This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[patch] Fix linux-nat.c {,lp->}status typo
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: gdb-patches at sourceware dot org
- Date: Sun, 18 Jul 2010 16:51:12 +0200
- Subject: [patch] Fix linux-nat.c {,lp->}status typo
Hi,
while reading the code I found this suspicious code, IMO a typo.
if (WIFSTOPPED (lp->status))
{
if (WSTOPSIG (lp->status) != SIGSTOP)
[...]
}
else if (WIFEXITED (status) || WIFSIGNALED (status))
[...]
STATUS value is not filtered by linux_nat_filter_event. LP->STATUS is.
That "status = 0;" would be better as some __deinitiailize_var (status) but
there is no such GCC extension.
The DEBUG_LINUX_NAT print could be possibly still STATUS, it depends on the
opinion IMO.
No regressions on {x86_64,x86_64-m32,i686}-fedora12-linux-gnu.
(I haven't tried to find an impact of this bug.)
Thanks,
Jan
2010-07-18 Jan Kratochvil <jan.kratochvil@redhat.com>
* linux-nat.c (linux_nat_wait_1): Reset STATUS after calling
linux_nat_wait_1. Use always LP->STATUS afterwards.
--- a/gdb/linux-nat.c
+++ b/gdb/linux-nat.c
@@ -3365,6 +3365,9 @@ retry:
lp = linux_nat_filter_event (lwpid, status, options);
+ /* STATUS is now no longer valid, use LP->STATUS instead. */
+ status = 0;
+
if (lp
&& ptid_is_pid (ptid)
&& ptid_get_pid (lp->ptid) != ptid_get_pid (ptid))
@@ -3373,7 +3376,7 @@ retry:
if (debug_linux_nat)
fprintf (stderr, "LWP %ld got an event %06x, leaving pending.\n",
- ptid_get_lwp (lp->ptid), status);
+ ptid_get_lwp (lp->ptid), lp->status);
if (WIFSTOPPED (lp->status))
{
@@ -3410,7 +3413,7 @@ retry:
lp->signalled = 0;
}
}
- else if (WIFEXITED (status) || WIFSIGNALED (status))
+ else if (WIFEXITED (lp->status) || WIFSIGNALED (lp->status))
{
if (debug_linux_nat)
fprintf (stderr, "Process %ld exited while stopping LWPs\n",