Fix chained hardware exception handling on Unix (#12316)

There is an issue when hardware exception occurs while handling another hardware exception. In such case,
the exception unwinding ends up in an infinite loop. It is caused by the kernel reusing the same location
for signal handler context.

The fix is to use the windows style context local variable in the common_signal_handler that contains
the right context - it is the original signal context converted to windows style context.

1 files changed, 2 insertions, 2 deletions

diff --git a/src/pal/src/exception/signal.cpp b/src/pal/src/exception/signal.cpp
index b82daca53d..6a046bd00b 100644
--- a/src/pal/src/exception/signal.cpp
+++ b/src/pal/src/exception/signal.cpp
@@ -119,7 +119,7 @@ struct sigaction g_previous_sigquit;
 struct sigaction g_previous_activation;
 #endif
 
-// Offset of the local variable containing native context in the common_signal_handler function.
+// Offset of the local variable containing pointer to windows style context in the common_signal_handler function.
 // This offset is relative to the frame pointer.
 int g_common_signal_handler_context_locvar_offset = 0;
 #endif // !HAVE_MACH_EXCEPTIONS
@@ -766,7 +766,7 @@ static bool common_signal_handler(int code, siginfo_t *siginfo, void *sigcontext
     native_context_t *ucontext;
 
     ucontext = (native_context_t *)sigcontext;
-    g_common_signal_handler_context_locvar_offset = (int)((char*)&ucontext - (char*)__builtin_frame_address(0));
+    g_common_signal_handler_context_locvar_offset = (int)((char*)&contextRecord - (char*)__builtin_frame_address(0));
 
     AllocateExceptionRecords(&exceptionRecord, &contextRecord);