Re: [RFA] linux-nat.c, linux_thread_alive, user errno instead of err

[Michael Snyder <msnyder@vmware.com>]

[-- Attachment #1: Type: text/plain, Size: 504 bytes --]

Pedro Alves wrote:
>>> --- linux-nat.c	1 Mar 2011 00:40:22 -0000	1.197
>>> +++ linux-nat.c	1 Mar 2011 23:50:34 -0000
>>> @@ -4045,7 +4045,7 @@ linux_thread_alive (ptid_t ptid)
>>>      fprintf_unfiltered (gdb_stdlog,
>>>  			"LLTA: KILL(SIG0) %s (%s)\n",
>>>  			target_pid_to_str (ptid),
>>> -			err ? safe_strerror (err) : "OK");
>>> +			err ? safe_strerror (errno) : "OK");
> 
> Not 100% correct.  target_pid_to_str is allowed to
> clobber errno (as any function call is).

OK.  How about this then?


[-- Attachment #2: errno2.txt --]
[-- Type: text/plain, Size: 908 bytes --]

Index: linux-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/linux-nat.c,v
retrieving revision 1.197
diff -u -p -u -p -r1.197 linux-nat.c
--- linux-nat.c	1 Mar 2011 00:40:22 -0000	1.197
+++ linux-nat.c	2 Mar 2011 18:00:39 -0000
@@ -4032,7 +4032,7 @@ linux_nat_xfer_partial (struct target_op
 static int
 linux_thread_alive (ptid_t ptid)
 {
-  int err;
+  int err, tmp_errno;
 
   gdb_assert (is_lwp (ptid));
 
@@ -4040,12 +4040,12 @@ linux_thread_alive (ptid_t ptid)
      running thread errors out claiming that the thread doesn't
      exist.  */
   err = kill_lwp (GET_LWP (ptid), 0);
-
+  tmp_errno = errno;
   if (debug_linux_nat)
     fprintf_unfiltered (gdb_stdlog,
 			"LLTA: KILL(SIG0) %s (%s)\n",
 			target_pid_to_str (ptid),
-			err ? safe_strerror (err) : "OK");
+			err ? safe_strerror (tmp_errno) : "OK");
 
   if (err != 0)
     return 0;