Add flags to generate debug symbols (#27)

* Add flags to generate debug symbols
* Merge main to akharche/enable_kernel_debug
* Add documentation about how to use gdb

Co-authored-by: Sergey Pokhodenko <sergey.pokhodenko@intel.com>

Author: akharche

DEBUGGING.md

@@ -0,0 +1,21 @@
+## Debugging with GDB
+
+Setting the debug environment variable `NUMBA_DPPY_DEBUG` (e.g. `export NUMBA_DPPY_DEBUG=True`) enables the emission of debug info to 
+the llvm and spirv IR. To disable debugging set this variable to None: (e.g. `export NUMBA_DPPL_DEBUG=`).  
+Currently, the following debug info is available:
+- Source location (filename and line number) is available. 
+- Setting break points by the line number.
+- Stepping over break points.
+
+### Requirements
+
+Intel GDB installed to the system  
+follow the instruction: https://software.intel.com/content/www/us/en/develop/tools/oneapi/components/distribution-for-gdb.html
+
+### Example debug usage
+
+```bash
+$ gdb -q python  
+(gdb) break sum.py:13     # Assumes the kernel is in file sum.py, at line 13  
+(gdb) run sum.py
+```

README.md

@@ -57,9 +57,13 @@ python numba_dppy/examples/sum.py
 
 ## How Tos
 
-Refer the HowTo.rst guide for an overview of the programming semantics,
+Refer the [HowTo.rst](HowTo.rst) guide for an overview of the programming semantics,
 examples, supported functionalities, and known issues.
 
+## Debugging
+
+Please follow instructions in the [DEBUGGING.md](DEBUGGING.md)
+
 ## Reporting issues
 
 Please use https://github.com/IntelPython/numba-dppy/issues to report issues and bugs.

numba_dppy/compiler.py

@@ -20,16 +20,19 @@
 import os
 from numba.core.compiler import DefaultPassBuilder, CompilerBase
 
-DEBUG=os.environ.get('NUMBA_DPPY_DEBUG', None)
+
+DEBUG = os.environ.get('NUMBA_DPPY_DEBUG', None)
 _NUMBA_DPPY_READ_ONLY  = "read_only"
 _NUMBA_DPPY_WRITE_ONLY = "write_only"
 _NUMBA_DPPY_READ_WRITE = "read_write"
 
+
 def _raise_no_device_found_error():
     error_message = ("No OpenCL device specified. "
                      "Usage : jit_fn[device, globalsize, localsize](...)")
     raise ValueError(error_message)
 
+
 def _raise_invalid_kernel_enqueue_args():
     error_message = ("Incorrect number of arguments for enquing dppy.kernel. "
                      "Usage: device_env, global size, local size. "
@@ -78,9 +81,11 @@ def compile_with_dppy(pyfunc, return_type, args, debug):
 
     typingctx = dppy_target.typing_context
     targetctx = dppy_target.target_context
-    # TODO handle debug flag
+
     flags = compiler.Flags()
     # Do not compile (generate native code), just lower (to LLVM)
+    if debug:
+        flags.set('debuginfo')
     flags.set('no_compile')
     flags.set('no_cpython_wrapper')
     flags.unset('nrt')
@@ -117,6 +122,7 @@ def compile_with_dppy(pyfunc, return_type, args, debug):
 def compile_kernel(sycl_queue, pyfunc, args, access_types, debug=False):
     if DEBUG:
         print("compile_kernel", args)
+        debug = True
     if not sycl_queue:
         # This will be get_current_queue
         sycl_queue = dpctl.get_current_queue()

numba_dppy/target.py

@@ -254,8 +254,12 @@ def sub_gen_with_global(lty):
     def declare_function(self, module, fndesc):
         fnty = self.call_conv.get_function_type(fndesc.restype, fndesc.argtypes)
         fn = module.get_or_insert_function(fnty, name=fndesc.mangled_name)
-        fn.attributes.add('alwaysinline')
+
+        if not self.enable_debuginfo:
+            fn.attributes.add('alwaysinline')
+
         ret = super(DPPYTargetContext, self).declare_function(module, fndesc)
+
         # XXX: Refactor fndesc instead of this special case
         if fndesc.llvm_func_name.startswith('dppy_py_devfn'):
             ret.calling_convention = CC_SPIR_FUNC