while Windows is loading)
+3. Crash