summary refs log tree commit diff stats
path: root/results/classifier/zero-shot/108/other/1074
diff options
context:
space:
mode:
Diffstat (limited to 'results/classifier/zero-shot/108/other/1074')
-rw-r--r--results/classifier/zero-shot/108/other/107433
1 files changed, 33 insertions, 0 deletions
diff --git a/results/classifier/zero-shot/108/other/1074 b/results/classifier/zero-shot/108/other/1074
new file mode 100644
index 000000000..d48c510f3
--- /dev/null
+++ b/results/classifier/zero-shot/108/other/1074
@@ -0,0 +1,33 @@
+debug: 0.734
+boot: 0.701
+files: 0.683
+graphic: 0.671
+device: 0.660
+semantic: 0.543
+other: 0.525
+PID: 0.417
+vnc: 0.376
+network: 0.347
+socket: 0.336
+performance: 0.254
+permissions: 0.240
+KVM: 0.171
+
+File under symlink gets corrupted when directory is mounted as FAT32 drive
+Description of problem:
+When mouting a directory as a FAT32 drive, the symlinked BOOTx64.EFI inside gets corrupted after booting it.
+Steps to reproduce:
+1. mkdir -p fat_dir/EFI/BOOT/
+2. ln -s BOOTx64.EFI fat_dir/EFI/BOOT/BOOTx64.EFI
+3. md5sum BOOTx64.EFI
+4. Run qemu with arguments like above.
+5. md5sum BOOTx64.EFI should print out different hash, confirming corruption.
+Additional information:
+[BOOTx64.EFI](/uploads/d0a6e899ec9331461179f8dc82fbc421/BOOTx64.EFI)
+
+The issue was not visible on earlier versions, but I don't know which one exactly was it.\
+I can only say, it was still working in April and it was possible that I was using Fedora 36 Beta.
+
+Copying the file instead of using a symlink can be used as a workaround.
+
+The binary should print some debug stuff, like avaliable memory regions and end with an infinite halt-loop.