On 7/4/14, 12:51 AM, Ansis Atteka wrote:
On Thu, Jul 3, 2014 at 12:57 PM, Ben Pfaff <b...@nicira.com> wrote:
On Sat, Jun 28, 2014 at 02:03:15PM -0700, Ansis Atteka wrote:
From: Ansis <ansisatt...@gmail.com>
Commit 8a9562 ("dpif-netdev: Add DPDK netdev.") reversed sequence
in which set_program_name() and proctitle_init() functions are
called. This introduced a regression where program_name and argv_start
would point to exactly the same memory (previously both of these
pointers were pointing to different memory locations because
proctitle_init() would have beforehand created a copy of argv[0]
for the set_program_name() call).
This regression on my system caused ovs-vswitchd monitoring process to
show up without process name:
... 00:00:00 : monitoring pid 26308 (healthy)
Ps output was lacking process name because following code was
using overlapping memory for source and target buffer:.
proctitle_set(const char *format, ...)
{
...
n = snprintf(argv_start, argv_size, "%s: ", program_name);
Overall C99 and POSIX standards state that behavior is undefined
if source and target buffers overlap.
Signed-off-by: Ansis Atteka <aatt...@nicira.com>
I like this regardless of the bug fix because it makes behavior more
consistent on Windows and Unix (program_name is now malloc'd in both
cases).
I would remove the 'if (program_name)' test around the free call,
since free(NULL) is a no-op anyway.
Acked-by: Ben Pfaff <b...@nicira.com>
Thanks. I removed the "if" statement and pushed this to master and branch-2.3.
It looks like with removing the "if" statement the free() call was moved
out of the "#ifdef _WIN32" section, which causes a warning (and the
build to fail when I have --enable-Werror) on Linux, with both Clang
3.4.2 and GCC 4.8.3.
-Lori
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
