fix some warning from compiler

[Shouren]

Problem

There are some warnings when building hami-core with gcc 11.4.0 in a container based on nvidia/cuda:12.3.2-devel-ubuntu22.04

root@d7a2fa755ecd:/libvgpu# rm -rf build && bash build.sh
...
/libvgpu/src/allocator/allocator.c: In function 'free_raw_async':
/libvgpu/src/allocator/allocator.c:201:24: warning: implicit declaration of function 'remove_chunk_async'; did you mean 'remove_chunk'? [-Wimplicit-function-declaration]
  201 |     unsigned int tmp = remove_chunk_async(device_overallocated,dptr,hStream);
      |                        ^~~~~~~~~~~~~~~~~~
      |                        remove_chunk
/libvgpu/src/allocator/allocator.c: In function 'allocate_async_raw':
/libvgpu/src/allocator/allocator.c:230:11: warning: implicit declaration of function 'add_chunk_async'; did you mean 'add_chunk_only'? [-Wimplicit-function-declaration]
  230 |     tmp = add_chunk_async(dptr,size,hStream);
      |           ^~~~~~~~~~~~~~~
      |           add_chunk_only
In file included from /libvgpu/src/multiprocess/multiprocess_memory_limit.h:22,
                 from /libvgpu/src/include/memory_limit.h:71,
                 from /libvgpu/src/multiprocess/multiprocess_memory_limit.c:21:
/libvgpu/src/multiprocess/multiprocess_memory_limit.c: In function 'set_gpu_device_sm_utilization':
/libvgpu/src/include/log_utils.h:21:24: warning: format '%lu' expects argument of type 'long unsigned int', but argument 10 has type 'unsigned int' [-Wformat=]
   21 |        fprintf(stderr, "[HAMI-core Info(%d:%ld:%s:%d)]: "msg"\n", getpid(),pthread_self(),basename(__FILE__),__LINE__,##__VA_ARGS__); }
      |                        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/libvgpu/src/include/log_utils.h:21:24: note: in definition of macro 'LOG_INFO'
   21 |        fprintf(stderr, "[HAMI-core Info(%d:%ld:%s:%d)]: "msg"\n", getpid(),pthread_self(),basename(__FILE__),__LINE__,##__VA_ARGS__); }
      |                        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

......

[ 29%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/memory.c.o
/libvgpu/src/nvml/hook.c: In function 'nvmlDeviceGetMemoryInfo_v2':
/libvgpu/src/nvml/hook.c:371:44: warning: passing argument 2 of '_nvmlDeviceGetMemoryInfo' from incompatible pointer type [-Wincompatible-pointer-types]
  371 |     return _nvmlDeviceGetMemoryInfo(device,memory,2);
      |                                            ^~~~~~
      |                                            |
      |                                            nvmlMemory_v2_t * {aka struct nvmlMemory_v2_st *}
/libvgpu/src/nvml/hook.c:317:73: note: expected 'nvmlMemory_t *' {aka 'struct nvmlMemory_st *'} but argument is of type 'nvmlMemory_v2_t *' {aka 'struct nvmlMemory_v2_st *'}
  317 | nvmlReturn_t _nvmlDeviceGetMemoryInfo(nvmlDevice_t device,nvmlMemory_t* memory,int version) {
      |                                                           ~~~~~~~~~~~~~~^~~~~~

......

[ 91%] Linking C executable test_runtime_host_register
/libvgpu/src/libvgpu.c: In function '__dlsym_hook_section':
/libvgpu/src/libvgpu.c:243:5: warning: 'cuModuleGetTexRef' is deprecated [-Wdeprecated-declarations]
  243 |     DLSYM_HOOK_FUNC(cuModuleGetTexRef);
      |     ^~~~~~~~~~~~~~~
In file included from /libvgpu/src/include/cuda_addition_func.h:4,
                 from /libvgpu/src/libvgpu.c:7:
/usr/local/cuda/include/cuda.h:7256:36: note: declared here
 7256 | __CUDA_DEPRECATED CUresult CUDAAPI cuModuleGetTexRef(CUtexref *pTexRef, CUmodule hmod, const char *name);
      |                                    ^~~~~~~~~~~~~~~~~
/libvgpu/src/libvgpu.c:244:5: warning: 'cuModuleGetSurfRef' is deprecated [-Wdeprecated-declarations]
  244 |     DLSYM_HOOK_FUNC(cuModuleGetSurfRef);
      |     ^~~~~~~~~~~~~~~
In file included from /libvgpu/src/include/cuda_addition_func.h:4,
                 from /libvgpu/src/libvgpu.c:7:
/usr/local/cuda/include/cuda.h:7290:36: note: declared here
 7290 | __CUDA_DEPRECATED CUresult CUDAAPI cuModuleGetSurfRef(CUsurfref *pSurfRef, CUmodule hmod, const char *name);
      |                                    ^~~~~~~~~~~~~~~~~~
[ 93%] Linking C executable ../../shrreg-tool
[ 93%] Built target test_runtime_host_register
[ 95%] Building C object src/CMakeFiles/vgpu.dir/utils.c.o
[ 95%] Built target shrreg-tool
[ 97%] Linking C shared library ../libvgpu.so
[ 97%] Built target vgpu
[100%] Linking CXX executable test_runtime_launch
[100%] Built target test_runtime_launch

Solution

Fix some code and added a cmake option(HOOK_CUDA_DEPRECATED_ENABLE) determine whether a deprecated cuda function should be hooked. For backward compatible, default value of HOOK_CUDA_DEPRECATED_ENABLE is set 1.

