[clang] [llvm] [AMDGPU] Convert AMDGPUResourceUsageAnalysis pass from Module to MF pass (PR #102913)

[Scott Linder via llvm-commits]

================
@@ -1755,6 +1755,56 @@ As part of the AMDGPU MC layer, AMDGPU provides the following target specific
 
      =================== ================= ========================================================
 
+Function Resource Usage
+-----------------------
+
+The function resource information (e.g., number of VGPRs) required depends on
+all of its callees' function resources. The expression to denote these
+resources should, therefore, be described as the propagative of its callees'
+equivalent expressions. Said expressions are generated and emitted (as symbols)
+by the compiler when compiling to either assembly or object format and should
+not be overwritten or redefined.
+
+The following describes all emitted function resource usage information:
+
+  .. table:: Function Resource Usage:
+     :name: function-usage-table
+
+     ===================================== ========= ========================================= ===============================================================================
+     Symbol                                Type      Description                               Example
+     ===================================== ========= ========================================= ===============================================================================
+     <function_name>.num_vgpr              Integer   Number of VGPRs used by <function_name>,  .set foo.num_vgpr, max(32, bar.num_vgpr, baz.num_vgpr)
+                                                     worst case of itself and its callees'
+                                                     VGPR use
+     <function_name>.num_agpr              Integer   Number of AGPRs used by <function_name>,  .set foo.num_agpr, max(35, bar.num_agpr)
+                                                     worst case of itself and its callees'
+                                                     AGPR use
+     <function_name>.numbered_sgpr         Integer   Number of SGPRs used by <function_name>,  .set foo.num_sgpr, 21
+                                                     worst case of itself and its callees'
+                                                     SGPR use (without any of the implicitly
+                                                     used SGPRs)
+     <function_name>.private_seg_size      Integer   Total stack size required for             .set foo.private_seg_size, 16+max(bar.private_seg_size, baz.private_seg_size)
+                                                     <function_name>, expression is the
+                                                     locally used stack size + the worst case
+                                                     callee
+     <function_name>.uses_vcc              Bool      Whether <function_name>, or any of its    .set foo.uses_vcc, or(0, bar.uses_vcc)
+                                                     callees, uses vcc or not
+     <function_name>.uses_flat_scratch     Bool      Whether <function_name>, or any of its    .set foo.uses_flat_scratch, 1
+                                                     callees, uses flat scratch or not
+     <function_name>.has_dyn_sized_stack   Bool      Whether <function_name> stack is          .set foo.has_dyn_sized_stack, 1
+                                                     dynamically sized
+     <function_name>.has_recursion         Bool      Whether <function_name>, or any of its    .set foo.has_recursion, 0
+                                                     callees, contains recursion
+     <function_name>.has_indirect_call     Bool      Whether <function_name>, or any of its    .set foo.has_indirect_call, max(0, bar.has_indirect_call)
+                                                     callees, contains an indirect call
+     ===================================== ========= ========================================= ===============================================================================
+
+Futhermore, three symbols are additionally emitted describing the compilation
+unit's worst case (i.e, maxima) ``num_vgpr``, ``num_agpr``, and
+``numbered_sgpr`` which may be referenced and used by the aforementioned
+symbolic expressions. These three symbols are ``max_num_vgpr``,
----------------
slinder1 wrote:

I didn't notice these symbols before, but do we want to also prefix these to reduce the chance for collisions? What about `.amdgcn.max_*`, to match the existing `.set .amdgcn.next_free_*` symbols?

https://github.com/llvm/llvm-project/pull/102913