summary refs log tree commit diff stats
path: root/results/classifier/zero-shot/108/other/1773743
diff options
context:
space:
mode:
Diffstat (limited to 'results/classifier/zero-shot/108/other/1773743')
-rw-r--r--results/classifier/zero-shot/108/other/177374345
1 files changed, 45 insertions, 0 deletions
diff --git a/results/classifier/zero-shot/108/other/1773743 b/results/classifier/zero-shot/108/other/1773743
new file mode 100644
index 000000000..5a06f8e04
--- /dev/null
+++ b/results/classifier/zero-shot/108/other/1773743
@@ -0,0 +1,45 @@
+performance: 0.750
+device: 0.740
+semantic: 0.739
+graphic: 0.724
+other: 0.716
+vnc: 0.657
+debug: 0.634
+network: 0.520
+permissions: 0.507
+socket: 0.506
+PID: 0.476
+files: 0.422
+boot: 0.277
+KVM: 0.215
+
+qemu-user -g xxx -E LD_PROFILE=xxx segfault
+
+Here is two simple steps to reproduce the bug:
+
+$ qemu-x86_64 -E LD_PROFILE=libc.so.6 -E LD_PROFILE_OUTPUT=. -g 12345 -L / /bin/ls
+
+(libc.so and /bin/ls might change on your system, in this case we just need a binary with a profilable needed library)
+
+In a other window launch:
+
+$ gdb
+(gdb) target remote :12345
+(gdb) c
+
+At this point qemu will segfault.
+
+It seems this problem is appends when sigprof passed to gdb.
+One way I have found to bypass this:
+patch gdbstub.c gdb_handlesig and ignore sig if
+sig == TARGET_SIGPROF
+(which means now I can't catch sigprof on gdb anymore)
+
+We were mis-parsing the 'vCont' packet in the gdb protocol, so when gdb told us "send a SIGPROF (0x1b)" we skipped the first digit in the hex signal number and interpreted it as "send an 0xb", which happens to be SIGSEGV.
+
+Should be fixed by this:
+https://<email address hidden>/
+
+
+https://gitlab.com/qemu-project/qemu/-/commit/3ddd9036389f5f577e09
+