gpu-compute: Fix Python/C++ object hierarchy discrepancies

The GPUCoalescer and the Shader classes have different base classes in
C++ and Python. This causes subtle bugs in SWIG and compilation errors
for PyBind.

Change-Id: I1ddd2a8ea43f083470538ddfea891347b21d14d8
Reviewed-by: Andreas Hansson <andreas.hansson@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2228
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Tony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Pierre-Yves Péneau <pierre-yves.peneau@lirmm.fr>
Reviewed-by: Bradford Beckmann <brad.beckmann@amd.com>

src/gpu-compute/shader.cc

@@ -50,7 +50,7 @@
 #include "mem/ruby/system/RubySystem.hh"
 #include "sim/sim_exit.hh"
 
-Shader::Shader(const Params *p) : SimObject(p),
+Shader::Shader(const Params *p) : ClockedObject(p),
     clock(p->clk_domain->clockPeriod()), cpuThread(nullptr), gpuTc(nullptr),
     cpuPointer(p->cpu_pointer), tickEvent(this), timingSim(p->timing),
     hsail_mode(SIMT), impl_kern_boundary_sync(p->impl_kern_boundary_sync),

src/gpu-compute/shader.hh

@@ -73,7 +73,7 @@ static const int LDS_SIZE = 65536;
 // Class Shader: This describes a single shader instance. Most
 // configurations will only have a single shader.
 
-class Shader : public SimObject
+class Shader : public ClockedObject
 {
   protected:
       // Shader's clock period in terms of number of ticks of curTime,