[Bug 750397] CRITICAL: Race condition in GstBus

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=750397

Vijay Jayaraman <vijay.jayaraman at echostar.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |vijay.jayaraman at echostar.co
                   |                            |m

--- Comment #1 from Vijay Jayaraman <vijay.jayaraman at echostar.com> ---
Hi,

I noticed a similar issue today where my application CPU usage shot up from
typically being 15% to 50-60%. On debugging further it looks like the following
debug shows up all the time when the CPU usage shoots up. 

0:06:59.338361149 21782   0x4b0600 INFO                 GST_BUS
gstbus.c:549:gst_bus_timed_pop_filtered:<bus1> we got woken up, recheck for
message
0:06:59.338832148 21782   0x4b0600 INFO                 GST_BUS
gstbus.c:549:gst_bus_timed_pop_filtered:<bus1> we got woken up, recheck for
message
0:06:59.339566815 21782   0x4b0600 INFO                 GST_BUS
gstbus.c:549:gst_bus_timed_pop_filtered:<bus1> we got woken up, recheck for
message
0:06:59.340383149 21782   0x4b0600 INFO                 GST_BUS
gstbus.c:549:gst_bus_timed_pop_filtered:<bus1> we got woken up, recheck for
message
0:06:59.340860000 21782   0x4b0600 INFO                 GST_BUS
gstbus.c:549:gst_bus_timed_pop_filtered:<bus1> we got woken up, recheck for
message

AS you can see, gst_bus_timed_pop_filtered wakes up almost every millisecond
and this causes the CPU usage to shoot up. When this happened, I tried pausing
my player and still this message fills up the log.

I tried patching gstreamer with the proposed patch and I had to use it in the
GST_BUS_ASYNC case as well and the issue seems to have disappeared.

Note: The original issue where the CPU shot up was not 100% reproducible but it
happened most of the time if I let the pipeline play long enough.

I am using gstreamer 1.4.0. Anybody else seeing this issue?

Thanks,
-Vijay

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.