add 18th iteration of classifier

1 files changed, 64 insertions, 0 deletions

diff --git a/results/classifier/118/hypervisor/1756927 b/results/classifier/118/hypervisor/1756927
new file mode 100644
index 000000000..d96f24c84
--- /dev/null
+++ b/results/classifier/118/hypervisor/1756927
@@ -0,0 +1,64 @@
+hypervisor: 0.849
+architecture: 0.844
+debug: 0.777
+arm: 0.772
+performance: 0.757
+device: 0.753
+x86: 0.713
+kernel: 0.688
+boot: 0.666
+mistranslation: 0.622
+ppc: 0.614
+risc-v: 0.562
+semantic: 0.554
+graphic: 0.551
+permissions: 0.546
+network: 0.531
+user-level: 0.525
+vnc: 0.523
+socket: 0.491
+VMM: 0.485
+PID: 0.475
+TCG: 0.474
+i386: 0.439
+register: 0.411
+assembly: 0.407
+files: 0.406
+KVM: 0.400
+virtual: 0.345
+peripherals: 0.244
+
+ARMv7 LPAE: IFSR doesn't have the LPAE bit in case of BKPT
+
+When a user application triggers a 'bkpt' instruction while LPAE is used, the bit [9] of IFSR is not correctly set during the prefetch abort exception.
+
+You'll find attached a minimal example to reproduce the issue (just run 'make all').
+The output I get is:
+
+supervisor
+user
+prefetch
+short-descriptor
+
+The last entry should read 'long-descriptor'.
+
+
+Qemu revision: 48ae1f60d8c9a770e6da64407984d84e25253c69
+Ubuntu verison: 16.04 LTS
+Cross Compiler: gcc linaro 6.3.1-2017.02-x86_64_arm-eabi
+
+
+
+I've just sent this patchset:
+http://<email address hidden>/
+which should fix this bug and a couple of others that I noticed with our debug exception handling while I was doing that.
+
+
+thanks Peter ! Any news on the review ?
+
+The patches are in master now.
+
+
+Hi Peter,
+we tested the fix and it work correctly now, thank you very much !
+