OpenE2K/binutils-gdb
12
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

* i386-nlmstub.c (handle_exception): Wait until the thread has

Browse Source 
been started before killing the NLM by pointing the PC at
  _exit().
This commit is contained in:
J.T. Conklin 1994-08-08 22:14:06 +00:00
parent 9623e82883
commit ac2068375f
1 changed files with 18 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
gdb/i386-nlmstub.c
View File

@ -627,7 +627,8 @@ handle_exception (frame)
{
int addr, length;
char *ptr;
static int killed = 0;
static int thread_killed = 0;
static int thread_started = 0;
static struct DBG_LoadDefinitionStructure *ldinfo = 0;
static unsigned char first_insn[BREAKPOINT_SIZE]; /* The first instruction in the program. */
@ -661,8 +662,12 @@ handle_exception (frame)
flush_i_cache ();
return RETURN_TO_PROGRAM;
case START_THREAD_EVENT:
thread_started = 1;
return RETURN_TO_PROGRAM;
case TERMINATE_NLM_EVENT:
if (!killed)
if (!thread_killed)
{
/* NetWare processes don't have an exit status so we
generate our own */
@ -731,6 +736,14 @@ handle_exception (frame)
break;
}
/* We point the PC at _exit() and continue to kill the NLM, but that
won't work until it's thread has been started. */
if (thread_started && thread_killed)
{
frame->ExceptionPC = &_exit;
return RETURN_TO_PROGRAM;
}
/* FIXME: How do we know that this exception has anything to do with
the program we are debugging? We can check whether the PC is in
the range of the module we are debugging, but that doesn't help
@ -850,8 +863,9 @@ handle_exception (frame)
at the start of _exit() and continue, while noting that
we've killed the process. */
killed = 1;
frame->ExceptionPC = &_exit;
thread_killed = 1;
if (thread_started)
frame->ExceptionPC = &_exit;
return RETURN_TO_PROGRAM;
case 'q': /* Query message */