GWorkspace stuck using CPU - thread event polling

22692 pts/7    Sl+  289:49                  |       \_ GWorkspace
22697 ?        Ss    10:38                  |           \_ 
/System/Tools/fswatcher --daemon --auto
22699 ?        Ss     0:00                  |           \_ 
/System/Tools/ddbd --daemon --auto

so it is GWorkspace itself

top
22692 multix    20   0  296768  59872  12908 S  10.9   0.2 290:03.40 
GWorkspace

but it is in state sleep, how can it consume so much CPU? (10% in a full 
load system with 12CPU is... almost a CPU)

So let me try GDB

gdb -pid 22692


(gdb) bt
#0  0x00007fc175c0fb42 in ?? () from /lib64/libc.so.6
#1  0x00007fc175c0451c in ?? () from /lib64/libc.so.6
#2  0x00007fc175c04561 in ?? () from /lib64/libc.so.6
#3  0x00007fc175c79b0a in poll () from /lib64/libc.so.6
#4  0x00007fc1762bf268 in poll (__fds=<optimized out>, __nfds=<optimized 
out>, __timeout=2)
     at /usr/include/bits/poll2.h:44
#5  -[GSRunLoopCtxt pollUntil:within:] (self=<optimized out>, 
_cmd=0x7fc1764d0180 <_OBJC_SELECTOR_TABLE+1184>,
     milliseconds=2, contexts=0x55d6ef96a480) at GSRunLoopCtxt.m:395
#6  0x00007fc1761e12d4 in -[NSRunLoop acceptInputForMode:beforeDate:] 
(self=0x55d6ef967d90,
     _cmd=0x7fc1764d01b0 <_OBJC_SELECTOR_TABLE+1232>, 
mode=0x7fc1764d0fb0 <_OBJC_INSTANCE_2>,
     limit_date=0x55d6f177c590) at NSRunLoop.m:1259
#7  0x00007fc1761e0f94 in -[NSRunLoop runMode:beforeDate:] 
(self=<optimized out>, _cmd=<optimized out>,
     mode=0x7fc1764d0fb0 <_OBJC_INSTANCE_2>, date=<optimized out>) at 
NSRunLoop.m:1339
#8  0x00007fc1769a4822 in -[GSDisplayServer(EventOps) 
getEventMatchingMask:beforeDate:inMode:dequeue:] (
     self=<optimized out>, _cmd=<optimized out>, mask=<optimized out>, 
limit=<optimized out>,
     mode=<optimized out>, flag=<optimized out>) at GSDisplayServer.m:1037
#9  0x00007fc171e9a20a in -[XGServer(X11Ops) 
getEventMatchingMask:beforeDate:inMode:dequeue:] (
     self=0x55d6ef63fec0, _cmd=<optimized out>, mask=4294967295, 
limit=0x55d6ef97f420,
     mode=0x7fc1764d0fb0 <_OBJC_INSTANCE_2>, flag=1 '\001') at 
XGServerEvent.m:2743
#10 0x00007fc1767dd9c5 in DPSPeekEvent (ctxt=<optimized out>, 
mask=<optimized out>, limit=<optimized out>,
     mode=<optimized out>) at 
../Headers/Additions/GNUstepGUI/GSDisplayServer.h:210
#11 -[NSApplication nextEventMatchingMask:untilDate:inMode:dequeue:] 
(self=0x55d6ef863ba0,
     _cmd=0x7fc176b125b0 <_OBJC_SELECTOR_TABLE+2928>, mask=4294967295, 
expiration=0x55d6ef97f420,
     mode=0x7fc1764d0fb0 <_OBJC_INSTANCE_2>, flag=1 '\001') at 
NSApplication.m:2208
#12 0x00007fc1767df8de in -[NSApplication run] (self=0x55d6ef863ba0, 
_cmd=<optimized out>) at NSApplication.m:1579
#13 0x000055d6c71cd987 in main (argc=<optimized out>, argv=<optimized 
out>, env=<optimized out>) at main.m:38

There is no direct GWorkspace class in this trace actually

which would say it is in poll runloop. A wild guess it is waiting for an 
event from X11. Keyboard or Mouse?
Or instead it is using X for polling a different event.

(gdb) info threads
   Id   Target Id                                      Frame
* 1    Thread 0x7fc171fa4500 (LWP 22692) "GWorkspace" 0x00007fc175c0fb42 
in ?? () from /lib64/libc.so.6
   2    Thread 0x7fc1700166c0 (LWP 22723) "GWorkspace" 
0x00007fc175c0fb42 in ?? () from /lib64/libc.so.6


shows me there are two threads.

If I switch to the second thread, I see:
[Switching to thread 2 (Thread 0x7fc1700166c0 (LWP 22723))]
#0  0x00007fc175c0fb42 in ?? () from /lib64/libc.so.6
(gdb) bt
#0  0x00007fc175c0fb42 in ?? () from /lib64/libc.so.6
#1  0x00007fc175c0451c in ?? () from /lib64/libc.so.6
#2  0x00007fc175c04561 in ?? () from /lib64/libc.so.6
#3  0x00007fc175c79b0a in poll () from /lib64/libc.so.6
#4  0x00007fc1762bf268 in poll (__fds=<optimized out>, __nfds=<optimized 
out>, __timeout=2147483647)
     at /usr/include/bits/poll2.h:44
#5  -[GSRunLoopCtxt pollUntil:within:] (self=<optimized out>, 
_cmd=0x7fc1764d0180 <_OBJC_SELECTOR_TABLE+1184>,
     milliseconds=2147483647, contexts=0x7fc168001800) at 
GSRunLoopCtxt.m:395
#6  0x00007fc1761e12d4 in -[NSRunLoop acceptInputForMode:beforeDate:] 
(self=0x7fc1680017b0,
     _cmd=0x7fc1764d01b0 <_OBJC_SELECTOR_TABLE+1232>, 
mode=0x7fc1764d0fb0 <_OBJC_INSTANCE_2>,
     limit_date=0x55d6ef97f420) at NSRunLoop.m:1259
#7  0x00007fc1761e0f94 in -[NSRunLoop runMode:beforeDate:] 
(self=<optimized out>, _cmd=<optimized out>,
     mode=0x7fc1764d0fb0 <_OBJC_INSTANCE_2>, date=<optimized out>) at 
NSRunLoop.m:1339
#8  0x00007fc1761df2c9 in -[NSRunLoop runUntilDate:] 
(self=0x7fc1680017b0, _cmd=<optimized out>,
     date=0x55d6ef97f420) at NSRunLoop.m:1369
#9  0x00007fc1703d9948 in +[ImageResizer connectWithPorts:] 
(self=0x7fc1703e0d80 <_OBJC_Class_ImageResizer>,
     _cmd=<optimized out>, portArray=<optimized out>) at Resizer.m:54
#10 0x00007fc176216d3e in nsthreadLauncher (thread=0x55d6f0613b30) at 
NSThread.m:1490
#11 0x00007fc175c07c13 in ?? () from /lib64/libc.so.6
#12 0x00007fc175c86f6c in ?? () from /lib64/libc.so.6


which shows too polling, again an event, in this case it is the 
ImageResizer thread.

49      serverObject = [[self alloc] init];
50      if (serverObject)
51        {
52          [(id)[serverConnection rootProxy] setResizer:serverObject];
53          [serverObject release];
54          [[NSRunLoop currentRunLoop] run];
55        }
56      [pool release];
57      [NSThread exit];


Is inter-thread DistributedObject stuck? something starved? I don't know...

Interesting is that is usually works and gets just sometimess suck, 
pointing out to some peculiar condition.

Riccardo