This PR applied and HOOK_CUDA_DEPRECATED_ENABLE set to 0

oot@d7a2fa755ecd:/libvgpu# rm -rf build && bash build.sh
CI_COMMIT_BRANCH:unknown
CI_COMMIT_SHA:af47a4c67b4642cfe1d2f6e20c151793caeefcd6
CMake Deprecation Warning at CMakeLists.txt:1 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


-- The C compiler identification is GNU 11.4.0
-- The CXX compiler identification is GNU 11.4.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Git hash is af47a4c
CMake Deprecation Warning at src/CMakeLists.txt:1 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Found CUDA: /usr/local/cuda (found version "12.3")
-- Configuring done
-- Generating done
-- Build files have been written to: /libvgpu/build
[  6%] Building C object src/multiprocess/CMakeFiles/multiprocess_mod.dir/multiprocess_memory_limit.c.o
[  6%] Building C object src/allocator/CMakeFiles/allocator_mod.dir/allocator.c.o
[  8%] Building C object src/nvml/CMakeFiles/nvml_mod.dir/nvml_entry.c.o
[  8%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/context.c.o
[  8%] Built target allocator_mod
[ 10%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/device.c.o
[ 12%] Building C object test/CMakeFiles/test_alloc.dir/test_alloc.c.o
[ 14%] Building C object src/multiprocess/CMakeFiles/multiprocess_mod.dir/multiprocess_utilization_watcher.c.o
[ 17%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/hook.c.o
[ 19%] Linking C executable test_alloc
[ 19%] Built target test_alloc
[ 19%] Built target multiprocess_mod
[ 21%] Building C object test/CMakeFiles/test_alloc_hold.dir/test_alloc_hold.c.o
[ 23%] Building C object test/CMakeFiles/test_alloc_host.dir/test_alloc_host.c.o
[ 25%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/event.c.o
[ 27%] Building C object src/nvml/CMakeFiles/nvml_mod.dir/hook.c.o
[ 29%] Linking C executable test_alloc_host
[ 31%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/memory.c.o
[ 34%] Linking C executable test_alloc_hold
[ 34%] Built target test_alloc_host
[ 34%] Built target test_alloc_hold
[ 34%] Built target nvml_mod
[ 36%] Building C object test/CMakeFiles/test_alloc_managed.dir/test_alloc_managed.c.o
[ 38%] Building C object test/CMakeFiles/test_alloc_pitch.dir/test_alloc_pitch.c.o
[ 40%] Building C object test/CMakeFiles/test_create_3d_array.dir/test_create_3d_array.c.o
[ 42%] Linking C executable test_alloc_managed
[ 46%] Linking C executable test_alloc_pitch
[ 46%] Linking C executable test_create_3d_array
[ 46%] Built target test_alloc_managed
[ 46%] Built target test_create_3d_array
[ 46%] Built target test_alloc_pitch
[ 48%] Building C object test/CMakeFiles/test_create_array.dir/test_create_array.c.o
[ 53%] Building C object test/CMakeFiles/test_host_register.dir/test_host_register.c.o
[ 53%] Building C object test/CMakeFiles/test_host_alloc.dir/test_host_alloc.c.o
[ 55%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/stream.c.o
[ 57%] Linking C executable test_create_array
[ 59%] Building C object src/cuda/CMakeFiles/cuda_mod.dir/graph.c.o
[ 61%] Linking C executable test_host_alloc
[ 63%] Linking C executable test_host_register
[ 63%] Built target test_create_array
[ 63%] Built target test_host_alloc
[ 63%] Built target test_host_register
[ 65%] Building C object test/CMakeFiles/test_runtime_alloc.dir/test_runtime_alloc.c.o
[ 68%] Building C object test/CMakeFiles/test_runtime_alloc_host.dir/test_runtime_alloc_host.c.o
[ 70%] Building C object test/CMakeFiles/test_runtime_alloc_managed.dir/test_runtime_alloc_managed.c.o
[ 70%] Built target cuda_mod
[ 72%] Linking C executable test_runtime_alloc
[ 74%] Building C object test/CMakeFiles/test_runtime_host_alloc.dir/test_runtime_host_alloc.c.o
[ 76%] Linking C executable test_runtime_alloc_host
[ 78%] Linking C executable test_runtime_alloc_managed
[ 78%] Built target test_runtime_alloc
[ 78%] Built target test_runtime_alloc_host
[ 78%] Built target test_runtime_alloc_managed
[ 80%] Building NVCC (Device) object test/CMakeFiles/test_runtime_launch.dir/test_runtime_launch_generated_test_runtime_launch.cu.o
[ 82%] Building C object test/CMakeFiles/test_runtime_host_register.dir/test_runtime_host_register.c.o
[ 82%] Built target python_test
[ 85%] Building C object src/CMakeFiles/vgpu.dir/libvgpu.c.o
[ 87%] Linking C executable test_runtime_host_alloc
[ 89%] Linking C executable test_runtime_host_register
[ 89%] Built target test_runtime_host_alloc
[ 91%] Building C object src/multiprocess/CMakeFiles/shrreg-tool.dir/shrreg_tool.c.o
[ 91%] Built target test_runtime_host_register
[ 93%] Building C object src/CMakeFiles/vgpu.dir/utils.c.o
[ 95%] Linking C executable ../../shrreg-tool
[ 95%] Built target shrreg-tool
[ 97%] Linking C shared library ../libvgpu.so
[ 97%] Built target vgpu
[100%] Linking CXX executable test_runtime_launch
[100%] Built target test_runtime_launch