[llvm] [Offload][OMPX] Add the runtime support for multi-dim grid and block (PR #118042)

[Shilei Tian via llvm-commits]

shiltian wrote:

@jplehr I tested it locally, but couldn't reproduce the failure.

```
➜ LIBOMPTARGET_INFO=63 ./main
omptarget device 0 info: Entering OpenMP kernel at unknown:0:0 with 9 arguments:
omptarget device 0 info: from(unknown)[49152]
omptarget device 0 info: to(unknown)[49152]
omptarget device 0 info: to(unknown)[49152]
omptarget device 0 info: firstprivate(unknown)[4] (implicit)
omptarget device 0 info: firstprivate(unknown)[4] (implicit)
omptarget device 0 info: firstprivate(unknown)[4] (implicit)
omptarget device 0 info: firstprivate(unknown)[4] (implicit)
omptarget device 0 info: firstprivate(unknown)[4] (implicit)
omptarget device 0 info: firstprivate(unknown)[4] (implicit)
omptarget device 0 info: Creating new map entry with HstPtrBase=0x000063ac60a736b0, HstPtrBegin=0x000063ac60a736b0, TgtAllocBegin=0x000079d64fa00000, TgtPtrBegin=0x000079d64fa00000, Size=49152, DynRefCount=1, HoldRefCount=0, Name=unknown
omptarget device 0 info: Creating new map entry with HstPtrBase=0x000063ac60a32500, HstPtrBegin=0x000063ac60a32500, TgtAllocBegin=0x000079d64fa0c000, TgtPtrBegin=0x000079d64fa0c000, Size=49152, DynRefCount=1, HoldRefCount=0, Name=unknown
omptarget device 0 info: Copying data from host to device, HstPtr=0x000063ac60a32500, TgtPtr=0x000079d64fa0c000, Size=49152, Name=unknown
omptarget device 0 info: Creating new map entry with HstPtrBase=0x000063ac60a676a0, HstPtrBegin=0x000063ac60a676a0, TgtAllocBegin=0x000079d64fa18000, TgtPtrBegin=0x000079d64fa18000, Size=49152, DynRefCount=1, HoldRefCount=0, Name=unknown
omptarget device 0 info: Copying data from host to device, HstPtr=0x000063ac60a676a0, TgtPtr=0x000079d64fa18000, Size=49152, Name=unknown
omptarget device 0 info: Mapping exists with HstPtrBegin=0x000063ac60a736b0, TgtPtrBegin=0x000079d64fa00000, Size=49152, DynRefCount=1 (update suppressed), HoldRefCount=0
omptarget device 0 info: Mapping exists with HstPtrBegin=0x000063ac60a32500, TgtPtrBegin=0x000079d64fa0c000, Size=49152, DynRefCount=1 (update suppressed), HoldRefCount=0
omptarget device 0 info: Mapping exists with HstPtrBegin=0x000063ac60a676a0, TgtPtrBegin=0x000079d64fa18000, Size=49152, DynRefCount=1 (update suppressed), HoldRefCount=0
"PluginInterface" device 0 info: Launching kernel __omp_offloading_10302_8ee2def_main_l28 with [2,4,6] blocks and [32,4,2] threads in SPMD mode
AMDGPU device 0 info: #Args: 10 Teams x Thrds:   48x 256 (MaxFlatWorkGroupSize: 1024) LDS Usage: 0B #SGPRs/VGPRs: 16/6 #SGPR/VGPR Spills: 0/0 Tripcount: 0
omptarget device 0 info: Mapping exists with HstPtrBegin=0x000063ac60a676a0, TgtPtrBegin=0x000079d64fa18000, Size=49152, DynRefCount=0 (decremented, delayed deletion), HoldRefCount=0
omptarget device 0 info: Mapping exists with HstPtrBegin=0x000063ac60a32500, TgtPtrBegin=0x000079d64fa0c000, Size=49152, DynRefCount=0 (decremented, delayed deletion), HoldRefCount=0
omptarget device 0 info: Mapping exists with HstPtrBegin=0x000063ac60a736b0, TgtPtrBegin=0x000079d64fa00000, Size=49152, DynRefCount=0 (decremented, delayed deletion), HoldRefCount=0
omptarget device 0 info: Copying data from device to host, TgtPtr=0x000079d64fa00000, HstPtr=0x000063ac60a736b0, Size=49152, Name=unknown
omptarget device 0 info: Removing map entry with HstPtrBegin=0x000063ac60a676a0, TgtPtrBegin=0x000079d64fa18000, Size=49152, Name=unknown
omptarget device 0 info: Removing map entry with HstPtrBegin=0x000063ac60a32500, TgtPtrBegin=0x000079d64fa0c000, Size=49152, Name=unknown
omptarget device 0 info: Removing map entry with HstPtrBegin=0x000063ac60a736b0, TgtPtrBegin=0x000079d64fa00000, Size=49152, Name=unknown
omptarget device 0 info: OpenMP Host-Device pointer mappings table empty
omptarget device 0 info: OpenMP Host-Device pointer mappings table empty
```

https://github.com/llvm/llvm-project/pull/118042